Objectivism, Subjectivism, and Relativism in Ethics: Volume 25, Part 1 (Social Philosophy and Policy) (v. 25)

SUBSTANTIVE MORAL THEORY*

By Philip Pettit

I. Introduction

Philosophy serves two functions in relation to the moral thinking thatall of us practice, philosophers and nonphilosophers alike. First of all, itprovides a metaethical commentary on the nature of moral thought, asthe methodology or the philosophy of science provides a commentary onthe nature of scientific thought. Second, it builds on the common pre-sumptions deployed in people’s moral thinking about moral issues, look-ing for a substantive moral theory that they might support. Metaethicsreceives a lot of attention among philosophers, but substantive moraltheory receives less. Philosophers pursue substantive theory, to be sure,but they often fail to distinguish this pursuit from metaethics. My essayseeks to repair this neglect.

The body of the essay is divided into three sections. In Section II,I give a general characterization of metaethics and show why there isroom and reason to seek out a substantive moral theory. In Section III, Ioffer an overview of some well-established styles of substantive theory. Inthe final section, I sketch out a theory that I find appealing myself.

II. From Metaethics to Substantive Moral Theory

A. The role of metaethics

True to its name, metaethics consists in higher-level reflection on thenature of moral talk and moral thought. It deals with questions of thefollowing kinds.

• The cognitivism question. Do moral predications have truth-conditions, as cognitivists claim? Do they serve to say how thingsare and not just play an expressive or related role in people’spsychologies? When I say “That’s the right option to take,” do Icharacterize the action in some way, or do I just give expression tothe fact that I approve of it?

* I was aided in preparing this essay by comments on an earlier draft by Tristram McPhersonand by conversations both with him and with Michael Smith. I am grateful for the writtensuggestions that I received from the editors of this volume and from Jan Narveson andMichael Huemer.

DOI: 10.1017/S0265052508080011© 2008 Social Philosophy & Policy Foundation. Printed in the USA. 1

• The realism question. If moral predications are truth-conditional,do they predicate real, bona fide properties, so that their truth-conditions are routinely fulfilled? Does the rightness I ascribe tocertain options, if cognitivism is correct, constitute a real property?Or is it as imaginary and unreal as phlogiston proved to be?

• The objectivism question. If the moral properties predicated arebona fide in character, do they answer to our intuitions about thecharacter of moral properties —intuitions to the effect that they arenot just properties of our subjective reactions, for example, and notproperties that are relative to a variable framework, cultural orotherwise? Does the rightness that belongs to certain options, ifrealism is correct, consist in a property of the options in them-selves, rather than in a relationship to us, and in a property that isaccessible in principle from different cultural and other standpoints?

Other, more specific questions come up in metaethics too, dependingon the line taken on the cognitivism issue. If we endorse cognitivism,there will be a question about how moral predicates come to designatethe properties or alleged properties they assign, and if we adopt expres-sivism, there will be an issue about how those predicates connect up withthe attitudes they express. These are questions in moral semantics, par-ticularly in the foundational semantics that seeks to explain how certainterms, in our mouths, latch onto the items they are said to denote orassign or express.1 And those semantic questions lead to correspondingquestions in moral epistemology, bearing on what sort of knowledge orevidence supports us in our making one or another moral predication.

Metaethics, understood in this way, plays a higher-level role in relationto moral or ethical thinking, similar to the role that methodology orphilosophy of science plays in relation to scientific theorizing. Metaethicsoffers a descriptive, insider-based picture of moral thinking, as such think-ing is practiced by ordinary people, and it goes on in a more critical styleto articulate, regiment, and perhaps revise that practice. The descriptiveaccount of how people think in the moral domain provides the criticalexercise with the data on which it works, and, unsurprisingly, there is alot more convergence on the data than there is on the critical revisionsproposed.

B. The data of metaethics

This convergence appears in the fact that a large number of the follow-ing assumptions are taken to be endorsed among the folk (that is, ordi-nary people) by different metaethical philosophies:

1 Robert Stalnaker, “Conceptual Truth and Metaphysical Necessity,” Philosophical Studies18 (2004).

2 PHILIP PETTIT

(1) Domain: Moral predications predicate thin values of rightnessand wrongness, goodness and badness, and thick values anddisvalues such as fairness and kindness, cruelty and inequity.Thin values do not have descriptive connotations, but thick val-ues do.

(2) Status: Moral predications are matters on which people agreeand disagree in apparently the same way that they agree anddisagree on matters of belief.

(3) Independence: The properties predicated are not properties of oursubjective reactions, and are not relative to a variable frame.They are intuitively objective.

(4) Internalism: To predicate rightness (wrongness) of an option is tobe disposed (not) to choose it, without reliance on any indepen-dent desire to do right (avoid wrong).

(5) Justifiability: A right option and only a right option is justifiable.There is no ground for resentment or indignation by others, andno ground for guilt on our own part.

(6) Grounding: A right option normally has rightness-making prop-erties involving thick values; a wrong option has wrongness-making properties involving thick disvalues.

(7) Supervenience: No difference in moral predication is defensibleunless there is a nonmoral difference between the items of whichthe moral property is predicated.

(8) Universal supervenience: The nonmoral differences that supportdifferences in moral predication do not privilege particular times,places, or persons in their particularity, only universal, replica-ble properties.

(9) Accessible supervenience: The nonmoral features that are of moralsignificance are accessible to reflection. People’s moral compe-tence will be put in question to the extent that they prove unableto cite the features that shape their moral judgments.

(10) Paradigms: There are certain independently fixed paradigms ofright and wrong options on which people broadly agree.

Short of being utterly iconoclastic, a metaethical theory will start frompresumptions of this kind, or at least from those presumptions that ittakes to be embodied in folk moral thought. It will provide a construal ofthose presumptions under which they provide answers to the questionsof cognitivism, realism, and objectivism. And it will provide a more orless limited endorsement of the presumptions, so construed —that is, ofthe views ascribed to the folk. There will be three levels, then, at which ametaethics may distinguish itself: in determining which of the candidatesare genuinely folk presumptions; in construing those presumptions sothat they take a particular line on the three questions distinguished; and

SUBSTANTIVE MORAL THEORY 3

in endorsing or rejecting the presumptions so construed, whether in wholeor in part.

C. Moral functionalism

My preferred metaethics assumes that all or most of the ten candidateslisted above are folk presumptions and argues that moral predications aretruth-conditional and that the properties they predicate are selected ascandidate semantic values (at least their selection is partly determined)2

by the fact that they satisfy such presumptions. Why will a propertydeserve to count as the semantic value of a given predicate —the propertythat the predicate assigns? It will deserve to do so because it has suitableconnections with paradigms and nonmoral features, with motivation andjustification and action, and with the properties that count on similargrounds as the semantic values of other moral predicates. Reference isfixed under this account by the functional profile that the referent of amoral term or concept is required to fit. For that reason, Frank Jacksonand I have described this account as a moral functionalism.3

On this account, then, a property will count as the property of justiceinsofar as (1) it is exemplified in certain procedures, (2) it makes sense ofwhy someone might choose actions that bear it, or might think they arejustifiable, and (3) its apparent presence in an action counts in ordinaryreasoning as a ground for thinking the action is the right choice. Justice isthat property, whatever it may be in itself, that meets those usage spec-ifications; it is individuated functionally by playing the role that theycharacterize.4 And as the referent of “justice” will be determined by itsplace in a network of such connections, so the same will be true of thesemantic values of other moral terms as well. Those terms or conceptswill have their referents determined simultaneously with the referentof “justice,” and the mastery of any one term —the identification of itsreferent —will go hand in hand with the mastery of others. Semanticcompetence will be achieved in a package deal.

2 Philip Pettit, “Descriptivism, Rigidified and Anchored,” Philosophical Studies 18 (2004):323–38.

3 See Frank Jackson and Philip Pettit, “Moral Functionalism and Moral Motivation,”Philosophical Quarterly 45 (1995): 20–40; reprinted in Frank Jackson, Philip Pettit, and MichaelSmith, eds., Mind, Morality, and Explanation (Oxford: Oxford University Press, 2004). Forfurther background, see Frank Jackson, “Critical Notice of Susan Hurley’s Natural Reasons,”Australasian Journal of Philosophy 70 (1992): 475–87; Frank Jackson, From Metaphysics to Ethics:A Defence of Conceptual Analysis (Oxford: Oxford University Press, 1998); and Philip Pettit,“Embracing Objectivity in Ethics,” in Objectivity in Law and Morals, ed. Brian Leiter (Cam-bridge: Cambridge University Press, 2001), 234–86.

4 Or, to gesture at a variant, it is the higher-order property of there being some propertypresent that meets those specifications; it may be the role property, as it is sometimes called,rather than the realizer property that plays the role. On this distinction, see Frank Jacksonand Philip Pettit, “Functionalism and Broad Content,” Mind 97 (1988): 381–400; reprinted inJackson, Pettit, and Smith, eds., Mind, Morality, and Explanation, 95–118.

4 PHILIP PETTIT

Moral functionalism, so characterized, endorses cognitivism and sug-gests a distinctive semantics for moral predicates. Moral predications aretruth-conditional, under this account, and the predicates they deploy pickout certain properties as their semantic values insofar as those propertiessatisfy the usage specifications given by the features listed, or some sub-set. Depending on exactly which subset is valorized, of course, the moralfunctionalism embraced may take one or another specific form; in thatsense, it directs us to a family of metaethical views rather than any par-ticular position. And as moral functionalism gives us a moral semantics,so too it will provide us with a moral epistemology. According to theimplied epistemology, we will know when to make one predication oranother, depending on the usage specifications that appear to be fulfilled.

Nonetheless, moral functionalism does not yet commit us on the issuesof realism and objectivism. According to the account that I find appealing,there will be suitable, naturalistic properties available, however disjunc-tive in character, that fit the required roles. And in that case moral prop-erties will be real and objective features of the world. Moral functionalismmight equally be squared, however, with a nonnaturalistic story about therealizers of the roles. And it might even be maintained consistently withan error theory about the realizers: a theory according to which moralthought is deeply misconceived, there being no properties available to fitthe roles that it projects for them.5

D. The folk’s schematic moral theory

The idea in moral functionalism is that we master the different moralpredicates by coming to understand, at one and the same stroke, what isrequired for each of them to have application —what is required for theproperty a predicate ascribes to be present. To achieve a mastery or under-standing of moral language, then, will be to have recognized how thoseproperties that answer to the moral terms relate to one another and howthey relate en bloc to evidence and action. And to take moral languageseriously, treating it as a way of representing how things are, will be tohave endorsed a theory —a formal or schematic theory —as to the exis-tence of properties that satisfy such a network of connections. Accordingto moral functionalism, there is no engaged moral talk or thought —noserious, literal usage of moral terms and concepts —without a commit-ment to this modicum of moral theory.6

5 For an example of an error theory, see J. L. Mackie, Ethics (Harmondsworth: Penguin, 1977).6 The approach represents an analytical functionalism in the moral area that parallels the

analytical functionalism that is often defended in the area of mind. It identifies differentmoral properties, at least in part, by the ways in which they are connected up in the practice,and in the implicit thought, of participants. Notice that the connections signaled involve theway people treat properties and the sorts of practices they follow. The connections aredifferent, but not so different, from the causal connections that figure in functionalismwithin the philosophy of mind. In particular, they are not themselves normative connections.


Is moral functionalism unusual or even unique in postulating a folkmoral theory of this schematic kind? I do not think so. There are twosalient ways in which a metaethics might break with the family of moral-functionalist doctrines. One is by holding that moral predications are nottruth-conditional but expressive; and the other is by holding that they aretruth-conditional but that the reference of moral terms or concepts isfixed on a purely direct basis, not by satisfaction —not even in part bysatisfaction —of suitable usage-specifications. Under either of these vari-ations, however, it appears that the metaethics in question will still haveto countenance the existence of a folk moral theory.

Under the first variation, where the truth-conditional character of moralpredication is denied, the folk will be said to regard moral predications aspurely expressive and to see moral agreement and disagreement as dif-ferent in character from agreement or disagreement in matters of belief.But this expressivist story does not relieve the folk of having to acceptsome or all of the other presumptions listed above (in subsection II.B).Presumably, it will remain the case that in order for people to becomecompetent in the non-assertoric practice that moral thinking allegedlyinvolves, they will still have to take guidance from such presumptions.They will have to recognize the role of thick in relation to thin values; theywill have to see the motivational connection and the justificatory role ofmoral predication; and they will have to employ such predications withsensitivity to requirements like supervenience and universalizability. Ifthey commit themselves seriously to moral talk and thought, then theywill still have to understand and adjust to this set of presumptions. Theywill have to endorse the theory that these presumptions constitute butnow in a different role from the role the theory plays under a moral-functionalist approach.

Under the second variation from a broadly moral-functionalist approach,moral terms and concepts can gain their semantic values on the basis ofexposure to suitable candidates, without a requirement that those candi-dates should satisfy any presumptions that the folk make about them.Suppose the folk use a moral term “t.” Assuming that error theory is false,the idea is that that term will pick up a property in the world as itsreferent on the basis of something about the property that may in no wayregister with users of the term. What selects it as the referent of the termmay be the fact that the property, unbeknownst to speakers, is at thecausal origin of a use of the term that they take as authoritative, or the factthat it uniquely co-varies with the general use of the term, or the fact thatamong the properties that co-vary with that use, it is a uniquely eligiblereferent, on some independent criterion of eligibility. The folk’s presump-tions about the property that they assign with the term “t” may be quitefalse, then; the reference is fixed by factors that may escape the notice ofspeakers, and thus the beliefs that the speakers form about the referentproperties may be quite misconceived.

6 PHILIP PETTIT

This direct theory of reference for moral terms and concepts is hardlya plausible theory for all the moral predicates there are. But if it is a theoryonly of some predicates, then it will leave moral functionalism, or some-thing like moral functionalism, in place for the remaining predicates.Thus, it will leave room for a version of the schematic moral theory thatmoral functionalists ascribe to the folk.

Nonetheless, even if the theory is said to apply to all moral predi-cates, it will still have to keep a role for such a folk theory. If thereferents of moral terms are fixed in a manner that puts them poten-tially beyond our ken, leaving us prey to endorsing nothing but false-hoods about them, then that opens the way for a radical, incrediblescepticism. We will each be held to engage unknown properties in ourassertions about different options or policies or persons and to ascribethose properties without any reliable basis —indeed, without any basisat all —for determining whether they actually apply. Making moral pred-ications will be like throwing darts in the dark. In order to guardagainst this outlandish scenario, those who opt for the direct theory ofreference for moral terms will have to tell a story according to whichthere are reliable guidelines and grounds available for the correct orpresumptively correct application of such terms. But where are thoseguidelines to come from, if not from a body of judgments and presup-positions that are ascribed to the folk and treated as fallible but reli-able? And where are we to find such folk commitments if not in thesorts of presumptions listed above? This approach will also have tocountenance a folk moral theory, though a theory that serves primarilyin an epistemological rather than a semantic role.

E. From schematic to substantive moral theory

If we hold a schematic moral theory, and if that theory guides us in ananalysis of moral issues, then reflection on where it guides us —reflectionon the pattern we are disposed to follow in our judgments —may enableus to articulate what I call a substantive moral theory. I will argue thisline, on the assumption that moral functionalism holds, and that it directsus to genuine properties, not illusions. This is for reasons of convenienceonly, however, since there will be a parallel line to argue under anymetaethics. The folk will be ascribed a schematic moral theory underevery approach, as we have seen, and in each case there will be a base forseeking out a substantive theory to which it points.

Suppose, then, that presumptions like those given in the earlier listserve for me as semantic markers or indicators of what is right or wrong,good or bad, just or unjust, and so on. And suppose that working withthese presumptively common guidelines, I argue for the rightness orwrongness, the goodness or badness, the justice or injustice, of one or


another action or policy. I will do so under the assumption that whilethere are trade-offs to be made and fine lines to be drawn in the courseof the exercise —it will involve the search for a reflective equilibriumbetween the presumptions I maintain and my intuitions about the casesadjudicated7 —still there is a correct and an incorrect way of extrapolatingto other cases, and a way of extrapolating on which I may expect a degreeof convergence with others.8

Assuming that the extrapolative, judgmental dispositions that the sche-matic presumptions underpin will guide me fairly determinately, there isan immediate prospect that I can work my way toward a substantivemoral theory. I will be able to explore this possibility, it appears, using afamiliar philosophical method of analysis: the method of cases. Equippedwith the judgmental dispositions, I can envisage any number of differentmoral issues and I can investigate where the schematic presumptionswould lead me across those possible, imagined cases. With those judg-ments of possible cases more or less secure, I can then look at the moralconfiguration they assume, in order to see whether there is any substan-tive commonality in the extension of rightness and wrongness, or good-ness and badness, or justice and injustice, and so on. The challenge will beto find a commonality that is more specific and more salient than thepattern given by the schematic presumptions, as I interpret and applythem.

According to any moral-functionalist metaethics, or at least one thatrejects wholesale error, the schematic presumptions associated with moralterms or concepts will serve like a net with which I can trawl the differentcases and issues, real and imagined, that I consider. That net will sift outsome options as instances of the right actions or policies to adopt in givencases, for example, and leave others as instances of actions or policies thatare wrong. The idea in pursuing a substantive moral theory is to seewhether there is a salient pattern to be found in such instances of right-ness or wrongness over and beyond the pattern that consists in the factthat they are caught or not caught by the net. That they are or are notcaught by the net is presumably due to their inherent character, andsubstantive moral theory is built on the hope that this character can be

7 On reflective equilibrium, see John Rawls, A Theory of Justice (Cambridge, MA: HarvardUniversity Press, 1971).

8 If I take myself to be guided by common markers of value and disvalue, after all, thenI will have to assume that that guidance is not peculiar to me. The common meaningsestablished for terms like “just” and “good” and “right” hold out the prospect of conver-gence in judgments of value, and so long as that possibility is live, I will find a point intalking with others, exchanging comments on what differences and continuities registerwith us, how different presumptions are weighted in relation to one another, how they areamended in the course of reflective equilibration, and so on. I may come to believe, ofcourse, that the meanings available leave issues of value indeterminate, and I may come totreat those issues, like various issues of baldness and color and the like, as vague at certainmargins. But I will presumably retreat to this sort of quietism only as a last resort; it amountsto throwing in the conversational towel.

8 PHILIP PETTIT

identified, using something like the method of reflection on cases, bythose who make moral judgments.9

There may not prove to be any interesting characterization availableat the level of substantive theory, of course, but that is not a reason fordespairing at the outset of the enterprise. At least, it is not a reason fordespairing, under the assumption that cases can be morally adjudi-cated on the basis of how they are described, and that adjudicationdoes not require causal exposure to the case assessed, or even to avivid simulacrum on stage, in a film, or in a novel. This assumption,which is implied in the claim that I described as accessible super-venience, is validated in the way we expect someone to be able toexplain a disparity in the moral judgment made over two cases bypointing to a descriptive difference between the cases —in particular, bypointing to it as a factor that justifies the disparity, not merely as aconjectural source of the disparity. Suppose some individuals report amoral judgment about one or another issue: a considered judgment,not just a dogged but not necessarily endorsed intuition. If they areunable to justify the judgment by reference to some descriptive char-acterization of the case, then we will doubt their skill at making moraljudgments. And if they are unwilling to acknowledge the need for suchdescriptive justification of the judgment, then we will doubt their veryunderstanding of what moral judgment involves.10

9 For an argument that there must be an inherent character there to explain the semanticcompetence of moral speakers, in particular their ability to catch on to the reference of aterm like “right,” see Frank Jackson, Philip Pettit, and Michael Smith, “Ethical Particularismand Patterns,” in Moral Particularism, ed. Brad Hooker and Margaret Little (Oxford: ClarendonPress, 2000). This argument challenges the radical version of particularism, according towhich there need be no independent pattern of any kind in the extension of a term like“right.” A more moderate form of particularism would defend the aesthetic model of moraljudgment mentioned below in note 10; this model holds that there need be no pattern in theextension of “right” that is accessible, at least in principle, to reflection on the part ofcompetent speakers. See Jonathan Dancy, Ethics without Principles (Oxford: Oxford Univer-sity Press, 2004).

10 This assumption is rejected under what may be described as the aesthetic model ofmoral judgment. When I make an aesthetic judgment on a painting, I must do so on thegrounds of how it impacts perceptually and causally on me. It would be absurd to have apainting described and then, on the basis of that description alone, to pass aestheticjudgment —to say, for example, that the painting must be very beautiful, or elegant, orwhatever. See Philip Pettit, “The Possibility of Aesthetic Realism,” in Pleasure, Preference, andValue: Studies in Philosophical Aesthetics, ed. Eva Schaper (Cambridge: Cambridge UniversityPress, 1983), 17–38. The aesthetic model of moral judgment would suggest that I can onlyform a moral judgment about an action on the basis of how it impacts experientially on me.There may be a substantive pattern to identify across the aesthetic judgments we defend (oracross moral judgments, if they are conceived as similar to aesthetic ones), but it will bethere to be identified only on the basis of empirical research; there will be no possibility ofrelying on reflection in order to bring it to light. Those who passed a positive judgment onthe drip paintings of Jackson Pollock, and a negative judgment on various imitations, werein no position to identify the substantive pattern that physicists have since claimed tounearth in Pollock’s paintings —a pattern of self-similarity at different grains of analysis. SeeRichard P. Taylor, “Order in Pollock’s Chaos,” Scientific American (December 2002): 117–21.


F. Looking for substantive moral theories

With the case made for why substantive moral theory, like metaethics,is a quarry that philosophers may pursue, we are in a position to look forsome examples of such theorizing in the tradition, and indeed amongcontemporary figures. I pursue that task in the next section.

The theory I ascribe to some thinkers may be conceived by them as anexercise in metaethics, not as a substantive moral theory. Consider utili-tarianism, Kantianism, contractualism, or virtue ethics. These are some-times taken as metaethical theses to the effect that insofar as they bear onrightness, the schematic presumptions in our earlier list, whether or notthe folk realize it, can really be boiled down to a single, favored thesis.The claims made under such approaches would be that the concept of theright is just the concept of that which maximizes happiness, or that whichcan be coherently willed as a universal law, or that which is immune toreasonable complaint, or that which the virtuous would choose.

It is not really plausible, however, to cast the claims in this way. For insupporting the claims, defenders generally appeal to background assump-tions about the rightness property, thereby displaying an adherence to adistinct schematic theory. They try to show that if we identify instances ofthe right on the basis of such schematic assumptions, then we will findthat the favored account applies across those judgments; the accountidentifies a substantive pattern in the extension of “right” and “wrong,”as those words are used under the guidance of the schematic assumptions.

One reason for the failure to recognize various theories as substantivemoral theories, not as contributions to metaethics, may derive from anuncertainty about the point of such theory. So what purpose is a substan-tive theory —say, a substantive theory of the right —likely to serve?

Such a theory may be comprehensive or restricted; it may claim toidentify a common character in all instances of rightness or just a char-acter common to instances of rightness in one or another area. In eithercase, its use will be twofold. In relation to moral analysis, it will providea basis on which judgments about particular cases can be facilitated orsupported. And in relation to schematic theory, it will provide somevindication of the presumption that the marks of rightness identify amore or less clearly patterned and learnable extension —an extension suchthat it is possible for those who are exposed to some proper subset ofinstances to extrapolate to further cases.

III. Five Substantive Moral Theories

There are at least five substantive theories of morality in the grandtradition. They focus on some central demands that morality intuitivelyimposes —some allegedly salient patterns in the distribution of the right-ness property —where these demands may include both strict imperatives

10 PHILIP PETTIT

and looser ideals. The theories identify the central constraints of moralitywith requirements, respectively, of nature, self-interest, benevolence, rea-son, and justifiability.

A. Nature

The identification of moral requirements with requirements of naturetakes a number of different forms and is common in classical and medi-eval thought. The Aristotelian equation of the demands of morality withwhat is required by nature if we human beings are to flourish and enjoyeudaimonia offers one version of the idea; this approach retains a placeamong neo-Aristotelians like Philippa Foot and Rosalind Hursthouse.11

Another version is the Stoic conception of laws of human nature that areavailable to our reason from the intimations of instinct, laws such that bybeing faithful to them we bring ourselves into line with the order of thecosmos. The medieval synthesis that was achieved by Thomas Aquinasbuilt on both of these foundations, identifying the prospect of humanfulfillment with obedience to laws whereby the natural instincts for self-preservation, procreation of the species, life in society, and knowledge ofgod could be reconciled and satisfied.

The metaphysics of naturally based requirements gave way, toward theend of the medieval period, to a metaphysics of natural rights, underpressures that are still much debated.12 Natural rights are supposed toplace demands, not on their holders, but on those who deal with theirholders. Absent an agreement to be treated otherwise, such rights aremeant to require that in dealing with one another we treat certain areas ofchoice, usually related to life, liberty, and property, as sacred; we alloweach other more or less total discretion within those boundaries.

The moral metaphysics associated with natural flourishing, naturallaw, and natural rights is not likely to appeal on a broad front in contem-porary discussion. There is no consensus about what human flourishinginvolves, or what natural instinct intimates, so that neither base identifiesa determinate set of requirements. And while many people talk the lan-guage of natural rights, this metaphysics also fails to deliver the sorts ofposits that are likely to satisfy us. If talk of rights is reduced to otherterms, as it frequently is, then it does not give us distinctive requirements;it becomes a mere facon de parler. And if it is not reduced, so that naturalrights are taken to make primitive, unanalyzable claims, then they ceaseto be obviously plausible entities to posit; they begin to look meta-

11 See Philippa Foot, Natural Goodness (Oxford: Oxford University Press, 2001); and RosalindHursthouse, On Virtue Ethics (Oxford: Oxford University Press, 1999).

12 See Annabel Brett, Liberty, Right, and Nature (Cambridge: Cambridge University Press,1998); and Richard Tuck, Natural Rights Theories: Their Origin and Development (Cambridge:Cambridge University Press, 1979).


physically dubious: “simple nonsense,” as Jeremy Bentham thought, even“nonsense upon stilts.” 13

B. Self-interest

In the history of ethical and political thought, the moral metaphysicsthat supplanted that of nature was one of self-interest. Sixteenth-centurythinkers such as Hugo Grotius and Thomas Hobbes continued to use thelanguage of natural law, but, especially in the work of Hobbes, naturallaws came to be more or less synonymous with maxims of self-interest, inparticular, self-preservation —in Hobbes’s words, “conclusions or theo-rems” that people could work out by reason “concerning what conducethto the conservation and defence of themselves.” 14

There is more determinate content to the idea of what is required forour self-interest and self-preservation than there is to the idea of what isrequired by our nature. But how could such requirements ever hope tocount as moral requirements? In particular, how could they justify whatis done in their name, especially when this may impose a risk of disad-vantage on others?

The answer that Hobbes gave, and the answer given by contemporaryHobbesians like David Gauthier, is that if the requirements of self-interestare adjusted for life with others, they begin to approximate intuitivelymoral demands.15 The idea is that self-interest will give each individuala reason to help establish certain principles whereby the common life ofmembers of the community can be organized to maximum mutual advan-tage and, even more important, a reason to abide by the principles accepted.

But still, there is a problem. Won’t people want to exploit others byfree-riding on their efforts, whenever this is in their self-interest? Indeed,won’t this be required when such exploitation maximizes personal advan-tage by giving the agent enough benefit to outweigh the damage thatothers suffer? It seems quite counterintuitive to identify moral require-ments with imperatives that allow, even mandate, such exploitation. Andthat problem is exacerbated by the fact that the pursuit or implementationof such prescriptions will often be self-defeating for the collectivity; ifeach individual acts on these prescriptions, then overall self-interest willnot be well served.16

Hobbes’s answer to this problem, which is scarcely persuasive, is thatfree-riding will not be in people’s self-interest, given the likelihood of

13 Jeremy Bentham, “Anarchical Fallacies,” in The Works of Jeremy Bentham, ed. J. Bowring(Edinburgh, 1843).

14 Thomas Hobbes, Leviathan, ed. E. Curley (Indianapolis, IN: Hackett, 1994), 15.41.15 David Gauthier, Morals by Agreement (Oxford: Oxford University Press, 1986). On the

Hobbesian approach more generally, see J. S. Kraus, The Limits of Hobbesian Contractarianism(Cambridge: Cambridge University Press, 1993).

16 Derek Parfit, Reasons and Persons (Oxford: Oxford University Press, 1984), part 1.

12 PHILIP PETTIT

detection. Someone “that breaketh his covenant,” being readily detect-able, “declareth that he may with reason do so,” and “can in reasonexpect no other means of safety than what can be had from his own singlepower.” 17 Gauthier’s answer, which is only a little more satisfying, is thatthe disposition to abide with an agreement made will generally be inpeople’s self-interest. While this may be correct, it doesn’t really addressthe problem: people are presumably able to act against their dispositions,especially when self-interest requires it.18

C. Benevolence

Despairing of the resources of self-interest, or at least of self-interesttaken on its own, the eighteenth-century British philosopher FrancisHutcheson introduced the idea that moral requirements are just the require-ments of benevolence.19 Hutcheson argued that people like others to behappy and do not like them to be unhappy, and that this benevolenceleads people to perform and approve of suitable actions, and to eschewand disapprove of others. Where Hutcheson took this to be an innatedesire, David Hume argued that it could be explained by reference to thesympathy we feel when others suffer what we think would be painful, orenjoy what we think would be pleasant.20 Exercised within the generalpoint of view in which we put aside our own particular concerns, sym-pathy, according to Hume, makes sense of the benevolence that Hutchesonpostulated.21

What we are offered here is essentially a utilitarian account of therequirements of morality, as indeed Hutcheson already advertised in 1725:“that action is best, which procures the greatest happiness for the greatest

17 Hobbes, Leviathan, 15.5. See K. Hoekstra, “Hobbes and the Foole,” Political Theory 25(1997): 620–54.

18 Gauthier argues, in later work, that rationality can oblige people to take a temporallyextended course of action (say, promising at one time and then actually delivering on thepromise later) even where promising-and-not-delivering has higher expected utility thanpromising-and-delivering. See David Gauthier, “Resolute Choice and Rational Delibera-tion,” Noûs 31 (1997): 1–25. See also Edward McClennen, Rationality and Dynamic Choice(Cambridge: Cambridge University Press, 1990). But this is an unorthodox and highlycontroversial view, since it introduces an artificial way of limiting the options that agentscan rationally consider; it rules out the option of reconsidering whether to deliver on thepromise, even when that possibility has become salient. It may, of course, be rational to havea policy of not in general reconsidering one’s options —a policy of not reconsidering exceptwhen the red lights go on. See Michael Bratman, Faces of Intention: Selected Essays on Intentionand Agency (Cambridge: Cambridge University Press, 1999). But this does not mean thatrationality rules out reconsideration, period.

19 Gilbert Harman, Explaining Value and Other Essays in Moral Philosophy (Oxford: OxfordUniversity Press, 2000), chap. 11.

20 David Hume, An Inquiry Concerning the Principles of Morals, ed. J. B. Schneewind (Indi-anapolis, IN: Hackett, 1983); David Hume, A Treatise of Human Nature, 2d ed., ed. P. H.Nidditch (Oxford: Oxford University Press, 1978).

21 Geoffrey Sayre-McCord, “On Why Hume’s ‘General Point of View’ Isn’t Ideal —andShouldn’t Be,” Social Philosophy and Policy 11, no. 1 (1994): 202–28.


numbers; and that, worst, which, in like manner, occasions misery.” 22 Theaccount was later given full formal dress in the work of Jeremy Benthamand other utilitarians. It identifies the requirements of morality, quitesimply, with the requirements associated with maximizing the expecta-tion of happiness, whether among human beings or among sentient crea-tures more generally. It retains a prominent place in contemporary moralphilosophy, figuring in the work of Peter Singer and many others.23

The value of benevolence is only one among a number of differentvalues —thick values —in our folk moral theory, and it is counterintuitiveto construe all moral requirements as dictates of benevolence. Equally,since benevolence is so demanding, it is counterintuitive to take every-thing that benevolence might dictate as a requirement of morality. Utili-tarianism would not substantiate our intuitive conception of moralrequirements so much as replace it with a novel conception; utilitarians,as they were seen to be in the nineteenth century, are philosophical radicals.

D. Reason

Reacting against the focus on sentiment and sympathy in Hume’s work,Immanuel Kant argued that moral demands were not requirements ofbenevolence but rather requirements of reason, under his particular con-ception of reason. He gave rise thereby to one of the major strands incontemporary moral theory. This strand is faithfully maintained in thework of many contemporary Kantians, such as Barbara Herman andChristine Korsgaard, but it is also represented in non-Kantian form by thework of other, very different thinkers, such as R. M. Hare.24

The Kantian claim is that the very idea of acting on reason imposesrequirements of a kind that can plausibly be identified with the require-ments of morality.25 There are many variations on the argument for thisidea, and many versions of each of those variations, but one central lineof thought is this: If I am in causal charge of what I choose, as I naturallyassume in any choice, then I must choose in accordance with a universallaw. Causation, Kant assumes, is essentially lawlike. It follows that if I failto choose in accordance with a universal law, then it is not I who is incausal charge of what is done but an alien, heteronomous force withinme. Choosing in accordance with a universal law would mean choosingin accordance with a maxim that can be coherently imposed as a universal

22 Quoted in Harman, Explaining Value, 186.23 Peter Singer, Practical Ethics, 2d ed. (Cambridge: Cambridge University Press, 1993).24 R. M. Hare, Moral Thinking: Its Levels, Method, and Point (Oxford: Oxford University

Press, 1981); Barbara Herman, The Practice of Moral Judgment (Cambridge, MA: HarvardUniversity Press, 1993); Christine Korsgaard, The Sources of Normativity (New York: Cam-bridge University Press, 1996).

25 Immanuel Kant, Groundwork of the Metaphysics of Morals (1785), trans. Mary Gregor(Cambridge: Cambridge University Press, 1998).

14 PHILIP PETTIT

law. Thus, if I am to vindicate the natural assumption that I am in causalcharge of what I choose, then I must choose in accordance with a maximthat I can will as a universal law. This is a constraint that binds any agentof reason unconditionally; it is a categorical imperative, as Kant calls it.

Kant’s view is that the requirements of morality just are the require-ments imposed by the categorical imperative of choosing in accordancewith coherently universalizable maxims. Let the maxim by which I act beuniversalizable and I will be acting morally; let it be nonuniversalizableand I will be acting immorally. Imagine that I steal someone else’s prop-erty, acting on the maxim of a thief. The Kantian idea is that I cannotcoherently will this maxim of stealing what belongs to another as a uni-versal law because if everyone acted on that maxim, then there would beno such thing as property and no such thing as stealing.

The Kantian proposal is hard to defend, at least in this version, since theissue of whether I can will a maxim as a universal law turns on how thatmaxim is described. I may not be able to endorse the maxim of stealingas a universal law, but I can endorse many more-specific versions of thatmaxim as universal laws. For example, there is no reason why I shouldnot be able to endorse, as a universal law, the maxim of stealing another’sproperty in very specifically characterized circumstances, such as those Ihappen currently to be in. I might endorse as a universal law the maximof stealing a book from someone who no longer needs it urgently, on theeve of an examination that requires me to have read that book.

R. M. Hare’s “prescriptivism” is an attempt to build on roughly thisKantian idea, avoiding the sort of problem I have just raised.26 Sayingthat an option is right, on Hare’s approach, is just prescribing it univer-sally, so that an action will be right just in case it is universally prescrib-able. But it is not enough for universalizability in this sense that universallyprescribing the option is a coherent strategy. What is also required is thatI be able to think that this is the option I would prescribe for me now, nomatter what position I occupied and no matter what impact I suffered.Universally prescribing the option must be a psychologically feasiblestrategy, as we might say, not just a coherent one. That is why Hare’s testof universalizability is supposed to avoid the indeterminacy problem.

Hare assumes that in prescribing an option for myself, without univer-salizing, I will always take into account the desires that I expect the optionto satisfy: I will prescribe the option on the grounds that it promisessatisfaction of those desires. He argues that when I prescribe the optionuniversally —that is, not just for me now, but for anyone in a relevantlysimilar situation —I will prescribe it on the grounds that it best providesfor the overall satisfaction of the desires of affected parties. Thus, on thisapproach, there is an even more specific commonality to right optionsover and beyond that of being universally prescribable. This is that the

26 See Hare, Moral Thinking.


options maximize expected desire-satisfaction —that is, that they have thecharacter prized by the preference-based form of utilitarianism. Hareoffers substantive theories at two levels; at one level, universal prescrib-ability is the common feature of right options; at a second level, utilitariandesire-satisfaction plays that role.

Hare’s approach suffers, for this reason, from all the problems associ-ated with utilitarianism, but it is also subject to a more damaging objec-tion.27 The argument for the determinacy of universal prescribability, andthus for Hare’s utilitarianism, depends crucially on the assumption thatin prescribing an option for myself, without universalization, I will alwaysdo so on the grounds that it promises maximal desire-satisfaction. Andthat assumption is very questionable.

It may well be true that whenever I prescribe an option for myself, I doso out of a corresponding desire, but that does not mean that I prescribeand choose it on the grounds that it will satisfy such a desire. It does notfollow, then, that in prescribing something universally I will prescribe iton the grounds that it does as well as possible by desire-satisfactionoverall. Socrates prescribed drinking the hemlock for himself on thegrounds that this is what respect for the law required, not on the groundsthat doing so would satisfy his desires, in particular his desire to showrespect for the law. Thus, he would have prescribed the choice of thehemlock for anyone in any situation where those grounds remained valid.And he would have prescribed it in that universal way, regardless ofwhether or not it happened to increase the net balance of desire-satisfaction.

E. Justifiability

A fifth and distinct conception of the requirements of morality wasprovided by Adam Smith, a friend and admirer of Hume.28 Taking up theHumean idea of sympathy, he developed it on different lines.29 We sym-pathize with the actions and reactions of others, Smith said, but onlyinsofar as we would display more or less similar responses in their cir-cumstances; if others underreact or overreact, by the criteria we employin an impartial viewpoint, then we disapprove. Given the notion of impar-tial sympathy, Smith tells a distinctive and persuasive story about howwe come to identify certain requirements for ourselves; these, by his

27 For a critique, see Philip Pettit, “Universalizability without Utilitarianism,” Mind 96(1987): 74–82; and Philip Pettit and Michael Smith, “Backgrounding Desire,” PhilosophicalReview 99 (1990): 565–92; reprinted in Jackson, Pettit, and Smith, eds., Mind, Morality, andExplanation, 131–53. See also Philip Pettit, “Preference, Deliberation, and Satisfaction,” inPreferences and Well-Being, ed. Serena Olsaretti (Cambridge: Cambridge University Press,2006), 131–53.

28 Adam Smith, The Theory of Moral Sentiments (1759), ed. D. D. Raphael and A. L. McFie(Indianapolis, IN: Liberty Classics, 1982).

29 Here I follow Harman, Explaining Value, chap. 11.

16 PHILIP PETTIT

account, are the requirements of morality. The elements in the story are asfollows:

(1) We sympathize with others just so far as their actions and reac-tions seem fitting to their circumstances, being responses we canimagine ourselves having;

(2) we learn that others sympathize with us in turn, forming positiveor negative appraisals, in the same impartial and critical manner;

(3) we learn to put ourselves in the shoes of others, and to see ourown actions and reactions as they would see them from theirimpartial, critical point of view;

(4) we are naturally motivated to seek the approval of others, and toavoid their disapproval; and so

(5) we come to be motivated to act and react in a manner deservingof more or less impartial sympathy.

According to this complex story, then, the requirements of moralityshould be identified with what is required, not for the sympathy that theignorant or biased may bestow, but for the impartial sympathy of others.The requirements of morality are those requirements that we have tosatisfy if, on the negative side, we are to avoid the reasonably basedcomplaints of others and if, on the positive side, we are to win theirreasonably based plaudits. Where Hutcheson and Hume had equated therequirements of morality with the requirements of benevolence, then,Smith was led from a similar starting point to take a rather different view.He identifies the requirements of morality with the conditions whoserealization would put us beyond the complaint of others, and within theambit of their approval. We might describe these, in a generic phrase, asrequirements of justifiability.

The requirements to which Smith directs us have great appeal. They arenot obscure, like natural law and natural rights. They are not counter-intuitively egocentric like the requirements of self-interest, nor counter-intuitively altruistic like the requirements of benevolence. And they donot suffer, on the face of it, from the indeterminacy problem associatedwith Kant. The justifiability feature is a very promising way of unifyingat least a good portion of those options that count intuitively —that is,count by our presupposed, schematic theory —as right.

Two contemporary theories build on broadly the same base as AdamSmith’s, at least as I see them. One is the contractualist approach of T. M.Scanlon.30 Right actions, for Scanlon, are those that are not wrong; andwrong actions are those against which others can raise complaint on thebasis of principles that no one could reasonably reject as principles for the

30 T. M. Scanlon, What We Owe to Each Other (Cambridge, MA: Harvard University Press,1998).


general regulation of social behavior. While Scanlon does not cite Smithas a predecessor, the approach represents something close to a negativeversion of the justifiability line. The requirements of morality are identi-fied as those whose fulfillment makes for immunity to reasonable com-plaints, though no mention is made of requirements whose fulfillmentwould serve to attract corresponding plaudits. Scanlon differs from Smithin giving us a quasi-contractual rather than a genealogical story as to whycertain complaints are reasonable: they are backed by principles that noone could reasonably reject as principles of social cooperation. But Scanlon’sapproach and Smith’s still clearly belong in the same family.

A second contemporary theory that shares a common base with AdamSmith’s is that of Michael Smith.31 Where the linking idea between AdamSmith and Scanlon is that of immunity to complaint, the linking idea hereis that of the impartial spectator. Adam Smith’s impartial spectator is theinternalized other, as we have seen; Michael Smith’s is the idealized self.The demands of morality, on this approach, belong among those require-ments that I would advise my actual self to satisfy in any situation, wereI, the advising self, to be beyond criticism —were I to be fully informed,for example, to have fully coherent and unified attitudes, and to be think-ing in a fully logical way. These are requirements, therefore, such thatsatisfying them ensures that I am justified; I pass muster with the impar-tial spectator of my ideal advising self.32

There is one salient problem that arises with Adam Smith’s theory, andit is inherited by the contemporary variants I have mentioned. The prob-lem is that we are given no reason to think that the demands of justifi-ability will have the stability that we naturally expect in any requirementsthat count as moral. Nothing that Adam Smith says provides a ground forexpecting stability across cultures, for example, in the standards to whichimpartial sympathy would hold us. Such an expectation would have tobe premised on an empirical claim that the psychological mechanismsidentified operate in the same way, and to at least a related effect, acrossdifferent contexts.

Scanlon is also exposed to a problem of stability. For what sorts ofconsiderations count as good reasons for rejecting a principle or makinga complaint? Scanlon emphasizes that they must be “personal” or agent-relative reasons that vary from individual to individual.33 For all that heshows, however, the sorts of personal reasons that count as good reasons

31 Michael Smith, The Moral Problem (Oxford: Blackwell, 1994).32 These are requirements of reason on two counts, and in that respect connect also with

the Kantian approach. First, they are requirements associated with the perfect, informedoperation of reason within the advising self. Second, they are requirements that the adviseeself, as a creature of reason, will want to satisfy. As between the self that does satisfy themand the self that doesn’t, there is no question about which is the more rational.

33 See Scanlon, What We Owe to Each Other; and Michael Ridge, “Saving Scanlon,” Journalof Political Philosophy 9 (2001): 472–81.

18 PHILIP PETTIT

for rejecting a principle or making a complaint may vary from cultureto culture.

Michael Smith’s proposal also faces a problem of stability, this timeacross individuals as well as across cultures. Why should people beexpected to converge on the requirements that their idealized selves wouldimpose, especially if they start from very different standpoints? Whyshouldn’t it be the case that my idealized self would recommend that Itake lines of action that your idealized self would reject? Why shouldn’tour idealized selves represent impartial spectators who work with differ-ent terms of appraisal?

In response to this problem, Michael Smith makes two points, onebearing on the possibility of convergence in what our idealized selvesrecommend, the other bearing on the necessity of such convergence. Thepossibility claim is that although our actual selves may have quite dif-ferent tastes and inclinations on certain fronts, our idealized selves mightstill agree that any person with the one set of tastes and inclinationsshould do such and such and any person with the other should do so andso. The necessity claim, which is much more important for the problemraised, is that anything that an idealized self would recommend for aperson of a particular mentality, in particular circumstances, has to besomething that any idealized self, no matter whose, would recommendfor that person; otherwise, it wouldn’t pass muster as a rationally com-pelling recommendation.

These claims, however, are far from conclusive. Someone with a par-ticular goal or allegiance —say, an allegiance to a particular religion —mayjust not be ready to regard that feature as a contingent circumstance ofpositioning such that there is a good, position-free question about how aperson, so positioned, should act. And, taking this view, such an agentwould naturally refuse to treat as idealized any counterpart who, notrecognizing the true faith or religion, lacked that allegiance. Asked todefend this line, the interlocutor might argue that an idealized counter-part who failed to recognize that particular faith or religion would justnot count as fully informed.

IV. A New Departure

In this final section, I sketch a sixth substantive theory that equates thecentral requirements of morality with the preconditions for the practice,surely more or less inescapable among human beings, of relating to oneanother in a co-reasoning, conversational mode. This approach breaksfrom those considered in Section III insofar as the obligations that itunifies on this basis are artificial in the sense in which Hume thought thatobligations of justice were artificial. The approach presupposes a dispo-sition on people’s part to generate a certain institution or practice and


identifies the relevant requirements by reference to that institution orpractice. The best way to present the approach in the brief compass avail-able is to reduce it to a number of propositions, without expanding inelaboration of these.34 My aim is not to give a full-dress defense of thetheory but rather to show that it is available, and how it fits in relation tothe other forms of substantive moral theory that I have surveyed.

A. Human beings are not just rational animals; they are reasoningcreatures as well

To be rational is to be a locus of evidence-responsive beliefs and desires,and to be disposed to act in the way that is required according to thosebeliefs, for the satisfaction of those desires. While various other animalsmay be rational without ever reasoning, we human beings are not amongthem: not only are we rational, or at least close to rational, we also reason.Reasoning involves intentionally asking oneself questions such that byanswering them, one may hope to guard against certain failures of ratio-nality in the evolution of one’s beliefs and desires; it consists in the sortof activity we naturally ascribe to Rodin’s figure of Le Penseur but neverto the dog or the cat. The questions typically considered bear on how fara certain body of evidence supports a proposition, whether two propo-sitions are consistent, and whether making a certain proposition true willbe a means of realizing another. By asking oneself questions of this kindabout propositions, and forming meta-propositional beliefs in response(beliefs about the relationships and properties of relevant propositions),one may hope to put extra checks on the ordinary process in whichintentional attitudes materialize and mutate. If one prompts oneselfto recognize that “p” and “q” are inconsistent, for example, that ought tocorrect or inhibit the belief that it is the case, or indeed the intentionto make it the case, that p and q.

B. Human beings co-reason with one another, as well asreasoning with themselves

Having access to language, human beings are able, uniquely amonganimals, to reason with one another as well as with themselves. They callon one another as advisers or informants or critics, and they try to guardagainst any irrationality in what they accept in common or attempt toachieve together. People who converse and co-reason in this way do for

34 For the central, motivating idea, see Philip Pettit and Michael Smith, “The Truth inDeontology,” in Reason and Value: Themes from the Moral Philosophy of Joseph Raz, ed. R. J.Wallace, Philip Pettit, Samuel Scheffler, and Michael Smith (Oxford: Oxford UniversityPress, 2004). And for a fuller account, see Philip Pettit, “Joining the Dots,” in Common Minds:Themes from the Philosophy of Philip Pettit, ed. Michael Smith, H. G. Brennan, R. E. Goodin,and F. C. Jackson (Oxford: Oxford University Press, 2007).

20 PHILIP PETTIT

one another what they try to do for themselves in personal reasoning.They lend one another their reason, as we might say, each serving in theforum of the other’s mind as an amicus curiae, a friend of the court. Theservice aspect of the activity comes out in the fact that although they mayinfluence one another significantly, neither will impose on the other. Theinformation or advice or guidance each provides is given on a take-it-or-leave-it basis; it is of the essence of co-reasoning that no one forces anoth-er’s hand.35

C. Co-reasoning requires the eschewal of force and coercionand related activities

Any social activity, whether it be checkers or chess or co-reasoning,puts constraints in place that all participants must honor, on pain of nolonger taking part in that activity; these are the constitutive rules orrequirements of the practice. As the counterpart of personal reasoning,co-reasoning rules out a variety of intuitively objectionable activities inthis manner. These are activities such that one cannot pursue them andclaim at the same time to be reasoning with others, as those persons mightreason with themselves. They include overtly forcing people to do some-thing by removing one or more options; overtly coercing people to dosomething by penalizing and thereby changing one or more of their options;and overtly purporting to have contracted or changed the options in thismanner. But they also include the covert activity of pretending to co-reasonwhile actually undermining the practice, whether through misinformingothers, or triggering some background manipulation of their ability tochoose, or relying on the intimidating effect of one’s power over them, orwatching over and riding herd on their performance —surveying how itgoes, while remaining disposed to resort to a corrective intervention inthe event that co-reasoning does not lead them in the desired direction.

D. These are presumptively moral requirements of respect

The relationship established in co-reasoning allows one person to influ-ence the choice of another but does not allow anyone to undermineanother’s power of choice; in that sense, its requirements enforce a regimeof mutual respect. In deliberative decision-making, I have to assume withrespect to each of the options among which I am deciding that it is within

35 Co-reasoning need not involve any degree of rigor or formality. It may often be whollyimplicit, as when you draw my attention to something and let me fill in the missing lesson.And it may be quite rhetorical in form. It may involve a story or a parable, in which youinvite me to see things another way; it may introduce ironic or sarcastic or mocking com-ment; and it may employ metaphor and image, and all the colors of persuasive overture.Such rhetoric may be needed in order to knock me out of my complacency, let me see howstupid my point of view is, and make your standpoint seem truly habitable.


my power; in choosing among three options, x, y, and z, I must be able tothink with respect to each option: “I can do that.” 36 If such can-do assump-tions were not true, or if I did not think they were true, there would be nopoint in deliberation. Co-reasoning is an inherently respectful form ofinteraction insofar as it leaves both the truth and the thinkability of can-doassumptions in place. It does not undermine the truth of any such assump-tion by reducing my capacity to choose, contracting my range of options,or changing an option in the choice-set, say by the imposition of a pen-alty.37 And it does not force me to reject any true assumption, misleadingme about how things stand. Co-reasoning may make me aware of anextra option, as when you inform me that someone (perhaps you your-self ) would be willing to offer a refusable reward for the choice of aparticular option, x. But this does not directly affect the truth or think-ability of the can-do assumption with respect to option x or y or z; itmerely adds the extra option, x+, of doing x and taking the reward.38

E. Co-reasoners purport to satisfy the requirements of respect

One can aspire to reason with others only on the basis of a sharedunderstanding of the practice and an acceptance of the demands it imposes.Thus, to present oneself as co-reasoning is to purport to endorse thosedemands as constraints on how one is going to behave toward those withwhom one means to reason. It is to renounce any resort to force or coer-cion and any reliance on resources of duplicity or intimidation or correc-tion, and to present oneself as embracing a relationship of respect forone’s partners. Those who purport to co-reason with others but then failin some way to practice respect are hoist by their own petard. They canbe indicted with failing to live up to the expectations that their presen-tation invited others to form; they can be condemned by criteria they areon record as endorsing.

F. Would-be co-reasoners also purport to satisfy the requirements of respect

Might a person subscribe to the requirements only when embarking onan episode of co-reasoning and not more generally? Not plausibly. I can

36 We may sometimes describe an option in a way that reaches out to a desired conse-quence, as when we think of it as hitting the target rather than firing the gun, but eachoption has to be accessible also as something we can just do.

37 For relevant arguments, see John Broome, Weighing Goods (Oxford: Blackwell, 1991);and Philip Pettit, “Decision Theory and Folk Psychology,” in Essays in the Foundations ofDecision Theory, ed. Michael Bacharach and Susan Hurley (Oxford: Blackwell, 1991); reprintedin Philip Pettit, Rules, Reasons, and Norms (Oxford: Oxford University Press, 2002).

38 See Pettit and Smith, “The Truth in Deontology.” The making of offers may not havethis cast outside what I later describe as the circumstances of respect, where everyone issufficiently well-off to be able to function properly in the local society. Extreme conditionsmight make it rational for someone to accept the offer of a slave contract, but that offer couldhardly be said to be respectful. See Philip Pettit, “Freedom in the Market,” Politics, Philos-ophy, and Economics 5 (2006): 131–49.

22 PHILIP PETTIT

hardly fail to honor these requirements in dealing with others and expectto have access to co-reasoning on an opportunistic basis. What assurancecould others have of my willingness to respect them in the course ofco-reasoning, if I present myself as someone ready to use force or coercionagainst them? What assurance could they have, indeed, if I present myselfas ready in this way to use force and coercion against others who occupya comparable position? In order to make it possible to resort to co-reasoningas opportunity requires, or inclination prompts —in order to have accessto the practice —I must show myself ready to respect those with whom Imight wish to co-reason, and any others who are likely to count as com-parators. I must purport to relate to my community as someone who isrespectful of others, disavowing any attempt to influence them other thanby appeal to co-reasoning or to overtures (say, of camaraderie and ban-tering) that are congruent with co-reasoning.39

G. The requirements of respect have a default status in social life

It is possible for someone to opt out of a regime of mutual respect,manifestly displaying an attitude of aggression, or an assumption of occu-pying superior status. However, there is no need to opt in, at least not incircumstances where resources are sufficiently rich and well distributedto enable everyone to function properly as a member of the society —thatis, not in what we might call the circumstances of respect.40 The de-fault assumption, absent hostility, will be that one is accessible to co-reasoning —at least to the extent that time and other constraints allow forsuch interaction —and that one eschews any resort to force or coercion ordeception or the like in dealing with others. If natural life is lived in theprecincts of battle, as Hobbes maintained, then civil life is lived in theprecincts of conversation and co-reasoning. By not displaying hostility,one acquiesces in the manifest fact that others assume one is accessible toco-reason and ready to honor the associated demands of respect. One isnaturally taken to endorse the requirements of respect as expectations towhich one may be held, criteria by which one may be judged.

H. The requirements of respect are defeasible requirements: excuses

That the requirements of respect have this default status, however, doesnot mean that they are nondefeasible, and that a breach is never excused.

39 The lesson is akin to that of the master and slave. What sensible slave would be honestin speaking his or her mind with a master? What sensible master would expect the slave tobe honest? What sensible slave, indeed, would expect to be expected to be honest? Noticethat if a class of masters is saliently distinguished, as under an apartheid regime, thenmasters may disrespect those in the slave class without disturbing their own reputations aspersons who respect other masters; slaves will not count as comparators of the masters. Thatis why it is important, not just to focus on the instantiation of respect, but also on thepromotion, as I describe it later, of respect-enjoyment.

40 On such functioning requirements, see Amartya Sen, Commodities and Capabilities (Amster-dam: North-Holland, 1985).


So long as the practice of co-reasoning is in place, one will be strictlyobliged by its rules.41 But often the practice will not clearly be in place,and then a breach of the rules will be excusable. No one can be bound todisplay respect for those who would do him harm, for example, sincethose others thereby reject the relevance of the practice. Nor, by a plau-sible extension, can anyone be bound to display respect for those whowould do harm to others who are party to the practice; no one can bebound to tell the truth about a victim’s whereabouts to a would-be mur-derer. Nor can anyone be bound to display respect under such diredistress —such “moral catastrophe” 42 —that the very practice of respect isin jeopardy; the circumstances of respect, as I called them, would not thenobtain. Disrespect is excused in cases like these, because the rules ofco-reasoning do not clearly apply.

I. The requirements of respect are defeasible requirements: justifications

But are there grounds on which a breach of the requirements of respectin these or in other cases might be positively justified, not just excused?Is there any value, accessible to all, whose promotion might argue thatwhereas satisfying those requirements is generally for the best, there arecases where it is not and where a breach may be justifiable? Arguably, thevery practice of respect points us to something that is valuable in thisway: namely, the status that people enjoy insofar as they command therespect of others —if you like, “respect-enjoyment.” The demands of sucha value might argue for the justification of disrespect in perverse caseswhere, by showing disrespect to some innocent persons, the extent towhich people enjoy respect overall is increased —at the limit, they may besaved from death or destitution. The recognition of such a value or targetmakes room for a consequentialist perspective on respect. One mightargue that while it is generally for the best to practice respect, there maybe cases where the promotion of respect-enjoyment (or, indeed, someother such value) justifies a temporary suspension of the practice, even indealing with an innocent party.43 The practice of friendship is not jeop-

41 Thus, one will not be able to argue that one is co-reasoning with another, on thegrounds that while one is violating the requirements of respect in some way, that is becausethis promises to maximize the respect of others for others. One might as well argue that oneis playing chess, on the grounds that while one is violating the rules of chess, that is becausesuch violation promises to minimize violations overall. The practice of co-reasoning requiresone to instantiate respect toward those with whom one co-reasons, not to promote suchrespect overall. This may not be for the best, but it will be required for being able to claimto co-reason.

42 See Robert Nozick, Anarchy, State, and Utopia (Oxford: Blackwell, 1974), 30n.43 Consequentialism might prescribe the honoring of the requirements of respect under

the proviso that this is for the best overall. Or, more plausibly, it might prescribe it under astricter proviso that imposes constraints on how far agents should check on what is for thebest (because such checking might itself be bad for respect-satisfaction). One such provisowould prescribe the honoring of the requirements except when there is independently

24 PHILIP PETTIT

ardized by the fact, acknowledged on all sides, that the demands offriendship are only compelling provided that they do not put innocentlife at risk, or create some similar horror. Neither should the practice ofrespect be jeopardized by the recognition of a parallel, consequence-basedproviso.

J. From ethics to politics

Not only is the emerging point of view consistent with a consequen-tialist perspective, but it is consistent with a consequentialism that wouldargue for the formation of a state —specifically, a state that would ensurethat the circumstances of respect are realized for all and that there aresufficient guards against the strong riding herd on the weak that respectis widely commanded and enjoyed. Thus, the cause of promoting respect-enjoyment overall may argue for a collective, political arrangement underwhich the weak are protected and empowered and a regime of mutualrespect is shaped so as to be for the best overall.44

If the argument encoded in the foregoing ten points is on the righttrack, then many moral obligations are really obligations under the codeof co-reasoning, the ethic of respect. This ethic is grounded in the fact that,uniquely among animals, we human beings are creatures of the word andcan relate to one another in the co-reasoning manner that speech makespossible. The constraints of the code are those associated with being, aswe say, on speaking terms with one another. They describe the onlypossible basis for civil relationships in which conversation is the norm —not coercion, and not command. And they make salient the inclusive idealof extending the enjoyment of respect and status to all. They do not coverthe treatment we should offer to animals, of course, nor any of a range of

salient reason to think that this is not for the best overall; conformity to the requirementswould become a default option under this proviso, but not an unconditionally compulsoryone. See Philip Pettit, “A Consequentialist Perspective on Ethics,” in Marcia Baron, MichaelSlote, and Philip Pettit, Three Methods of Ethics: A Debate (Oxford: Blackwell, 1997). One wayof seeing T. M. Scanlon’s contractualism is as a theory that endorses, broadly, the require-ments of respect in dealings between the respectful, and that argues for a distinctive,nonconsequentialist form of modulation for other cases: the right way to behave in thosesecond-best cases is in accord with the principles for such cases that we might expect toprove reasonably unobjectionable among people respectfully debating with one another. SeePhilip Pettit, “Can Contract Theory Ground Morality?” in James Dreier, ed., ContemporaryDebates in Moral Theory (Oxford: Blackwell, 2005).

44 I take it that the sort of regime envisaged would maximize freedom in the republicansense of nondomination. See Philip Pettit, Republicanism: A Theory of Freedom and Government(Oxford: Oxford University Press, 1997). Another approach that might be taken to politics,starting from the practice of respect, is to ask in contractarian style about the sort of globalregime that people might be led to adopt as a result of reasoning with one another in asituation of mutual respect. Yet another approach would be to ask about the global sort ofregime that would emerge as a result of the different local arrangements that we mightexpect people to make with one another in co-reasoning contracts. The first approach is inthe spirit of Rawls, the second in the spirit of Nozick.


duties to ourselves, but they account for a good part of what we naturallythink of as the moral realm.

How does this substantive theory of morality —this theory of a centralsubset of moral obligations —relate to the theories already canvassed? Themost striking point of contrast is that, in Hume’s terms, it is a theory ofartificial obligation and virtue. It presupposes an artificial, cultural prac-tice, one of conversation and co-reasoning, and presents the relevantobligations as preconditions for the survival and prosperity of that prac-tice. The obligations are not presented as preinstitutional, inescapablelaws of nature (or self-interest or benevolence or reason or justifiability),but as requirements of an institutional practice in which co-reasoning andrespect are at the center. They are laws, in a word, of civility.

Their institutional character means that the laws of respect only havehypothetical or conditional status. They identify what is required of peo-ple if they are to have access to the practice of co-reasoning and to all theassociated benefits. But this conditionalization is relatively innocuous,since it involves only a very small “if.” The practice of co-reasoning is atthe very center of human life and has an inescapable hold on us, at leastin the normal case where it serves to promote the value associated withrespect-enjoyment. If the laws of respect are requirements for access tothat practice, as argued here, then they are conditional on a goal that isclose to unrejectable. Unlike most hypothetical imperatives, they are tiedto an aspiration that human beings can scarcely disavow, and thus theyhave the assured status that we would expect moral requirements —asdistinct, say, from requirements of etiquette —to display.45

Substantive moral theories, as I have emphasized, seek out unifyingpatterns among the requirements that pass as moral demands, by the testsassociated with our schematic theory of rightness. While a number oftheories might each have something to be said in their favor, especially ifthey focus on different subsets of demands, they are clearly in competi-tion. So how, finally, does the theory sketched here compare with the fivetheories we looked at earlier?

The laws of respect identify conditions for human flourishing, andconnect on that front with the laws of nature. But they identify conditionsthat it could scarcely be in anyone’s interest to violate, at least not in anygeneral way, and thus they also make contact with the tradition that linksmorality with egoistic interests. How do the laws of respect relate to theremaining three theories we surveyed? These laws point us to require-ments of benevolence that emphasize the evil of disrespect and the neu-tral ideal of prompting respect-enjoyment overall. They constitute idealsof reason in the sense of prescribing a mode of treatment that is salientlyuniversalizable. And they connect with the idea of justifying oneself to

45 See Pettit and Smith, “The Truth in Deontology,” and Mark Schroeder, “The Hypo-thetical Imperative?” Australasian Journal of Philosophy 83 (2005): 357–72.

26 PHILIP PETTIT

others insofar as the regime they put in place generally allows each indi-vidual to exercise only the sort of influence that others should welcome;this would facilitate individuals’ rationality without restricting their powerof choice.

These observations suggest that the theory sketched here is on the righttrack, picking up a pattern that would explain why the other approacheshave each seemed plausible to so many. If such a substantive theory issound, then it means that there are two different standpoints from whichwe can survey the terrain of moral obligation. From the standpoint of theschematic theory described earlier (in Section II), moral obligations presentthemselves as requirements to which we are pointed by the battery ofconnections and paradigms that establish the very reference of terms like“right” and “wrong.” From the standpoint of the substantive theory devel-oped in this section, a core set of obligations stand out as requirementsthat derive from the code of co-reasoning, the ethic of respect, on whichthe very possibility of civil life is premised. Morality, under this picture,has a dual aspect and can be investigated in philosophy on either of itstwo fronts.

Philosophy, Princeton University


VIRTUE AND NATURE*

By Christopher W. Gowans

I. Introduction

That human beings are rational animals is such a platitude in so much ofWestern moral philosophy that it may seem unlikely that additional insightis to be gained from further reflection on the meaning and implications ofthis phrase. Traditionally, by and large, it has been supposed that ratio-nality is our most important feature. In recent years, however, it has beensuggested by some advocates of virtue ethics rooted in Aristotle that arenewed emphasis on, and understanding of, the fact that we are animals,as well as rational, offers a promising avenue for defending an objectivejustification of morality. The best-known proponents of this approach arePhilippa Foot and Rosalind Hursthouse.1 Since they each depict theirposition as a form of ethical naturalism, their shared outlook may becalled Neo-Aristotelian Ethical Naturalism (hereafter NAEN). Many ofthe facts about our animal nature highlighted by Foot and Hursthouse arecertainly relevant to ethical deliberation, and to this extent their newoutlook is a welcome contribution. Nonetheless, my thesis in this essay isthat NAEN is inadequate because, by its own standards, it does notprovide a naturalistic justification of its ethical commitments.

In what sense does NAEN purport to be a form of ethical naturalism?In moral philosophy, naturalism ordinarily is taken to preclude any appealto the supernatural (for example, to God), and Foot and Hursthouse areethical naturalists in this respect. Beyond this, ethical naturalism usuallyimplies that (a) there is some significant connection between moral valuesand natural facts, where (b) the natural facts include only facts counte-nanced by contemporary science (including, in particular, psychologyand evolutionary biology). Proponents of NAEN certainly affirm (a). At aminimum, they believe that the justification of virtues is, in some impor-

* I would like to thank my colleague John Davenport as well as the other contributors tothis volume, and its editors, for helpful comments on an earlier draft of this essay. I wouldalso like to express my appreciation to Fred D. Miller, Jr., and the Social Philosophy andPolicy Center for inviting me to contribute to this volume.

1 The main sources are Philippa Foot, Natural Goodness (Oxford: Clarendon Press, 2001),and Rosalind Hursthouse, On Virtue Ethics (New York: Oxford University Press, 1999), esp.part III. Foot and Hursthouse both appeal to a crucial idea in Michael Thompson, “TheRepresentation of Life,” in Rosalind Hursthouse, Gavin Lawrence, and Warren Quinn, eds.,Virtues and Reasons (Oxford: Clarendon Press, 1995), 247–96. For a similar position, seeAlasdair MacIntyre, Dependent Rational Animals: Why Human Beings Need the Virtues (Chi-cago: Open Court, 1999), esp. 78.

DOI: 10.1017/S026505250808002328 © 2008 Social Philosophy & Policy Foundation. Printed in the USA.

tant sense, dependent on facts about human nature and the circumstancesof human life. For example, Foot says that “the grounding of a moralargument is ultimately in facts about human life.” 2 Sometimes it seems tobe suggested, more strongly, that some moral facts are natural facts. Accord-ing to Foot, “the fact that a human action or disposition is good of its kindwill be taken to be simply a fact about a given feature of a certain kind ofliving being.” 3 With respect to (b), matters are less clear. Foot andHursthouse make no appeal to evolutionary biology. Instead, what playsa crucial role in their arguments is a set of statements about plants andanimals —and about human beings as a kind of animal —called “natural-history sentences” or “Aristotelian categoricals,” which are said to be truedescriptions of objective facts in the natural world.4 This distinguishesNAEN from many other forms of ethical naturalism, and it raises a ques-tion about what kind of natural facts Foot and Hursthouse have in mind(about which I will have more to say below).

However, my critique of NAEN is not that it is inadequately scientific.Rather, my main argument is that it cannot account for the concerns ofmoral universalism, the view that each human being has moral worthand thus deserves significant moral consideration.5 After explaining themain contours of NAEN (in Section II), I explore ways in which it mightdeal with moral universalism, and I argue that each of these ways isinadequate (in Section III). I then broaden the discussion (in Section IV)and maintain that the ends concerning other persons proposed by NAENseriously underdetermine the virtues that are said to promote these ends.My conclusion (in Section V) is that those attracted to an Aristotelianvirtue ethics would be wise to abandon its naturalism, at least beyond aminimal and fairly uncontroversial appeal to some facts about humannature and circumstances.

II. The Basic Argument of Neo-AristotelianEthical Naturalism

NAEN was first formulated by Foot and later developed by Hursthouse.Their positions are close, but not identical.6 I will proceed by explaining

2 Foot, Natural Goodness, 24.3 Ibid., 5.4 See ibid., 29.5 Foot and Hursthouse both put forward their accounts in a rather tentative spirit, and

Hursthouse says that justice is a gap in her theory. Nonetheless, I believe NAEN does nothave the resources to deal with the issues I raise.

6 Hursthouse’s understanding of Foot is based, not on Natural Goodness (which had notyet been published), but primarily on two earlier essays, “Rationality and Virtue” (1994),and “Does Moral Subjectivism Rest on a Mistake?” (1995), both reprinted in Philippa Foot,Moral Dilemmas and Other Topics in Moral Philosophy (Oxford: Clarendon Press, 2002), 159–74,189–208. Hursthouse’s own position is also expressed in a more recent paper, “On theGrounding of the Virtues in Human Nature,” in Jan Szaif, ed., Was ist das für den MenschenGute? Menschliche Natur und Güterlehre (Berlin: De Gruyter, 2004), 263–75.

VIRTUE AND NATURE 29

and commenting on Foot’s main line of argument, noting Hursthouse’samendments along the way. In the end, however, my critique will con-centrate on Hursthouse’s more elaborate position. Their argument dividesinto three phases: a set of claims about the evaluation of living things,especially animals; application of this framework to the evaluation ofhuman beings; and discussion of the difference human rationality makesto this application.

Phase one. Foot begins by stating that ‘good’ and ‘bad’ are “attributiveadjectives” whose criteria of application depend on the noun (or nounexpression) they modify.7 This appears true in a wide variety of contextsin which ‘good/bad X’ is meaningful on account of some human activity,interest, or concern. However, according to Foot, there are some thingsthat may be evaluated as good or bad, not by reference to any humanperspective, but simply in virtue of the kinds of things they are. In par-ticular, living things such as plants and animals may be evaluated on thebasis of standards that are implied by the nature of their species.

Foot’s explanation of this is as follows. The life cycle of a member of aspecies includes development, self-maintenance, and reproduction. Tele-ological statements (the “natural-history” statements or “Aristotelian cat-egoricals”) explain the function of something —a part, characteristic, oroperation —in this life cycle. For example:

(F) “[T]he male peacock displays its brilliant tail in order to attract afemale during the mating season.” 8

Statements such as F are said to be neither universal nor statistical gen-eralizations. Rather, F explains an operation by reference to its function inreproduction. It asserts that the purpose of a male peacock’s raising its tailis to attract a female. This is not taken to mean that the peacock has thispurpose nor that it was consciously designed for it.

Teleological statements such as F are said to be factual: they are deter-mined by the nature and life cycle of the species (including its needs,capacities, and natural habitat). Hence, their truth-value does not dependon the needs or wants of other species, including human beings. More-over, these statements are necessary for properly describing and under-standing the natural history of a species.

According to Foot, these teleological statements imply evaluative ornormative statements about individual members of the species. For exam-ple, F entails:

(E1) An individual male peacock needs to or should be able to displayits tail during the mating season.

7 She takes this point from P. T. Geach, “Good and Evil,” Analysis 17 (1956): 33–42.8 Foot, Natural Goodness, 31.

30 CHRISTOPHER W. GOWANS

And F also entails:

(E2) An individual male peacock that is able (unable) to display itstail during the mating season is good (defective) in this respect.

Hence, the teleological statements provide a basis for judging naturalgoodness and defectiveness in living things. An individual’s having theforms of goodness appropriate to its species contributes to its living agood life for a member of its species, a life of proper development, self-maintenance, and reproduction (though whether it succeeds in doing thisalso depends on other factors). In conclusion, Foot says at this stage of herargument, “the norms that we have been talking about so far have beenexplained in terms of facts about things belonging to the natural world.” 9

There are, she says, “patterns of natural normativity.” 10

Hursthouse accepts the essentials of Foot’s position in these respects,but she develops it with respect to higher social animals such as wolvesand horses. Hursthouse summarizes her view as follows:

Teleological Framework. A good social animal (of one of the more sophis-ticated species) is one that is well fitted or endowed with respect to(i) its parts, (ii) its operations, (iii) its actions, and (iv) its desires andemotions; whether it is thus well fitted or endowed is determined bywhether these four aspects well serve (1) its individual survival, (2)the continuance of its species, (3) its characteristic freedom from painand characteristic enjoyment, and (4) the good functioning of itssocial group —in the ways characteristic of the species.11

The Teleological Framework is said to provide a basis for objective eval-uations of certain animals. For example, similar to E2, these animals “aredefective —something is wrong with them —when they do not want to eator reproduce” (“the continuance of its species” refers specifically toreproduction).12

Much of my argument below focuses on item (4) in the TeleologicalFramework as Hursthouse applies it to human beings. She has little to sayabout what constitutes a social group. The phrase is used to make adistinction between animals that live rather solitary lives, such as tigers,and those that live more social lives. The latter belong to a social group(such as a wolf pack). In the case of animals, these groups vary enor-mously in character across different species, and among human beingsthere are many different kinds of social groups. Hursthouse does not

9 Ibid., 36–37.10 Ibid., 38.11 Hursthouse, On Virtue Ethics, 202. I have added the title.12 Ibid., 200.


discuss these differences. Presumably, she supposes that social groups arefairly small, in comparison with the species as a whole, and involve somenotion of membership such that members of the group act in variousways for the well-being of other members of the group (and perhaps for“the group as a whole”).13

An obvious worry about Foot’s and Hursthouse’s position arises fromquestions about its status vis-à-vis evolutionary biology. Foot speaks explic-itly of Aristotelian categoricals and necessities, and she states that theterm ‘function’ in her account is used in an “everyday” sense rather thanin the “technical” sense of evolutionary biology.14 Moreover, Hursthousesays that “the non-ethical evaluations of living things that I have outlinedare ‘Aristotelian’ rather than Darwinian.” 15 These comments may invitethe complaint that Foot and Hursthouse are relying on an Aristotelianapproach that has been refuted by Darwinian biology.16 However, theyclearly believe that teleological statements such as F, as well as evalua-tions such as E1 and E2, are part of the natural histories of animals asunderstood by contemporary science. Hursthouse says that these evalu-ations of living things are “scientific” and are employed in botany, zool-ogy and ethology.17 In fact, natural histories do contain many statementsof this kind (albeit expressed in a variety of ways). Moreover, Foot andHursthouse need not, and do not, deny that F states a scientific fact thatis explained by natural selection.18 Hence, they are best interpreted asintending to present an account that is compatible with evolutionarybiology, and as hoping to avoid as much as possible engaging in debatesabout the proper role and understanding of concepts such as teleology,function, purpose, and design in biology.19

It is doubtful that a full defense of their position can avoid this engage-ment. Their claim is that statements such as F, E1, and E2 are intelligibleand essential to properly understanding living organisms. A critic, at thisstage of the argument, would need to show that statements of this kindcould be eliminated without loss of descriptive or explanatory power inbiology and related sciences. Foot and Hursthouse appear committed tosupposing that these statements could not be eliminated by this standard.There is a sense, then, in which they think a form of evaluation is essentialto understanding plants and animals, and I am willing to accept this

13 See ibid., 201–2.14 Foot, Natural Goodness, 32 n. 10. See also ibid., 40 n. 1.15 Hursthouse, On Virtue Ethics, 257.16 For example, see Philip Kitcher, “Biology and Ethics,” in David Copp, ed., The Oxford

Handbook of Ethical Theory (New York: Oxford University Press, 2006), 164–65.17 See Hursthouse, On Virtue Ethics, 202, 229.18 See Foot, Natural Goodness, 92.19 For examples of these debates, see Colin Allen, Marc Bekoff, and George Lauder, eds.,

Nature’s Purposes: Analyses of Function and Design in Biology (Cambridge, MA: The MIT Press,1998). Hursthouse enters these discussions a bit more than Foot; she expresses doubts aboutwhether Darwinian standards could replace Aristotelian ones (see Hursthouse, On VirtueEthics, 258).


claim here for the sake of argument. It is important to note, however, thatthe evaluations E1 and E2 only identify species-dependent forms of good-ness and defectiveness. There is no suggestion, for example, that peacockreproduction is good simply speaking or from the standpoint of nature assuch (whatever that might mean).20 Reproduction is only a characteristicgood of peacocks in virtue of the nature of the species. Foot and Hursthousecan agree with a critic who claims that, though reproduction plays acentral explanatory role in evolution, there is no overall (not simply species-dependent) sense in which it is good —say, from the standpoint of natureitself, or of evolutionary theory —that an individual peacock (or indeedany peacock) reproduces. Of course, for zookeepers, animal breeders,farmers, gardeners, and the like, the reproduction of living things is some-times a good thing (and sometimes a bad thing). But these judgmentsdepend on their specific interests; they are not dictated by nature itself.However, given their interests, Foot and Hursthouse maintain, evalua-tions based on natural facts such as F take on importance.

Phase two. The next step in Foot and Hursthouse’s argument asserts thatevaluations of human actions and dispositions have the same conceptualstructure as the aforementioned evaluations of plants and animals. Themeaning of ‘good’ and related terms is the same in both cases. Evalua-tions of human beings, Foot says, including moral evaluations, “can onlybe understood in these terms.” Hence, she regards “moral evil as ‘a kindof natural defect’.” 21 She acknowledges that there is a major difference:human beings, unlike plants and other animals, have a rational will andperform voluntary actions. However, she believes it is essential to startwith the purported similarity. Though human good is “different fromgood in the world of plants or animals,” she says, “there is a ‘natural-history story’ about how human beings achieve this good as there isabout how plants and animals achieve theirs.” 22

The first part of this story is that, despite cultural diversity, there is a“quite general account of human necessities, that is, of what is quitegenerally needed for human good.” 23 For example, these necessities includefood, housing, and clothing as well as relationships of love and friendship.

The second part of the story is that we need virtues such as industri-ousness, loyalty, and kindness to attain these necessities. Hence, Footsays, following P. T. Geach, “virtues play a necessary part in the life ofhuman beings as do stings in the life of bees.” 24 There is thus a commonstructure of evaluation of all living things: teleological (“in order to”)statements provide a factual basis for evaluations about what is good,needed, or should be the case. Hence, Foot appears committed to claim-

20 See Foot, Natural Goodness, 50.21 Ibid., 5.22 Ibid., 51.23 Ibid., 43.24 Ibid., 35. Cf. ibid., 44.


ing that, just as F is true and implies E1 and E2, so (in an exampleemphasized by Foot):

(F*) A human being is trustworthy in order to (among other things)promote the exchange of goods needed for material well-being

is true and implies:

(E1*) A human being needs to or should be trustworthy

and:

(E2*) A human being who is trustworthy (untrustworthy) is good(defective) in this respect.

F* explains a characteristic of human beings by reference to the charac-teristic’s function in self-maintenance: among human beings, given ournature and circumstances, self-maintenance typically requires the exchangeof goods, and this in turn requires trustworthiness. As with F, F* is sup-posed to be a particular kind of factual statement —a teleological state-ment. It is neither a universal nor a statistical generalization. It depictsone form of being virtuous and says that the function of being virtuous inthis respect is the fulfillment of certain human needs. However, thoughfactual, F* is said to imply evaluations such as E1* and E2*. This is theconceptual parallel Foot thinks is essential to establish.

Once again, Hursthouse accepts Foot’s basic argument and elaborates:since human beings are sophisticated social animals, the aforementionedTeleological Framework provides a basis for ethical evaluations of us. Forexample, Hursthouse says:

Human beings who are good in so far as they are courageous defendthemselves, and their young, and each other, and risk life and limb todefend and preserve worthwhile things in and about their group,thereby fostering their individual survival, the continuance of thespecies, their own and others’ enjoyment of various good things, andthe good functioning of the social group.25

Hursthouse thinks similar accounts can be given of other virtues such ascharity, honesty, generosity, loyalty, justice, and trustworthiness.

Phase three. The final step in the argument for NAEN concerns what isdistinctive about human beings. Foot does not believe that the conceptualparallel stated above means that human good is the same as the good of

25 Hursthouse, On Virtue Ethics, 209.


plants and other animals. “Human good,” she says, “is sui generis.” 26 Themain difference is that human beings have a rational will: we act volun-tarily for reasons. This gives rise to greater diversity and complexity.More important, this means that we have a capacity to understand andthat, for any proposed action X, we can always ask, “Why should I do X?”An answer to this question purports to give us a reason for action basedon what is judged to be good. According to Foot, “Kant was perfectlyright in saying that moral goodness was goodness of the [rational] will.”However, Kant was mistaken, she says, in not recognizing that “the eval-uation of human action depends on [in addition to abstract practicalreason] essential features of specifically human life.” 27

Since human beings can understand the teleological facts and theirimplications, we can answer the question “Why should I do what a trust-worthy person would do here and now?” by appealing to, among otherthings, the truth of F* and the fact that F* implies E1* and E2*. Morespecifically, we can understand the facts about human necessities and ourcircumstances that warrant us in saying that “in giving a promise onemakes use of a special kind of tool invented by humans for the betterconduct of their lives, creating an obligation that (although not absolute)contains in its nature an obligation that harmlessness does not annul.” 28

Rationality does not change the fact that human goodness is a form ofnatural goodness.

For Hursthouse, the fact that human beings, unlike other animals, arerational, does not modify the Teleological Framework cited earlier by wayof addition or subtraction (there are the same four aspects and four ends),but it does transform the framework in ways similar to those envisionedby Foot. Rationality gives rise to greater diversity as well as to the capac-ity to understand the rationale of the virtues and the possibility of ratio-nally revising the virtues we currently accept. Moreover, rationality enablesus to choose whether to act virtuously or not.

Hursthouse says that “our characteristic way of going on . . . is a ratio-nal way” —that is, “any way we can rightly see as good, as something wehave reason to do.” 29 Nonetheless, the nature of our species imposessignificant restrictions on what we have reason to do. “Human beings areethically good,” she says, “in so far as their ethically relevant aspectsfoster the four ends appropriate to a social animal, in the way character-istic of the species.” The Teleological Framework, she adds, “really doesconstrain, substantially, what I can reasonably maintain is a virtue inhuman beings.” Hence, in determining whether a character trait is avirtue, we have to consider whether it “would foster or be inimical to

26 Foot, Natural Goodness, 51.27 Ibid., 14.28 Ibid., 51. I take this to mean that, for Foot, the fact that breaking a promise brings about

no harm does not by itself annul the obligation to keep the promise.29 Hursthouse, On Virtue Ethics, 222. Cf. ibid., 228.


those four ends.” 30 In ethical evaluation, Hursthouse claims, “we evalu-ate ourselves as a natural kind, a species which is part of the naturalbiological order of things.” 31 These comments suggest that she acceptsthe following criterion:

A character trait C is a virtue only if (a) C promotes in a substantialway at least one of the four ends, and (b) C does not significantlyinhibit the four ends.

With respect to (a), it seems too strong to say that C is necessary for anend or that C relates to all four ends. A virtue may be very important evenif not necessary for advancing an end, and different virtues may promotedifferent ends. What Hursthouse requires is that C furthers at least one ofthe ends to some substantial extent. With respect to (b), it seems toostrong to say that any inhibition of the ends would prevent a charactertrait from being a virtue. Otherwise, courage could not be a virtue, sinceit can require us to risk our life. But some reference to the negative impacton ends is surely needed. Otherwise, a character trait that promoted oneend, but virtually precluded all the others, would be a virtue. For exam-ple, extreme self-centeredness might promote self-preservation but seri-ously damage other ends. The expressions ‘promotes in a substantialway’ and ‘does not significantly inhibit’ are obviously vague and in needof interpretation, but they are sufficiently meaningful for Hursthouse’spurpose and my discussion. The criterion says that (a) and (b) are nec-essary conditions for a character trait’s being a virtue. It does not say, andI am not supposing that Hursthouse thinks, that these conditions aresufficient.

Beyond this, it seems clear that Hursthouse supposes that the Teleo-logical Framework makes a contribution to the correct conception of thevirtues. This is evident in her comment, quoted above, that people aregood “in so far as their ethically relevant aspects foster the four ends,” aswell as in her discussion of Peter Singer (considered below). For example,suppose two competing conceptions of courage both met conditions (a)and (b), but the first promoted the ends in a more substantial way thanthe second, or inhibited them to a lesser extent. Surely, in this circum-stance, the first conception has a better claim to be a virtue than thesecond, all other things being equal, because the first better fulfills con-ditions (a) and (b) than the second. Unless there were some special con-sideration not evident here (for instance, that once an end is fulfilled to acertain point, it does not matter if it is fulfilled to a greater extent), it

30 Ibid., 224. The four ends are individual survival, continuation of the species, charac-teristic freedom from pain and characteristic enjoyment, and the good functioning of thesocial group.

31 Ibid., 226.


would be odd to accept the criterion, but reject this consequence. Hence,it is reasonable to suppose that Hursthouse is committed to the following:

Teleological Criterion. A character trait C is a virtue only if (a) C pro-motes in a substantial way at least one of the four ends, and (b) Cdoes not significantly inhibit the four ends; and if two character traits,C1 and C2, in competition for the status of virtue, both meet con-ditions (a) and (b), but C1 promotes the ends in a more substantialway than C2, or inhibits them to a lesser extent, then, all otherthings being equal, there is more reason to regard C1 as a virtuethan C2.

The intuitive idea of Hursthouse’s version of NAEN, then, is that thosecharacter traits that best promote and least inhibit the four ends arevirtues.

The Teleological Criterion provides a criterion for determining whetheror not a character trait is a virtue. But which character traits fulfillthis criterion? Hursthouse is also, at least tentatively, committed to thefollowing:

Justification Thesis. The standard list of virtues meets the TeleologicalCriterion.

According to Hursthouse, the “standard list” of virtues includes many ofAristotle’s virtues (including courage, temperance, and justice), plus somenon-Aristotelian virtues such as charity and benevolence. These are thevirtues Hursthouse is prepared to defend. The Justification Thesis saysthat, for each virtue on the list, the virtue promotes in a substantial wayat least one of the four ends and does not significantly inhibit the fourends (and presumably meets these conditions better than any competingconception of the virtue).

The details of Hursthouse’s Teleological Framework, with its specifi-cation of four aspects and four ends, and the employment of this frame-work in the Teleological Criterion as well as the Justification Thesis, aremore substantial than Foot’s presentation of NAEN. For this reason, I willfocus primarily on Hursthouse’s position in the critique that follows.There is another important difference between Hursthouse and Foot. Atthe outset of her discussion of naturalism, Hursthouse says her aim is toshow the feasibility of establishing the objectivity or rationality of neo-Aristotelian virtue ethics in accordance with John McDowell’s “Neurathianprocedure.” 32 According to this procedure, “validation must take place

32 Hursthouse, On Virtue Ethics, 165. She cites John McDowell, “Two Sorts of Naturalism,”in Hursthouse et al., eds., Virtues and Reasons, 149–79. McDowell’s view is part of a broaderphilosophical project expressed in several places in his work. It is unclear to what extentHursthouse endorses this project, and I interpret her here on the basis of what she (ratherthan McDowell) says.


from within an acquired ethical outlook, not from some external ‘neutral’point of view.” 33 The acquired outlook is that of the virtuous agent (asunderstood by Hursthouse). The Neurathian approach rejects the plau-sibility of justification on the basis of a foundation of non-ethical beliefssuch as those provided by science or logic. This is taken to imply thatthere is no aspiration to convince “anyone whose ethical outlook or per-spective is largely different from” that of the virtuous agent.34 At the sametime, the Neurathian account is not merely to re-express or rationalize thevirtuous agent’s perspective: it “may serve to provide rational credentialsfor our beliefs about which character traits are the virtues.” 35 This isbecause, by allowing (in W. V. O. Quine’s terminology) “plank-by-plank”rebuilding of our ship while at sea, there is “the possibility of radicalethical reflection, the critical scrutiny of one’s ethical beliefs which couldbe genuinely revisionary.” 36 In fact, it could eventually be shown that thevirtuous agent’s ethical outlook is completely wrong, and this possibilityis said to provide us with a suitable notion of ethical objectivity. There isnothing in Foot’s account resembling this Neurathian understanding ofjustification.

Hursthouse neither defends nor develops her Neurathian approach,and we are mostly left to speculate on its implications. The possibility ofplank-by-plank rebuilding, even to the point of eventually replacing everyplank, is itself no guarantee of objectivity. An arbitrary rule (“replaceevery seventeenth plank”) repeatedly applied, or a random procedure,could have this result, but would contribute nothing to objectivity. At aminimum, rebuilding must be in response to relevant facts, and a keyquestion is whether Hursthouse’s account relies on such facts. Her mainclaim is that the justification of central features of her position is —or maybe —epistemically available to a virtuous agent, but not necessarily toothers. This should not be rejected out of hand. Loving and sympatheticpersons sometimes see facts that others do not. Perhaps a case could bemade for saying that virtuous persons see facts that others —for example,those who are morally corrupt, deeply egotistical, or flatly amoral —do not. However, since Hursthouse does not make this case, skepticalsuspicions may be hard to suppress. As we move from the Teleological

33 Hursthouse, On Virtue Ethics, 165. In light of this, David Copp and David Sobel appearmistaken in supposing that Hursthouse is relying on a “morally neutral investigation ofanimal nature.” See their essay “Morality and Virtue: An Assessment of Some Recent Workin Virtue Ethics,” Ethics 114 (2004): 537. However, as I note below, they do raise a legitimatequestion.

34 Hursthouse, On Virtue Ethics, 193. It has been argued that this aspect of Hursthouse’sposition disqualifies it from being a significant form of ethical naturalism. See William Rehgand Darin Davis, “Conceptual Gerrymandering? The Alignment of Hursthouse’s Natural-istic Virtue Ethics with Neo-Kantian Non-Naturalism,” The Southern Journal of Philosophy 41(2003): 583–600.

35 Hursthouse, On Virtue Ethics, 193. She adds that the procedure “is not intended toproduce motivating reasons” (194).

36 Ibid., 165.


Framework to the Teleological Criterion to the Justification Thesis, theperspective of the virtuous agent seems increasingly important. The Tele-ological Framework is based on an examination of social animals, and itwould seem that, if it were correct, any sufficiently informed and impar-tial observer should be able to see this (as we might put it, only epistemicas opposed to ethical virtues would be necessary) —and Hursthouseappears to assume as much in her presentation of it. In contrast, acceptingthe Justification Thesis would seem to require a detailed knowledge of thevirtues, and it might plausibly be said that only a virtuous agent couldpossess this knowledge. What about the Teleological Criterion? Why shouldthe four ends be central criteria of virtue? For my purpose, this is thecrucial question for Hursthouse’s version of NAEN. If the purported factssupporting acceptance of the Teleological Criterion could only be appre-hended by the virtuous (by her lights), then Hursthouse’s defense wouldbe plausible only on the basis of an epistemology that shows the virtuousto be in an epistemically superior position in the relevant respects. But itis not clear that this is her position. She certainly tries to render theTeleological Criterion plausible to a diverse philosophical audience that isnot presumed to be fully virtuous by her standards, and in this respecther aspiration to objectivity does not differ markedly from Foot’s.

III. The Challenge of Moral Universalism

Why, then, should we accept the Teleological Criterion? Even if weagree that ‘good’ and ‘bad’ are attributive adjectives whose criteria ofapplication are fixed by the kind of thing they refer to, and even if weaccept the Teleological Framework as a proper framework for evaluatingsocial animals, we can still ask whether this framework is relevant, andindeed especially relevant, for judging human virtues. This question mightbe pursued from different directions. We might wonder why the factspresupposed in the Teleological Framework are the morally relevant onesinsofar as there are other facts, equally scientific, that might be relevant aswell (or, at least, relevant from a different moral perspective).37 Or, begin-ning with the Teleological Framework and the Teleological Criterion asthey are presented by Hursthouse, and reflecting on our considered moralconvictions, we might ask whether we should accept the TeleologicalCriterion, and hence the Justification Thesis, in light of these convictions.My argument presses a particular form of this second question.38

The difficulty that arises for proponents of NAEN centers on the thoughtthat each human being is morally significant. Various forms of this thoughthave been expressed, in diverse ways, in numerous and very different con-

37 This is the question stressed by Copp and Sobel, “Morality and Virtue,” 534–37.38 Much of the discussion of Foot, and especially Hursthouse, is devoted to arguing that

a virtuous agent would tend to live a life of genuine happiness (eudaimonia). I am notconcerned with this aspect of their position.


texts. For example, the Universal Declaration of Human Rights begins withthe “recognition of the inherent dignity and of the equal and inalienablerights of all members of the human family.” 39 Among philosophers,Immanuel Kant declared that “the human being and in general every ratio-nal being exists as an end in itself,” and he took this to mean that everyhuman being possesses “an inner worth” or “dignity” that entitles him orher to moral consideration.40 In the utilitarian tradition, Jeremy Benthamdeclared, in the dictum cited by John Stuart Mill, “everybody to count forone, nobody for more than one.” 41 These last two statements are expres-sions of the Enlightenment (arguably, all three are), but there are much olderformulations of the idea that each human being has moral importance. Forinstance, the Christian command to love one another meant that we wereto love all persons, as is made clear in the story of the Samaritan who helpedthe beaten man on the road and in Paul’s statement that “there is neitherJew nor Greek.” 42 Similar beliefs are also central to Buddhist traditions,especially in the aspiration of the Bodhisattva to seek enlightenment out ofcompassion for all beings. According to His Holiness the Dalai Lama, weshould “respect and appreciate the sameness of ourselves and others ashuman beings.” 43 In the ancient Greco-Roman world, such ideas are espe-cially associated with the Stoics. For example, Hierocles urged us to haveconcern for “the whole human race.” 44

These statements do not all express the very same thesis (much lesswould they be supported by the same rationale). Nonetheless, there isconsiderable overlap among these statements, and it is this overlap, vaguethough it may be, that is important for my purpose. Let us say that theyall agree, more or less, in affirming:

Moral Universalism. Each human being has moral worth or standing,and hence deserves serious moral consideration.

Moral Universalism does not entail an impartialist morality according towhich there is one and only one moral principle that states that eachhuman being deserves the same moral consideration. Though some of the

39 Universal Declaration of Human Rights, in United Nations Department of Public Infor-mation, ed., The United Nations and Human Rights, 1945–1995 (New York: United NationsReproduction Section, 1995), 153, first paragraph.

40 Immanuel Kant, Groundwork of the Metaphysics of Morals, trans. Mary Gregor (Cam-bridge: Cambridge University Press, 1998), 428, 435 (standard Prussian Academy pagination).

41 John Stuart Mill, Utilitarianism, ed. Oskar Piest (Indianapolis, IN: Bobbs-Merrill Com-pany, 1957), 76 (chap. 5, par. 36).

42 Galatians 3:28. The story of the Samaritan is in Luke 10:25–37.43 Tenzin Gyatso, The Fourteenth Dalai Lama, “Hope for the Future,” in Fred Eppsteiner,

ed., The Path of Compassion: Writings on Socially Engaged Buddhism (Berkeley, CA: ParallaxPress, 1988), 3.

44 See A. A. Long and D. N. Sedley, eds., The Hellenistic Philosophers: Translations of thePrincipal Sources with Philosophical Commentary (Cambridge: Cambridge University Press,1987), 349.


aforementioned sources may have accepted such an impartialist morality,this is a much stronger position than Moral Universalism. An adherent ofMoral Universalism need not maintain that each person deserves the verysame moral consideration nor that this is the only moral principle. None-theless, it does make a moral difference whether or not Moral Universal-ism is accepted. For those who accept it, all human beings —including“strangers” with whom we have no direct or close relations, or withwhom we are seriously at odds —deserve significant moral considerationin virtue of being human beings. For example, on this view, proper treat-ment of prisoners of war is always a moral issue. However, for those whodo not accept Moral Universalism, this need not be the case. Some ofthese persons may think that some human beings have moral standingand others do not. This is a common view and will be the main alternativeat issue here. Nonetheless, not accepting Moral Universalism may alsomean thinking that no one has moral standing or not taking a position onthe question, even implicitly.

It is plausible to suppose that, for many persons, Moral Universalism isa considered moral conviction. My contention is that it cannot be shown,within the framework of NAEN, that a virtuous agent would necessarilyact in accord with Moral Universalism. We are told by Foot and Hursthousethat a virtuous person is just, charitable, kind, etc. I will argue that, if aperson with these virtues is understood to presuppose or act in agree-ment with Moral Universalism, then these virtues do not meet the Tele-ological Criterion. Since the Justification Thesis says that these virtues domeet this criterion (they are on Hursthouse’s standard list), my argumentwill be that this thesis is false.

Some proponents of virtue ethics in the Aristotelian tradition may grantthat a virtuous agent need not act in accordance with Moral Universal-ism.45 However, there is every indication that, for Foot and Hursthouse,a virtuous agent thinks that, or at least acts as if, all human beings havemoral standing (though Foot and Hursthouse are not inclined to articu-late general principles of this kind).46 Foot speaks of “one who recognizes

45 John Cottingham agrees that Moral Universalism does not naturally arise out of theperspective of Aristotelian virtue ethics. See his “Partiality and the Virtues,” in Roger Crisp,ed., How Should One Live? Essays on the Virtues (Oxford: Oxford University Press, 1996),68–70, 75–76.

46 In a discussion of applied ethics, Hursthouse says we should not think about the“moral status” of a fetus in reflecting on the morality of abortion, or of animals in deliber-ating about the morality of eating animals. See Rosalind Hursthouse, “Applying VirtueEthics to Our Treatment of the Other Animals,” in Jennifer Welchman, ed., The Practice ofVirtue: Classic and Contemporary Readings in Virtue Ethics (Indianapolis, IN: Hackett Publish-ing Company, 2006), 136–55. I am not claiming that Hursthouse or Foot would explicitlyendorse Moral Universalism, but the remarks I cite in this paragraph are sufficient to showthat their virtuous agent acts in accordance with this position. My argument is that thevirtues, as understood in these remarks, cannot be warranted by the Teleological Criterion.In this respect, my argument could be made without bringing Moral Universalism into thepicture, though I think it is important to keep it in.


the claim of any human being to a certain kind of respect.” 47 However,implicit recognition of Moral Universalism is especially true of Hurst-house’s virtuous agent, and it is her position that I feature in this discus-sion. To see this, let us focus on charity, a virtue emphasized by theNeo-Aristotelians (though not by Aristotle) that is clearly related to issuesraised by Moral Universalism. Charity involves helping other persons.But which persons? Hursthouse says, perhaps with the Samaritan in mind,that charity imposes a requirement to help a “wounded stranger” alongthe road.48 More importantly, in a discussion of how racism distorts char-ity, she speaks of charity with respect to “our fellow human beings” and“other groups.” 49 In view of these remarks, my critique is internal in thissense: the neo-Aristotelian naturalistic framework cannot provide a basisfor the virtues as Foot and Hursthouse understand them.50

Let us suppose, then, something such as the following: A charitableperson would recognize that there is a requirement (perhaps overridableby other moral considerations) to help any human being in serious need,including a stranger, when in a position to do so. According to the Tele-ological Criterion, charity, understood as implying this requirement, must(a) promote in a substantial way at least one of the four ends and (b) notsignificantly inhibit the four ends; and it must meet these conditionsbetter than any competing virtue. Here is what Hursthouse says about thenaturalistic justification of charity:

I have read that, amongst the social animals, both wolves and ele-phants have patterns of action that resemble our charitable and benev-olent acts, and again it seems plausible to say that the patterns playsimilar roles in the different forms of life. Charity directed to theyoung and helpless particularly serves the continuance of the spe-cies; directed more widely it serves the good functioning of the socialgroup by fostering the individual survival, freedom from pain, andenjoyment of its members, and also by fostering its cohesion.51

On the model of other social animals, charity in this passage is under-stood to promote two ends directly: the raising of our young children(this is the meaning of continuance of the species) and, “more widely,” thewell-being of our social group. Hursthouse also suggests that a person’s

47 Foot, Natural Goodness, 103. Cf. ibid., 114.48 See Hursthouse, On Virtue Ethics, 6, 36.49 Ibid., 118.50 I focus on charity primarily because it is a virtue that is accepted by Foot and Hursthouse

and is relevant to Moral Universalism. The main issue, however, is the moral standing ofhuman beings as such. Hence, my critique could be made even with respect to a minimalform of Moral Universalism that states that a virtuous agent would recognize a prima facierequirement not to harm any human being. I am indebted to several other contributors tothis volume for pointing this out.

51 Ibid., 209.


charity promotes his or her own individual survival, but mostly in indi-rect ways. There is no reference in this discussion to charity directed tothe well-being of human beings generally speaking. This is not surpris-ing, since none of the four ends refers to human beings in general: thefour ends involve only oneself, one’s children, and one’s social group.

It might be observed that charity in the broad sense, as directed towardall human beings, would nonetheless promote the four ends (since one’schildren and social group are included in all human beings). However,even if this were true, it would not follow that such charity could bejustified by the Teleological Criterion, because charity in the narrow sense,as directed only to one’s children and social group, would better serve thefour ends than charity in the broad sense. This appears to be the tacitassumption in the passage quoted above, and that this is Hursthouse’sassumption seems to be confirmed by her critique of Peter Singer’s viewthat “completely impersonal benevolence” toward all sentient beings is avirtue.52 She argues that impersonal benevolence cannot be a virtue becauseit would substantially interfere with the two ends featured in the justifi-cation of charity: continuance of the species (raising our own children)and the good functioning of our social group. This point is offered as anexample of her claim, already quoted, that the structure of the four ends“really does constrain, substantially, what I can reasonably maintain is avirtue in human beings.” 53

It is evident that charity (or benevolence) might be construed morenarrowly or more broadly as we move from one’s children, to one’s socialgroup, to all human beings, and finally to all sentient beings. Hursthousethinks, or is committed to thinking, that teleological facts about humannature dictate at least the approximate point in this sequence to whichcharity should extend. But there is an ambiguity in her position. She oftenspeaks of charity as extending to all human beings, but no further. How-ever, in her justification of charity quoted above she extends it only toone’s children and one’s social group. Moreover, charity in this narrowersense appears to serve the four ends much better than charity in thebroader sense because it does not dilute or interfere with our concern forthese ends. Hence, by the Teleological Criterion, the narrow sense ofcharity is more justified than the broad sense. Indeed, there is nothingexplicit in the Teleological Criterion that encourages us to extend charitybeyond our children and our social group, and thus there is nothing tosupport the idea that in some respects our charity should reach any otherhuman beings.

It might be said, in Hursthouse’s defense, that the Teleological Cri-terion could be interpreted or modified so as to justify charity as sheusually understands it (that is, so that all human beings, and not only

52 See ibid., 224–26.53 Ibid., 224.


members of one’s social group, fall within its scope). There are two gen-eral approaches that might be taken in this endeavor. One is to leave thecriterion as it stands and to give an indirect argument on this basis for thebroader understanding of charity. The other is to formulate the criterionso as to directly include the well-being of all other members of the humanspecies as one of our ends. There are obstacles to both of these approaches.

Hursthouse suggests that racists and sexists often have misunderstand-ings about other races and sexes. The implication is that correcting thesemisunderstandings would provide a reason for abandoning racism andsexism. In a similar way, it might be argued, proper understanding wouldreveal that there are no morally relevant differences between our ownsocial group and other social groups. Hence, since we have reason to careabout our own social group, we have comparable reason to care aboutother groups as well. On the basis of this line of thought, we could beshown to have reason to care about all human beings, irrespective ofsocial group.

The difficulty with this argument is that, according to the TeleologicalCriterion, the fact that this group is our own social group, and theseothers are not, is a morally relevant difference. Group membership iswhat counts, and it does not matter that there are other similarities.Concern for our social group is not inferred from a set of non-indexicalproperties possessed by the group such that, were the same propertiespossessed by another group, concern for that group should also be inferred.Concern for our group is said to be based on the fact that the natural endsof our species are normative and the well-being of one’s specific group isone of those ends. Though the expression ‘social group’ is not defined byHursthouse, it is obviously used to express the idea that human beings,like other social animals, are not solitary creatures and naturally livetogether in fairly small and local groups. What is morally relevant isprecisely membership in some particular group. Hence, there is no basisfor extending concern to all human beings in the way envisioned by thisargument.

In this connection, it is worth considering a suggestion, offered as afriendly amendment to Hursthouse, that fair treatment of those outsideour community could be warranted by appeal to a broader notion ofcommunity than was originally maintained in the Aristotelian tradition.According to Karen Stohr and Christopher Heath Wellman, “given thefacts of economic, social, and environmental interdependence of culturesand communities in our world today, we must also think about ourselvesin the context of a global community.” 54 An initial difficulty with thisproposal (somewhat reminiscent of Hierocles’ suggestion that we extendour circles of concern to the whole human race) is that it stretches the

54 Karen Stohr and Christopher Heath Wellman, “Recent Work on Virtue Ethics,” Amer-ican Philosophical Quarterly 39 (2002): 69.


notion of community or group, as originally employed by Hursthouse,to the breaking point. Once all human beings become members of one’sgroup, the distinctive features of membership in a particular group willbe lost —features presupposed, for example, in Hursthouse’s critique ofSinger and in her discussion of loyalty. Moreover, the facts of inter-dependence are not sufficient to establish a community or group withnormative significance. Though various forms of interdependence arepresupposed by groups in Hursthouse’s use of the term, much more isinvolved. For example, comparatively small groups often involve bondsof affection and concern that arise from a common place, history, andensemble of mores. Indeed, it is usually thought to be an advantage of theAristotelian tradition that these are stressed. By their nature, such bondscannot be extended indefinitely. Hence, there is a clear qualitative differ-ence between the local groups presupposed in Hursthouse’s TeleologicalCriterion and any conception of a global community.

Another indirect argument is based on the claim that different socialgroups would all benefit from mutual cooperation. Realization of thisgives each group a reason to cooperate with other groups, and a com-mitment to mutual cooperation provides a basis for the belief that thereis reason to help any human being, irrespective of social group.

Even if this approach went some distance in addressing the concerns ofMoral Universalism, it is not likely to go far enough, because it providesonly a tenuous, instrumental justification of concern for human beingsbeyond our social group. In conflicts in which other groups are disin-clined to cooperate, or worse (e.g., actively opposed to cooperation, orthreatening violence), the fact that all groups would benefit from mutualcooperation does not give us a reason to cooperate. If our only primaryconcern is for the well-being of our own group, then in these circum-stances there is no basis for concern about other groups. In general, onthis approach, as reasonable expectations that other groups will cooperatedecline, so too does our reason to cooperate. However, for many propo-nents of Moral Universalism, it is especially in circumstances of conflictand noncooperation that it is important to have a firm conviction thateach human being, irrespective of group membership, deserves seriousmoral consideration. Given the strong propensity to protect our owngroup (a propensity that Hursthouse highlights), it may be said, theremust be a basis for believing that each human being has moral worth orstanding, regardless of social group, that does not depend on how coop-erative human beings actually are. For example, many would argue thatrespect for human rights should persist even in warfare.

There is another difficulty with the cooperation argument. Similar argu-ments, based on the benefits of mutual cooperation, have been given toshow why each individual, presumed to be only self-interested by nature,nonetheless has reason to have concern for other individuals. A distinc-tive feature of NAEN is to take a different approach than these argu-


ments: namely, that we are social by nature and thus there is no need toshow why self-interested individuals nevertheless have reason to coop-erate. This is part of the point of appealing to the good functioning ofsocial groups as one of our natural ends.55 But if this claim is inadequatebecause we are, as it were, insufficiently social by nature (by nature, ourconcern only extends to our own group), and an appeal to the benefits ofmutual cooperation needs to be made anyway, then it is unclear why weshould accept this mixed approach rather than a pure mutual-cooperationapproach. Either way, a mutual-cooperation argument needs to be sound,and the mixed approach has the additional burden of establishing thatconcern for our social group is an end we have by nature.

If an indirect approach will not work, perhaps a direct one will. Itmight be argued that the Teleological Criterion could be modified so asto include as an end the well-being of all members of the species(either as an additional, fifth end or as a modification of the social-groupend), and that this would overcome the difficulties with the indirectapproaches.

The key question here is: What is the basis, within the framework ofNAEN, for understanding the Teleological Criterion in this way? In theoriginal argument, the fact that the four ends were ends for social animalsin general appeared to play a crucial role in the claim that, since we arealso social animals, these are ends for us as well. But it is implausible tosuppose that social animals in general are concerned about all membersof their own respective species. Since they are social animals, they areconcerned about the well-being of some —typically, very few —membersof their species (and there is great variation in the forms this may take).But this does not mean that they are concerned with all members of theirspecies, and there is no reason to think this is true of social animalsoverall. They may well be indifferent or even hostile to members of theirspecies outside their social group. Hence, no argument of this form couldbe credible.

In her critique of Singer, Hursthouse seems to imply, at one point, thatpartiality toward or caring about members of our own species, in contrastwith members of other species, is part of our biological nature.56 How-ever, even if this were true, it need mean only that we have concern forsome members of our species —for example, our relatives, friends, andneighbors —that we do not have for members of other species. (The roleof pets in many people’s lives suggests that even this claim would requirequalification.) Such partiality is consistent with an absence of concern formembers of our species beyond those just mentioned.

55 According to Hursthouse, “that ‘man’ is by nature an entirely self-centred egoist mustsurely be a view that could only come about through its proponents overlooking the factthat if their mothers had not cared for them for many years in their infancy they would nothave survived” (On Virtue Ethics, 252–53). Cf. Foot, Natural Goodness, 16.

56 See Hursthouse, On Virtue Ethics, 225–26.


If the well-being of human beings as such were to be included amongour ends, the argument for this would have to appeal to some differencebetween human beings and other social animals —presumably, for Footand Hursthouse, our rationality. It would have to be shown that humanbeings, as rational, have as an end (among others) the well-being of allother members of their species. This claim need not go as far as Kant, forexample, who argued that we should have concern for any rational being,human or otherwise (were other rational beings to exist). The proponentof NAEN would only need to show that a characteristic end for humanbeings as rational is the well-being of any member of their own species.The point might be put this way: Just as animals, in virtue of their dis-tinctive nature, have ends that plants do not have, so human beings, invirtue of their distinctive nature —that is, their rationality —have an endthat neither plants nor animals typically have. Unlike them, human beingshave as an end the well-being of all members of their own species.

At this stage, it is not clear to what facts proponents of NAEN couldappeal. In other cases —for example, self-preservation, raising children,and living in social groups —there are some evident facts about humannature that might be thought to lend prima facie plausibility to the claimthat our ends should include reference to these. From what we knowabout human beings, it is difficult to envision a human society that isindifferent to these ends (even though particular individuals sometimesare). In this sense, it might be said, there is something “natural” aboutthese ends. However, it is hard to see what fact about human nature couldplay an analogous role in an argument for the claim that our ends shouldinclude the well-being of human beings generally. It is not difficult toenvision a society that is indifferent to this end (i.e., a society in whichconcern for the well-being of others extends only as far as relatives, friends,and neighbors). In any case, the extent to which human beings actuallyaccept this end is not thought to be what is primarily relevant. The ques-tion is whether or not concern for all human beings is a natural end fora species of rational, social animals such as ourselves. By way of com-parison, it might be maintained that communication via language is anatural part of human life in view of our nature as rational, social ani-mals. Hence, it might be said, linguistic communication is a natural endand thus we cannot imagine a human society without language. What isneeded is a similar argument that establishes that the well-being of allhuman beings is a natural end for rational, social animals.57 Moreover, inorder to be consistent with naturalism, rationality in such an argumentwould need to be seen simply as a natural fact about us, and not, forexample, as something that reveals our participation in the divine or in acosmic order (as the Stoics supposed). In short, even if it could be shown

57 Hursthouse does not think our use of language itself has ethical significance. See ibid.,219.


that our natural ends include the well-being of some human beings, thechallenge for proponents of NAEN whose virtuous agent acts in accordwith a form of Moral Universalism is to show that these ends also includethe well-being of all human beings.

It might be suggested that Hursthouse’s Neurathian procedure is rel-evant here. For it could be argued that a virtuous agent —someone who isjust, charitable, etc. —comprehends that the well-being of all human beingsis one of our ends even though there is no fact that the virtuous agentcould point to that would show this to be true from a neutral perspectiveor from other ethical points of view. In effect, only the virtuous, as under-stood by NAEN, can perceive the relevant fact, whatever it may be.58

However, merely to assert this is surely to succumb to the danger, advertedto by Gary Watson, that this would “no longer ground moral judgmentbut rather express it.” 59 The point of the Teleological Criterion is explic-itly to avoid this danger and provide, as Hursthouse says, “rational cre-dentials” for the virtuous person’s belief that character traits such asjustice and charity (as Hursthouse understands them) are in fact virtues.To achieve this goal and remain properly naturalistic, there must be aplausible story that takes us from a version of the Teleological Criterionto the conclusion that the virtuous agent has concern for all human beings —and this story must be available to anyone with proper epistemic quali-fications, however these may be understood, and not simply to thosewho, in effect, already accept it.

This issue may be approached in a somewhat different way. Hursthouseaccepts Bernard Williams’s contention that objectivity in ethics requiresthat ethical disagreements be rooted in factual disagreements. In partic-ular, she supposes that, within the framework of virtue ethics and natu-ralism, disagreements about the virtues can be resolved by reference tofacts about human nature and the circumstances of human life.60 In viewof this, let us imagine that the following dispute arises among those whoaccept Hursthouse’s overall approach and are at least moderately virtu-ous (they are all on board the H.M.S. Virtue Ethics, we might say, and theyare debating about whether and how to replace some planks). All agreewith the Teleological Criterion and its general implications for virtuessuch as justice and charity with respect to one’s own children and one’ssocial group (henceforth, I will refer to these two together as “the localgroup”). What they disagree about is the implications of these virtues forhuman beings beyond the local group (perhaps they have only recently

58 It might be said that Moral Universalism is a constitutive part of “the moral point ofview,” and that a virtuous agent, but not necessarily others, apprehends this conceptualtruth. However, even if this claim could be defended, it would be a point independent ofNAEN; and there is no reason to attribute it to Hursthouse or Foot.

59 Gary Watson, “On the Primacy of Character,” in Daniel Statman, ed., Virtue Ethics: ACritical Reader (Washington, DC: Georgetown University Press, 1997), 67.

60 See Hursthouse, On Virtue Ethics, 241–42.


interacted with people outside their local group in significant ways). Noone thinks there should be equal concern for all human beings; they allbelieve there should be some special concern for the local group. In thedebate that ensues, four basic positions emerge:

(1) The strong partialists say there should be concern for members ofthe local group and no one else.

(2) The weak partialists say there sometimes should be concern for allhuman beings, but only if this concern in no way harms thewell-being of the local group.

(3) The weak universalists say there occasionally should be concernfor all human beings even though this harms the well-being ofthe local group.

(4) The strong universalists say there often should be concern for allhuman beings even though this harms the well-being of the localgroup.

These positions are vague, and, in a real debate, they would be elaboratedby reference to specific sorts of concerns. Nonetheless, the range of out-looks indicated by the four positions has clear relevance to some of thedifferent views taken in contemporary debates about issues raised byMoral Universalism vis-à-vis virtues such as charity and justice. Of course,there may be a truth about which position most humans beings overallare inclined to accept. However, no statistical fact of this kind is thoughtto be relevant according to proponents of NAEN.61 Hursthouse is com-mitted to the view that there is some teleological fact (or facts) abouthuman nature that provides genuine guidance for resolving this contro-versy and shows which position across this spectrum is at least approx-imately correct. Moreover, this fact is available (at least) to those on boardthe H.M.S. Virtue Ethics. What is elusive, for the reasons already given, iswhat this fact might be. We may suppose that all parties to the disagree-ment acknowledge that members of the local group and other humanbeings are all members of the same species. Those inclined to positions 2,3, and 4 take this similarity to have increasingly greater moral weight,while those inclined to position 1 take it to have no moral weight. Sup-pose, for example, that the weak-universalist understanding of charityand justice is correct. What fact establishes this that all the partialists, aswell as the strong universalists, are failing to understand?

It might be said that debates such as this would be resolved, not bydirectly appealing to any teleological facts, but by pointing out that, forexample, in some circumstances ignoring the well-being of human beingsoutside one’s social group would be callous (or some other rather specificterm from the vocabulary of the virtues and vices). Hursthouse says that

61 See ibid., 223.


“reasons for acting are the reasons people with the relevant charactertrait do, or would, give . . . not the fact that the character trait in ques-tion sustains any of the four ends.” For example, loyal people stick bytheir friends because they are their friends, not because this fosters“the good functioning of the social group.” 62 Nonetheless, it remainsthe case for Hursthouse that loyalty is a virtue (at least in part) becauseit promotes the well-being of the social group. Likewise, presumably,callousness is a vice because it undermines one or more of the fourends, and the question remains why, in terms of these ends, it wouldbe callous to ignore the well-being of human beings outside one’s socialgroup. Hursthouse insists that there is no direct path from the Teleo-logical Criterion to the determination of right actions. The TeleologicalFramework is said to provide a basis for determining which charactertraits are virtues, but it requires practical wisdom ( phronêsis) on thepart of the virtuous to determine, for example, what would be callousin a given situation. This is fair enough so far as it goes. However, inassessing NAEN, everything then depends on elaboration of the delib-erations of the virtuous. If it turns out that these deliberations arelargely dependent on culturally informed understandings of the virtueswith no real anchor in the four ends, then the Teleological Criterionsubstantially underdetermines the virtues.

IV. The Normativity of Natural Ends

Let us now broaden the discussion and ask more generally to whatextent the Teleological Criterion is helpful in thinking about the virtues.I have already argued that it is not helpful in reflecting on issues raisedby Moral Universalism for virtues such as charity. Consider next a dif-ferent issue, one that is directly discussed by Hursthouse and brieflymentioned by Foot —the morality of homosexual ways of life. BothHursthouse and Foot believe that a virtuous agent could live an activehomosexual life and not reproduce or raise children.63 Hence, they do notthink that homosexual activity as such is morally wrong. Does NAENsupport their position and do so in the right way?

Hursthouse herself raises the question of whether the end of continu-ance of the species (that is, reproduction) implies that “practicing homo-sexuals” are not virtuous insofar as they fail to promote this end. Inresponse, she claims that forms of sexuality are not themselves character

62 Ibid., 235.63 See ibid., 214–15. Hursthouse’s discussion of homosexuality was criticized by Brad

Hooker in “The Collapse of Virtue Ethics,” Utilitas 14 (2002): 36–37. She responded to thiscritique in “Virtue Ethics vs. Rule-Consequentialism: A Reply to Brad Hooker,” Utilitas 14(2002): 43–46. Foot mentions homosexuality in passing in Natural Goodness, 109, and in“Rationality and Goodness,” in Anthony O’Hear, ed., Modern Moral Philosophy: Royal Insti-tute of Philosophy Supplement 54 (Cambridge: Cambridge University Press, 2004), 11.


traits and that the relevant character trait for addressing this issue istemperance. Hence, those who engage in homosexual activity would notbe virtuous only if they were intemperate. Naturalism itself does notimply that such persons are intemperate, she says, and she thinks there isno reason to believe they are (for example, by being “all wildly, willfully,promiscuous”).64 The fact that homosexual activity cannot result in repro-duction is irrelevant.

In one respect, this is a perplexing response, at least in light of Aristo-tle’s remark that the temperate person “finds no pleasure at all in thewrong things.” 65 Many people think the pleasure produced by homosex-ual activity is of the wrong sort, and (leaving religious considerationsaside) they often think this because they believe it is, in some sense,“contrary to nature.” It might be thought that NAEN provides a rationalefor this belief. After all, Hursthouse and Foot both think there are actionsthat are absolutely prohibited so that (at least ordinarily) a virtuous agentwould not perform them.66 Presumably, these prohibitions have theirwarrant in the fact that the prohibited actions in some way counter orundermine the natural ends of our species. This is what gives rise to thesuspicion that NAEN could provide a basis for the aforementioned cri-tique of homosexual activity.

In fact, however, this suspicion is misguided. Practicing homosexualscan and do reproduce and raise children. As long as they did this, theywould be promoting the end of continuance of the species. The fact thatthey also engage in homosexual activity need not undermine this end.(Similarly, eating food that has no nutritional value need not be contraryto the end of health, so long as one also eats food that has nutritionalvalue.) However, this is not the response Hursthouse gives, for she thinksthat her outlook does not imply that each individual is required to repro-duce and raise children —and Foot agrees. Moreover, they both suggestthat a celibate person could be a virtuous person in some circumstances.67

Hence, the end of continuance of the species is regarded as optional forindividuals, and this acknowledgment is the feature of the discussion ofhomosexuality that I take to be important.

Are other ends optional as well? Hursthouse’s example of a celibatecontemplative implies that some enjoyment (and also freedom from pain?)might be optional. Still, individual survival could hardly be optional inthis sense: Any way of life presupposes that individual survival is impor-tant (even though a virtuous person might sometimes risk his or her life,or perhaps even intentionally end it in special circumstances). What aboutthe good functioning of one’s social group? It is hard to see how a person

64 Hursthouse, On Virtue Ethics, 215.65 Aristotle, Nicomachean Ethics, trans. Terence Irwin, 2d ed. (Indianapolis, IN: Hackett

Publishing Company, 1999), 1119a14.66 See Foot, Natural Goodness, 49; and Hursthouse, On Virtue Ethics, 58.67 See Foot, Natural Goodness, 42; and Hursthouse, On Virtue Ethics, 215.


who is kind, charitable, and just could fail to have this as an end, and bothFoot and Hursthouse clearly think that this end is not optional in the waythat reproduction is. This reveals a striking asymmetry in their account.They are both fond of saying that, even as there is something wrong witha “free-riding wolf” that enjoys the benefits of the hunt without partici-pating in it, so there is something wrong with a human being who is notcharitable or just.68 They also think there is something wrong with ani-mals that do not reproduce,69 but they do not say in this case that likewisethere is something wrong with human beings who choose not to repro-duce. Hence, in the determination of the virtues, our natural ends associal animals are appealed to in a highly selective way. What determinesthe selection, pretty clearly, is an assumed ethical outlook according towhich everyone is expected to be charitable and just in some respects (atleast) in the sense of promoting the well-being of one’s social group, butnot everyone is expected to be temperate in the sense of sometimes engag-ing in sexual intercourse and reproducing.

When we combine this result with the earlier conclusion concerningMoral Universalism, we can see that the Teleological Framework is doingrather little work in providing a rationale for the virtues. In particular, thetwo ends that are other-regarding —continuance of the species, and thewell-being of the social group —significantly underdetermine, if they donot undermine, virtues such as temperance, charity, and justice that areunderstood to relate to these ends. The Teleological Framework is not thesubstantial constraint Hursthouse claims it is. A straightforward applica-tion of the Teleological Criterion would suggest that virtuous humanbeings would have character traits that (among other things) lead them toreproduce, raise children, and promote the interests of their own socialgroup. If we think (as many, including Foot and Hursthouse, do) thatvirtuous persons need not conceive and raise children, and should beconcerned about human beings well beyond their social group, then weare relying on considerations that have no basis in the TeleologicalFramework.

This is not an isolated point: it is a persistent feature of Hursthouse’sapproach. Here, briefly, is a final example. Hursthouse says that in situ-ations in which meat is produced by factory farming and is not necessaryfor survival, the virtuous would be vegetarians because eating meat wouldbe callous and cruel (among other things) in light of the suffering ofnonhuman animals raised for food.70 Elsewhere, she says that “vegetari-anism is required” in many circumstances by the “virtues of temperance

68 See Foot, Natural Goodness, 16; and Hursthouse, On Virtue Ethics, 196. Hursthouse saysthat a virtuous contemplative need not live a life that is social (see ibid., 228), but that isdifferent from having the good functioning of one’s social group as an end, as a charitableperson presumably does.

69 See Foot, Natural Goodness, 29–30; and Hursthouse, On Virtue Ethics, 200.70 See Hursthouse, “Applying Virtue Ethics,” 141–43.


and compassion.” 71 There is no reference to the Teleological Criterion inarguments for these claims, and, indeed, there appears to be no connec-tion at all between these vices and virtues, taken to have this implication(concern for the suffering of nonhuman animals), and the four ends,which make no reference to nonhuman animals. Hursthouse says that herargument relies on an “ordinary understanding” of these virtue and viceterms. But this only brings out the significant gap between the very thinsense of the virtues that plausibly could be sustained by the TeleologicalCriterion and the very thick sense that is assumed in particular discus-sions such as this.

V. Conclusion: Minimal Naturalism

We need not suppose, in light of this critique, that the facts of humannature adduced by Foot and Hursthouse are irrelevant to our ethicaloutlook. But their relevance is much more limited and indirect than Footand Hursthouse envision. There are clearly some facts about us, rooted inour animal nature and our circumstances in the physical world, as well asin our capacity for rational reflection, that are important for ethical delib-eration. For example, there are facts about our basic needs and vulnera-bilities (physical and mental), our interdependence (on one another aswell as on our physical environment), our tendencies toward conflict andcapacities for cooperation, our sexuality, reproduction, and developmentfrom birth to maturity, and our forms of social relations and kinds ofenjoyments. Some of these facts pertain to virtually every human being.In other cases, the relevant facts concern ranges of “typical” needs, capac-ities, and the like. At some elementary level, these facts are obvious andvirtually beyond controversy. No one would deny that human beingsneed to eat and are vulnerable to a variety of physical ailments. Of course,ethical reflection requires a great deal more detail than this, and, in morespecific accounts, controversies are likely to ensue. However, this neednot mean that there are no relevant facts about which people disagree northat there is no prospect of a more or less objective understanding of thesefacts.

These facts about human nature sometimes have great ethical signifi-cance. For example, the need for nutrition is a presupposition of ethicalissues concerning hunger. To a large extent, however, the importance ofsuch facts is to help give shape to the ethical issues that confront us. Assuch, in some ways these facts do constrain what acceptable resolutionsto these issues could be like. An adequate ethical theory needs to take intoaccount the facts of human nature that give rise to our ethical concerns.To this extent, an adequate ethical theory needs to be, as we might put it,

71 Hursthouse, “Virtue Ethics vs. Rule-Consequentialism,” 48. Cf. Hursthouse, On VirtueEthics, 227.


minimally naturalistic. This is not to say that every theory currently onoffer adequately meets this criterion. Nonetheless, even Kant —who mightbe thought problematic in this regard —accepts the criterion (on the evi-dence of discussions in The Metaphysics of Morals), though his success inproperly applying it is rather mixed.72

One of the dangers of even this minimal naturalism is that there is a strongpropensity to assign to various purported features of human beings the sta-tus of “natural,” with the suggestion that what is natural is fixed or partof an order we must respect, when in fact beliefs about the features in ques-tion reflect mere convention or prejudice. Appeals to nature have an unfor-tunate history of disguising, for example, an unreflective acceptance of thestatus quo, or a complacent acquiescence to the thought that we humanbeings are too limited or weak or sinful to live up to ethical demands oraspirations that we might otherwise have good reason to accept. There isneed to take into account what really are facts of human nature, but thereis also need to exercise great care in ascertaining just what these facts are.Minimal naturalism is but one fallible element in ethical reflection.73

The proponents of NAEN obviously aim to be minimally naturalistic,and to this extent their efforts are to be applauded. But in their claim thatthere are teleological features of human nature that significantly constrainwhat the virtues are, their ambitions clearly go beyond minimal natural-ism. Though Foot and Hursthouse allow for considerable diversity amongvirtuous agents, they both think that the teleological features they iden-tify establish rather definite contours for the virtuous life. This is certainlythe main purport of Hursthouse’s Teleological Criterion. In this they havethe precedent of Aristotle, whose function argument is evident in thenot-too-distant background of their discussions.74 But this precedent haslittle inherent weight. In view of the issues I have raised here, thoseattracted to a virtue theory, even one that is Aristotelian in some familiarrespects, are not well-advised to follow Aristotle’s naturalism beyondwhat minimal naturalism implies. The main problem is not that Aristotlerelies on an outmoded teleological biology that has been replaced byevolutionary biology. The difficulty is that the relevant facts of humannature are too sparse, and the diversity and richness of actual ethicalpractices are too substantial, for us to suppose that an understanding of

72 For example, Kant refers to “our need to be loved (helped in case of need) by others”in an argument for the conclusion that “the happiness of others is . . . an end that is also aduty.” See Immanuel Kant, The Metaphysics of Morals, trans. Mary Gregor (Cambridge:Cambridge University Press, 1991), 393 (standard Prussian Academy pagination).

73 In some respects, though not all, minimal naturalism is closer to what Julia Annas calls,somewhat oddly, “a stronger form of naturalism” in Annas, “Virtue Ethics: What Kind ofNaturalism?” in Stephen M. Gardiner, ed., Virtue Ethics, Old and New (Ithaca, NY: CornellUniversity Press, 2005), 22–29.

74 See Aristotle, Nicomachean Ethics, I.7. The function argument purports to establish thatthe human good is a life of virtue on the basis of the fact that the distinctive function (ergon)of the human species is to live in accordance with reason.


these facts could do the main work, and oftentimes even significant work,in providing an objective rationale for one specific ethical outlook.

The concerns raised by Moral Universalism might be, and have been,addressed in a variety of ways. I have argued that the Teleological Crite-rion is unhelpful in addressing these concerns and that it provides littlesupport for Foot’s and Hursthouse’s own understanding of virtues suchas charity and justice. The claim that human beings are social animals nodoubt expresses a biological truth, but this truth is too blunt an instrumentto provide an effective criterion for understanding the virtues. Hursthousehas suggested that she would give up her naturalism before she would giveup her understanding of the virtues.75 Since her understanding of the vir-tues is generally more sound than her version of naturalism, she would bebetter off seeking a different approach to justifying the virtues.

What, then, is the alternative? A virtue theory that draws on Aristotle —for example, by emphasizing his understanding of character, happiness,emotions, friendship, and the like —might be compatible with a numberof metaethical outlooks (though probably not all). Insofar as proponentsof such a position are committed to rejecting non-objective accounts ofmorality, as Foot and Hursthouse are, such proponents might turn in thedirection of intuitionism, or perhaps a reflective-equilibrium account, andmake their case for objectivity on those grounds. Some of Hursthouse’scomments on the Neurathian model might be understood in these terms.A different approach would be to follow Aristotle in important respects,but abandon the claim that the basic commitments of a virtuous agentadmit of objective justification. Perhaps a form of relativism could beembraced instead. Yet another approach would be to combine an objec-tive account with some relativist elements, as Foot herself, many yearsago, suggested might be possible.76 Whatever approach is taken, many ofthe purported advantages of an Aristotelian virtue theory might well beretained without any appeal to naturalism in the form of the TeleologicalCriterion or a close approximation of it.

Philosophy, Fordham University

75 See Hursthouse, On Virtue Ethics, 211; and Hursthouse, “Virtue Ethics vs. Rule-Consequentialism,” 52. In the article, a response to criticism of the book, Hursthouse appearsto downplay and modify the importance of the Teleological Criterion. She continues to beimpressed by the purported similarities between ethical evaluations and evaluations ofsocial animals. But she says that, in view of her Neurathian outlook, the Teleological Cri-terion “is ‘justificatory’ only in a pretty thin sense” (“Virtue Ethics vs. Rule-Consequentialism,”50), and that “it would be better to think of it as an explanatory, in contrast to a justificatory,”principle (ibid., 52). In the book, the criterion is clearly offered as the central element in ajustification project (see On Virtue Ethics, 164, 166, 193–94).

76 See Foot, “Morality and Art” (1970), and “Moral Relativism” (1979), reprinted in herMoral Dilemmas, 5–19, 20–36. In the introduction to this volume, Foot said she was reprintingthese essays “only hesitantly” in view of her current acceptance of NAEN (ibid., 2). Cf. Foot,Natural Goodness, 23–24. In this connection, see David B. Wong, Natural Moralities: A Defenseof Pluralistic Relativism (New York: Oxford University Press, 2006), esp. chap. 2.


THE IMPORTANCE OF METAPHYSICAL REALISMFOR ETHICAL KNOWLEDGE*

By Douglas B. Rasmussen

Kant’s glory . . . is to say that the very fact that we cannot separateour own conceptual contribution from what is “objectively there” isnot a disaster. . . . Similarly . . . Kant rejects the idea that we havesomething analogous to the medieval “rational intuition” with respectto moral questions. And again he argues that this is no disaster, thaton the contrary it is a Good Thing. The whole Kantian strategy . . . isto celebrate the loss of essences. . . .

—Hilary Putnam, The Many Faces of Realism

The fundamental reason that I myself stick to the idea that there areright and wrong moral judgments and better and worse moral out-looks . . . is not a metaphysical one. The reason is simply that that isthe way that we —and I include myself in this “we” —talk and think,and also the way that we are going to go on talking and thinking.

—Hilary Putnam, Renewing Philosophy

What do you mean, “We’re” in trouble?

—Indian scout to General Custer

I. Introduction: Essentialism without Realism

Recent works in ethics evidence a resurgence in the belief that an appealto human nature is crucial for an understanding of the central compo-nents of the moral life.1 This resurgence represents in many cases a height-ened appreciation for the wisdom of the Aristotelian tradition and that

* This essay has benefited from the helpful suggestions of Douglas J. Den Uyl, HarryDolan, David Gordon, Fred D. Miller, Jr., Ellen Paul, Tibor R. Machan, Jan Narveson, and theother contributors to this volume.

1 I have the following works in mind: Julia Annas, The Morality of Happiness (New York:Oxford University Press, 1993); Douglas J. Den Uyl, The Virtue of Prudence (New York: PeterLang, 1991); Philippa Foot, Natural Goodness (Oxford: Clarendon Press, 2001); RosalindHursthouse, On Virtue Ethics (Oxford; Oxford University Press, 1999); Anthony Lisska,Aquinas’s Theory of Natural Law: An Analytic Reconstruction (Oxford: Clarendon Press, 1996);Michael Thompson, “The Representation of Life,” in Rosalind Hursthouse, Gavin Lawrence,and Warren Quinn, eds., Virtues and Reasons: Philippa Foot and Moral Theory (Oxford: ClarendonPress, 1995), 247–96; and Warren Quinn, Morality and Action (Cambridge: Cambridge Uni-versity Press, 1993), esp. chap. 11.


tradition’s view that there is not in all cases some great ontological dividebetween what is and what is valuable. Particularly, there has been thegrowing realization that such a divide need not exist if one’s ontologyaffords a place for some form of natural teleology2 and essentialism.3 Yetto see the importance of natural teleology and essentialism for dealingwith the question of whether there is a basis in the “is” for grounding an“ought,” it is necessary to look at this issue from the context of ontologicaland epistemological realism.4 Indeed, the entire debate between thosewho hold that there is an unbridgeable gulf between the way things areand what ought to be, and those who do not, only has a point if thisdebate is ultimately about (1) the fundamental character of reality —mostimportantly, human nature —and (2) our ability to know it. It is then forthese reasons most important to consider the foundations of an argumentthat is antithetical to the Aristotelian tradition and holds not merely thatontological and epistemological realism are unnecessary for ethical knowl-edge but that their downright rejection is required if the so-called dividebetween what is and what is valuable is to be avoided.

Martha C. Nussbaum illustrates this argument most vividly in heradvocacy of what she calls “internalist” essentialism.5 This version ofessentialism rejects the claim that “there is some determinate way the

2 Natural teleology holds that for at least some class of entities, usually living ones, thenatures of those entities are also those entities’ ends or functions. Though natural teleologyis often thought of as requiring a worldview in which the natures of things are seen as reflect-ing the will of a deity, this need not be so. As Michael Thompson notes, “I think we are veryfar from the category of intention or psychical teleology. . . . Natural-teleological judgementsmay . . . be said to organize the elements of a natural history: they articulate the relations ofdependence among the various elements and aspects and phases of a given kind of life. . . .[E]ven if the Divine Mind were to bring a certain life-form into being ‘with a view to’ securingan abundance of pink fur along the shores of the Monongahela, this would have no effect onthe natural-teleological description of that form of life.” Thompson, “The Representation ofLife,” 293–94. See also Douglas B. Rasmussen, “Human Flourishing and the Appeal to HumanNature,” Social Philosophy and Policy 16, no. 1 (1999): 1–43, esp. Section V.

3 In its most general sense, essentialism holds that to be is to be something, and thus thatthere is something about a thing without which that thing would neither exist nor be thatthing. See the works cited in notes 115 and 121 below.

4 Ontological realism holds that there are beings that exist and are what they are inde-pendent of and apart from our cognition; epistemological realism holds that we can come toknow, though not without great difficulty, both the existence and nature of these beings. SeeRoger Trigg, Reality at Risk: A Defense of Realism in Philosophy and the Sciences (Sussex: Har-vester Press; and Totowa, NJ: Barnes and Noble Books, 1980), for a discussion of these views.

5 Martha C. Nussbaum, “Human Functioning and Social Justice: In Defense of Aristote-lian Essentialism,” Political Theory 20, no. 2 (May 1992): 202–46. Nussbaum claims that herrejection of ontological and epistemological realism has its origins in Aristotle’s understand-ing of “appearances.” She claims that “we can have truth only inside the circle of theappearances, because only there can we communicate, even refer, at all.” Martha C. Nussbaum,The Fragility of Goodness: Luck and Ethics in Greek Tragedy and Philosophy (Cambridge: Cam-bridge University Press, 1986), 257. Christopher Long notes, however, that Nussbaum failsto do justice to Aristotle’s view: “Aristotle operates with a naturalistic conception of therelationship between being and language that allows him to recognize that our very speak-ing about beings reveals something of the nature of these beings themselves.” Long, “SavingTa Legomena: Aristotle and the History of Philosophy,” The Review of Metaphysics 60, no. 2(December 2006): 251–52.

THE IMPORTANCE OF METAPHYSICAL REALISM 57

world is apart from the interpretive workings of the cognitive faculties ofliving beings,” 6 and, more particularly, it rejects the claim that the natureor essence of a human being is “part of the independent furniture of theuniverse.” 7 Instead, an internalist essentialism holds that a deep exami-nation of

human history and human cognition from within . . . reveals a more orless determinate account of the human being, one that divides essen-tial from accidental properties. . . . Separating these two groups ofproperties requires an evaluative inquiry: for we must ask, whichthings are so important that we will not count a life as a human lifewithout them? Such an evaluative inquiry into what is deepest andmost indispensable in our lives need not presuppose an externalmetaphysical foundation, clearly: it can be a way of looking at our-selves, asking what we really think about ourselves and what holdsour history together.8

In other words, an internalist essentialism is an essentialism withoutrealism; that is to say, it does not make the ontological claim that thingsare what they are independent of and apart from how people think aboutor interact with them.

Part of Nussbaum’s motivation for rejecting the realist ontologicalclaim is tied to her acceptance of the now common, and ultimatelyKantian, view that a realist epistemological claim —namely, that peoplecan know the character of things as they really are or, as Nussbaumputs it, the “real structure as it is in itself” —is not sustainable.9 Thereis no access to reality independent of human cognitive activity or apartfrom human history and interpretation (that is, without the use of con-ceptual systems), and since this is so, “the hope for a pure unmediatedaccount of our human essence as it is in itself . . . is no hope at all buta deep confusion.” 10

Of course, Nussbaum’s plan is not to throw in the essentialist toweland give up the fight to the proponents of extreme relativism or sub-jectivism.11 She is against those who would say there are no norms fordefending an account of what a human being essentially is. Nonethe-less, her tack is to defend internalist essentialism without recourse toeither realist ontological or realist epistemological claims; the essential-ism she seeks to defend is, as she puts it, an essentialism without“metaphysical realism.”

6 Nussbaum, “Human Functioning and Social Justice,” 206.7 Ibid.8 Ibid., 207–8.9 Ibid., 206.

10 Ibid., 207.11 Ibid., 212.

58 DOUGLAS B. RASMUSSEN

Once we rid ourselves of thinking that we need metaphysical real-ism and its illusory promise of providing some transcendent groundingfor our evaluative judgments, Nussbaum thinks that we can get onwith the business of reasoning in which we are already engaged.“We have everything that we always had all along: the exchange ofreasons and arguments by human beings within history, in which,for reasons that are historical and human but not the worse for that,we hold some things to be good and others bad, some arguments tobe sound and others not sound.” 12 In support of her thesis, she appealsto Hilary Putnam’s argument that the demise of metaphysical realismmay even raise the cognitive status of ethical evaluation.13 Nussbaumstates:

For the metaphysical realist frequently made a sharp distinctionbetween fact and value, believing that truth of the sort the realistis after is available in the scientific realm but not in the realm ofvalue. Bringing science inside human history makes what we alreadybelieved to be in there look better, not worse —because its claimsare no longer contrasted sharply with claims that look “harder”and more “factual.” Thus, the polarity between scientific fact andsubjective ethical values on which much of neoclassical economicsrests is called into question by the collapse of realism —from theside of science, to be sure, but this reopens the whole question ofthe relationship between ethics and science and makes it possibleto argue, as Putnam does, that ethics is no worse off than anyscience.14

Thus, not only is metaphysical realism unnecessary for explaining whichaccounts of human nature are better than others, but the demise of meta-physical realism, as argued by Putnam, provides an avenue by which thefact/value dichotomy can be overcome and, in turn, another way ofunderstanding ethical evaluations. Indeed, there is the promise of a bettermeans of defending the cognitive status of such evaluations than whatmetaphysical realism affords.

In this essay, I plan to consider whether the alleged demise of meta-physical realism actually does provide a better way of defending thecognitive status of ethical judgments. It will be my contention that therejection of a realist ontology and epistemology does not help to establishthe claim that ethical knowledge is possible. More specifically, I will argue

12 Ibid., 212–13.13 Nussbaum also mentions Donald Davidson, W. V. O. Quine, and Nelson Goodman as

critics of metaphysical realism. I will not be considering these thinkers because, for the mostpart, they have not used their criticisms of metaphysical realism as a basis upon which tobuild an account of ethical knowledge.

14 Nussbaum, “Human Functioning and Social Justice,” 213–14.


that Putnam’s argument does not succeed in making a case for ethicalknowledge. In fact, his account of the procedures by which our valuationsare warranted ultimately begs the question in a number of crucial ways.Moreover, his account prejudices the moral and political discussion incertain ideological respects. Finally, though Putnam has apparently mod-ified to some extent his approach to the issue of realism in recent years,15

I will suggest that these modifications are not fundamental and do nothelp to advance the case for ethical knowledge, and that ultimatelyNussbaum’s use of Putnam’s argument works against the revival ofAristotelianism in contemporary ethics.

I will proceed as follows. In Section II, I will consider the essentialsof Putnam’s argument. This will involve examining his reasons for reject-ing metaphysical realism and espousing conceptual relativism, his viewof truth and objectivity, and, most importantly, his account of howvalues become objective or valuable through a procedure that makesfull use of human intelligence or, as he also describes it, by an appealto the criteria of “idealized inquiry.” In Section III, I will examine whetherthe criteria of idealized inquiry do indeed provided a basis for ethicalknowledge. Here I will begin by considering whether Putnam’s accountof idealized inquiry achieves his ultimate political goal (namely,showing that a social-democratic political/legal order is the precondi-tion for the full use of human intelligence), or whether a classical-liberal political/legal order might instead be necessary for the full useof human intelligence. This question will also involve the issue of whethera classical-liberal order might better respect the autonomy of personsthan a social-democratic one. My point will be simply that the criteriaof idealized inquiry fail to provide a way for choosing between thesetwo opposed forms of political order. I will next consider whetherPutnam’s account of idealized inquiry packs into its criteria assump-tions about ethics and human nature that are very controversial andwhether these criteria can, without deeper claims about human nature,withstand a series of basic objections. Overall, the aim of Section IIIwill be to indicate that Putnam cannot succeed in making a case forethical knowledge without returning to an account of human naturethat involves the very viewpoint he has rejected —that is, metaphysicalrealism. In Section IV, I will consider, but only in a limited manner,whether Putnam’s arguments for the demise of metaphysical realismare in fact compelling and whether they might not rest on basic ambi-guities in his thought. Finally, in Section V, I offer my overall evalua-tion of Putnam’s argument.

15 In part 1 (“Sense, Nonsense, and the Senses: An Inquiry into the Powers of the HumanMind”) of Hilary Putnam, The Threefold Cord: Mind, Body, and World (New York: ColumbiaUniversity Press, 1999), Putnam discusses his earlier mistakes regarding realism, but he stillmaintains his rejection of metaphysical realism. Part 1 first appeared in The Journal ofPhilosophy 91, no. 9 (September 1994): 445–517.


II. Finding Facts in a World of Values

A. Metaphysical realism and conceptual relativity

Hilary Putnam does not think that metaphysical realism is false. Rather,he thinks it is unintelligible,16 and he thinks this is so because he believesthat it makes no sense to claim to know reality as it is “in itself.” 17 Thereason this makes no sense is that “we have no access to unconceptual-ized reality.” 18 Furthermore, attempts to get a precise or clear formulationof metaphysical realism do not succeed, for when we try to do so, suchformulations “become compatible with strong forms of ‘antirealism.’ ” 19

Thus, they do not succeed “because there is no real content there to becaptured.” 20

Putnam’s overall strategy in dealing with metaphysical realism is oneof trying to provide a reductio ad absurdum by showing that when appliedto various issues metaphysical realism leads to antinomies. Indeed, hisbottom-line position is a version of conceptual relativism that is a miti-gated form of Kantianism. Putnam writes:

If objects are, at least when you get small enough, or large enough,or theoretical enough, theory-dependent, then the whole idea oftruth’s being defined or explained in terms of a “correspondence”between items in a language and items in a fixed theory-independentreality has to be given up. The picture I propose instead is not thepicture of Kant’s transcendental idealism, but it is certainly relatedto it. It is the picture that truth comes to no more than idealized rationalacceptability.21

Putnam’s picture of truth as nothing more than idealized rational accept-ability —as well as his adjustments22 to this picture —will be consideredlater, but for now, we need to concentrate on his conceptual relativism.Putnam ultimately holds that referring to something is only accom-plished internally through a system of practices —that is, a conceptual

16 Hilary Putnam, “The Question of Realism,” in James Conant, ed., Words and Life (Cam-bridge, MA, and London: Harvard University Press, 1994), 300.

17 Ibid., 302.18 Ibid., 297.19 Ibid., 303.20 Ibid.21 Hilary Putnam, “A Defense of Internal Realism,” in James Conant, ed., Realism with a

Human Face (Cambridge, MA, and London: Harvard University Press, 1990), 41 (emphasisadded).

22 Putnam has more recently qualified this picture. He no longer defines truth as war-ranted assertibility. Rather, he does not think that a definition of truth is necessary, thoughhe claims that he can examine the concept of truth in relation to other semantic and epis-temological concepts. See Putnam, The Threefold Cord, part 1, and Putnam, “Are Values Madeor Discovered?” in Putnam, The Collapse of the Fact/Value Dichotomy and Other Essays (Cam-bridge, MA, and London: Harvard University Press, 2002), 107.


scheme. It is only within a system of practices that a reference is fixed oran object described.23 The ability to use signs in the right ways in theright circumstances consists in following the practices of a communityof users. For example, using “horse” but not “cat” to refer to horses, orsaying “There is a brown bear” but not “There is a bearish brown,” isonly possible and situationally appropriate within a conceptual scheme.Further, to employ an example24 Putnam often uses to illustrate hisconceptual relativity,25 consider a world with these objects: x1; x2; andx3. Do three objects exist in this world or do seven (x1; x2; x3; x1 + x2;x1 + x3; x2 + x3; x1 + x2 + x3) exist? For Putnam, such a question is,apart from a system of practices regarding the use of the word “exist,”silly. He writes:

Signs do not intrinsically correspond to objects, independently ofhow those signs are employed and by whom. But a sign that isactually employed in a particular way by a community of userscan correspond to particular objects within the conceptual scheme ofthose users. “Objects” do not exist independently of conceptualschemes. We cut up the world into objects when we introduce oneor another scheme of description. Since the objects and signs arealike internal to the scheme of description, it is possible to saywhat matches what.26

Reality itself does not fix the reference to any term, even terms such as“reality” and “exist.” There are no metaphysically privileged referencesor descriptions. All references and descriptions are context-sensitive andinterest-relative. They occur only in a given human situation and relativeto our aims or goals. As Putnam states in his Carus Lectures, “the trail ofthe human serpent is over all.” 27

Accordingly, the problem of how mind or language “hooks on to theworld,” which is how Putnam describes the problem of intentionality,

23 Contrary to some deconstructionists, Putnam does not see his view as threatened bysome regress of interpretation, because he thinks, with Wittgenstein, that we can follow arule without first having an interpretation attached to it. See Hilary Putnam, “Skepticismabout Enlightenment,” in Putnam, Ethics without Ontology (Cambridge, MA, and London:Harvard University Press, 2004), 120.

24 This example does not note some of the qualifications Putnam usually makes, but thisdoes not affect its usefulness here. See Hilary Putnam, “A Defense of Conceptual Relativity,”in Putnam, Ethics without Ontology, 38–40.

25 This also involves the issue of conceptual pluralism. Putnam notes: “The whole ideathat the world dictates a unique ‘true’ way of dividing the world into objects, situations,properties, etc. is a piece of philosophical parochialism. But just that parochialism is andalways has been behind the subject called Ontology.” Ibid., 51.

26 Hilary Putnam, Reason, Truth, and History (Cambridge: Cambridge University Press,1981), 52 (emphasis in original).

27 Hilary Putnam, The Many Faces of Realism: The Paul Carus Lectures (La Salle, IL: OpenCourt, 1987), 16.


only makes sense for him within a system of practices.28 It is not that oneappeals to the intentional character of our signs or concepts to explain howit is that our practices are able to refer to or describe some independentlyexisting reality. Indeed, Putnam thinks such a view is part of what he callsthe “magical theory of reference,” where “noetic rays” connect signs toreferents.29 Rather, our practices are fundamental.30 They are what maketalk about such things as “reference” and “reality” intelligible in the firstplace. Thus, the question of how thought or language hooks on to theworld, if understood as some underlying and transcendent metaphysicalproblem, is not genuine and is simply something we should get over.

Despite some of his own formulations of his position, Putnam does nothold that references, descriptions, and quantifications are all merely dif-ferent ways of providing form or structure to an independently existingreality. Putnam rejects the cookie-cutter metaphor in which our concep-tual contribution gives form or structure to the “dough,” the indepen-dently existing reality. The cookie-cutter metaphor assumes that, apartfrom our conceptual schemes, we can speak of the dough that “exists,”but this is the very thing that Putnam’s conceptual relativity denies. “Oneperson’s ‘existence’ claim may be another person’s something else.” 31 Wenot only cannot say what exists or what the facts are independent of ourconceptual choices, we cannot even speak of a reality to which we giveform or structure. As I have noted, for Putnam our conceptual choices orpractices are fundamental.

Perhaps the best statement, then, of what Putnam takes to be the centraland basic problem with metaphysical realism is as follows:

What is wrong with the notion of objects existing “independently”of conceptual schemes is that there are no standards for the use ofeven logical notions apart from conceptual choices. . . . The alter-native to this idea is not the view that, in some inconceivable way,it’s all just language. We can and should insist that some facts arethere to be discovered and not legislated by us. But this is some-thing to be said when one has adopted a way of speaking, a lan-guage, a “conceptual scheme.” To talk of “facts” without specifyingthe language to be used is to talk of nothing; the word “fact” nomore has its use fixed by Reality Itself than does the word “exist”or the word “object.” 32

28 Hilary Putnam, “Aristotle after Wittgenstein,” in Conant, ed., Words and Life, 63.29 Putnam, Reason, Truth, and History, 51.30 For an examination of this claim, see Douglas B. Rasmussen, “Realism, Intentionality,

and the Nature of Logical Relations,” Proceedings of the American Catholic Philosophical Asso-ciation 66 (1992): 267–77. See also Laird Addis, Natural Signs: A Theory of Intentionality(Philadelphia, PA: Temple University Press, 1989), 87–94.

31 Putnam, The Many Faces of Realism, 34.32 Ibid., 35–36 (emphasis in original).


There is no way to understand reference as pertaining to “reality” ifthat term means something that is transtheoretic/transconceptual/translinguistic.33

Putnam does not hold, however, that one cannot be a realist; indeed, hedifferentiates himself from what he calls Richard Rorty’s “extreme lin-guistic idealism which teeters on the edge of solipsism.” 34 But on Putnam’sview, one must be a realist who understands that referring to realities thatare what they are independent of the practices that constitute our con-ceptual schemes simply makes no sense.35 Ultimately, it is Putnam’s aimto show how it is possible to develop a “deliberate naïveté” or return towhat William James called “natural realism.” 36 Putnam understands thisview as common-sense realism37 and as being part of Ludwig Witt-genstein’s view that ordinary language provides a safe-haven from theraging conflicts over scientific and ontological commitments. As a result,Putnam thinks, contrary to Rorty, that there is a metaphysically innocentway of saying that our words represent things outside themselves.38 Yet,for Putnam, the journey from “the familiar to the familiar,” 39 which heregards as the true path of philosophy, still involves conceptual relativity,and it requires that we understand the meaning of “truth” in a mannerthat in no way involves a commitment to metaphysical realism.

B. Conceptual relativity and getting it right

Putnam forthrightly states that “conceptual relativity . . . holds that thequestion as to which . . . ways of using ‘exist’ (and ‘individual,’ ‘object,’etc.) is right is one that the meanings of the words in natural language,that is the language that we all speak and cannot avoid speaking everyday, simply leave open.” 40 Such conceptual relativity does indeed meanthat the right way of using terms is a matter of convention. Yet this is not

33 Putnam also notes that there is “the common philosophical error of supposing that theterm reality must refer to a single superthing instead of looking at the ways in which weendlessly negotiate —and are forced to renegotiate —our notion of reality as our language andour life develop.” Putnam, “Sense, Nonsense, and the Senses,” 9 (emphasis in original).

34 Putnam, “The Question of Realism,” 306. See Richard Rorty, Philosophy and the Mirrorof Nature (Princeton, NJ: Princeton University Press, 1979).

35 Putnam initially described this form of realism as “internal realism,” but he has morerecently described it as “pragmatic realism” or “realism with a small ‘r.’ ” See his “Reply toGary Ebbs,” in The Philosophy of Hilary Putnam, Philosophical Topics 20, no. 1 (Spring 1992):353.

36 Putnam, “Realism without Absolutes,” in Conant, ed., Words and Life, 284.37 Putnam states that there is “common-sense realism: the realism that says that moun-

tains and stars are not created by language and thought, and yet can be described bylanguage and thought. . . . [The] metaphysical kind of realism is ‘incoherent.’ ” Putnam,“The Question of Realism,” 303.

38 Putnam, “Are Values Made or Discovered?” 101.39 Putnam, “The Question of Realism,” 300. He also references John Wisdom, “Metaphys-

ics and Verification,” Mind 47, no. 188 (October 1938): 452–98.40 Putnam, “A Defense of Conceptual Relativity,” 43 (emphasis in original).


mere convention. When it comes to characterizing our knowledge of whatis the situationally appropriate use of a term, it would be a mistake to seePutnam as claiming that we can analytically discern the factual from theconventional. Rather, what is factual is relative to certain alternatives, andwhat is conventional is relative to certain others —and these alternativeschange. There are no unadulterated factual or conventional componentsof our knowledge.41 Putnam also understands “convention” to be a set ofpractices that are a response to problems that require a solution. Thealternatives arise because of the problems. Just as the problems we facecan vary —and thus the alternatives as well —the contrast in our accountof what is factual and what is conventional in our knowledge can vary.But the aim of our conventions remains one of meeting some difficulty orfulfilling some goal. Just as the convention to drive on the right-hand sideof the road in the United States and the convention to drive on the left inthe United Kingdom are solutions to a coordination problem, so are theconventions that determine what is the appropriate use of terms. Never-theless, this does not mean for Putnam that it is necessary to deny “thattruth genuinely depends on the behavior of things distant from thespeaker. . . . [Still, it is necessary to acknowledge that] the nature of thedependence changes as the kinds of language games we invent change.” 42

Indeed, he claims that

we have practices of interpretation. Those practices may be context-sensitive and interest-relative, but there is, given enough context —given, as Wittgenstein says, the language in place —such a thing asgetting it right or getting it wrong. There may be some indeterminacyof translation but it isn’t a case of “anything goes.” 43

For Putnam there is, then, a notion of truth, or, as he prefers, of “gettingit right,” that can be used to protect knowledge from the threats of extremeskepticism and relativism.

Putnam holds that just as Quine showed that the whole idea of classi-fying every statement as either “factual” or “conventional” is a hopelessconfusion,44 the same holds true for the fact/value dichotomy. Values areas much a part of the fabric of “facts” as conventions are, and they are justas ineradicable. It is impossible to determine what the facts are apart froman appeal to epistemic values. There is no way of telling what the facts areor what is true apart from such values as coherence, simplicity, predictive

41 Ibid., 45.42 Putnam, “The Question of Realism,” 309.43 Hilary Putnam, “The Craving for Objectivity,” in Conant, ed., Realism with a Human

Face, 122 (emphasis in original).44 W. V. O. Quine, “Two Dogmas of Empiricism,” in Quine, From a Logical Point of View

(Cambridge, MA: Harvard University Press, 1961), 20–46.


efficacy, preservation of past doctrines, beauty, naturalness, and so on.45

Our sense-experience does not provide us with some immediate access toreality but rather only becomes informative through our employment ofconceptual schemes, and since these schemes are themselves practicesthat we adopt in order to find solutions to problems, they already carrythese values with them. These values are presupposed by the very pos-sibility of our learning from experience and acting in the ways in whichwe do. As Putnam states:

Without the cognitive values of coherence, simplicity, and instrumen-tal efficacy we would have no world of facts, not even facts aboutwhat is so relative to what. And these cognitive values . . . are simplypart of our holistic conception of human flourishing.46

Thus, the fundamental question is not “How do we find values in a worldof facts?” Rather, it is “How do we find facts in a world of values?” 47

It is important to understand, however, that Putnam is not claiming tooffer merely an instrumental justification of these epistemic values. Instead,the very contention that we will, on the whole, come closer to truth aboutthe world by choosing theories that exhibit these epistemic values thanwe would by simply relying, for example, on the authority of some tra-dition is itself guided by these values. Putnam admits that this is circularjustification, but he thinks that we should make the most of it. To say thatthese epistemic values enable us to more correctly describe the worldthan any alternative set of epistemic values “is something we see throughthe lenses of those very values. It does not mean that those values admit an‘external’ justification.” 48 Those values are part of the practices by whichwe determine what is rational or warranted, and this is where the searchfor justification ends.

Of course, Putnam is arguing that we need to free ourselves from ourfixation with a powerful belief —namely, the idea that objectivity rests

45 Hilary Putnam, “The Entanglement of Fact and Value,” in Putnam, The Collapse of theFact/Value Dichotomy, 31; and Putnam, “Pragmatism and Moral Objectivity,” in Conant, ed.,Words and Life, 154.

46 Hilary Putnam, “Beyond the Fact/Value Dichotomy,” in Conant, ed., Realism with aHuman Face, 139 (emphasis in original).

47 Putnam does not deny that one can distinguish between facts and values; rather, hedenies that there is some metaphysical divide between them. See Putnam, “The EmpiricistBackground,” in Putnam, The Collapse of the Fact/Value Dichotomy, 19. Further, he states: “I . . .find it convenient to use the term valuings as a general term for value judgments of everysort. . . . [M]y position isn’t simply that ‘valuings are not descriptions’; my position is thatsome valuings, in fact, some ethical valuings, are descriptions (though not of anything‘nonnatural’), and some valuings are not descriptions. Valuings do not contrast simply withdescriptions; there is an overlap, in my view, between the class of descriptions and the classof valuings.” Putnam, “‘Ontology’: An Obituary,” in Putnam, Ethics without Ontology, 74(emphasis in original).

48 Putnam, “The Entanglement of Fact and Value,” 33 (emphasis in original).


fundamentally on a correspondence with some metaphysical reality. Tocounter this picture of objectivity and to help us overcome our craving forsome sort of ontological resting point, he notes Quine’s argument thatmathematics would be no less successful if we did not interact with theabstract entities some philosophers posit to account for its success.49

Putnam claims that the justification for accepting the truth of set theoryrests not on its correspondence with some metaphysical reality but in itsindispensability to mathematics, and since mathematics is one of theparadigms of knowledge, metaphysical worries about the existence ofsets do not carry much weight. Thus, Putnam wants us to see that whenthe chips are down, what we need in order to continue our practicestrumps the ontological concerns of “first philosophy.” Justification fun-damentally amounts to continuing the practices that allow us to solve ourproblems.

If we can only come to see that practice is primary in philosophy,50 thenthis realization will in turn help us to understand objectivity. As Putnamclearly states,

it is time we stopped equating objectivity with description. There aremany sorts of statements —bona fide statements, ones amenable tosuch terms as “correct,” “incorrect,” “true,” “false,” “warranted,”and “unwarranted” —that are not descriptions, but that are underrational control, governed by standards appropriate to their partic-ular function and context.51

Once we free ourselves from our craving for an objectivity that requirescorrespondence to metaphysical objects, then we can come to see theindispensable role that epistemic values play in our pursuit of the truth.

Putnam does differentiate ethical values from epistemic ones, but bothtypes of values are part of the set of practices by which we determinewhat is “objective.” According to what Putnam calls “the pragmatistenlightenment,” 52 ethical values are part of the ways in which humansdeal with problematic situations. These values are not concerned withsome special problem that exists in splendid isolation apart from the restof life’s problems. Rather,

49 Putnam, “Pragmatism and Moral Objectivity,” 153.50 Ibid., 154. Putnam does qualify this claim, however. He notes that the emphasis on

practice does not mean that we should “get rid of metaphysics once and for all.” In fact, hestates that “revisionary metaphysics is not always bad.” Ibid., 159–60. I will discuss theimportance of this qualification in Sections III and IV. Yet it is important to realize that, forPutnam, “revisionary metaphysics” includes not only those metaphysical views that wouldoverturn or replace our everyday practices and common-sense views, but also those meta-physical views that seek to explain such practices and views —that is, metaphysical viewsthat attempt to offer something more primary than practice.

51 Putnam, “The Entanglement of Fact and Value,” 33 (emphasis in original).52 Hilary Putnam, “The Three Enlightenments,” in Putnam, Ethics without Ontology, 96.


any human problem at all, insofar as it impacts our collective orindividual welfare, is . . . “ethical” —but it may also be at the sametime aesthetic, or logical, or scientific, or just about anything else; andif we solve a problem and cannot say, at the end of the day, whetherit was an “ethical problem” in the conventional sense of that term,that is not at all a bad thing. Thinking of logic, as Dewey did, as thetheory of inquiry and not as a branch of mathematics that happens tobe taught in philosophy departments, and of ethics as the relation ofinquiry to life —so the same book, e.g., Dewey’s Logic, viewed oneway is a text in logic (or in epistemology, even if Dewey disliked theword) and viewed another way is a book about social ethics —is, Ibelieve, the right way, indeed the only way, to open up the whole topicof ethics, to let the fresh air in. And that is an essential part of what“the pragmatist enlightenment” calls for.53

Thus, ethical values are for Putnam part of the process of inquiry, but hemeans “inquiry” in the widest sense of the term —namely, human deal-ings with problematical situations. Therefore, it is in the practices thatconstitute inquiry in this broad sense that we find ethical as well asepistemic values.

The mutual interpenetration of facts and ethical values is shown, Putnamnotes, in how many of our experiences are entangled with thick ethicalconcepts. For example, such concepts as “cruel,” “brave,” “generous,”“elegant,” “skillful,” “weak,” “vulgar,” and so on are counterexamples tothe fact/value dichotomy. He argues that we cannot successfully dividetheir meaning into descriptive and prescriptive components. If we try, forexample, to provide the “descriptive meaning component” of “cruel”without using “cruel” or a synonym, we find that we cannot do so.54

Alternatively, if we think that “brave” simply means “not afraid to risklife and limb,” then we would not be able to understand the distinctionbetween foolhardiness and bravery.55 If we will but look to our actualexperiences and not the world of sense data we inherited from the Britishempiricists, then we will find that value and normativity permeate ourexperiences.

In answer to the question of whether we make or discover values,Putnam also follows what he understands as John Dewey’s view. Putnamstates, “We make ways of dealing with problematical situations and wediscover which ones are better and which worse.” 56 Understanding whichways are better and which ways are worse is not, however, a matter ofsimply consulting our experience for Putnam. Though our experience is

53 Ibid., 107 (emphasis in original).54 Putnam, “The Entanglement of Fact and Value,” 38.55 Ibid., 39–40.56 Putnam, “Are Values Made or Discovered?” 97.


not “neutral” and comes to us “screaming with values,” 57 it would be anerror to suppose “that merely being valued, as a matter of experientialfact, suffices to make something valuable.” 58 Putnam writes:

Dewey’s answer to the question, “What makes something valuable asopposed to merely being valued?” in a word, is criticism. Objectivevalue arises . . . from the criticism of our valuations. Valuations areincessant and inseparable from all our activities, including our “sci-entific” ones, but it is by intelligent reflection on our valuations,intelligent reflection of the kind Dewey calls “criticism,” that weconclude that some of them are warranted while others are unwar-ranted. (Philosophy, by the way, is described by Dewey as criticism ofcriticism!)59

Putnam is aware, however, that with this response, he must now answerthe further question: What are the criteria by which we decide that somevaluations are warranted and some are unwarranted? Moreover, he isaware that he must do so in ways that do not require him to reintroducemetaphysical realism.

C. Idealized rational acceptability and the democratization of inquiry

Putnam continues to follow what he understands as Dewey’s view anddistinguishes three parts of Dewey’s answer to the question of whatwarrants our valuations.60

First, it should be realized that we are never in a position of having tojustify how we reason from an “is” to a most basic “ought,” because injudging the outcome of an inquiry we never come to it simply having astock of factual judgments and no value judgments. We always bringmany valuations that are not the specific object of dispute to bear on ourjudgment of the outcome of the inquiry. Warranting our valuations is likerebuilding a wooden ship at sea one plank at a time.61 We warrant ourvaluations by showing their ability to withstand criticism that arises fromproblems encountered, not by comparing them (either singly or wholly)

57 Ibid., 103.58 Ibid.59 Ibid. (emphasis in original).60 Ibid., 103–6.61 Putnam would not want the “sea” in this analogy to be interpreted as a reality that

exists apart from our conceptual practices —not even a Kantian “noumenal reality.” The sortof Kantianism that Putnam endorses considers metaphysical realism, even as a noumenalreality, unintelligible. See Putnam, “Model Theory and the ‘Factuality’ of Semantics,” inConant, ed., Words and Life, 361–62. The ship analogy originates with Otto Neurath. See OttoNeurath, “Protocol Sentences” (trans. George Schick), in A. J. Ayer, ed., Logical Positivism(New York: The Free Press, 1959), 201. This essay first appeared in German as “Protokollsätze,”Erkenntnis 3 (1932/1933).


to some metaphysical reality.62 The process of warranting is not free fromthe values that are inherent in the practices by which we have learned tosolve our problems.

Second, the authority of philosophy to judge the process of warrantingvaluations (that is, the authority of “criticism of criticism”) does not comefrom some store of philosophical knowledge or methods that provide acriterion or set of criteria. There is no appeal to metaphysical realities thatprovide standards for judging our valuations, no special access to whatdoes and does not warrant our valuations. The authority philosophy doeshave, however, is the authority of intelligence.

Third, what the authority of intelligence amounts to is what we havelearned63 about inquiry from the conduct of inquiry in general, and thisapplies to value inquiry in particular.64 What we have learned is to makefull use of human intelligence, and this in turn establishes the makeup thatinquiry must have if it is to determine what valuations are warranted.Thus, one warrants valuations by engaging in inquiry that meets the setof criteria that are the precondition for the full application of humanintelligence.

According to Putnam, to meet the criteria for the full employment ofhuman intelligence, inquiry

(1) must be a cooperative activity and not done in isolation. It mustinvolve working together to actively intervene with and manip-ulate the environment in order to form and test ideas. This in turninvolves trying different solutions to problems or at least consid-ering ways of solving problems others have tried and reflectingon the consequences;

(2) must treat its judgments as fallible —as open to being revised orfalsified;

(3) must respect autonomy, symmetric reciprocity, and follow prin-ciples that are much akin to those that characterize Jürgen Haber-mas’s “discourse ethics.” These principles are:(a) communicating honestly and authentically, trying to win by

the force of the better argument and not by manipulation orinvolvement in relations of hierarchies of dependence;

(b) not excluding persons affected by any proposed action fromparticipating in the discussion;

(c) not preventing the consideration of any proposal or the expres-sion of any person’s attitudes, wishes, or needs;

(d) neither assuming there will be no disputes over which valu-ations are warranted nor trying to eliminate all such disputes;

62 Problems, for Putnam, are part of the practices in which we are engaged.63 See Putnam, “The Three Enlightenments,” 89–108.64 Putnam, “Pragmatism and Moral Objectivity,” 170–77.


(e) keeping the communication going by not allowing coercionor violence or total refusal to discuss; and

(4) must insist on applying the criteria noted in (1)–(3) to more andmore institutions and relationships.65

For Putnam, as well as for Habermas, these criteria provide an account ofidealized inquiry or rational acceptability that we have learned from thedevelopment of science, and thus they constitute the modern conceptionof rationality. Since facts and values are interdependent and interpenetrat-ing, these criteria show us how to proceed in determining whether our judg-ments, be they of “fact” or of “value,” are warranted. Nonetheless, thecriteria do so without any appeal to metaphysical realities. They are, rather,the norms that we have learned from our practices. Our ability to applyintelligence fully to solving human problems presupposes these norms.

It should be emphasized that Putnam’s understanding of idealizedrational acceptability differs from that of Habermas in two significantways. First, Putnam does not, as he once did, regard truth as being definedas warranted assertibility under ideal conditions. He thinks that it ispossible for the truth about some subject matter not to be verifiable evenunder such conditions.66 Putnam comes to this realization, however, notfrom any acceptance of some metaphysical definition of truth (in fact, hedoes not think a definition of truth is necessary) but rather from anunderstanding of the practices of science and common sense. He findssuch a contingent relationship between truth and warranted assertibility“deeply imbedded in the world views of both science and commonsense.” 67 Yet Putnam is also anxious to combat the idea “that truth canalways transcend warranted assertibility under ideal (or good enough)conditions.” 68 He holds that idea to be as illicitly metaphysical as theideas of metaphysical realism and insists, “Truth cannot be so radicallynon-epistemic.” 69 Thus, he does not accept the idea that the norms thatcharacterize idealized rational acceptability are sufficient to determinethe truth in every case, but he does think that these norms suffice whentalking of familiar objects such as furniture in a room or when dealingwith the subject matter of ethics and law. He writes:

There is no reason to suppose that one cannot be what is called a“moral realist” in meta-ethics, that is, hold that some “value judg-

65 Ibid. See Jürgen Habermas, “Discourse Ethics: Notes on a Program of PhilosophicalJustification,” in Habermas, Moral Consciousness and Communicative Action, trans. ChristianLenhardt and Shierry Weber Nicholson (Cambridge, MA: The MIT Press, 1990), 43–115.

66 Putnam, “Are Values Made or Discovered?” 106–9.67 Hilary Putnam, “Values and Norms,” in Putnam, The Collapse of the Fact/Value Dichot-

omy, 124.68 Putnam, “Are Values Made or Discovered?” 107 (emphasis added).69 Ibid., 108.


ments” are true as a matter of objective fact, without holding thatmoral facts are or can be recognition transcendent facts. If some-thing is a good solution to a problematical human situation, thenpart of the very notion of its being a good solution is that humanbeings can recognize that it is. We need not entertain the idea thatsomething could be a good solution although human beings are inprinciple unable to recognize that it is. That sort of rampant Plato-nism is incoherent.70

For Putnam, truth in ethics and law does not transcend warranted assert-ibility under ideal (or “good enough”) conditions.

Still, if the norms that characterize these ideal conditions are to work inethics and law in determining (as well as we as fallible and limitedhuman beings can expect) what is true, then another perspective is needed.This involves the second way in which Putnam’s understanding of ide-alized rational acceptability differs from Habermas’s understanding —namely, in Putnam’s employment of an Aristotelian perspective when itcomes to understanding these norms. Putnam states that “in ethics weneed both Aristotelian and Kantian insights.” 71 As a result, there is forPutnam neither a norm/value nor a reason/inclination dichotomy. Heargues that it is impossible to understand the norms of idealized inquiryapart from valuations. The “‘maxims’ and ‘laws’ that we impose uponourselves by universalizing them, themselves contain value terms, in par-ticular the so-called ‘thick ethical words’ . . . such as ‘kind,’ ‘cruel,’ ‘imper-tinent,’ ‘sensitive,’ ‘insensitive,’ and so on.” 72 These terms are not neutraldescriptions of mere natural psychological impulses but are a part of ourattempts to achieve —and have our lives express —the goods, virtues, andintellectual skills that we take to be constitutive of human flourishing.These terms are the lenses of value concepts that give idealized inquiry apoint. Accordingly, we need to see the entire process of inquiry as func-tioning within the context of the development of human capabilities.

Thus, Putnam holds, contrary to Habermas, that our value judgmentsregarding goods and virtues are capable of being warranted. They mustnot be treated in a noncognitive manner if the norms of idealized inquiryare not to be empty. We need to employ the value concepts that aregenerated from our practices that make up the human activities of meet-ing opportunities and solving problems:

Contrary to the Kantian picture, in our moral lives we cannot and donot get by with a vocabulary obtained by supplementing a starklynaturalistic vocabulary with a single moral notion (the notion needed

70 Ibid., 108–9 (emphasis in original).71 Putnam, “Values and Norms,” 134.72 Ibid., 118.


to indicate that one is “imposing the form of law on psychologicallygenerated incentives,” say ought). Without our human manifold ofvalues, there is no vocabulary for norms . . . to be stated in.73

When it comes to making value judgments regarding the issues andproblems that arise when one seeks an admirable and worthwhile life, ifthere is no such thing as a correct answer, then to what does the effort oftrying to determine better reasons through idealized inquiry amount?Putnam’s answer to this question is ultimately “nothing,” for this effortwould not belong to our practices of living. We would have neither rea-son nor motivation to be involved.74

Putnam does not assume that there is a single conception of humanflourishing or that there might not be disputes about what ought to bedone. Moreover, he does not see it as philosophy’s task to produce a“final” system of ethics or set of rules of conduct. He does think, how-ever, that moral philosophy can participate in an examination and eval-uation of what first-order rules should govern our conduct by subjectingour valuations to criticism. This requires, as I noted before, engaging ina process of inquiry that meets the norms that express what it is tomake full use of human intelligence. That is to say, we need to exam-ine our valuations according to the norms of idealized inquiry or ratio-nal acceptability.

Such a process of inquiry also constitutes for Putnam “the democrati-zation of inquiry.” The norms of such inquiry not only tell us what isnecessary for warranting our valuations; they tell us what form of sociallife is necessary for the full application of human intelligence. Thus, if weunderstand “democracy” as social life that is organized in accordancewith the norms of idealized inquiry, then “democracy is . . . the precon-dition for the full application of intelligence to the solution of humanproblems.” 75 Moreover, Putnam makes it quite clear that this sense of“democracy” is to be understood as both “deliberative” and “social.”Taking what he understands as Dewey’s position, Putnam holds that thecitizens should do more than vote every so many years for legislators andother government officials. They should participate in the actual decision-making process.76 Further, Putnam holds that this process should issue ina social democracy:

73 Ibid., 119 (emphasis in original).74 For a criticism of Habermas from a point of view that is more or less Aristotelian but

does not assume the primacy of practice, see Douglas B. Rasmussen, “Political Legitimacyand Discourse Ethics,” International Philosophical Quarterly 32 (March 1992): 17–34.

75 Hilary Putnam, “A Reconsideration of Deweyan Democracy,” The Southern CaliforniaLaw Review 63 (1990): 1671.

76 Such direct participation does not require, however, that they refrain from calling uponexperts to assist them in their decision-making.


That our communities . . . should be social democracies follows fromthe fact that huge inequalities in wealth and power that we permit toexist effectively block the interests and complaints of the mostoppressed from serious consideration, and thus prevent any seriousattempt at the solution of such problems as the alleviation of stub-born poverty, or deeply entrenched unemployment, or the inferioreducational opportunities afforded to the children most in need ofeducation from ever getting off the ground.77

Finally, Putnam also makes it clear that a “liberal democracy” that pro-vides the framework for the protection of “negative liberty” will notsuffice, because it does not allow for “serious consideration” of the inter-ests and needs of those with lesser wealth and power. He quotes Deweyat length to drive this point home. Here is the salient part of Dewey’sstatement:

The notion that men are equally free to act if only the same legalarrangements apply equally to all —irrespective of the differences ineducation, command of capital, and that control of the social envi-ronment which is furnished by the institution of property —is pureabsurdity, as the facts have demonstrated. . . . This “free” action oper-ates disastrously as far as the many are concerned. The only possibleconclusion, both intellectually and practically, is that the attainmentof freedom conceived as power to act in accord with choice dependsupon positive and constructive changes in social arrangements.78

Trying to provide people with a political/legal order that merely protectstheir so-called negative liberty is not enough. Only a political/legal orderthat seeks to provide them with “positive liberty” is fundamentally con-sistent with a social system based on the demands of democratized inquiry.

III. Questioning Idealized Inquiry

In Section II, we learned that Putnam considers metaphysical realismunintelligible because there is no way to talk of what is real apart from aconceptual system. Further, we learned that on Putnam’s view conceptual

77 Putnam, “The Three Enlightenments,” 105.78 Putnam, “A Reconsideration of Deweyan Democracy,” 1696. Dewey also states: “Since

actual, that is effective [as opposed to original or native or ‘natural’], rights and demands areproducts of interactions, and are not found in the original and isolated constitution ofhuman nature, whether moral or psychological, mere elimination of obstructions is notenough. The latter merely liberate force and ability as that happens to be distributed by pastaccidents of history.” Ibid. The quotation is drawn from John Dewey, “Philosophies ofFreedom,” in H. M. Kallen, ed., Freedom in the Modern World (New York: Coward-McCann,1928), 249–50.


systems carry with them both epistemic and ethical values, and it is by anappeal to these values that we determine what it is “to get things right.”In fact, it is by appeal to these values that we determine “what is real.”Regarding the question of whether we construct or discover values, Putnamfollows Dewey in holding that “we make ways of dealing with problem-atical situations and we discover which ones are better and which worse.” 79

Nevertheless, it is ultimately by a set of procedures that make full use ofhuman intelligence that we determine what values are indeed objective orvaluable. It is by use of this set of procedures, which is called “idealizedinquiry,” and not by reference to metaphysical reality (particularly not byreference to anything about what human beings are) that we determinewhat is true for the ethical and political/legal orders. The basic issue forthis section, then, is whether this method of inquiry succeeds in deter-mining what is true or whether, in order to defend its very criteria andmake sense of itself, it must ultimately appeal to something beyond itself.I will proceed by raising a series of questions that begin at the politicallevel, work down through normative ethics, and come finally to the veryfoundations of idealized inquiry.

A. Whose freedom? Which way of expressing human intelligence?

Putnam makes no bones about the political implications of his accountof idealized inquiry; it requires a social-democratic political/legal order.Yet how does Putnam deal with the view that the process of democraticinquiry must itself be subject to strict limitation? How does he respond tothe view that holds that since individuals are ends and not merely means,there is a moral side-constraint on the political/legal order which pro-hibits it from sacrificing or using individuals in order to achieve ends notof their own choosing?80 Particularly, how does Putnam respond to Rob-ert Nozick’s claim that

whether it is done through taxation on wages or on wages over acertain amount, or through seizure of profits, or through there beinga big social pot so that it’s not clear what’s coming from where andwhat’s going where, patterned principles of distributive justice involveappropriating the actions of other persons. Seizing the results ofsomeone’s labor is equivalent to seizing hours from him and direct-ing him to carry on various activities. If people force you to docertain work, or unrewarded work, for a certain period of time, theydecide what you are to do and what purposes your work is to serveapart from your decisions. This process whereby they take this deci-sion from you makes them a part-owner of you; it gives them a prop-

79 Putnam, “Are Values Made or Discovered?” 97.80 Robert Nozick, Anarchy, State, and Utopia (New York: Basic Books, 1974), 30–31.


erty right in you. Just as having such partial control and power ofdecision, by right, over an animal or inanimate object would be tohave a property right in it.81

Putnam accepts, of course, the notion that democratized inquiry shouldtreat people as ends and not merely as means.82 Yet he considers Nozick’suse of that principle to be an example of a priori ethical argumentationthat “prove[s] too much,” 83 because this argument requires that there beno welfare spending —that, for example, taxation cannot even be used tofund orphanages. Putnam states that Nozick’s position “contradicts themoral outlook of the whole Western tradition. . . . The idea that there aretrade-offs between rights to property, protection of the poor and helpless,and other interests of the community has long been central to our moralpractice.” 84 Thus, what is wrong with an argument such as Nozick’s isthat it assumes that we can have moral knowledge of what ought to bedone apart from the system of practices that have developed in responseto the issues of living —be they individual or social issues. Further, suchan argument assumes that instances of ethical decision-making call forfinal solutions to moral problems, as opposed to adjudications or read-ings of what needs to be done that are part of an ongoing process of acommunity of inquirers.

Putnam thinks that ethical decision-making according to the norms ofidealized democratic inquiry should be understood in terms of twometaphors —adjudication, which is taken from law, and reading, which istaken from interpreting a text —and he thinks that this communitarianprocess requires the rejection of an “absolute perspective” in ethics andan acceptance of openness and nonfinality.85 The problem with the argu-ments of such analytic philosophers as Nozick is that they are “curiouslyevasive or superficial about the relation of the premises of these argu-ments to the ideals and practices of any actual moral community.” 86

Thus, whether individuals have the right not to have their lives, resources,and actions used for or directed to purposes for which they have notconsented is to be determined by the norms of the idealized inquiryprocess. The freedom to partake in this process is paramount for Putnamand overrides any claim (be it to negative liberty or anything else) that isnot first subjected to this process; and since the freedom to partake in thisprocess requires the positive liberty that a social democracy promotes, the

81 Ibid., 172 (emphasis in original).82 Putnam, “Values and Norms,” 115. Putnam lists respect for autonomy among the

criteria of idealized inquiry; see Putnam, “Pragmatism and Moral Objectivity,” 173.83 Hilary Putnam, “How Not to Solve Ethical Problems,” in Conant, ed., Realism with a

Human Face, 180.84 Ibid.85 Ibid., 181–83.86 Ibid., 180.


fundamental right not to have one’s life, liberty, or property directed orused without one’s consent is ruled out from the beginning.

Nonetheless, Putnam does not think that all issues remain problematic.Human slavery, racism, and male chauvinism are now “just plain wrong.” 87

Thus, it is possible for judgments, even ethical ones, to become settled andno longer subject to democratic deliberation. Yet, if this is so, then the ques-tion can be asked: What prevents the process of idealized inquiry from com-ing to accept what Nozick suggests —namely, that the use of taxation forthe achievement of distributive justice is itself a form of human slavery andthus just plain wrong? Maybe this process of inquiry has not yet reachedan adequate understanding of all that is involved in treating human beingsas ends and not merely as means. Indeed, maybe the right to private prop-erty will come to be seen as the fundamental expression of this very idea.Moreover, what prevents this process from concluding that respect for theprinciple that people are ends and not merely means requires that strictlimits be placed on what valuations are to be adjudicated and interpreted?

Putnam might reply: “Nothing prevents this, but that does not changethe initial judgment that Nozick’s position is too extreme and out of stepwith the Western moral tradition and practices of actual moral commu-nities.” However, it now appears that such a reply begs the question. Itassumes that having the ability to partake in the inquiry process —that is,having a social democracy which aims to achieve positive liberty —is thefundamental expression of the idea that people are ends and not merelymeans. Yet this is the very point that is at issue. Nozick’s position is, infact, a challenge to the idea that participation in this inquiry process is theparamount value. Is such participation really the fundamental expressionof human intelligence?

The questions do not stop here. Nozick’s position can be interpreted asan even more fundamental challenge —namely, are the criteria of idealizedinquiry something fixed once and for all? Are we clear about their mean-ing? Do we know, for example, what it means to respect autonomy or notengage in coercion? Given Putnam’s account of the criteria, Nozick’s posi-tion loses. However, might it not be necessary to modify or reinterpret thesecriteria in order to test Nozick’s position? In other words, could not theopenness of the idealized inquiry process and its willingness to considerany proposal require that it change some of its own criteria? If the processof validating our values is like rebuilding a wooden ship at sea one plankat a time, then there seems to be no reason not to consider such a changewhen it comes to examining the very criteria of idealized inquiry. Thus, there

87 Ibid., 183. Interestingly enough, Putnam never really explains how the process ofidealized inquiry concludes that slavery, racism, and male chauvinism are wrong. We knowwe believe they are wrong, but why should we think that our belief conforms to the resultof idealized inquiry? For additional concerns regarding Putnam’s account of democraticinquiry, see Matthew Festenstein, “Putnam, Pragmatism, and Democratic Theory,” The Reviewof Politics 57, no. 4 (Fall 1995): 693–721.


seems to be no reason to assume that idealized inquiry must continue touphold social democracy and not embrace a liberal democracy in whichindividuals have a moral space or territory that strictly limits the scope ofdemocratic deliberation. And finally, since we are asking fundamental ques-tions, what prevents the inquiry process from realizing that there are stan-dards beyond the inquiry process itself to which inquiry must appeal, andthat practice is not always or necessarily primary?

Putnam’s reply to this sort of questioning is to wave the bloody-shirt ofmetaphysical realism and to accuse anyone who asks such questions ofassuming what is not the case —namely, “that there is a workable philo-sophical notion of ‘objective fact’” 88 and that there is some standardbeyond human practices to which to appeal. This reply will be consideredin the final section of this essay, but it is worthwhile now to note what wasreferenced above. Putnam does not hold that practice always trumps overphilosophical reflection. The emphasis on practice does not mean that weshould get rid of metaphysics finally or that revisionary metaphysics isalways bad. “We owe many insights, in science and morals and in poli-tics, to various kinds of revisionary metaphysics.” 89 But if some meta-physical revision of our practices is okay, and some is not, the obviousquestion is: How do we determine which is which? Of course, Putnam’sanswer is to stay within the circle: the process of idealized inquiry itselfmust work out this question.

I will have more to say about this reply later, but what are we toconclude regarding Putnam’s claim that a social democracy must be aresult of the commitment to idealized inquiry? Has he shown this? Has heshown that a liberal democracy that makes negative liberty the para-mount value of the political/legal order is morally inferior to a socialdemocracy that makes the promotion of positive liberty the chief aim?Has he shown that the sort of freedom espoused by the classical-liberaltradition and its practices could not be the expression, or at least anexpression, of what the criteria of idealized inquiry demand? It certainlydoes not seem so. Indeed, it even seems that Putnam’s account of thecriteria of idealized inquiry begs the question against a position such asNozick’s. Much remains to be determined. The only thing we may safelyconclude is this: we do not know whose account of freedom or which wayof expressing human intelligence is to be preferred or why.

B. Ayn Rand and universalizability, or Asking questions you were alwaysafraid to ask

In holding that idealized inquiry could not conclude that negative lib-erty should be the paramount value for the political/legal order, Putnam

88 Putnam, “How Not to Solve Ethical Problems,” 184 (emphasis in original).89 Putnam, “Pragmatism and Moral Objectivity,” 160.


seems as guilty of a priorism as Nozick. If idealized inquiry is inspired bywhat it takes to make full use of human intelligence, then should not thepossibility be considered that a classical-liberal political order (and notsocial democracy) is the precondition for the use of human intelligence?Particularly, how does Putnam know that free and open markets mightnot be, as F. A. Hayek argued,90 the basic social mechanism by which touse human knowledge in coordinating and determining the best andmost useful ways to use resources, direct human effort, and solve prob-lems? Could not free and open markets themselves be a vital expressionof human intelligence? Indeed, it seems that Putnam’s account of thecriteria that determine the full use of human intelligence stacks the deck.It discounts the view that much of the most important knowledge that weuse in seeking human flourishing is not found in a community of know-ers examining a common issue, but is instead diversified in many indi-viduals and concerned with only particular and contingent matters. DespitePutnam’s claim that there must be a role for human flourishing in ideal-ized inquiry, he offers little or no discussion of how such diversifiedknowledge is brought to bear on determining what is best and mostuseful. In short, what is singularly lacking in Putnam’s work is any dis-cussion of the virtue of practical wisdom.91

When it comes to understanding the nature of ethics, Putnam’s accountof the criteria of idealized inquiry stacks the deck in even more funda-mental ways. His remarks about the relationship of the views of AynRand to those of Habermas reveal this clearly. He states: “Habermascertainly believes that one must act so as to treat the other always as anend, and not as a mere means. This is what the rational egotist violates.(Ayn Rand’s rejection of altruism is precisely a rejection of this formula-tion of the categorical imperative.)” 92 Putting aside such questions aswhether Rand might be more accurately identified as an egoist (ratherthan an egotist) or whether rejecting altruism is indeed equivalent torejecting concern for others, what is important to see is that Putnamconsiders Rand’s views as beyond the ethical pale. Her views fail to meetwhat Putnam regards as the minimum prerequisite for the ethical life —

90 F. A. Hayek, “The Use of Knowledge in Society,” American Economic Review 35, no. 4(September 1945): 519–30.

91 Practical reason is the intellectual faculty employed in guiding conduct, and practicalwisdom is the excellent use of practical reason. Practical wisdom is more than mere clev-erness or means-end reasoning. It is the ability of an individual at the time of action todiscern, in particular and contingent circumstances, what is morally required. It involves theintelligent management of one’s life so that all the necessary goods and virtues are coher-ently achieved, maintained, and enjoyed in a manner that is appropriate for the individualhuman being. It is the intellectual virtue of a neo-Aristotelian conception of human flour-ishing. See Fred D. Miller, Jr., “Rationality and Freedom in Aristotle and Hayek,” ReasonPapers, no. 9 (Winter 1983): 29–36. See also Douglas B. Rasmussen, “Capitalism and Moral-ity: The Role of Practical Reason,” in Robert W. McGee, ed., Business Ethics: Social Respon-sibility in Business (New York: Quorum Books, 1992), 31–44.

92 Putnam, “Values and Norms,” 115.


that is, having a community of human beings who consider the ends of othersto be as important as their own.93 Indeed, it is clear that Putnam notonly thinks this, but also feels it, because he resorts to personal attack. Hesays of Rand that she is an “influential if amateurish philosophizer —Icannot call her a philosopher.” 94 Rand’s views are not part of the prac-tices that constitute ethical thinking, talking, and acting as far as Putnamis concerned.

It is important to note here that Rand does indeed hold that eachhuman being is an end in her- or himself and not a means to the ends ofothers.95 Moreover, she holds that one should not treat others as meremeans, but the point is that she also holds that someone else’s humanflourishing should not be as important for a person as that person’s own.The development of one’s own life and character is each person’s primaryethical obligation. There is thus a distinction drawn between recognizingthat each human being is an end in her- or himself and requiring that onevalue the lives of others as much as one values one’s own. Thus, it maybe that her views are not as easily dismissed as Putnam seems to believe.My purpose here is not to engage in an evaluation of Rand’s ethicalviews. Rather, my purpose is to show that Putnam is actually assumingsomething that is quite controversial in his account of the criteria ofidealized inquiry, and that he is stacking the deck against all ethicaltheories that see ethics as resulting from an individual’s basic concernwith her or his own flourishing or self-perfection.96

Specifically, Putnam is assuming that the Categorical Imperative or theprinciple of universalizability97 requires that one should approach all moraljudgments in an impersonal fashion. This means that all morally salient

93 Ibid. Putnam regards Kant’s great achievement in moral philosophy to be the Categor-ical Imperative. He understands by this “the idea that ethics is universal, that insofar asethics is concerned with the alleviation of suffering, it is concerned with everyone’s suffering,or if it is concerned with positive well-being, it is with everybody’s positive well-being.”Hilary Putnam, “Ethics without Metaphysics,” in Putnam, Ethics without Ontology, 25. For acritique of this type of reasoning, see Douglas B. Rasmussen and Douglas J. Den Uyl, “InSearch of Universal Political Principles: Avoiding Some of Modernity’s Pitfalls and Discov-ering the Importance of Liberal Political Order,” in Stephen Elkin and Stephen Simon, eds.,Seeking Common Principles of Justice: The Prospects, Challenges, and Risks of Universalism in aDivided World (University Park: Pennsylvania State University Press, 2008).

94 Putnam, “Values and Norms,” 114.95 Ayn Rand, “The Objectivist Ethics,” in Rand, The Virtue of Selfishness: A New Concept of

Egoism (New York: New American Library, 1964), 27.96 I hasten to add that these theories need not be understood as necessarily egoistic, at

least in the usual sense of that term, for it is quite possible for the welfare of other persons,though not every person, to be an essential feature of, and not a mere means to, one’s ownflourishing. Moreover, an ethics of human flourishing or self-perfection neither denies theprofoundly social character of human life nor assumes an atomistic perspective. See Doug-las B. Rasmussen and Douglas J. Den Uyl, Norms of Liberty: A Perfectionist Basis for Non-Perfectionist Politics (University Park: Pennsylvania State University Press, 2005), 127–43; andDouglas J. Den Uyl and Douglas B. Rasmussen, “The Myth of Atomism,” The Review ofMetaphysics 59, no. 4 ( June 2006): 841–68.

97 He also uses the term “symmetric reciprocity.”


values, reasons, and rankings are construed as “agent-neutral,” and whatmakes them so is that they do not involve as part of their description anessential reference to the person for whom the value or reason exists or theranking is correct. As Douglas J. Den Uyl notes, “For any value, reason orranking V, if a person P1 is justified in holding V, then so are P2–Pn underappropriately similar conditions. . . . On an agent-neutral conception it isimpossible to weight more heavily, or indeed to weight at all, V, simplybecause it is one’s own value.” 98 On an agent-neutral view, when it comesto describing a value, reason, or ranking, it does not ethically matter whosevalue, reason, or ranking it is. In other words, it does not matter ethicallyspeaking whose human flourishing it is.

In contrast to this view of morality, there is the agent-centered view thattreats morally salient values, reasons, and rankings as fundamentallypersonal. What makes them so is that they are ultimately “agent-relative,”and they are agent-relative if and only if their distinctive presence inworld W1 is a basis for some person P’s ranking W1 over world W2, eventhough they may not be the basis for any other person’s ranking W1 overW2. Simply put, there are no morally salient values, reasons, or rankingsfull stop. Rather, they are always and necessarily values, reasons, andrankings for some person or other. Human flourishing is always andnecessarily human flourishing for some person or other.

Universalizability as a logical matter does not mean or require imper-sonalism. Let us say that values, reasons, and rankings V1, for a person P1,are agent-relative, and further let us say the same holds true, respectively,for values, reasons, and rankings V2–Vn for persons P2–Pn. Conduct basedon such agent-relative values, reasons, and rankings can be universalizedas follows: if V1 provides a basis for P1 to act, so does V2 provide a basisfor P2 to act, as well as Vn for Pn. One cannot claim that V1 provides P1with a legitimate reason to act without acknowledging that V2 provides P2with a legitimate reason to act, and so on. What is universalized here isthe knowledge that something’s being a value, reason, or ranking forsome person is what provides a basis for that person’s conduct. Yet noth-ing about universalization requires that what is a value, reason, or rank-ing for one person must be so for another.99

It might be objected, however, that the account of universalizabilitybeing used here is exceedingly thin and that Putnam might respond thatthe conception of universalizability used by idealized inquiry is thicker —

98 Den Uyl, The Virtue of Prudence, 27. See also Eric Mack, “Moral Individualism: AgentRelativity and Deontic Restraints,” Social Philosophy and Policy 7, no. 1 (1989): 81–111.

99 However, agent-relativity alone does not require that P1 be an ethical egoist, for it isperfectly possible for P1’s morally salient values, reasons, or rankings to be agent-relativeand entirely altruistic. Yet this possibility does not show that these values, reasons, andrankings are agent-neutral, because benefiting others must still be a value, reason, or rank-ing for P1, and not necessarily for P2–Pn. Further, there is still the question of whether it isdesirable for P1 to be entirely altruistic. See Rasmussen and Den Uyl, Norms of Liberty,134–36; and Rasmussen, “Human Flourishing and the Appeal to Human Nature.”


namely, universalizability involves rules of conduct that all rational indi-viduals would agree upon. If we understand universalizability in thisway, then we might be able to see that the ethical life cannot be under-stood in a personal and agent-relative manner. Of course, the responsehere is simply: What is the basis or standard that rational individuals usein determining which rules to agree upon? If there is no basis or standard,then what is it about these individuals that justifies the adjective “ratio-nal”? If there is a basis or standard (let us say some value, reason, orranking V), then we still have the question of whether V should be con-sidered in an agent-neutral or an agent-relative manner. We are back towhere we started. Clearly, this gambit will not work without much deeperpremises.

There is also the question of just what makes V desirable or choice-worthy for someone, but Putnam cannot answer that question apart fromthe idealized inquiry process. There can be no appeal to the nature ofhuman beings to understand human flourishing, for that would involveus in metaphysical realism. Human flourishing must itself be understoodin terms of the criteria of idealized inquiry, and thus, given Putnam’sapproach, there is no way to warrant valuations of personal flourishing.There is no way to see ethics as a practice that is concerned primarily withone’s own human flourishing or self-perfection.

Indeed, Putnam is quite explicit about this. He holds that one must beeducated into the ethical life and believes that this involves a transfor-mation of one’s understanding of one’s good such that one thinks interms of “we” and not simply “me” and does not look outside the prac-tices that constitute “ethical living” for reasons to be ethical.100 If this isso, however, this impersonalist rendering of universalizability creates aproblem for Putnam’s claim that practice is primary. Why should oneever agree to follow these so-called ethical practices? Beyond name-calling, what can Putnam say in reply to those who ask why one ought tobe moral in an impersonal sense? Since the very intelligibility of speakingof ethical matters in this impersonal way requires the personal to whichit is contrasted, it is perfectly coherent to ask why it would be good,worthy, or appropriate for me to adopt these “ethical” practices. So, whatcan Putnam say?

Of course, Putnam can say that following certain practices is not merelya matter of choosing to do something. Rather, the practices are actuallypart of our understanding of ourselves as well as how we think and act,and as such are not so easily overcome. However, this only shows justhow radical someone may need to be in order to get us to reconsider orrevise certain practices. Why isn’t someone like Rand, who asks questionsabout the very legitimacy of certain practices, an example of someonewho is offering a revisionary metaphysics or, more precisely, a revisionary

100 Putnam makes this point in many places, but see Ethics without Ontology, 3, 29, and 102.


way of thinking about ethics? Putnam’s account of ethical knowledgereally has no way to deal with someone like Rand. Alternatively, hisapproach to ethics seems particularly blind to those ethical traditions andpractices that see the ethical life as primarily a project of personal devel-opment, and, as such, his approach seems to imply that ethical practicesare unconnected to the lives of real people.

C. The human capabilities approach, or Legislating for human flourishing

It might seem unfair or simply inaccurate to accuse Putnam of hav-ing a blind spot to an ethics of personal development. In fact, part ofhis motivation for attacking ethical noncognitivism101 is to express hissolidarity with Martha Nussbaum’s and Amartya Sen’s attempts to makeuniversal development of human capabilities the aim of both politicaltheory and welfare economics.102 In other words, Putnam wants tomake a case for moral knowledge so that he can make the develop-ment of human capabilities the standard by which to legitimate polit-ical projects and economic institutions. In turn, he sees this as furtheringhis social-democratic ideals and advancing a version of egalitarian lib-eralism. And while there is certainly nothing wrong in desiring to assistor help as many people as possible with the development of theirhuman capacities, there is an issue that Putnam’s approach to ethicalknowledge passes over in silence —specifically, why should one assumethat developing human capabilities is the aim of the political/legalorder? More generally, what is it that justifies moving from the ethicalorder to the political/legal order?

Simply put, no great issue seems to be at stake for Putnam in deter-mining what it is that justifies making an ethical concern a political/legalone. Following Dewey, as well as Hegel, he holds that “we are communalbeings from the start,” and even the idea, as a thought experiment, thatwe are beings who belong to no community is “utterly fantastic.” 103

Accordingly, as I have indicated, Putnam sees the ethical process as largelya communitarian affair, and thus the discussion is already in terms ofwhat “we” ought to do. The ethical and the political, if they are not

101 Ethical noncognitivism holds that moral claims are not knowledge claims.102 See the following works by Martha C. Nussbaum: Frontiers of Justice: Disability, Nation-

ality, Species Membership (Cambridge, MA: Harvard University Press, 2006); Women andHuman Development: The Capability Approach (Cambridge: Cambridge University Press, 2000);“Non-Relative Virtues: An Aristotelian Approach,” in Martha C. Nussbaum and AmartyaSen, eds., The Quality of Life (Oxford: Clarendon Press, 1993), 242–69; “Human Functioningand Social Justice”; and “Aristotelian Social Democracy,” in R. Bruce Douglass, Gerald Mara,and Henry Richardson, eds., Liberalism and the Good (New York: Routledge, 1990), 203–52.See the following works by Amartya Sen: Development as Freedom (New York: Anchor Books,2000); “Capability and Well-Being,” in Nussbaum and Sen, eds., The Quality of Life, 30–53;Inequality Reexamined (Cambridge, MA: Harvard University Press, 1992); and Ethics andEconomics (Oxford: Blackwell, 1987).

103 Putnam, “The Three Enlightenments,” 107.


identical for Putnam, are certainly coextensive. There is certainly no prin-cipled difference between them.

Clearly, this is a major gap in Putnam’s thinking. If we were ancientGreeks living in a polis, it might be reasonable to see no difference betweenone’s community or society and the political/legal order, but this is cer-tainly not the case for humankind now. It is one thing to belong to acommunity or society and quite another thing to be subject to the laws ofa political/legal order. Thus, just what is it, if anything, that legitimates apolitical/legal order? Furthermore, assuming there is some connectionbetween the ethical order and the political/legal order, what is the natureof the connection between the two? That is to say, what is it that justifiesmoving from the judgment that X-ing is right (or wrong) and ought (orought not) to be done to the judgment that X-ing ought to be politically/legally required (or prohibited)? These two judgments are not semanti-cally equivalent, and a political philosopher needs to know what justifiesconnecting the two and thus how political and legal institutions might belegitimated.104 Indeed, this is the datum explanandum of political philosophy.

However, it may be that Putnam agrees with Alasdair MacIntyre thatwe should not think in terms of political/legal orders or in terms ofnation-states but rather in terms of much smaller political units wherethe search for human flourishing is supposedly a much more cooper-ative affair and where everyone aims at the same goal.105 But even ifthis were so, there would still be a gap in Putnam’s thinking. Attempt-ing to develop human capabilities through laws that are supposed toapply to everyone in a community raises the problem of preferring oneform of flourishing to others. How can Putnam seek to develop humancapabilities through universal laws and not ultimately end up creatinga system that, as a matter of principle, requires the sacrifice of oneform of flourishing to another —that is to say, a system that treats peo-ple as mere means?

This is, of course, the well-known problem of “liberal neutrality.” How-ever, it is important to see this issue clearly, or the difficulty it presents forPutnam will not be fully appreciated. This issue does not concern whether,as a result of the everyday conduct of individuals in various communitiesand societies, certain forms of human flourishing are easier to achievethan others are, or even whether some are in fact impossible to achieve.Indeed, if that were the issue, then the attempt to have a society that is“neutral” to all the various forms of flourishing would be an illusory goal,because there are always more factors involved in the actual practice ofachieving flourishing than can be legislated or planned. Such a goal would

104 The point here is not that a political conversation is separate from an ethical one, orfrom any other one for that matter. Rather, the point is that there is a distinction —a differ-ence in the issue that is being addressed.

105 Alasdair MacIntyre, Dependent Rational Animals: Why Human Beings Need the Virtues(Chicago and La Salle, IL: Open Court, 1999), 99–118.


be a case of hubris born of an ethical rationalism that fails to see thecomplexity and particularity of individual moral lives.

However, trying to achieve “liberal neutrality,” in the foregoing sense,is not the issue. The issue is one of trying to determine what political/legal principles, if any, will provide parameters for conduct in societythat, as a matter of principle, will not require that any form of flourishingbe preferred structurally to another. It is thus an issue of structural neu-trality. Can one even find a basis for such political/legal principles, andwhat might they be? Douglas J. Den Uyl and I have expressed this issuein a set of questions, which we have called “liberalism’s problem”:106

How . . . is the appropriate political/legal order —the order that pro-vides the overall structure to the social/political context —to be deter-mined? What is its ethical basis? Since the structure provided by thepolitical/legal order will rule over all equally, how can the univer-salism of political/legal structural principles square with the pluralismand self-direction required by human flourishing? Hence, how is itpossible to have an ethical basis for an overall or general social/political context —a context that is open-ended or cosmopolitan —that will not require, as a matter of principle, that one form of humanflourishing be preferred to another? How, in other words, can thepossibility be achieved that various forms of human flourishing willnot be in structural conflict?107

To what, if anything, in an account of human flourishing can one appealto solve this problem? To the extent that Putnam seeks to advocate apolitical/legal order that applies to everyone in order to develop capa-bilities for human flourishing, he needs to answer this question.

Putnam is, of course, correct to hold that political and ethical problemscannot be solved without there being some moral knowledge, and thusthe “justificatory neutralism” of much contemporary liberal theory issimply a nonstarter. There must be moral knowledge. Putnam is alsocorrect to hold that many of the underlying metaphysical and epistemo-logical views of analytic philosophy are insufficient to provide a basis forthis knowledge. Moreover, he realizes that neither the capabilities forflourishing nor human flourishing itself are Platonic Forms. They come inmany variations. However, he does not seem to realize fully that even ifso-called generic goods and virtues, or, more precisely, the capabilities forpursuing these goods and virtues, must in fact be present in every form

106 The reason we label it so is that liberalism has been, largely, the only political traditionto recognize the fundamentality and importance of this problem. Yet this does not mean thatliberalism has always done so coherently. See Rasmussen and Den Uyl, Norms of Liberty,chap. 3.

107 Ibid., 271 (emphasis in original).


of flourishing, their exact character and role in each person’s life will varyfrom individual to individual. Individuals are more than mere reposito-ries for these capabilities or generic goods and virtues. The circumstances,interests, histories, endowments, and, most importantly, choices of indi-viduals give shape, form, and value to these goods and virtues. Individ-uals do more than simply locate these goods and virtues or capabilities inspace and time. Thus, attempting in the name of human flourishing tofind an ethical basis for political/legal principles that will apply to allequally and achieve structural neutrality requires that Putnam deal withthe individuality of the practices that constitute human flourishing.So-called generic goods and virtues or human capabilities do not providea solution. The point here is not only that human flourishing is agent-relative in its fundamental character but that it is individualized as well.108

Finding a solution to liberalism’s problem thus cannot be achieved at theprice of adopting an impersonalist reading of human flourishing.109 Regret-tably, Putnam never addresses this problem. Not only do Putnam’s cri-teria of idealized inquiry make no principled distinction between an ethicalconcern and a political/legal one, they also take little or no notice of theagent-relative and individualized character of human flourishing. Thus,my earlier charge that Putnam’s approach to ethics seems unconnected tothe lives of real people seems both fair and accurate.110

D. Questioning the criteria of idealized inquiry

Despite Putnam’s admission that practice does not always trump otherconcerns and that there is a place for revisionary metaphysics, it is quiteclear that he is going to insist that any ethical or political dispute beadjudicated or interpreted by an appeal to the criteria of idealized inquiry.After all, one cannot appeal to the nature of things. It is only through ouridealized inquiry procedure that we may determine what beliefs andvaluations are warranted. This much is clear. However, are the criteria

108 Human flourishing involves an essential reference to the individual for whom it isgood as part of its description. Further, it is not merely achieved and enjoyed by individuals.It is itself individualized. See Rasmussen and Den Uyl, Norms of Liberty, 132–38.

109 There are two related points here. First, to hold that human flourishing is “objective”does not mean (or require) that it amounts to the same activity for each person in practice.Human flourishing is not one-size-fits-all. Thus, it is not sufficient either for determining theaim of the political/legal order or for developing a standard for welfare economics. (On itsinsufficiency for welfare economics, see David Gordon’s review of Putnam’s book TheCollapse of the Fact/Value Dichotomy, “The Facts of Economic Life,” The Mises Review [Winter2002], http://www.mises.org/misesreview_detail.asp?control=217&sortorder=auth.) Sec-ond, structural neutrality is a function of recognizing real diversity. Ethical impersonalismis a function of dismissing or flattening real diversity.

110 Real people are engaged in a quest for answers and solutions. This means that whatthey say or conclude is often less important than the basis for their conclusions. This bringsus back to the idea that there is something more than human practices when it comes tofinding answers and solutions.


themselves so clear? Are they helpful in addressing basic ethical issues?We have already seen that the criterion of freedom or noncoercion thatPutnam employs is very much a matter of dispute, as is the claim thatevery valuation must be subject to examination. Indeed, do we reallywant to hold that every relationship must be evaluated in terms of thesecriteria? Further, we have seen that the Categorical Imperative or princi-ple of universalizability does not necessarily carry with it an impersonal-ist ethical perspective, and thus cannot be used as a basis for rejecting apersonal or agent-relative view of the ethical life. It offers no way tochoose between personalism and impersonalism. More importantly, theview of ethics that idealized inquiry generates also ignores the long-standing traditions of ethics that make the development of one’s ownself-perfection or flourishing the central aim of the ethical life, and thusthis view separates ethical practices from actual human life.

The problems with the criteria of idealized inquiry do not end here.There is also the question of whether the criteria are themselves adequatefor attaining ethical knowledge or whether they stand in need of some-thing more. What follows is a series of interrelated questions. Each ofthese questions could be developed in detail, though I do not developthem here. The questions should be sufficient, however, to indicate someof the grave difficulties that idealized inquiry faces.

1. What if the process of idealized inquiry had not concluded thatslavery and racism were “just plain wrong”? Wouldn’t these ways oftreating human beings nonetheless still be wrong? If truth is not definedby warranted assertibility under ideal conditions, as Putnam now admits,why must the criteria for warranted assertibility under ideal conditionsbe the last word in determining ethical truths? If such criteria might notsuffice in the scientific arena, then might they not also be insufficient inthe ethical arena? Despite his intention to the contrary, is it possible thatPutnam is maintaining a fact/value dichotomy, or at least some version ofit? Is there a polarity between scientific fact and ethical value after all?

2. Since the strategy behind the rejection of metaphysical realism was toshow, as Nussbaum put it, “that ethics is no worse off than any sci-ence,” 111 it certainly seems that Putnam does not want to maintain thatthe standards for truth in science are different from those in ethics. But ifso, then what is the basis for ethically dismissing the practices of slaveryand racism? Are the practices of slavery and racism just plain wrongbecause the process of idealized inquiry judges them to be, or does theprocess of idealized inquiry judge the practices of slavery and racism tobe just plain wrong because of what these practices are?112

111 Nussbaum, “Human Functioning and Social Justice,” 214.112 It might be replied that Putnam’s decision not to define truth as warranted assertibility

under ideal conditions is based on understanding that the contingent relationship betweentruth and warranted assertibility is deeply imbedded in the worldviews and practices ofboth science and common sense. Thus, Putnam seeks to take advantage of metaphysical


3. Putnam regards metaphysical realism as unintelligible; thus, he can-not appeal to the nature of slavery and racism as they are apart from theconceptual scheme employed by a community of knowers. Yet, if there isno metaphysical difference between what these practices are and how acommunity of knowers conceptualizes them, then what these practicesare (which includes how they are valued) will vary with different com-munities of knowers. What, then, can Putnam say to a community thatdoes not employ the same conceptual scheme or share the same valua-tions as Putnam’s? Are slavery and racism just plain wrong only for us inour community?

4. Putnam wants to insist that without a sense of community we cannotbegin to adjudicate ethical problems successfully.113 Nevertheless, uponwhat is this sense of community supposed to be based? This question isnot grounded on any atomistic conception of human beings but rather onthe recognition that human sociality comes in various forms and at numer-ous levels. So, among the many social practices that make up the humanform of life, what determines which practices are relevant for having acommunity that has the requisite sense of “ethical” practices? What enti-tles us to expect that the communities that develop will be based onethical values that “we” find legitimate?114

5. If Putnam is to remain consistent, he cannot appeal to some set ofhuman practices that express values that are valuable apart from ideal-ized inquiry. This he regards as Platonism. He might avail himself of atranscendental argument that attempts to show that certain epistemicvalues are necessary for the very possibility of our speaking of what is“true” or “objective,” but even if we suppose that such a gambit works,there remains an ultimate difficulty: Putnam never actually shows thatthese epistemic values carry with them any substantive ethical values.What he shows is that human experience is shot through with valuations,but he does not show that the valuations are valuable in the sense thatthey are worthy of being chosen or followed. In other words, Putnam mayindeed show that there is no ontological chasm between facts and values,but he does not show that values are valuable in the sense that theythemselves ought to be valued.

6. The ultimate value for Putnam seems to be human intelligence, andhe is concerned that we make full use of it. Particularly, he is concerned

realism’s view of truth while at the same time maintaining the primacy of practice. I discussthe problems with this maneuver in Section IV.

113 Putnam, “How Not to Solve Ethical Problems,” 185. See also Putnam, “The ThreeEnlightenments,” 101–4.

114 Furthermore, how is “we” idealized? Why idealize along the lines of educated cos-mopolitans rather than middle-class folks with mediocre educations, or the uneducated, orthose who have to live with the results, and so on? It is never clear how this is done, except,of course, to assume simply that universality and consistency are the standards. Yet whymust these values be the standards? Why must science and the values of modernity be ourguide? Obviously, these questions are not just theoretical any longer, if they ever were.


that we make full use of what we have learned from the process ofinquiry, especially what we have learned from scientific methods andprocedures. This is what is supposed to guide us, and this is what makesthe criteria of idealized inquiry so important. They are the second-orderprinciples that we use to judge our first-order rules as well as conduct.Nonetheless, for what purpose is the full use of intelligence to be employed?If we answer simply “employing intelligence even more,” we have onlywidened the circle and explained nothing. We can quite intelligibly ask,“What is intelligence for?” Is it for anything that we choose? Are anychoices better than others are? And if so, why? Does it matter? Intelli-gence is a valuable tool, but we do not know its function. The criteria ofidealized inquiry provide no answer.

7. It might be said, however, that intelligence is for knowledge. Thepursuit of knowledge is certainly a fundamental human practice. How-ever, does this show that knowledge, even scientific knowledge, is thestandard by which human conduct and institutions are to be judged?Does this even show that we ought to pursue knowledge?

8. Can we not also ask, “What is knowledge for?” Knowledge is, asPutnam admits, part of a holistic conception of human flourishing, and hefurther admits that human flourishing is an activity that is not uniformlythe same. Yet, if this is so, then where are we to rank knowledge when itcomes to other human practices such as pursuing leisure, beauty, plea-sure, or friendship? And even if the pursuit of knowledge is more basicthan these other pursuits, does it follow that pursing knowledge ought inevery case to be valued more than these pursuits? Certainly, it is possiblefor there to be forms of flourishing, as well as communities and societies,in which knowledge is not given as much importance as the values ofleisure, beauty, friendship, or pleasure. Therefore, there might be personsand societies whose practices do not value the criteria of idealized inquiryas much as we do, and thus the criteria of idealized inquiry may not haveas privileged a place as Putnam supposes.

This section has sought to engage in a “criticism of criticism” thatshows that something more than the criteria of idealized inquiry is neededif there is to be ethical knowledge. If this criticism has succeeded, then aquestion asked earlier is more pertinent than ever. What, if anything,prevents the idealized inquiry process from concluding that there arestandards beyond itself to which it must appeal? The answer is simply“nothing.” Nothing prevents there being a transcendent standard. To saythat human practices are necessary for and inclusive of all explanations ofethical knowledge does not show that these practices are a complete andsufficient explanation of ethical knowledge. Moreover, the criteria of ide-alized inquiry do not themselves prevent the recognition of standardsbeyond the idealized inquiry process. There is nothing about the impor-tance of cooperative activity, the admission of fallibility, making one’sbeliefs open to revision, universalizability, honesty, sincerity, and non-


coercion that is incompatible with there being some standard beyondinquiry itself to which it can appeal. Thus, it is time to consider what elseis necessary for ethical knowledge. This is, of course, a tall order and isnot something that can be completed here, but now that the adequacy ofthe criteria of idealized inquiry as a basis for moral knowledge hasbeen called into question, the first step can be taken —namely, a return torealism.

IV. Reconsidering the Demise of Metaphysical Realism

A. Getting clear on metaphysical realism and Aquinas’s distinction

It should be recalled that metaphysical realism involves both an onto-logical thesis and an epistemological thesis. The ontological thesis is thatthere are beings that exist and are what they are independent of and apartfrom anyone’s cognition. The epistemological thesis is that these beingscan be known, more or less adequately, often with great difficulty, but stillknown as they really are. A number of things need to be said about eachof these theses as well as Putnam’s approach to them.

To hold that there are beings that exist and are what they are indepen-dent of and apart from anyone’s cognition of them does not require thatthe natures of these beings are universals existing either ante rem or inrebus. In other words, there is no need to assume that ontological realismrequires the existence of universal natures either in some supersensiblereality or as metaphysical parts or elements of individual beings. It isperfectly possible that while all individuals are natured, all natures areindividualized, and it is therefore possible to reject the idea that realitycomes already conceptually divided for us without rejecting the idea thatbeings are what they are apart from our conceptual processes.115

The debate regarding the status of universals has a long and importantphilosophical history, and much of Putnam’s aversion toward metaphys-ical realism seems to be born of his desire to avoid “Platonism” withregard to mathematical entities. Much depends here, of course, on one’sview of abstraction,116 but it is strictly speaking not necessary to affirm

115 See Douglas B. Rasmussen, “Quine and Aristotelian Essentialism,” The New Scholas-ticism 58 (Summer 1984): 316–35; and Douglas B. Rasmussen, “The Significance for Cogni-tive Realism of the Thought of John Poinsot,” American Catholic Philosophical Quarterly 68(Summer 1994): 409–24. See also John P. O’Callaghan, Thomistic Realism and the LinguisticTurn (Notre Dame, IN: University of Notre Dame Press, 2003), 194–98 and 257–74; and PaulA. Boghossian, Fear of Knowledge: Against Relativism and Constructivism (Oxford: ClarendonPress, 2006), 35–38.

116 I cannot discuss this in detail here. For some important insights regarding the “mod-erate realist” approach to abstraction, see Joseph Owens, Cognition: An Epistemological Inquiry(Houston, TX: Center for Thomistic Studies, 1992), 139–65; and Owens, “Common Nature:A Point of Comparison between Thomistic and Scotistic Metaphysics,” Mediaeval Studies 19(1957): 1–14. See also Thomas Aquinas, On Being and Essence, 2d rev. ed., trans. ArmandMaurer, C.S.B. (Toronto: The Pontifical Institute of Mediaeval Studies, 1968).


such Platonism in order to affirm ontological realism. Nonetheless, thisbrings us to the vexing issue of epistemological realism.

To affirm that we can know the natures of things requires that werecognize at the start a distinction made by Aquinas. He states:

For although it be necessary for the truth of a cognition that thecognition answer to the thing known, still it is not necessary that themode of the thing known be the same as the mode of its cognition.(Summa contra Gentiles, II, 76)

This distinction allows us to acknowledge many of the important pointsoften advanced against epistemological realism by its critics, withouthaving to give up the claim that we can nonetheless come to know thingsmore or less adequately.

The general point made by critics of epistemological realism is, ofcourse, that human cognition has a nature. Yet Aquinas’s distinction holdsthat cognition need not be without a nature in order for it to be able to beof or about the natures of things. Cognition is not something miraculous;it has a certain modality and proceeds in a certain manner. Thus, oneshould not suppose, without argument, that if cognition has a specificnature, it impossible to know the natures of things. There is no need tosuppose, at least initially, that epistemological realism requires that weget “outside our skins” or adopt “a view from nowhere.”

I turn now to various points often made by critics of epistemologicalrealism. Given our understanding of Aquinas’s distinction, we can showthat these criticisms are not so devastating as they might initially appearto be.

1. In nearly every case, we must employ some conceptual scheme orlanguage, use some set of words, or engage in some set of social practicesin order to know anything. Nevertheless, that does not show that con-ceptual schemes, languages, or social practices determine the natures ofthings.117 Jumping to such a conclusion would trade on the same ambi-guity found in Berkeley’s well-known claim that since beings cannot bethought of as existing apart from a thinking mind, their existence andnature are impossible without a thinking mind. Of course, one cannotthink of beings without thinking (for us, the relevant terms are “conceiv-ing” or “saying”), but that is some distance from claiming that the exis-tence and nature of beings are dependent on or determined by our thinking.Likewise, it is some distance from claiming that the natures of things aredetermined by our conceptual scheme, language, or system of social prac-tices. That we must use concepts and words or engage in social practices

117 Of course, the natures of many things have been shaped by human purposes. SeeRasmussen, “The Significance for Cognitive Realism of the Thought of John Poinsot,” 411n. 6.


in knowing something does not show that they determine the ultimatecharacter of what we know.

2. That we employ conceptual schemes, languages, or social practicesin cognition is undeniable, but it is unnecessary to treat these as “thirdthings” that we must somehow know first before we gain access toreality. Such a view of cognition, the so-called “way of ideas” thatDescartes and Locke are often charged with introducing into modernphilosophy, had some currency in the seventeenth and eighteenth cen-turies, but even then, thinkers such as John Poinsot and Thomas Reidchallenged it. The view quickly degenerated into skepticism. It is onlyone model of the cognitive relation, and not a very good one. Indeed,according to at least one version of Wittgenstein’s private languageargument, it is a nonstarter.118

3. We do not know all of reality at once in all its detail. We are not God,and so we must start somewhere. Our needs, interests, and concerns, aswell as our historical, social, and cultural circumstances, at least initiallydetermine the breadth and the depth of our examinations of reality. Yetthese admissions do not preclude epistemological realism. Our knowl-edge of reality just is partial, incomplete, and capable of being revised. Itdoes not, cannot, and need not operate from some absolute perspective.

4. Human beings are measurers of things. Indeed, cognition is for themost part not something that happens to us but something that we do.Concepts, classifications, theories, hypotheses, judgments, evaluations,and so on do not exist independently of or apart from human cognitiveeffort. They are our cognitive tools, and we are responsible for theirexistence. However, this does not mean or show that what they are about,or what they classify, explain, or assert as true or good, is simply a matterof convention.119 Some of our claims are still better than others are. None-theless, it is not necessary to hold that the structure of logic is the same asthe ontological structure of existence in order to defend this claim. Logiccan be understood as an organon or tool of knowledge and need not beseen as an ontology.

118 See the following works by Sir Anthony Kenny: The Legacy of Wittgenstein (Oxford:Blackwell, 1984), chap. 5; The Metaphysics of Mind (Oxford: Clarendon Press, 1989); TheUnknown God: Agnostic Essays (London and New York: Continuum, 2004), chap. 11; andWittgenstein, rev. ed. (Oxford: Blackwell, 2006), chap. 10. John McDowell seems to share thisview of Wittgenstein as well, because he suggests that Wittgenstein’s private languageargument should not be understood as upholding the primacy of interpretive practices.Instead, it should be understood as providing reasons for abandoning the “master thesis” —that is, the view that whatever one has in one’s mind when one knows just “stands there”and needs some interpretation to relate it to the world. See John McDowell, “Meaning andIntentionality in Wittgenstein’s Later Philosophy,” Midwest Studies in Philosophy 17 (1992):40–52.

119 Obviously, a crucial issue is whether sense perception only becomes intelligible byhaving form imposed on it by our conceptual schemes or whether sense perception becomesintelligible by having the form it implicitly carries discovered by our conceptual tools. For animportant defense of the latter view, see Thomas A. Russman, A Prospectus for the Triumphof Realism (Macon, GA: Mercer University Press, 1987).


5. We are fallible, and while this does not mean, as J. L. Austin oncenoted,120 that we should infer from this alone regarding any particularclaim that we might be mistaken, we must understand that the entireepistemological enterprise only exists because we are fallible beings. Infal-lible beings would have no need to consider the standards, methods,procedures, and criteria that are involved in knowing. Discovering thatwe are not God does not invalidate epistemological realism.

Obviously, each of these points could be made the subject of a book-length examination, but they are at least sufficient to show that many ofthe arguments that Putnam (and Nussbaum) use against metaphysicalrealism (which I noted in earlier sections of this essay) are subject toserious challenge and indeed are quite dubious.

However, it might be objected that in the foregoing remarks I haveactually adopted much of Putnam’s pragmatism; for I have acknowl-edged the role of human interests and needs in human cognition, and Ihave admitted that human knowing is fallible and subject to revision. Icould respond that these observations, and others like them, have longbeen part of the epistemological realist’s repertoire.121 Indeed, Putnamseems to have recognized this recently.122 Yet such a debate would beneither very fundamental nor very interesting. Let us instead cut to thechase and ask, “What is it that makes metaphysical realism, at least asdescribed in this section, superior to the so-called ‘common sense’ or‘pragmatic’ realism endorsed by Putnam?”

B. “Early” and “later” Putnam and rejecting the primacy of practice

The answer to this question requires that we deal with both an “early”and a “later” Putnam, for his views on these matters have apparently notstood still. The essential thing about the “early” Putnam is that he endorsesthe primacy of practice, and the essential thing about the “later” Putnamis that he at least appears at times to have some doubts about its primacy.Let us begin by considering the “early” Putnam.

The “early” Putnam’s appeal to the primacy of practice as an accountof human knowledge does not succeed, because it cannot differentiatebetween the explanandum, what it seeks to explain (that is, human knowl-edge) and the explanans, the explanation (namely, practice). It fails toexplain the very phenomenon of cognition. Here is why. For any expla-

120 J. L. Austin, “Other Minds,” in Austin, Philosophical Papers, 2d ed. (London: OxfordUniversity Press, 1970), 98.

121 For a few contemporary examples, see Baruch Brody, Identity and Essence (Princeton,NJ: Princeton University Press, 1980); Tibor R. Machan, “Epistemology and Moral Knowl-edge,” The Review of Metaphysics 36, no. 1 (September 1984): 23–49; Edward Pols, RadicalRealism: Direct Knowing in Science and Philosophy (Ithaca, NY, and London: Cornell UniversityPress, 1992); and Henry B. Veatch, Two Logics: The Conflict between Classical and Neo-AnalyticPhilosophy (Evanston, IL: Northwestern University Press, 1969).

122 Putnam, “Sense, Nonsense, and the Senses,” 22–24.


nation E, it only counts as an explanation if it can be distinguished fromthe thing explained, T. But E is distinct from T only if one can see that Eis distinct from T. Therefore, it is fallacious to derive E from T. So, if ourexplanation of human knowledge holds that the objects of knowledge aresimply constructs of practice, while at the same time we hold that know-ing is also a practice, then we have fundamentally failed to differentiatebetween our explanandum and our explanans. We have explained nothing.123

Putting this point in a linguistic vein and using it against a certainreading of Wittgenstein that Putnam so often employs in defending theprimacy of practice, we can ask the following question, “What is it aboutour practices that allows us to make a mark, sound, or movement intosomething that conveys linguistic significance?” To respond that it issimply our form of life, our conventions and rules —that is, our practices —only repeats what we want explained and adds nothing fundamental toour understanding of the phenomenon. There is a difference between amark, sound, or movement as such and as conveying linguistic significance.What is it about the human form of life and its practices that makes itpossible for us to turn marks and sounds into words? Saying that we justdo amounts to nothing more than spinning our intellectual wheels. Lan-guage has indeed gone on holiday here, but what has left the job is not thequestion being asked but the proposed answer. Practice simply does notsuffice.124

It is, of course, true that not all words have the same function, and it isa mistake to suppose that for every word there must be a thing to whichit corresponds. Further, and as I noted earlier, we must be careful when itcomes to arguments that seek to populate reality with “abstract entities.”There is always the threat of Platonism, and this is important not only forontology but for ethics as well. When it comes to understanding human

123 It might be replied that this argument works only if we cannot imagine two differentconversations —one which is the explanandum and one which is the explanans. If so, thenmaybe another way of putting the same objection is to say that for any conversation C, ifpractice is primary, then there is no way in principle to tell whether this conversation is anexplanandum or an explanans. I owe this point to Douglas J. Den Uyl.

124 Putnam would object, of course, that I have now asked one question too many. Butwhy? The reason seems to be this: When it comes to explaining linguistic significance, acrucial premise behind the thesis that practice is primary is the claim that all other proposedexplanations actually presuppose what they are supposed to explain. That is, they presup-pose our linguistic practices. Putnam (at least, the “early” Putnam) believes this to be aparticularly devastating objection to any account of linguistic significance that would havea role for concepts or mental powers to play. Indeed, this is how he interprets Wittgenstein.See the works cited in notes 28 and 29. Yet what one finds in Wittgenstein is not an objectionto concepts or mental powers per se but an objection to a certain conception of them:namely, one in which they are treated as “intermediaries” or “third things” that somehowoccupy mental space and are the basis for some “private language.” As I noted earlier, thereare really no good reasons to adopt such a view of concepts or mental powers, and no goodreasons for thinking that the thesis that practice is primary follows from a rejection of a“third thing” view of concepts or mental powers. See also notes 30, 115, and 118, andDouglas B. Rasmussen, “Rorty, Wittgenstein, and the Nature of Intentionality,” Proceedingsof the American Catholic Philosophical Association 57 (1983): 152–62.


flourishing, one need not understand it as some good that exists in iso-lation from the needs and interests of individual human beings. As Iargued earlier, human flourishing can be understood in a personalist oragent-relative fashion. However, none of this shows that we must becontent with “explaining” our knowledge, our world, our good, or our-selves ultimately in terms of practices. It is still the case that we needsomething for our thoughts and words to be about and something for ourpractices to be for. Ultimately, we need an overall context in which tooperate, and this is why metaphysical realism continues to exert its legit-imate and powerful pull.

It is here, however, that the careful reader of Putnam might well say,“Well done! You have really made Putnam’s basic point! It may be a pointthat he has been struggling to make for many years, but it is nonethelesswhat he holds.” So let us now consider the “later” Putnam. We can beginby noting this lengthy observation:

William James . . . employs the example of someone choosing how todescribe beans that have been cast on the table. The beans can bedescribed in an almost endless variety of ways depending on theinterests of the describer. And James asks, Why should not any suchdescription be called true? James insists that there is no such thing asa description that reflects no particular interest at all. And he furtherinsists that the descriptions we give when our interests are not theo-retical or explanatory can be just as true as the ones we give when ourinterests are “intellectual.” . . . A traditional realist philosopher mightreply to James as follows: “If that is all you are saying, then I do notsee that any of your fulminations against philosophers who believein a ‘ready-made world’ are in order. . . . Suppose you decide toclassify the beans by color, or by whether they are next to beans of thesame size, or in some other way. The reason that such a classificationis possible, and can be extended to other similar collections of beansin the future, is that there are such properties as colors, sizes, adja-cency, etc. Your beloved ‘interests’ may determine which combina-tions of properties you regard as worth talking about, or even leadyou to invent a name for things with a particular combination ofproperties if there is no such name already in the language, but itdoes not change the world in the slightest. The world is as it isindependently of the interests of any describer.” 125

Putnam follows this statement with the remark, “I do not, myself, sidecompletely with James, nor do I side completely with the traditionalrealist critic. I agree with the critic that the world is as it is independentlyof the interests of the describers. James’s suggestion that the world that

125 Putnam, “Sense, Nonsense, and the Senses,” 5–6 (emphasis in original).


we know is to an indeterminate extent the product of our own minds isone I deplore.” 126

What are we to make of this remark? Has Putnam totally reversedhimself?127 Does the “later” Putnam now hold that it is intelligible to talkof a world that exists and is what it is independent of and apart from ourinterests and practices? Moreover, how does this admission affect hisarguments against the fact/value dichotomy and on behalf of moral cog-nitivism? After all, these arguments depended in many ways on rejectingthe idea that we are supposed to find a basis for our valuations in thenatures of things. For Putnam, it is not human nature that grounds ethicalknowledge, but our practices —our ways of thinking and talking.

This is indeed a complicated matter. However, the statement aboutWilliam James cited above comes from a work that was published beforePutnam’s recent ethical works, The Collapse of the Fact/Value Dichotomy andOther Essays (2002) and Ethics without Ontology (2004). Thus, it is doubtfulthat Putnam regards his apparent endorsement of traditional realism asaffecting the plausibility of his ethical views. Further, Putnam insists thatmetaphysical realism is committed to a view that he regards as a sheerflight of imagination:

The metaphysical fantasy is that there is a totality of Forms, or Uni-versals, or “properties,” fixed once and for all, and that every pos-sible meaning of a word corresponds to one of the Forms or Universalsor properties. The structure of all possible thought is fixed in advanceby the Forms.128

Thus, he still wants to distinguish his position from metaphysical realism.How does Putnam do this?

Putnam’s major concern here seems to be that there are properties“fixed once and for all.” This concern can be understood in a number ofways. First, it may simply mean a rejection of the existence of what wasearlier called “universal natures” —that is, the idea that reality comesalready conceptually divided for us. But it is certainly possible to be ametaphysical realist and not accept universals as part of one’s basic ontol-ogy. Second, Putnam’s concern may be to avoid the idea that the forms ofour knowledge claims and the ways in which they can be adequate toreality have already been determined. He seems clearly to want to rejectthat view, because he states:

It is true that a knowledge claim is responsible to reality, and, in mostcases, that reality is independent of the speaker. But reflection on

126 Ibid., 6.127 See Colin McGinn, “Can You Believe It?” The New York Review of Books 48, no. 6 (April

12, 2001): 71–75.128 Putnam, “Sense, Nonsense, and the Senses,” 6.


human experience suggests that neither the form of all knowledgeclaims nor the ways in which they are responsible to reality is fixedonce and for all in advance, contrary to the assumptions of the tra-ditional realist.129

This is a most ironic statement, because despite what Putnam says about“the traditional realist,” this is not something that is generally denied bytraditional realists. Indeed, Putnam’s point here is but a version of thedistinction made by Aquinas that I discussed earlier.130 Third, Putnammay be objecting to the claim that all the properties are known or thatthey all exist. Yet a metaphysical realist need not deny that knowledgeprogresses or that new things come to be. Fourth, Putnam may want toreject the idea that there actually are a totality of properties that haveexisted in the past, exist in the present, and will exist in the future, andthat these are what we have been, are, and will be in the process ofdiscovering. That is, he may want to reject the idea that there are beingsthat exist and are what they are independent of and apart from ourinterests and practices. If this is what he wants to reject, then despite thestatement cited above, he is not abandoning the primacy of practice. Sowhat is his position?

Putnam states that “the notion that our words and life are constrainedby a reality not of our own invention plays a deep role in our lives and isto be respected.” 131 Note that this statement does not say that there existsa reality not of our own invention that constrains our words and lives.Rather, it says that the notion of a reality not of our invention that con-strains us plays a deep role in our lives. To say this is not to give up theprimacy of practice. The real remains dependent on practice.132 Putnam issubtle, and he is adopting a strategy described earlier: namely, he is trying

129 Ibid., 7.130 Aquinas also states: “[I]t ought to be said that it is not necessary to assume diversity

in natural things from the diversity of intelligible characters or logical intentions whichfollow upon our manner of understanding, since the intelligible character of one and thesame thing may be apprehended in diverse ways.” Aquinas, Summa Theologiae I. 76.3, ad 4.

131 Putnam, “Sense, Nonsense, and the Senses,” 9.132 As I have already noted, Putnam states that “we have no access to ‘unconceptualized’

reality.” Yet this statement is followed shortly by this remark: “But it doesn’t follow thatlanguage and thought do not describe something outside themselves, even if that somethingcan only be described by describing it (that is by employing language and thought); and, asRorty ought to have seen, the belief that they do plays an essential role within language andthought themselves and, more importantly, within our lives.” Putnam, “The Question ofRealism,” 297 (emphasis in original). Two things should be noted here: (1) Putnam does notsay that whatever language and thought describe outside themselves has a character ornature apart from the describing; and (2) he explains the belief that there is somethingoutside language and thought that they do describe as a belief that comes from withinlanguage and human practices. Finally, Putnam has more recently remarked that he findsthe explanatory value of appealing to the form of something to be either “tautological ornonsensical.” Hilary Putnam, “Comments on John Haldane’s Paper,” in James Conant andUrszula M. Zeglen, eds., Hilary Putnam: Pragmatism and Realism (London and New York:Routledge, 2001), 108.


to adopt an attitude of “deliberate naïveté” and return to “natural real-ism.” Yet how does one cross over the bridge to natural realism given thephilosophical waters that have flowed beneath it? How does one adoptwhat Putnam now calls a “second naïveté”? 133 The only way Putnam cando so is by discrediting metaphysical questioning and making practiceultimate. But now this must be done on one hand while not letting theother hand be part of the process. Practice must somehow be both pri-mary and not primary. Thus, Putnam has succeeded in showing not theantinomy of realism but rather the antinomy of practice.134

V. Conclusion

The problem of finding a basis for moral knowledge is one of thecentral issues of philosophy, and as I noted at the beginning of this essay,there has recently been a resurgence in the belief that an appeal to humannature is crucial for understanding and defending moral knowledge. Thisessay has not examined this resurgence; rather, it has considered MarthaNussbaum’s claim that Hilary Putnam’s argument against metaphysicalrealism provides a better way of defending moral knowledge.

Putnam’s arguments do a magnificent job of destroying those remain-ing elements of logical positivism in analytic philosophy that assumethere is some ontological chasm between facts and values, but Putnamnever provides an account of ethical knowledge that can adjudicate betweencompeting views in ethics and politics. Moreover, the account of idealizedinquiry that Putnam presents leaves many gaps. It has no way to accom-modate an ethics of human flourishing or self-perfection that sees per-sonal development as the central aim of ethics. It makes no principled

133 Putnam, “Sense, Nonsense, and the Senses,” 15. David Macarthur argues that naturalrealism is not merely common sense and that its success hinges on its ability to provide amore satisfying account of perception that can withstand skeptical threats motivated by thetraditional realist/antirealist dispute. See David Macarthur, “Putnam’s Natural Realism andthe Question of a Perceptual Interface,” Philosophical Explorations 7, no. 2 ( June 2004): 167–81.

134 Putnam does not clearly distinguish between (1) those metaphysical views that wouldoverturn or replace our everyday practices and common-sense views and (2) those meta-physical views that seek to explain more deeply such practices and views. As a result, hedoes not fully appreciate that it is not necessary to reject metaphysics in order to show thatrealism needs no defense. Moreover, he does not fully realize that it is only by engaging inmetaphysics of this latter sort and showing the errors of those who think that they can eitherdeny or escape realism that a “second naïveté” can be achieved. As John Haldane has noted,“Realism with a human face requires the support of a metaphysical skull” (Haldane, “Real-ism with a Metaphysical Skull,” in Conant and Zeglen, eds., Hilary Putnam: Pragmatism andRealism, 97). Indeed, it has not been for lack of metaphysical investigation that philosophersin the Aristotelian-Thomistic tradition have argued that realism is perennial. As EtienneGilson states: “The first step on the path of realism is to perceive that one has always beenrealistic; the second is to perceive that, whatever one does to become otherwise, one willnever succeed; the third is to ascertain that those who do pretend to think otherwise, thinkin realistic terms as soon as they forget to act their part.” Gilson, “Vade Mecum of a YoungRealist,” in Roland Houde and Joseph P. Mullally, eds., Philosophy of Knowledge: SelectedReadings (Chicago, Philadelphia, New York: J. B. Lippincott, 1960), 386.


distinction between an ethical concern and a political/legal one and takeslittle or no notice of the agent-relative and individualized character ofhuman flourishing. Further, idealized inquiry never comes to grips withwhat ultimately makes a valuation valuable and so never shows whatmakes something worthy of valuing. There is thus a failure to explainwhat it is that obligates one to pursue some value or end. Putnam’saccount of idealized inquiry begs the question when it comes to ground-ing basic ethical obligations. Though there is nothing in the criteria ofidealized inquiry that prevents it from appealing to standards beyonditself, it is the attempt to do without deeper metaphysical commitmentsand make practice primary that is the source of its insufficiency.

Not only does the rejection of metaphysical realism not provide a betterway of defending moral knowledge, the claim that metaphysical realismis in demise is itself dubious. The arguments advanced by Putnam againstmetaphysical realism are by no means compelling, and in various waysthey reveal a fundamental misunderstanding of the position as well as afailure to make the distinctions necessary to avoid basic ambiguities.Overall, the suggestion that we might find a better way of establishingmoral knowledge by rejecting metaphysical realism fails. It simply doesnot work. Thus, the attempt by Nussbaum to develop an ethics of humanflourishing by embracing Putnam’s rejection of metaphysical realism ulti-mately undercuts what is distinctive and compelling about the Aristote-lian ethical tradition. It makes fundamental ethical values something thatwe construct rather than something grounded in our nature. Yet, evenmore, the attempt to do without metaphysical realism hampers the progressthat has been made in contemporary ethics by distracting us from ourefforts to understand human nature and reality —efforts that are vital tothe ethical enterprise.

Philosophy, St. John’s University


WHY MORAL JUDGMENTS CAN BE OBJECTIVE*

By Tibor R. Machan

I. The Desirability of Objectivity

Are we able to make objective moral judgments? By “objective” I mean“supportable by reference to facts about the world in contrast to what wemay feel, desire, wish, or hope for vis-à-vis those facts.” 1 By “moraljudgments” I mean, broadly, “judgments pertaining to how human beingsought to conduct themselves, how their institutions should be set up, andso forth.” 2

This perennial philosophical —yet at once popular3 —topic needs oftento be revisited because it is central to human life. Judgments about howpeople conduct themselves, about the institutions they devise —in short,about whether they are doing what is right or what is wrong —are ubiq-uitous. Even if one complains that people are judgmental or that theymistakenly think they can make true moral judgments or assessments(when, in fact, that is supposedly unjustified), one is making judgmentsthat one intends to have taken seriously, to be correct, even as one isrejecting the wisdom of doing so. “You shouldn’t think moral judgmentsare objective” is itself a moral judgment, albeit one pertaining to a specialundertaking, one of a restricted scope. Those making it seem to treat it asobjectively true.4

In short, moral skeptics or relativists, too, criticize those who disagreewith them, making note of their opponents’ supposedly misguided think-ing about this metaethical issue. And they expect that they are correctwhen they do so —objectively right, that is.

Some may hold that such judgments are no more than expressions ofwhat is desirable —that is to say, they express what some desire. But that

* I wish to thank Professors Randall R. Dipert, Kenneth Lucey, Aeon Skoble, the othercontributors to this volume, and its editors, for their very helpful comments on an earlierversion of this paper. Of course, all of what appears here is my responsibility.

1 For a fuller explanation of the nature of objectivity, see Tibor R. Machan, Objectivity:Recovering Determinate Reality in Philosophy, Science, and Everyday Life (Burlington, VT: AshgatePublishing Co., 2004).

2 Moral judgments here also cover political judgments, in other words.3 Whether objectivity is possible, for instance, in journalism, self-understanding, or psy-

chotherapy are issues of popular interest and not solely the concern of professionalphilosophers.

4 Truth is objective in any case. To add “objectively” to “true” serves to signal that noother kind of truth is possible. There can be subjective feelings or experiences or phenom-ena, but there can be no subjective truths. As to whether criticisms in philosophy qualify asethical or moral, it seems to me that they do, as part of professional ethics.


is not what most people mean when they make such judgments —theytake these judgments to be true and, thus, objective. The mere expressionof a desire does not amount to a criticism.

Yet such dialectical points are not sufficient to make out the case for theobjectivity of moral judgments —that is, to defend the position that it ispossible to gain knowledge of (and demonstrate) the truth or falsity ofmoral judgments as we do with other judgments (for example, those inphysics, biology, engineering, or drivers’ education). All that these dia-lectical points make evident is that denying the objectivity of moral andother evaluations leads to self-referential inconsistency; then, perhaps,what follows is only that we ought to remain silent on the topic (asWittgenstein appears to have believed).5

One promising approach in defense of objectivity in ethics is to workout a sound, naturalistic idea of what the human good amounts to and togo on to demonstrate that certain ways of acting advance this good whileothers thwart it. If one can show that there are such ways of acting, thiswould satisfy the requirement of objectivity, since whether certain waysof acting advance the good or not is a matter capable of being established,shown to be true based on evidence and argument. In this essay, I plan todeploy a neo-Aristotelian naturalism in order to keep the “is-ought” gapat bay and place morality on an objective footing.6 I will do this with theaid of the ideas of Ayn Rand —as well as, but only by implication andassociation, those of Martha Nussbaum and Philippa Foot.

It is important to mention at the outset that in defending the objectivityof moral judgments, one is not defending the universal applicability of allof them, except where identical circumstances obtain. Thus, if in fact Mrs.George ought to send her child to take piano lessons, it does not follow

5 See Ludwig Wittgenstein, “A Lecture on Ethics,” in Philosophical Review 74 (1965): 11–12,and the last sentence of Wittgenstein’s Tractatus Logico-Philosophicus (New York: Harcourt,Brace and Company; London: K. Paul, Trench, Trubner & Co., 1922): “Whereof one cannotspeak, thereof one must be silent.” As Arto Tukiainen says, “Wittgenstein regarded the pointof the Tractatus as ethical. His purpose was to show that there cannot be any meaning-ful ethical sentences.” So Wittgenstein is supposed to have denied that ethics is objective.See http://examinedlifejournal.com/articles/printerfriendly.php?shorttitle=wittmoralval&authorid=54. But see also Tibor R. Machan, “Heretical Essay in Wittgenstein’s Meta-Ethics,”Analysis and Metaphysics (forthcoming).

6 Among those who have made this attempt within the naturalist, neo-Aristotelian schoolare Henry B. Veatch, For an Ontology of Morals: A Critique of Contemporary Ethical Theory(Evanston, IL: Northwestern University Press, 1971); John Cooper, Knowledge, Nature, and theGood: Essays on Ancient Philosophy (Princeton, NJ: Princeton University Press, 2004); MarthaNussbaum, The Therapy of Desire: Theory and Practice in Hellenistic Ethics (Princeton, NJ:Princeton University Press, 1996); and Philippa Foot, Natural Goodness (New York: OxfordUniversity Press, 2001), as well as several others. Only Ayn Rand identifies herself explicitlyas an Objectivist.

One of my own early efforts to defend a naturalist-objectivist metaethics appears in TiborR. Machan, “Epistemology and Moral Knowledge,” Review of Metaphysics 36, no. 3 (Sep-tember 1982): 23–49. I try to sketch the appropriate conception of human knowledge insupport of this task in Tibor R. Machan, “Education and the Philosophy of Knowledge,”Educational Theory 20, no. 3 (1970): 253–68.

WHY MORAL JUDGMENTS CAN BE OBJECTIVE 101

that everyone should. It is irrelevant that the truth of such a claim isuniversally ascertainable; the point is that it doesn’t have to apply uni-versally, to all moral agents, although there may be some moral judg-ments, very basic ones, that are universally applicable in this sense. (Ihave elsewhere defended this point.)7

Nor is objectivity about ethics the same as impartiality or neutrality: thatsomeone ought morally or ethically to advance his or her child’s welfarecan be the case without whoever knows this having to be indifferent to thematter. “Objective” in this context, then, means —at the minimum —thatjudgments are supported by facts that are themselves independent ofhuman wishes, hopes, desires, or similar states of mind not concerned withgrasping facts.8 As the philosopher Raimond Gaita has put the point:

Hardly anyone will deny that it requires disciplines of mind andcharacter to “see things as they are” as opposed to how they appearto be and, especially in the case of psychological phenomena, how formany and subtle reasons we fantasize them to be. Nor are manypeople likely to deny that we must often distance ourselves from oursubject so that our fears, fantasies and affections do not interfere withour sense of what is objectively the case.9

Thus, within the limitations of my purpose (to show that objectivemoral judgments are possible), I will focus primarily on showing that AynRand and similar neo-Aristotelians have made a very promising case forplacing ethical or moral judgments on an objective footing.10 Rand calledher philosophy “Objectivism” because she defended the view that thereis an independent reality (including ethical or moral reality) that we cancome to know objectively, that is, based on evidence and reasoning thatyield understanding without distorting what is being understood. Theprinciples of ethics, for an Objectivist, are principles no less objectivethan, say, the rules of some game or sport. In important respects, theserules are similarly conditional, though not relative —which is to say, theypresuppose some facts (e.g., the choice to live) but are not predicated on

7 Tibor R. Machan, “Why It Appears That Objective Ethical Claims Are Subjective,”Philosophia 26, nos. 1–4 (1997): 1–23. A recent critique of the is-ought gap is Hilary Putnam’sThe Collapse of the Fact-Value Dichotomy (Cambridge, MA: Harvard University Press, 2002).Putnam’s critique, not unlike Martha Nussbaum’s and Philippa Foot’s efforts, comes severaldecades after Rand briefly laid out her case for the objectivity of value judgments. (See note11 below.)

8 I discuss a wide range of issues related to objectivity in Machan, Objectivity.9 Raimond Gaita, A Common Humanity: Thinking about Love and Truth and Justice (London:

Routledge, 1998), 248.10 There is a subtle difference between what Rand means by “objective” when it comes to

ethics and what she means by the term in other, more straightforwardly epistemological,contexts. For more on this, see Jason Raibley, “Rand on the Objectivity of Values” (2002),http://www.ios.org/events/advsem02/JRObjectivity.pdf.

102 TIBOR R. MACHAN

contingencies (such as where one lives, the weather, one’s cultural back-ground, etc.). The idea is that it can be objectively shown, for example,that “If one is to live the good human life, one ought to be honest.”Furthermore, Rand argued both that objective ethical knowledge is pos-sible and that the “is-ought” gap is bogus.

For Rand, “objective” means “neither revealed nor invented, but asproduced by man’s consciousness in accordance with the facts of reality, asmental integrations of factual data computed by man —as the products ofa cognitive method of classification whose processes must be performedby man, but whose content is dictated by reality.” 11 That is to say, realityexists independently of us, and for us to apprehend or grasp it correctly,we must do the mental work she describes in her book on epistemology(from which the foregoing quotation is drawn). In relation to morality,however, this reality involves a certain kind of relationship between whoand what we are and what it is that enhances living for such a being aswe are. So while objective, morality as Rand, and indeed many others,understand it is agent-relative.

II. Objectivism and Objective Ethical Judgments

In her essay “The Objectivist Ethics,” 12 Rand argues against DavidHume that moral judgments can be validated or shown to be well founded.She rejects the passage in which Hume denies that one can deduce fromtrue statements about what is the case certain true statements about whatone ought to do. Rand goes on to develop her case, as we have alreadygleaned above, that human life is normatively pregnant, as it were, mean-ing that it is bound up, through and through, with a normative dimen-sion.13 The concept of value, Rand argues, is tied intimately to the conceptof life, so when something is a living organism there is no escaping

11 Ayn Rand, Introduction to Objectivist Epistemology (New York: New American Library,1970), 54 (emphasis added to indicate the source of objectivity apprehended by humanconsciousness).

Ayn Rand preceded Martha Nussbaum and Philippa Foot in making a case in recentphilosophy for an objectivist, neo-Aristotelian metaethics. Both Nussbaum and Foot pub-lished their discussions of these metaethical matters three decades after Rand came out withher discussion of the topic in her essay “The Objectivist Ethics,” in her book The Virtue ofSelfishness: A New Concept of Egoism (New York: New American Library, 1961). Incidentally,this title is often misunderstood to indicate that Rand is defending some normative versionof Hobbes’s psychological egoism. In fact, her ethical egoism is more in line with Aristotle’seudaimonism, whereby ethics per se stands in service of living a good human life. Thus thesubtitle’s reference to “a new concept of egoism.”

12 Ibid.13 Accordingly, it would be a mischaracterization of Rand’s view to suggest that she

proposes to derive moral conclusions from nonmoral facts. The facts from which she pro-poses to derive moral conclusions do contain normative elements, but they are in no waydisqualified as facts for that reason alone. The qualification of these facts as “moral” or“ethical” is, in crucial respects, similar to the qualification of other facts as “biological” or“sociological”; the qualification locates these facts ontologically.


certain normative considerations about such a being. Living beings, forexample, do well or badly. “Epistemologically, the concept of ‘value’ isgenetically dependent upon and derived from the antecedent concept of‘life.’ To speak of ‘value’ as apart from ‘life’ is worse than a contradictionin terms. It is only the concept of ‘Life’ that makes the concept of ‘Value’possible.” 14 (This, by the way, is a point also implicit in the works ofMartha Nussbaum and Philippa Foot; see The Therapy of Desire15 andNatural Goodness,16 respectively.) Because of the human capacity for choice,however —for freely deciding to do one thing or another or yet another —inthe lives of human beings this normative component is transformed intoan ethical or moral one. Thus, not only is it possible to consider whether ahuman being is doing well or badly, it is also possible to consider whetherone is choosing and acting responsibly in doing well or badly.

Rand does not try to establish that one can provide deductive argumentsfor moral judgments. She does argue, however, that moral judgments canbe validated or derived from other judgments that are not explicitly moral(though they would normally be normative or value-laden). These deri-vations are not deductions but inferences, more akin to inductions orconceptual implications than to formal deductions —for example, “If oneacts prudently, generously, honestly, and/or courageously, one will (forthe most part) live a successful, good, or happy life.”

This is how best to understand what Rand proposes in opposition toHume, based on her understanding of the nature of concepts and defi-nitions and how they work in propositions and arguments.

We can glean what Rand set out to do if we consider that in the sciencesthere would also be a gap (call it the “is-must” gap) if we were to expectthat scientific reasoning needs to be deductive. But that isn’t so. Scientificarguments that rest on research and testing —and lead, optimally, to objec-tive knowledge of scientific facts, laws, or principles —involve inferenceswhose conclusions are true not beyond a shadow of a doubt but beyonda reasonable doubt. Hume could have no complaint about such reason-ing, and as Rand understood moral reasoning, Hume might not have hadgrounds for complaint about what she set out to defend. It has the fol-lowing form: “It is good for a human being to thrive, and the choice to beprudent (or courageous, generous, and so forth) is required for thriving,so it is good for one to choose to be prudent.” 17 The precise nature of this

14 Rand, “The Objectivist Ethics,” 18.15 Martha Nussbaum, The Therapy of Desire (Princeton, NJ: Princeton University Press,

1994).16 Philippa Foot, Natural Goodness (New York: Oxford University Press, 2001).17 Thus, Eric Mack is correct when he writes that “[b]eing an objectivist with respect to

claims about the effectiveness of means to chosen ends no more removes an advocate of thechoice doctrine from the basic subjectivist camp than it removes Hume from that camp”(Mack, “More Problematic Arguments in Randian Ethics,” The Journal of Ayn Rand Studies 7[Spring, 2006]: 301). It is not this that removes the choice doctrine from the basic subjectivistcamp; rather, it is the nature of the end chosen, namely, the ultimate end of human life. For

104 TIBOR R. MACHAN

choice is, of course, complicated; it would ordinarily involve the cultiva-tion of the virtues that constitute and help promote one’s happiness.18

III. Beyond the “Is-Ought” Gap

The famous “is-ought” gap that Hume identified expresses the philo-sophical claim that a conclusion that contains moral terms such as “oughtto” or “ought not to” cannot be deduced from premises that lack thosemoral terms, because a valid deductive argument can only have in itsconclusion components that are fully supported by its premises. If thepremises fail to give complete support to the conclusion, the conclusionis not valid. Here is how Hume put the point:

In every system of morality which I have hitherto met with, I havealways remark’d, that the author proceeds for some time in the ordi-nary way of reasoning, and establishes the being of a God, or makesobservations concerning human affairs; when of a sudden I amsurpriz’d to find, that instead of the usual copulations of proposi-tions, is, and is not, I meet with no proposition that is not connectedwith an ought, or an ought not. This change is imperceptible; but is,however, of the last consequence. For as this ought or ought not,expresses some new relation or affirmation, ’tis necessary that itshould be observ’d and explain’d; and at the same time that a reasonshould be given, for what seems altogether inconceivable, how thisnew relation can be a deduction from others, which are entirelydifferent from it. But as authors do not commonly use this precau-tion, I shall presume to recommend it to the readers; and am per-suaded that this small attention would subvert all the vulgar systemsof morality, and let us see, that the distinction of vice and virtue is notfounded merely on the relations of objects, nor is perceiv’d by rea-son. . . .19

Hume discredits all rationalist —or deductivist —efforts to prove moraljudgments true. He does this by arguing that they beg the question becausesuch a proof would have to have premises that already contain the crucial

Hume, that end appears to be anything one might desire, whereas in reality, Rand claims,it is only the one ultimate value that is possible for us: human life. Any other choice wouldamount, in time, to the annihilation of ends.

18 The happiness at issue here is not what so many contemporary social scientists take toamount to happiness —cheerfulness, even giddiness. See, for example, Jean-Paul Pecqueur,The Case Against Happiness (Farmington, ME: Alice James Books, 2006). On my own view,happiness is the self-awareness of living successfully.

19 David Hume, A Treatise of Human Nature (Garden City, NY: Doubleday and Company,1961), 423.


term, “ought” or “ought not,” thus simply pushing the problem back astep in each case, leaving the moral judgment ultimately unprovable.20

In response to Hume’s passage, Rand says:

In answer to those philosophers who claim that no relation can beestablished between ultimate ends or values and the facts of reality,let me stress that the fact that living entities exist and function neces-sitates the existence of values and of an ultimate value which for anygiven living entity is its own life. Thus the validation of value judg-ments is to be achieved by reference to the facts of reality. The factthat a living entity is, determines what it ought to do. So much for theissue of the relation between “is” and “ought.”21

She writes that the passage from Hume “purports to mean that ethicalpropositions cannot be derived from factual propositions —or that knowl-edge of that which is cannot logically give man any knowledge of whathe ought to do. And wider: it means that knowledge of reality is irrelevantto the actions of a living entity and that any relation between the two is‘inconceivable.’ ” 22

However, what Hume actually appears to have contended is not whatRand claims he did, namely, that “knowledge of reality is irrelevant to theactions of a living entity and that any relation between the two is ‘incon-ceivable’,” or that “ethical propositions cannot be derived from factualpropositions.” Rather, what Hume argued is that it is impossible to deduce

20 For a recent effort to address Hume’s concerns, see Mark H. Bickhard, “Process andEmergence: Normative Function and Representation,” Axiomathes 14, nos. 1–3 (2003): 121–55. Bickhard also reads Hume as saying that one cannot derive “ought” from “is,” whereasHume said that one cannot deduce “ought” from “is.”

21 Rand, “The Objectivist Ethics,” 19 (emphasis in the original). Did Rand claim to havederived an “ought” from an “is” in this essay, as Mack states (“More Problematic Argu-ments,” 297)? Let me reiterate that Rand is using the term “validation” in her discussion ofvalue judgments. This idea of “validation” is not the same as the Humean idea of deduction;nor, it seems, is it the same as Mack’s idea of derivation. What Rand seems to me to be doingis arguing (in some legitimate fashion or other) that moral judgments can (objectively) beshown to be true. She does not develop in detail just what such validation consists of. Isurmise, from the full context of her approach to demonstrating the truth of her ideas, thathers is a type of conceptual rather than syllogistic “proof.”

At the risk of sounding heretical, Rand’s form of validation smacks of the kind of rea-soning we find in Wittgenstein or in criminal prosecution, akin to the argument to the bestexplanation. So her showing that moral claims can be objective rests on a different idea ofwhat it takes to validate such claims from what Hume appears to have regarded as requiredfor such a task (and then denied is possible). See Tibor R. Machan, Ayn Rand (New York:Peter Lang, 2001), 57ff.

It is also necessary to note here that when Rand and others in this tradition of naturalisticmetaethics and ethics defend a principle, they do this “for the most part” and not “neces-sarily in every conceivable case.”

22 Journals of Ayn Rand, ed. David Harriman (New York: Dutton, 1997), 683 (emphasis inthe original).

106 TIBOR R. MACHAN

what one ought and ought not do from what is or is not the case. Thedifference is not at all negligible.

IV. Not All Reasoning Is Deductive

A deduction is deemed by many philosophers to be a formal logicaloperation, capable of involving only concepts with closed, final defini-tions. On this view, deductions are treated as timeless proofs. This is atradition of thinking about knowledge along lines suggested by Plato,with his theory of forms or the natures of things, and continued, espe-cially, by Descartes. Both of these very influential philosophers appearedto suggest that for us to have knowledge of anything, this knowledgemust be capable of being stated as a necessary truth. That, in turn, suggeststhat whenever reasoning is performed invoking facts that human beingsknow, this reasoning must amount to a species of logical deduction com-parable to that which occurs in formal logical (indeed, symbolic) opera-tions. In such reasoning, the operative terms and propositions are symbolicof closed, final definitions and propositions comprised of terms with suchdefinitions. Such reasoning is, in consequence, timelessly valid, and what-ever truth might be achieved by means of it would also amount to time-less, necessary truth.

However, when logic is deployed to produce proofs involving facts thatmay imply other facts, this is not a formal (let alone symbolic) deductionbut another type of proof —another way to derive a factual propositionfrom other factual propositions.23 David Hume did not discuss the dif-ference between this type of derivation and formal deduction, and as Iargue in my book Ayn Rand,24 Rand’s Objectivist epistemology holds outa credible promise for the success of such a derivation but does notestablish that the strict formal deduction Hume thinks is inconceivable isactually available in establishing or proving the truth of any moralpropositions.25

23 Ken Lucey has objected with the rhetorical question, “Does Machan think a fact is suchthat it can have implications?” Yes. For example, the fact that there is an ashtray on a tablein a restaurant implies that smoking is permitted there. Or the fact of someone’s racismimplies the appropriateness of moral contempt for that someone.

24 Machan, Ayn Rand, chap. 2.25 Rand can be interpreted as holding that the kind of strictly formal deductions that

others think exclusively deserve the term are simply very broad, symbolic models of logicalreasoning. So she would argue that deductions do, in fact, obtain between judgments offacts involving concepts that are themselves contextually (as distinct from timelessly) defined,possessing essential attributes that make logical deductions possible. Her objection to Hume,then, is that Hume failed to see that concepts such as “ought to” and “ought not to” can bederived from definitions of “human goodness” as, in part, essentially involving choices andultimate values. In other words, Rand would argue that a sound, valid theory of humangoodness deductively yields moral conclusions, conclusions regarding what one ought todo, so long as by “deduction” one does not mean logical arguments involving closeddefinitions of concepts. For more, see Machan, Ayn Rand, and Machan, Objectivity.


V. Conceptualization Versus Deduction

The idea of concept formation clarifies why Rand does not accept eitherHume’s or G. E. Moore’s idea of what it is to know something and howto define a concept.26 Put plainly, what is at issue here is the claim that theway we acquire knowledge is by developing and organizing our ideasbased on awareness we have of the world for the time being, by means ofour perceptual organs, guided by axiomatic concepts (which I have dis-cussed elsewhere and which are central to Rand’s Objectivism).27 In a bitmore detail, the process goes on roughly as follows.

One detects various similarities and differences by means of the sen-sory organs; one recognizes that nothing like that is usually, normallypossible unless something exists that is being perceived, and then onecarefully, parsimoniously, arranges a system of ideas in which principlesof logic are followed, all the while keeping in continued focus the initialdifferences and similarities, thus learning what it is that exists. The resultis a system of sound —well-grounded —ideas that best (but not necessar-ily finally) capture for us aspects of reality that we set out to understand.28

When we know reality, moreover, we do not know it by grasping itforever. We know it at a given time, not for all times. One reason it isn’teasy to give an account of knowing reality is that knowing isn’t like otheraccomplishments with which we are familiar. The way human beingsknow is unique in the world; they know conceptually, not perceptually asmost other living beings do, and they have the capacity to reflect on theirconceptual knowledge —to become self-aware —a process that needs to beidentified and understood without the benefit of easy analogies. But someanalogies come closer than others. For example, we know somewhatanalogously to how we grasp —or grab —something as well as we can, butnot without the possibility of improvement, and we can observe ourgrasping —or grabbing —as well. All efforts to understand the nature ofhuman knowledge must admit that there is nothing quite like it in the restof nature and that the best we can do is depend on our ordinary, non-systematic familiarity with cases of knowing, something that runs the risk

26 I examine both Hume’s and Moore’s criticism of “naturalism” in Tibor R. Machan,Human Rights and Human Liberties (Chicago: Nelson-Hall, 1975), and Machan, Individuals andTheir Rights (Chicago: Open Court, 1989). For both Hume and Moore, definitions appear toamount to necessary truths, whereas in the neo-Aristotelian, Randian epistemological tra-dition they are understood to be contextual and open-ended. See Rand, Introduction toObjectivist Epistemology.

27 Machan, Ayn Rand, chap. 2, and Rand, Introduction to Objectivist Epistemology.28 By “in which the principles of logic are followed” I am referring to the policy of being

consistent, avoiding contradictions, and avoiding all the more specialized fallacies involvedin bad reasoning. I do not, however, have in mind adhering to the strict requirements offormal or symbolic logic, wherein one aims for purely deductive reasoning. For example,substantive reasoning (including moral reasoning) need not result in conclusions that arelogically impossible to deny. For more on this, see Tibor R. Machan, “Another Look at LogicalPossibility,” The Personalist 51, (1970): 246–49. See also Douglas B. Rasmussen, “LogicalPossibility: An Aristotelian Essentialist Critique,” The Thomist 47 (October 1983): 513–40.

108 TIBOR R. MACHAN

of begging the question. In a certain respect, an attempt to give an accountof human knowing involves starting out on an entirely new endeavor.29

In any case, when we draw conclusions in everyday life (about whetherto open a door when trying to go through it, or about how to build abridge or a helicopter, or how to construct an argument for objective

29 A fatal flaw of much of philosophical reflection about what it is to know has been (fromPlato through Descartes to our own time) that knowledge has often been understood torequire timeless certainty. If one knows, it must be impossible to even conceive that one iswrong. Knowledge must be absolute, perfect, incorrigible, finished, and, as it is sometimesput, in the final analysis. For a discussion of a different view of knowledge, see Adam Leite,“Is Fallibility an Epistemological Shortcoming?” The Philosophical Quarterly 54, no. 215 (April2004): 232–51.

The Platonist view (not necessarily Plato’s view) of what knowledge must be is an impos-sible ideal. This comes across when, for example, Socrates notes that he knows only that heknows nothing. Socrates is arguably invoking the idealist view of perfect knowledge ofForms and disclaiming having achieved it. Given the prevalence of the Platonist view ofknowledge, the result has been a great variety of more or less skeptical ideas about knowl-edge, such as the prominent view that we cannot really know anything, or if we can,perhaps, it’s just an approximation; maybe we can have probable or approximate or fallibleknowledge. In contrast, I would say that sometimes we probably know this or that; some-times we have approximate knowledge; sometimes our knowledge is fallible. At othertimes, however, we simply know, without qualification. It is only the model of ideal (Pla-tonic) knowledge that inclines one to place the qualifiers before “knowledge” in everyinstance.

Currently, Richard Rorty’s communitarian view of knowledge is widely respected, accord-ing to which one knows in terms of one’s group and objectivity is not possible but solidarityis. For the full story, see Richard Rorty, Objectivity, Relativism, and Truth (Cambridge: Cam-bridge University Press, 1991). All this has been especially influential with regard to knowl-edge about right and wrong, good and evil, producing some disturbing practical results incrucial areas such as the way personal misconduct, professional malpractice, crime, andterrorism are widely understood and discussed. For more on this, see Tibor R. Machan,“Terrorism and Objective Moral Principles,” International Journal of World Peace 4, no. 4(October–December 1987): 31–40. (The ideas of this paper are incorporated in Machan,Objectivity.)

What Rand has proposed is that human beings, if they do the hard work, can obtainknowledge just fine and dandy. There is, of course, ample evidence of this throughout thesciences, in technology, and, let’s not forget it, in ordinary life. But what is this humanknowledge? As the name of her system makes evident, the key to knowledge is objectivevalidity (meaning that knowledge must be “reasonable or justifiable in the circumstances”).As she put the point in her Introduction to Objectivist Epistemology, 46:

Objective validity is determined by reference to the facts of reality. But it is man whohas to identify the facts; objectivity requires discovery by man —and cannot precedeman’s knowledge, i.e., cannot require omniscience. Man cannot know more than he hasdiscovered —and he may not know less than the evidence indicates, if his concepts anddefinitions are to be objectively valid.

Ayn Rand’s position is, I believe, the one that Gilbert Harman expressed well: namely,that we must “take care not to adopt a very skeptical attitude nor become too lenient aboutwhat is to count as knowledge” (Gilbert Harman, Thought [Princeton, NJ: Princeton Uni-versity Press, 1976], 145). Another prominent contemporary philosopher who has advancedthis understanding of human knowledge is J. L. Austin, in his essay “Other Minds,” inAustin, Philosophical Papers (Oxford: Clarendon Press, 1961). Arguably, in his On Certainty(New York: Harper Torchbooks, 1969), Ludwig Wittgenstein also suggests such a view ofknowledge.


ethics), the conclusions are not strictly speaking deduced —as formal logi-cians would characterize deductions —but are conceptually inferred fromsuccessive facts that are known contextually. Moreover, the state of ourknowledge is subject to change. We may learn that what we knew once(e.g., regarding history or biology) is now understood somewhat differ-ently. Given this, it is also possible to imagine that in the future, the stateof knowledge will change once again. That is to say, the premises ofsubstantive (including moral) arguments are not finally certain in the waythat those of formal, logical arguments are; nonetheless, these premises,and the conclusions they support, can be objective and contextuallycertain.30

The reason this is different from pure logical deduction is that suchdeductions are formal and symbolic, and thus not dependent on actualconcepts, but only on symbols of concepts. As such, strictly logical deduc-tions are timeless. Let us look first at a simple syllogism from term logic,and then at one from propositional logic. We will see why they canmislead us about the nature of logical argumentation about substantivematters. First:

All As are Bs.All Bs are Cs.So, all As are Cs.

This argument does not pose the problem that A is open-ended, notfinally identified —or, to put it slightly differently, that the concept of A(unlike the concept “human being” or “apple” or “lion”) isn’t finallyclosed. That is because A is not a concept but a formal symbol. Now let’sconsider an argument from propositional logic:

Given that P implies Q,and that Q implies R;therefore, if P then R.

As I have noted, A (or B or C) is not a concept but a symbol for one. Thepoint to take to heart is that the way symbols behave in formal argumentsmust not be confused with how concepts would behave. Nor are P, Q,and R propositions —they simply stand in for them. But to appreciatethe nature of substantive reasoning, it is necessary to explore the natureof propositions and concepts, not only their symbolic representations informal logic texts.

30 Thus, if one reasons, “I just left and locked my car in the faculty parking lot, and it isreasonable to hold that no one has moved it, so that is where my car is now, so that is whereI will find it upon my return,” this is not quite the same sort of formal reasoning as “Sincep strictly implies q, and q strictly implies r, and since p, therefore r.”

110 TIBOR R. MACHAN

Compare the formal symbolic syllogism with the following: “All humanbeings are biological entities; all biological entities are mortal; so all humanbeings are mortal.” The concepts here are not finally locked in, so it isintelligible to propose that the conclusion does not follow since the sec-ond premise might be false. The mortality of animals is not a logical truth,as one might put it in terms of contemporary analytic philosophy. But sowhat?

Indeed, science fiction writers create plausible enough stories by oftendenying various parts of such syllogisms. The conclusions are, thus, for-mally valid only if the statement of what is an animal or biological entityis treated as closed or finally true, the last word on the subject. But thensome will object that definitions merely stipulate and do not confirmreality’s unchangeable nature.

Purely formal deductive arguments do not have these problems. This iswhy most reasoning does not involve deductive inferences but concep-tual (logical) ones, in which a very significant role is played by theoriesand definitions. Rand made this a central part of her epistemology.

In any case, Hume’s view about our inability to deductively proveethical claims has had an enormous impact on the social sciences and onmoral philosophy. Nonetheless, Hume himself did not appear to haveembraced the full-blown ethical and political skepticism to which hisargument is thought to have given rise. Hume, after all, argued forcefullyin support of many normative claims. What he did not do, however, is layout a cogent explanation for how such support could be provided apartfrom deductive inferences. I would argue that this was due to Hume’s (aswell as many other philosophers’) embrace of the idea that unless some-thing is necessarily true, it cannot count as bona fide, certain humanknowledge, knowledge that can be defended against skepticism.

Thus, by affirming what to many appeared an unbridgeable gap betweenfactual and value judgments, Hume’s anti-rationalism laid the foundationfor positivism. This is the view that while what are called empirical factsare something we can know about, values are not within the province ofthe knowable.

This may be said to be a major reason why social scientists havemostly kept away from making value judgments. With respect to suchjudgments, social scientists invoke the “is-ought” gap, saying, in essence,that talking about values would be unscientific, since values are notsubject to factual confirmation. Since the hard sciences had always beenclosely associated with the idea that factual judgments can be con-firmed, the social sciences, in order to carry “the mantle of science,”were fashioned to mimic them. The method by which the evidenceand reasoning of the hard sciences was supposed to proceed —datagathering and unbiased analysis —needed to be followed in the socialsciences, and this appeared to preclude dealing with values, includingmorality and politics.


Rand, like many other philosophers, gave no sign of taking into accountthe distinction between Hume’s claim that we cannot deduce moral con-clusions from factual premises and the possibility of this leaving it openthat we could derive or infer them from factual premises in some otherway. Rand took the former claim to be affirming the impossibility ofrational moral judgments, thus showing a tendency toward rationalism,even though in her epistemological works she disavowed it.

VI. Morality Rightly Understood

Along with Martha Nussbaum, Philippa Foot, Henry Veatch, John Coo-per, and others, Ayn Rand was a naturalist; all of these theorists, in fact,developed what can be regarded as Aristotle’s biocentric metaethics (aswell as ethics). Such an approach resembles the foundations of medicine:the biocentric approach develops a conception of what can be construedas broad healthfulness and rests on it a set of guidelines for “healthy,”flourishing living (which, of course, extends beyond the narrowly bio-logical aspect of a person to the entire, robust living human being involved).Once such an understanding is achieved, various practices are identifiedas promoting a generally healthful life for the kind of living entities atissue, namely, human beings. These practices then are identified as moralvirtues, guiding the acting agent toward living and achieving a goodhuman life where that is a matter of choice, akin to how medicine teachesus how to lead healthful lives.31

This approach is, of course, applicable to all living things, so far as whatis of value to them is concerned, but with human beings the fact that theyconduct themselves by choosing to do so (to embark on their human lives)introduces the ethical or moral element (i.e., the self-responsible, self-directed element).32 Morality amounts, then, to the principles of the self-directed fulfillment of the purpose of human flourishing.

Let me now turn to the issue of what would be the most generalunderstanding of the good life for human beings. To get a clue on howthis would best be answered, let us consider how a general understand-ing of the goodness, excellence, or well-being of any living thing wouldhave to be understood.

31 Can flourishing and justice conflict? I would say they can, in very exceptional cases —when, as John Locke is supposed to have said, “politics is impossible.” See H. L. A. Hart,“Are There Any Natural Rights?” Philosophical Review 64 (1955): 175–91. Flourishing forhuman beings comes about through the practice of the virtues, including justice.

32 The precise way one makes this choice is a complicated matter because it happensincrementally in one’s life, starting with tiny decisions reaffirmed over and over again, aswell as neglected now and then (and, by some, even frequently). For more on this, see TiborR. Machan, Human Rights and Human Liberties (Chicago: Nelson-Hall Publishing Co., 1975),94ff. See also Tibor R. Machan, “Rand and Choice,” in The Journal of Ayn Rand Studies 7, no. 2(Spring 2006), 257–73.

112 TIBOR R. MACHAN

A good (or healthy) life of a gazelle or worm, generally considered —as,say, from the viewpoint of someone who studies such living things andhas occasion to identify specimens that are good versus mediocre or badones —involves knowing the kind of being at issue.33 To be able to iden-tify the specimens that are doing well, one needs to know what kind ofliving thing is being evaluated. The condition of such a living beingcannot be assessed without knowing (even if only roughly) what is itsnature, what are the characteristic attributes that could be in better orworse shape for the life of such an entity. Birds that get on with their livesby flying long distances, to use a simple example, require healthy wings,and if they lack them, they would not be good specimens of such birds.

To assess how well a human being is doing in terms of its physical ormedical condition, one would have to be well informed about the centralphysical constituents —organs, faculties, anatomical and physiologicalfeatures —of such a being and how well they are doing and are likely todo in the future. That is how a good (versus mediocre or poor) specimenwould be identified. This is what is meant by calling this a naturalistapproach —taking as one’s starting-point knowledge of the nature of thething being assessed or evaluated.34

When it comes to broader or more robust attributes than those bearingonly on the physical or medical condition of a living entity, it is even morenecessary to a have a clear grasp of what kind of being we have in mindto evaluate. Given the complexity of human beings, a clear grasp ofhuman nature, the kind of biological entities humans are, is required inorder to assess how well they are doing at living a human life.35

VII. Human Nature and the Objective Human Good

The history of thought is, of course, replete with attempts to identifyhuman nature —that is, to learn what a human being is as such. Theattempt has been made difficult because what has been deemed to countas a success in this undertaking has been influenced largely by the Pla-tonist notion of what the nature of something is, and by the Cartesiannotion of what it is to know something. Both notions offer up impossibleideals, and both, therefore, encourage skepticism.36

Definitions, which purport to state the nature of something, can none-theless be understood along lines that Rand and others suggest, namely,

33 Such an examination may take place in a survey of such animals or other living thingsafter some natural calamity, such as a severe drought or flood.

34 I consider in some detail the oft-discussed problems with such naturalism —with theidea of an objective “nature” —in Machan, Individuals and Their Rights, chap. 1.

35 In some cases, seriously impaired persons can excel in certain very significant respects(e.g., astrophysicist Stephen Hawking), though no one could reasonably deny their beingimpaired.

36 For more, see Machan, Objectivity, chap. 1.


contextually.37 The idea is that a definition of something has just thedegree of rigor or exactitude that it is reasonable to expect within theontological region in which the kind of thing defined is located. To besure, some things may be given highly formal definitions along linesrequired in geometry or some other formal systems such as mathematics.Yet when it comes to human beings, which are biological entities, thedefinition is open-ended and can admit of some borderline grayness; forexample, such a definition need not fully fit a damaged specimen or aninfant.38 We might say that the definition picks out the normal caseswithout barring extraordinary ones.39

Although it is accurate enough (within this conception of what a defi-nition must be) to state that human beings are best understood as the sortof beings that are rational animals, there is no need for it to be true thatevery member of the species is at all times fully capable of rationality (e.g.,while asleep, as an infant, or as a member in extreme old age afflicted withsenility). All these instances can be accounted for as belonging within thekind in question without needing to fulfill the criteria of membership as asquare would fulfill its definition in geometry. (Recall here J. L. Austin’sdiscussion of goldfinches, where he shows that what counts as knowingwhat a goldfinch is does not include predicting what it will always be!)40

To put it another way, experience tells us that rationality is typically avital part of the life of human beings. While some appear to do withoutexercising much rationality, they are typically parasitic on others whoexercise it routinely. This neo-Aristotelian conception of definitions, devel-oped by Rand and approximated by some others, is as good as it gets inwhat we require to think reasonably about human affairs,41 and can serveas the standard for moral evaluation of human actions and institutions.

VIII. Objective Ethics and Rational Individuals

For our purposes, special heed must be paid to the fact that humannature implies the individuality of each human being. A rational or think-ing animal engages its rational capacity by choice, volitionally, as a matterof its free will. The point is evident on the basis of several considera-tions —the explanatory force of the idea, introspection, scientific analysis,

37 Rand, Introduction to Objectivist Epistemology, chap. 5.38 For more on this, see Tibor R. Machan, Putting Humans First: Why We Are Nature’s

Favorite (Lanham, MD: Rowman and Littlefield, 2004). See also Austin, “Other Minds.”39 This is not the place to develop a full case for the contextualist position. The position

faces many challenges, especially with respect to how a contextual definition tracks theontological features of what is being defined, the nature of the thing.

40 Austin, “Other Minds.”41 For more along these lines, see Hilary Putnam, Words and Life (Cambridge, MA: Har-

vard University Press, 1994), esp. Part I, “The Return of Aristotle.” See also Douglas B.Rasmussen, “Human Flourishing and the Appeal to Human Nature,” Social Philosophy andPolicy 16, no. 1 (1999): 1–43. I too have addressed the issue in Machan, Individuals and TheirRights, 68–83.

114 TIBOR R. MACHAN

etc.42 Not only is each human being an essentially unique (and, in afundamental respect, irreplaceable) individual, but once a human beinghas activated (or has failed to activate) his or her rational capacity, resultswill emerge that extend the individuation process even further, in variouscreative, productive, destructive, or obstructive ways.43

The reason this is vital to our concern here has to do with what isimplied by the central place of individuality in human nature. The centralplace of individuality implies that although ethics can be objective, it israrely universal: different human individuals ought (and ought not) to dodifferent things, ought to conduct themselves differently. While no con-tradictory moral judgments can be part of ethics or morality, there areenormous differences in how people should act. (We can draw an analogyhere to the field of medicine, where there are some general principles ofhealthful conduct, but where particular judgments about cures, dosages,etc., will vary enormously.)

Along these lines —very much in the spirit of Aristotle’s metaethicalapproach, but with some variations from the specific understanding heproposed —while certain moral virtues are indeed to be practiced by every-one so as to live a morally good human life, there will be different waysof exercising these virtues. Some persons will need to practice certain (butnot all) virtues quite vigilantly, while other persons will need to practicedifferent ones more vigilantly (e.g., soldiers may need to be more coura-geous than, say, accountants, while accountants may need to be moreprudent than soldiers).

In all cases where we are concerned with what is the right thing to do,the answer can be ascertained objectively, based on facts and reasoning,but facts pertaining to who one is will matter alongside facts pertainingto what one is —indeed, considerations of the latter will place limits onthose of the former. What career one ought to choose, what political partyone should support, what kind of diet one should follow, and myriadother questions can be answered in this way, well supported by relevantfacts and principles.44

42 For more on this, see Edward Pols, The Acts of Our Being (Amherst, MA: University ofMassachusetts Press, 1983); Roger W. Sperry, Science and Moral Priority (New York: ColumbiaUniversity Press, 1983); and Tibor R. Machan, Initiative —Human Agency and Society (Stanford,CA: Hoover Institution Press, 2000).

43 For more, see Tibor R. Machan, Classical Individualism: The Supreme Importance of EachHuman Being (London: Routledge, 1998). An Al Capone or an Adolf Hitler becomes indi-viduated in destructive or obstructive ways.

44 Arguably, many so-called “self-help” books are, in fact, books of ethics, but they havenot been treated as such under the influence of the prominent conception of moral philos-ophy according to which morality pertains only to interpersonal conduct. For a good cri-tique of this position, see W. D. Falk, “Morality, Self, and Others,” in Morality and theLanguage of Conduct, ed. Hector-Neri Castaneda and George Nakhnikian (Detroit: WayneState University Press, 1965). Proper consideration of how one ought to act as the individualone is will be constrained by general principles based upon one’s membership in the humanspecies: e.g., rationality, respect for human rights, etc.


Much else could be said about all of this, but I want now to turn to aparticular issue that arises in connection with Rand’s idea of the objec-tivity of morality or ethics.45 She ties this objectivity to a choice a humanbeing makes to live his or her life, and this has provoked some to suggestthat she has, in fact, given away the ballgame and caved in to subjectivism.46

IX. Randian Objectivity Rests on Choice

Eric Mack and Douglas B. Rasmussen have both argued that sinceRand rests the moral principles she believes should guide us on a priorchoice to live, her substantive ethics —the set of virtues she considersbinding on human beings —is conditional, not categorical, and thus sub-jective rather than objective. Since I share the Randian idea that withouta prior choice there is no grounding for ethics, and since I also believe thather ethics has objective foundations, it will be important to spend sometime discussing these criticisms.47

If, as I believe, Rand is right that we have free will, so that there is achoice that we can (continually) make to live and to think,48 then prior to

45 Some have objected that Rand does not manage to derive ought claims from true claimsabout what is the case. See, for example, David Friedman, “Some Problems with Ayn Rand’sDerivation of Ought from Is,” posted on the newsgroup humanities.philosophy.objectivism(reportedly circa November 1996). The criticism rests on a formalist or Platonist conceptionof definitions and concepts. (See also Robert Nozick, “On the Randian Argument,” ThePersonalist 52 [Spring 1971].) As such, it misunderstands Rand and her neo-Aristoteliannaturalist metaethics and the conception of definitions she adopts. (See also Putnam, Wordsand Life.)

Actually, Friedman focuses not on ethics per se but on politics, complaining (as doesMack) that Rand cannot show, based on her ethical egoism, that one ought to respectanother’s right to life, liberty, property, etc. For a very good discussion of why Rand’s ethicalegoism is often misunderstood along similar lines, see Robert White, “A Study of the EthicalFoundations of Ayn Rand’s Theory of Individual Rights” (Auckland, New Zealand: Uni-versity of Auckland, a thesis submitted 2005). White shows, among other things, that Rand’segoism is peculiar in that it sees ethics as such as egoistic or eudaimonist.

46 Eric Mack, “Problematic Arguments in Randian Ethics,” The Journal of Ayn Rand Studies5, no. 1 (Fall 2003); and Douglas B. Rasmussen, “Rand on Obligation and Value,” The Journalof Ayn Rand Studies 4, no. 1 (Fall 2002).

47 Douglas Rasmussen makes the following challenging point about this issue: “Rand’sderivation of an ‘ought’ from an ‘is’ seems of limited value: if I choose to live, then I oughtto do such and such, but since there can be no obligation without this choice, there isnothing, either logically or morally, that obligates me to choose to live and thus no reasonto be moral. Possibly, there was something to Hazel E. Barnes (1967) including a chapter on‘Objectivist Ethics’ in her book, An Existentialist Ethics. Morality seems to be based on anirrational or arational commitment —the very thing Rand vehemently rejects” (Rasmussen,“Rand on Obligation and Value,” 75). The claim, however, that this makes Rand’s view “oflimited value” assumes that “objective” must mean exactly what Rand denies it must mean,namely, “intrinsic” —that is, that for a moral standard to be objectively true, it would haveto amount to a set of categorical imperatives and could not amount to a set of conditionalimperatives. For more, see Machan, “Rand and Choice.”

48 Or, to put it as I would, it is possible for us (that is, for each normal human individual)to keep up the initiative to live and to think. Let me stress that the choice to live is afundamental or first choice, or is best so conceptualized. It may be envisioned as beingmade, in the initial stages of one’s life, haltingly, implicitly, gradually, over and over again,

116 TIBOR R. MACHAN

such a choice we are at most surviving physically (fully dependent onothers) with a mind that is at most dormant, as yet unfocused, and thusnot ready to be aware of any alternatives among which an individualmight select one and reject the others. And if no commitment has yet beenmade to keep focusing one’s mind and thus keep identifying what theright course of conduct amounts to, there cannot be any responsibility.Ignorance may not be an excuse in the law, but total lack of engagementin reality is an excuse in morality.

I should like to suggest that the parable of the Garden of Eden mightilluminate what is at issue here: Prior to eating from the tree of knowl-edge, Adam and Eve had the potential to be fully human. But only afterembarking upon gaining knowledge by eating the apple from the tree didthis potential become realized and lead to having the personal responsi-bility to choose right from wrong. The rest of the animals, however,remained in a state of innocence, as it were, because gaining knowledgeremained out of their reach. Arguably, every human being replays thisdevelopment from being an amoral infant/animal to a developing human/moral agent.

David Kelley has provided some promising analysis here, spelling outthe foregoing point in less metaphorical terms:

For humans, the choice to live plays the same role that hard-wiredinstinct plays for lower animals. Our own choice is the source of ourcommitment to life; it is what gives us a lock on life as a goal. If thecommitment is not there, if I do not actually value my life, then mylife cannot be a value for me. It is not something I act to gain or keep.

Of course I am still subject to the facts of reality, [which include thefact] that life is the only thing capable of serving as an ultimate value.If I try to make something else my highest value —say golf, or choc-

expressing (as one might put it) the will to live, to be the human being one can become. Thechoice —or initial emergence of conceptual consciousness —I am considering is a logical firststep in action and thus cannot be motivated by some desire or knowledge. It is not a choicein the sense of a selection process, going into force with prior information at hand; rather,it is the exertion of an initial effort by a rational agent who at that stage of its developmentlacks other decisive prompters to action. This initiative is, as it were, the act of free will. (Formore on this perennial topic, see Machan, Initiative —Human Agency and Society.) As onecommentator has noted, what I am defending is a view of morality that is reminiscent of itscharacterization as “a system of hypothetical imperatives.” Only in this case, as the com-ment goes, “the hypothetical can only really go one way. We might think of the decision tolive as one that is inconceivable to reject. . . .” Only the decision is unlike others, since it isnot considered but commences or initiates any and all considerations. I should note, also, thatalthough the choice to live commits one to flourishing, it does not necessitate it. One canstart off with the choice to live but then default on what it implies, namely, to think and actvirtuously. But a consistent alternative to choosing to live with this implicit commitment toaiming to flourish still would amount to implicitly choosing to wither away. Indeed, if onechooses to live —to survive as a human being —instead of to perish, the implicit choice doesinvolve attempting to flourish. For human beings, the choice to live is not merely the choiceto survive —as one survives a plane crash by merely not dying in it —but the choice to liveas human beings can.


olate eclairs, or service to the proletarian revolution —just becausethey give me the most pleasure or inner satisfaction, then I am actingon subjective whim, in denial of the facts of reality. It would be OKon Hume’s or Sartre’s theory, because they are indeed subjectivists.Objectivism does not permit this sort of arbitrary and unconstrainedchoice. But neither does it commit the opposite fallacy of intrinsicismby ignoring the fact and the role of human choice. If I do not chooseto live, if it is a matter of indifference to me whether I live or die, thenfrom a moral standpoint there is nothing more to be said.49

The precise account of the prior state of mind —prior to the time when,gradually or by leaps and bounds, the individual commences the think-ing he is equipped to do and needs to do —is complicated by the fact thatit is yet impossible to administer a test for it. At that stage of humandevelopment, there is no self-consciousness by which what is going on inthe minds of infant human beings could be brought to light.50 Indeed,because it is hypothesized that in such a state the crucial process ofconceptual thinking has not yet commenced, and because it would haveto be such conceptual thinking that brings to light the introspective datathat would enable us to observe or take note of what goes on prior to thechoice to live and to think, the only way to grasp what goes on prior tothat choice is by a kind of rational reconstruction. This would involve aconsideration of what must have preceded that choice, given what we knowof those who have indeed developed, grown up, and made progress intheir human living and carried on accordingly.

Might it be the case, however, that the infant who fails to make thechoice to live and to think is in fact doing something that amounts to“lazy thinking”? Might it not, after all, be a moral failure not to take upthe task of gradually but relentlessly becoming aware? Yet, prior to whenone makes the choice to think, no alternatives as to what one might do

49 David Kelley, “Choosing Life” (unpublished paper, 2005).50 Neuroscientists, of course, have made observations of the brain’s behavior in infancy

that correlate with what (in later life) turn out to be mental activities, thus suggesting thatsuch activities do take place in infancy and perhaps even before. Bernard Baars, author ofIn the Theater of Consciousness (London: Oxford University Press, 1996), made some inter-esting observations to me via e-mail on this topic: “My friend Stan Franklin, who is amathematician/computer scientist, talks about ‘autonomous agents.’ Humans are nothing ifnot autonomous agents —not in a mystical sense, but in a very specific and causal sense. Oneof the ways we are autonomous is in terms of substitutability of resources. On the level offood, we like to eat meat, but if that runs out, potatoes will do. So there are options. In termsof human relationships, we’d like to have Julia Roberts as our playmate, but there are otherfish in that sea. In terms of making a living, we’d all like to be paid for our books, but . . .(etc.). I think that’s one of the keys to autonomy, substitutability of resources. Another isflexibility in acquiring knowledge. Humans are by far the best learners in the animalkingdom, obviously. But acquired knowledge also shapes who we are and how we defineour purposes and interests. Gerald Edelman . . . makes a big thing about the distinctivenessof the individual human brain. His Neural Darwinism gives a conceptual account ofindividuality from solid biological evidence.”

118 TIBOR R. MACHAN

could have come into focus, and the process of selection could not yethave commenced. But after the choice (or, as I prefer to put it, taking theinitiative) to think has occurred, even if in an incremental, small-step-by-small-step fashion, then to choose to reject the alternative to remain infocus and continue to think can be seen to involve an evasion of some-thing accessible and of enormous value —a case of self-betrayal, in light ofthe original choice to live.51

It is worth noting that the political philosopher Leo Strauss, someonevery different from Rand but also, arguably, within the same philosoph-ical tradition, suggests a similar way of understanding the birth (emer-gence) of morality. For Strauss, the good life for man is

simply the life in which the requirements of man’s natural inclina-tions are fulfilled in the proper order to the highest possible degree,the life of a man who is awake to the highest possible degree, the lifeof a man in whose soul nothing lies waste.52

In the Randian sense of objective morality, the principles of ethics emergeand are true because they further the goal one has acknowledged one hasin life, something fundamental that the agent has decided (at the momentconceptualization has commenced): namely, to embark on living a humanlife. One may view this decision as a very basic commitment —an oathtaken, as it were —to embark on living a human life. The principles onethen becomes aware of, following one’s taking of this oath, are akin to theobjective principles of, say, mechanical engineering that become requiredonce an engineer chooses to embark on building a bridge or a skyscraper.Given the initial choice, these principles are necessary and clearly iden-tifiable as such.53 Even as one embarks on playing some game, such asbridge, one commits oneself to following certain rules that are objective,though they are conditional upon the existence of the game (which didnot need to exist).

51 The word “choose” can mislead because it is often used to mean “select,” and forselections there need to be several alternatives. This fundamental choice, however, is of the“on versus off” type, so “taking the initiative,” which involves moving from rest to motion,more accurately characterizes what may well be going on here.

52 Leo Strauss, Natural Right and History, 2d ed. (Chicago: University of Chicago Press,1970), 127.

53 For a detailed discussion, see Machan, “Rand and Choice.” At this point, I should makeclear that the choice not to live, made at the initial stage of one’s life (that is to say, the choicenot to initiate one’s living) is premoral. One must embark upon life before one is bound bythe principles that guide it. Yes, one’s flourishing human life is potentially one’s end, but itis not one’s end in fact without one’s having embarked upon living. I should also note thatnot choosing to live, not initiating one’s life process, as it were, is not the same as embarkingupon suicide, which is done when one has lived for quite a while and is certainly open tomoral criticism by that stage of one’s life. Should one decide to commit suicide —maybejustifiably —the method used would also be subject to moral evaluation.


Of course, the Randian version of objective morality assumes free willand assumes that there is a point when a young human being summonshimself or herself to the task of thinking. If one rejects these assumptions,which neither Mack nor Rasmussen does, one is opting for a radicallynon-Randian approach to ethics.

X. Why Is Morality Objective?

As I have already noted, by “objective” I mean that the claims aboutwhat one ought to do or ought not to do are capable of being shown tobe true by reference to our very long history of experience with and studyof facts that are open to any rational mind’s apprehension (although notbeyond a shadow of doubt).54 The fact that someone has chosen to embarkon living a human life is one such (very general) fact. Other such factsinclude what the agent’s life requires to flourish, namely, the practice ofvarious virtues in the context of who that individual is, what his or heroptions and resources are, etc. Indeed, a Bill Gates would have to practicethe virtue of generosity in very different ways from someone at the pov-erty line: the specific judgment of how much they each ought to give tocharity would be different, though the general judgment that peopleought to act generously would hold true.

Without the choice to embark on living one’s human life, the self-binding nature of ethics would be missing. Why should one act as ethics —the virtues of living a good human life —would require one to act? Theanswer is that one should act this way because one has freely accepted thefact that one’s primary goal is to live as a human being (i.e., to live in away that is fitting for the kind of organism one is). Thus, one has com-mitted oneself to what this kind of life requires, on analogy to the moreparticular case of being bound by the ethics of a given profession that onehas taken up as a matter of free choice. For instance, only those notconscripted into the military can be said, properly, to be bound by mili-tary ethics, not those who were drafted and made no commitment to liveby the principles of military life (although this isn’t what, misguidedly,military law requires from conscripts). The same goes, arguably, for anyother profession.55

In summary, then, what I have tried to show is this: In a naturalistmetaethics such as what we find in Ayn Rand’s “The Objectivist Ethics” —aswell as in Martha Nussbaum’s Therapy of Desire and in Philippa Foot’s

54 This is why objecting to conclusions on these matters on the grounds that it is logicallypossible (i.e., that there is no formal contradiction in the denial) that they aren’t so isirrelevant. As Wittgenstein puts it, “The reasonable man does not have certain doubts”(Wittgenstein, On Certainty, 29e). That is, some doubts are unreasonable.

55 For more on this point, see Tibor R. Machan and James E. Chesher, A Primer on BusinessEthics (Lanham, MD: Rowman and Littlefield, 2003). See also the introductory essay, “Ethicsand Its Uses,” in Tibor R. Machan, Commerce and Morality (Lanham, MD: Rowman andLittlefield, 1988).

120 TIBOR R. MACHAN

Natural Goodness —an objective ethical position can be developed basedon a sound understanding of human nature (that is, an understandingderived from experience and the study of and reflection upon humannature). In Rand’s and my version, however, two vital issues are centralto the substantive ethics that emerges: namely, individuality and freedomof the will.56

What comes out of this inquiry is a system of moral virtues to guideone in living a good or excellent human life to which one has made acommitment as a rational individual, a commitment that binds one to actaccording to judgments based on such a system of virtues. The particu-lars, of course, will be quite cumbersome to identify and will require, inmost cases, knowledge of the individual agent who is to act in one oranother way (“local” knowledge). But this is no different from how, inattending properly, prudently to one’s physical well-being, one will needto know one’s body, as well as general principles of sound medicine,nutrition, and the like.

It seems to me that this approach to ethics preserves a reasonableconception of objectivity for ethical judgments. It is also an approach thatreflects quite faithfully how most people make their day-to-day ethicaldecisions in life (especially if they aren’t guided by various skepticallyconceived metaethical theories). What may be important to mention here,although the development of the idea would require considerably morespace than I have at my disposal, is that the concept “objectivity” coulditself be contextual rather than what it is widely taken to be, namely, justone sort of process or method involved in gaining knowledge.57 Objectiveknowledge of A will most likely reflect in certain important respects theontology of A —what type or even kind of being A is. Obtaining objectiveknowledge, in short, will not amount to an identical process in each casebecause what this knowledge is of will be reflected in what it is to knowit objectively. To put it somewhat differently, since objectivity concerns themeans by which knowledge is obtained —the type of evidence and argu-ment required to gain knowledge of the world as it is —it makes goodsense to suppose that if different facts about the world provide differentsorts of evidence and argument by which to learn of them, the objectivityat issue will itself be different in these cases.58

This explains, in the most general terms, why objectivity in ethics is notidentical to objectivity in, say, chemistry or psychology. Objectivity is in

56 Martha Nussbaum, in contrast, seems to be concerned with the human good withoutgiving a central role to choice. I discuss Nussbaum in Tibor R. Machan, LibertarianismDefended (Burlington, VT: Ashgate, 2006), chap. 19, “Two Philosophers Skeptical of NegativeLiberty.”

57 Of course, if (in fact, in the realm of ontology) everything is the same type or kind ofbeing (e.g., matter-in-motion, sense data, spirit, or whatever), then objectivity will involvejust what it takes to properly grasp this one type or kind of being.

58 In this account, truth is a property of judgments or statements, not of abstractly existingpropositions. An unknown or unknowable truth is, accordingly, nonsense.


part related to what type or kind of being one knows. In ethics, sinceobjectivity involves coming to know the guidelines for human conduct(the conduct of a volitionally conscious living being), it follows that thevery objectivity of the knowledge will reflect this. Accordingly, unlike in,say, chemistry, in ethics objective knowledge is conditional upon whetherthe being that is to be guided by the principles of ethics has made thechoice to commence or embark on its life. Objective knowledge is also,following Aristotle, a matter of “for the most part,” not akin to principlesin geometry but more akin to those in medicine. The point of such anethics is to provide guidelines toward achieving the good life of theorganism in question —in this case, of a human being.

If this is the right way to understand objectivity as far as moral knowl-edge is concerned, what are we to make of an influential viewpoint con-sidered by Thomas Nagel in connection with Bernard Williams’s ethicalposition?59 Nagel notes that Williams ridicules the idea that “if liberalismis correct, it must apply to all those past people who were not liberals:they ought to have been liberals, and since they were not, they were bad,or stupid, or something on those lines.” 60 The suggestion from Williamsis clearly that the idea of objective morality is misguided because moraljudgments are not universalizable.61

Yet the contextual theory of knowledge that Rand develops and othershave elaborated would address the issue with which Williams is con-cerned (according to Nagel’s succinct statement of Williams’s point). Cer-tainly Williams’s viewpoint is a very plausible one: few would insist thatthe interpersonal ethics or political institutions of ancient Greece ought tohave conformed to modern principles of right conduct or of justice. Thisis what is so insightfully developed by the political theorist Hanna F.Pitkin in her analysis of the concept of justice as both stable and variablethroughout human history.62 Thus, to repeat a point I have made else-where, “certain principles and truths do endure despite all the variations.For example, although there are options in how children might be raised,some basic ideas do remain true in any human community, any era. It isunderstood that parents have the responsibility to prepare their childrenfor adulthood —a basic and universal moral assumption. . . .” 63 Of course,just what will constitute such preparation in the fourth century b.c., as

59 Thomas Nagel, “The View from Here and Now,” London Review of Books 28, no. 9 (May11, 2006). (For my purposes, Nagel’s summary captures the critical point that Williamsvoices.)

60 Ibid., 10.61 For more on this, see Machan, “Why It Appears That Objective Ethical Claims Are

Subjective.”62 Hanna F. Pitkin, Wittgenstein and Justice: On the Significance of Ludwig Wittgenstein for

Social and Political Thought (Berkeley: University of California Press, 1972). See also Tibor R.Machan, “Can There Be Stable and Lasting Principles?” International Journal of Social Eco-nomics 32, no. 1/2 (2005): 218–27.

63 Machan, “Can There Be Stable and Lasting Principles?” 225.

122 TIBOR R. MACHAN

compared to the twenty-first century a.d., will differ, so the moral claimsvalid at each time will also differ, even though they can be equally trueand though they are based on lasting and stable moral principles. More-over, we can multiply examples by considering how we ought to act ininnumerable circumstances.

The point Nagel attributes to Williams —that “personal ethics and polit-ical morality cannot aspire to the kind of objective validity that is areasonable aim for science and mathematics” 64 —is not accurate. In sci-ence, there is variation based on historical context, not unlike in ethicsand political morality, given that the facts available for cognizance atdifferent points in the history of science vary in both the natural and thesocial sciences. Mathematics is different —although even there some sim-ilarities arise, based on context65 —mainly because mathematics is a for-mal discipline, like logic.

XI. Summing Up for Objectivity

This has been a short discussion of an issue that arises in metaethics.The issue has been addressed via Ayn Rand’s attempt to defend theobjectivity of ethics. I have attempted to briefly but accurately show whyRand believed that ethical knowledge is objective —that human beingscan know what is morally right and wrong —based on her understandingof “objective” as “neither revealed nor invented, but as produced byman’s consciousness in accordance with the facts of reality, as mental inte-grations of factual data computed by man —as the products of a cognitivemethod of classification whose processes must be performed by man, butwhose content is dictated by reality.” 66

Ayn Rand called her philosophy Objectivism because she wanted to beknown as someone who defended our ability to know the world as it is,not as distorted or constructed by our minds, not as we wish it to be, etc.Subjectivism and relativism propose these other positions and Rand wasagainst them; indeed, she considered those who promoted them, such asImmanuel Kant, vicious.

When it came to one of the most controversial areas of human knowl-edge, namely, ethics or morality, Rand insisted that she was an objectivisthere as well, unlike, say, the economists Milton Friedman and Ludwigvon Mises, who denied that we can know (objectively) what is right orwrong. Friedman, for example, wrote:

The liberal conceives of men as imperfect beings. He regards theproblem of social organization to be as much a negative problem of

64 Nagel, “The View from Here and Now,” 10.65 For an example of how context matters in formal disciplines, see Imre Lakatos, Math-

ematics, Science, and Epistemology, ed. John Worrall and Gregory Currie (New York: Cam-bridge University Press, 1978).

66 Ayn Rand, Introduction to Objectivist Epistemology, 54.


preventing “bad” people from doing harm as of enabling “good”people to do good; and, of course, “bad” and “good” people may bethe same people, depending on who is judging them.67

Yet Rand’s ethical objectivism isn’t all that simple. She isn’t a straight-forward realist who believes that ethical principles are evident in theworld independently of human existence. This is what is implied byobjectivity in, say, the natural sciences. If a principle of physics or chem-istry is objective, that means the principle is part of the world regardlessof whether anyone knows it to be so. For Rand, ethical objectivism is notlike this. The reason is that without human beings or some beings verymuch like them, there would be no principles of ethics or virtues. ForRand, ethics emerges with human (or possibly some very similar) lifeonly. Once we have human beings living their lives, ethics does apply tothem, and this is knowable by us as an objective fact. Nonetheless, thereis something conditional about this type of objectivity, namely, the factthat human life is a free option to each individual.

Rand advanced the view that only when one makes the choice to liveand think (which are practically the same thing, as she understood them)does one become bound by the principles that must guide human living.This idea has proven to be quite controversial. Do people generally makea choice to live and think? Or are they pretty much living and thinkingbeings without any choice to be such ( just as they have no choice aboutwhether to be born)? If the latter, then it would seem that the principlesof ethics are binding on them independently of any choice, as a necessaryand not conditional aspect of their lives. Why would Rand think that thisisn’t the case?

Before answering that question, it bears noting that conditional prin-ciples can be every bit as objective as unconditional ones. For example, letus take it that the principle of gravitation is unconditional and necessary(although in some very broad metaphysical sense it might well be con-ditional: if no objects with mass existed, there would be no gravitation).Let us take the principles of a sport or game in which, while the game(say, baseball or bridge) need not have existed at all, once it does, there areprinciples or standards of conduct that are (conditionally) necessary. Inbridge or baseball, if one wants to get something done —make a bid or hita home run —there are certain ways one must go about doing it; not justany way will do. Of course, the rules of the game could be changed, butsimilarly, if (improbably) human nature were to change, the principles ofethics would change as well. And this could be objectively ascertained.

Rand believed that human beings have free will, so whether they willtake up the task of living and thinking is up to them. They initiate these

67 Milton Friedman, Capitalism and Freedom (Chicago: University of Chicago Press, 1962),12.

124 TIBOR R. MACHAN

tasks; they are not instinctual or automatic. Rand held, in consequence,that whether a person is bound by the principles of ethics is in the lastanalysis up to him, although the alternative would be annihilation (andthus effectively irrelevant). As she put the point:

Life or death is man’s only fundamental alternative. To live is hisbasic act of choice. If he chooses to live, a rational ethics will tell himwhat principles of action are required to implement his choice. If hedoes not choose to live, nature will take its course.68

As with many aspects of Ayn Rand’s philosophy, her views on theobjectivity of ethics and on free will are only sketchily laid out by her.Only in the field of epistemology did she advance a fairly detailed andin-depth theory. Nonetheless, as sketched by her and related here, thereseems to be nothing amiss in the Randian idea of objectivity in ethics.Once we commit to living our human lives —and there is no other option,other than fading out of existence —we are freely but firmly bound (byour own choice to live our human lives) to the principles of ethics whichare knowable by us. (In the special areas of human life, the same holds:it is only once one decides to become a doctor or a lawyer, for example,that the ethics of the profession become binding.)

In the end, then, I believe I may now conclude that it is most likely truethat objectivity is possible in ethics, provided that knowledge is con-ceived in an essentially non-Platonist, non-idealist fashion and providedthat “objective” is not taken to mean that moral values are intrinsic.69

Business Ethics, Argyros School of Business and Economics,Chapman University

68 Ayn Rand, Philosophy: Who Needs It? (Indianapolis, IN: Bobbs-Merrill, 1982), 118. Butsee also her hero’s statement in her novel Atlas Shrugged (New York: Penguin Putnam, 1992):“This, in every hour and every issue, is your basic moral choice: thinking or non-thinking,existence or non-existence, A or non-A, entity or zero” (1020). This suggests that even whatRand takes to be the premoral choice is subject to moral evaluation. I take her here, however,to be referring to the choices one makes following having made the basic choice that, so to speak,constitutes one’s entry into the moral universe. One might see it, again, on analogy withsomeone having made the choice to be a doctor, after which he or she is committed topaying close attention to medical matters. But prior to that choice, no such obligation exists.

69 This is a crucial point stressed by Ayn Rand throughout her discussion of objectivity inall her writings, but especially in her Introduction to Objectivist Epistemology.


THE IMPORTANCE OF THE SUBJECT IN OBJECTIVEMORALITY: DISTINGUISHING OBJECTIVE

FROM INTRINSIC VALUE*

By Tara Smith

I. Introduction

Why does the claim of objectivity in morality spook us so —more sothan in other realms? Claims of objectivity hardly enjoy unanimous con-fidence in other spheres, yet in many areas, people are much more readyto accept that knowable facts govern and we can definitively answer ourquestions (in chemistry, computers, baking, biking, and so on). With eth-ics, however, more people are hesitant to say, “These views are right.”Obviously, not everyone is reticent. Many people, sometimes fueled byreligious ardor, are unequivocally devoted to the certainty of their moralviews. Nonetheless, we do find more people who are hospitable to claimsof objectivity in other realms to be decidedly less sure-footed when itcomes to ethics. Why is this?

Numerous factors, no doubt, contribute.

1. People disagree a great deal about moral issues. Standards andpractices across cultures and within the same culture often divergedramatically. Many people resign themselves to such disagree-ment as a permanent, unalterable fact of life and conclude that ittestifies to the absence of objective moral truths.

2. It is difficult, in ethics, to prove that you are right. From the factthat it is difficult to prove that a moral position is correct, someconclude that it is impossible to prove a position correct and thatthere are no right answers.

3. People are uncertain of how to measure moral judgments; they arenot sure what yardsticks are properly used to assess competingclaims. This, in turn, results from widespread uncertainty aboutthe purpose of ethics and exactly what ethics is. The prevailingview is that morality governs social relations and is a means ofgetting along with other people. Its goal (at least, among thosewho think that it has one) is to facilitate a certain ease in social

* Thanks to Ellen Paul and Harry Dolan for helpful editorial comments on an earlier draft.The essay has also benefited greatly from questions raised by the other contributors to thisvolume, as well as from conversations with Marc Baer, Greg Salmieri, Onkar Ghate, LeonardPeikoff, Darryl Wright, and Allan Gotthelf.


relations, to promote the smooth functioning of society. Such easeand smooth functioning are themselves liable to be taken to meandifferent things, however. And this haziness in understanding whatethics is leaves us without a clear standard for evaluating differingcodes of ethics or more specific moral claims.1

4. Ethics is emotional. It occupies touchy territory. Those who dis-agree often care fervently about the issues in dispute. People arewary of challenging views that are so deeply felt, since doing socan seem like an attack on the person who advocates those views.The overlay of ethics with emotion, moreover, often leads peopleto regard differing thoughts about moral issues as merely expres-sions of emotions.

5. Ethics is abstract. The claim that something is good or bad or rightor wrong stands atop layers of reasoning.2 Yet the basic subjectmatter of ethics —how each of us ought to lead his life —leadspeople to expect knowledge of moral truths, if such exists, to bemore readily available than knowledge in other abstract areaswhere the need for specialized and advanced study is plain, suchas medicine or engineering. How to lead one’s life, people pre-sume, shouldn’t be rocket science. If it is difficult to understandmoral truths, that must be because no such truths exist.

6. Finally, skepticism about the credentials that qualify those whoprofess to know moral truths fuels doubts about objectivity. Who’sto say what’s right or wrong?

This is far from an exhaustive catalog, and it ignores the interplay ofvarious factors that explain people’s reluctance to embrace moral objec-tivity. What I wish to focus on are the philosophical roots of this attitude.Philosophically, I think that the issue primarily comes down to two fac-tors: misunderstanding of the function of ethics, and misunderstandingof the nature of moral objectivity. In this essay, I shall address the latter.I will elaborate on Ayn Rand’s conception of moral objectivity because Ibelieve that it offers an illuminating perspective on the traditional debatebetween subjectivism and objectivism that can help to move that debateforward.

The opposing camps in this debate are long entrenched. The more thatis written in defense of each side, it sometimes seems, the more deeplyentrenched advocates’ commitments to their respective positions become.And the alternative between the two sides seems clear as day. Essentially,moral objectivists affirm moral truths that are independent of any human

1 While some philosophers in recent years have defended ethics as a means of achievingeudaimonia or personal flourishing, this is not a common idea outside the academy, andmany within as well as outside academia reject such a notion as a characterization ofprudence rather than of morality.

2 This claim is admittedly controversial.

THE IMPORTANCE OF THE SUBJECT IN OBJECTIVE MORALITY 127

beliefs or attitudes concerning the content of those truths; subjectivistsdeny this, maintaining that moral truth is always dependent on the viewsof a particular person or persons. (I will speak of “subjectivism” through-out to encompass all forms of subjectivism and relativism.) Yet in fact,Rand maintains, the defense of objective value has been dominated byadvocates of intrinsic value, which is not the same thing. Confusing thesetwo conceptions of value bolsters the credibility of subjectivism and dis-torts the entire debate. Genuine objectivity is rarely on the table. Randdistinguishes between three central alternatives: subjectivism, intrinsi-cism, and objectivism. In what follows, I shall focus on the distinctionbetween objectivism and intrinsicism so as to clarify the distinctive natureof the moral objectivity that she espouses. Doing so will reveal a crucialelement that is usually overlooked: the importance of the subject in objec-tive morality.

My presentation of Rand’s view is sympathetic, but it is not a fulldefense of her moral theory. While I will naturally need to present certainpoints of argument to convey the plausibility of relevant portions of hertheory, my mission here is to present a perspective that can enhance ourunderstanding of the traditional debate. Regardless of whether one acceptsRand’s larger moral theory, if she is right about the frequent confusion ofobjective with intrinsic value, her trichotomy offers a more accurate pic-ture of the alternatives and thereby places us in a better position to assessthe relative merits of moral objectivism and subjectivism.3

The essay is divided into three main parts. I shall first present Rand’sview of moral intrinsicism; next, elaborate on her own view of the natureof moral objectivity; and lastly, underscore certain features marking thedistinctiveness of her view. Throughout, I will speak primarily of theobjectivity of value because value serves as the foundation for all othermoral judgments, in Rand’s view. I shall also use “ethics” and “morality”interchangeably.

II. The Intrinsicist Model of Moral Objectivity

When people contemplate the objectivity of ethics, very often whatthey have in mind is the notion that value or propriety resides in certainthings. They conceive of objective value as roughly some sort of substanceor property that exists “out there in the world,” independently of humanattitudes. Value is considered a type of external existent, and the questionfor objectivists and subjectivists to settle is whether any of this hypoth-esized stuff actually exists.

3 For more in defense of her view, see Tara Smith, Viable Values: A Study of Life as the Rootand Reward of Morality (Lanham, MD: Rowman and Littlefield, 2000); Tara Smith, Ayn Rand’sNormative Ethics: The Virtuous Egoist (New York: Cambridge University Press, 2006); andLeonard Peikoff, Objectivism: The Philosophy of Ayn Rand (New York: Dutton, 1991), esp.chap. 7.

128 TARA SMITH

Rand maintains that many accounts of objective ethics actually thusreflect intrinsicism, the view that “the good resides in some sort of reality,independent of man’s consciousness. . . .” 4 Intrinsicist theories hold that“the good is inherent in certain things or actions as such, regardless oftheir context and consequences, regardless of any benefit or injury theymay cause to the actors and subjects involved. It is a theory that divorcesthe concept of ‘good’ from beneficiaries, and the concept of ‘value’ fromvaluer and purpose —claiming that the good is good in, by, and of itself.” 5

What —or who —is Rand talking about?Plato represents the paradigm of the intrinsicist view. A champion of

objective morality, Plato holds that all moral goodness reflects the Formof the Good. Every Form, on his account, is an essence or universal thatexists on its own —in a realm apart from the realm of particulars. All of theindividual things that people encounter in everyday experience are whatthey are insofar as they partake of the relevant Form. (The relationship ofpartaking is notoriously elusive.) What is important for our purposes isthat it is only due to the freestanding existence of something that isgoodness, on Plato’s view, that anything else can be good or possess anyof the derivative moral properties (such as being right, wrong, permissi-ble, virtuous, vicious, etc.).6

This intrinsicist outlook, in various permutations, continues to the presentday. Consider several more-recent presentations of the objectivity ofmorality.

Michael Huemer, an intuitionist advocate of moral objectivity, soundsthoroughly Platonic in writing that “[m]oral claims are assertions about aclass of irreducible, objective properties, which cannot be known on thebasis of observation.” 7 Geoffrey Sayre-McCord describes moral objectiv-ists as holding that the truth conditions of moral claims “make no refer-ence to anyone’s subjective states or to the capacities, conventions, orpractices of any group of people. Underlying objectivism is the sense,well articulated by [W. D.] Ross, that ‘it is surely a strange reversal of thenatural order of thought to say that our admiring an action either is, or iswhat necessitates, its being good. We think of its goodness as what weadmire in it, and as something it would have even if no one admired it,something that it has in itself.’ ” 8

4 Ayn Rand, “What Is Capitalism?” in Rand, Capitalism: The Unknown Ideal (New York:Signet, 1967), 22.

5 Ibid. Rand offers a fuller statement of the three views (subjectivism, objectivism, andintrinsicism) there. See also Peikoff, Objectivism, 254.

6 In dialogues other than the Republic, Plato sometimes departed from this account andportrayed value as conditional on a person’s making proper use of it. See, for instance,Euthydemus 281ff., Laws II.661, Laws I.641c, and Gorgias 470e. See also Mark LeBar’s discus-sion of this in his essay in this volume. Thanks to Mark LeBar for reminding me of this.

7 Michael Huemer, Ethical Intuitionism (New York: Palgrave MacMillan, 2005), 99.8 Geoffrey Sayre-McCord, “Introduction: The Many Moral Realisms,” in Sayre-McCord,

ed., Essays on Moral Realism (Ithaca, NY: Cornell University Press, 1988), 19–20. The material


Ross eloquently articulates the intrinsicist model again when he affirmsthat “the moral order” expressed in propositions claiming that somethingis prima facie right “is just as much a part of the fundamental nature ofthe universe . . . as is the spatial or numerical structure expressed in theaxioms of geometry or arithmetic.” 9 G. E. Moore’s contention that abeautiful world has more value than an ugly one, regardless of whetheranyone occupies or observes these worlds, is another example of objectivevalue conceived as residing in select things.10 More recently, ThomasNagel echoes Moore by claiming that when a culture dies out, “somethingvaluable has gone out of existence. This cannot be explained by the harmto existing individuals, all of whom will have other things to do and otherways to flourish. . . . [I]ts disappearance would be a loss nonetheless,though a loss to no one. . . .” 11

Yet another statement of the self-sufficiency of objective morality comesfrom Russell Shafer-Landau, who contends that moral facts exist “inde-pendently of what any human being, no matter his or her perspective,thinks of them.” “Moral obligations constitute reasons for everyone to actas they direct, regardless of whether such reasons bear any relations toone’s existing commitments.” 12

According to Rand, these depictions of moral objectivity suggest thatthe alternative to subjectivism —to value’s existing “in people’s heads” —isvalue’s existing “out there,” lodged within particular things. This is amistake, however. In fact, objective value relies on contributions of both

quoted is from W. D. Ross, The Right and the Good (New York: Oxford University Press, 1930),89; emphasis mine.

In exactly the same vein, Ronald Dworkin holds that it is natural to say that we want tolook at a Rembrandt because it is wonderful, and not that it is wonderful because we wantto look at it. Ronald Dworkin, Life’s Dominion (New York: Knopf, 1993), 72. For statementsin which Dworkin clearly divorces a thing’s intrinsic value from its being good for anyone,see ibid., 69–71.

9 W. D. Ross, “What Makes Right Actions Right?” in Wilfred Sellars and John Hospers,eds., Readings in Ethical Theory (New York: Appleton Century Crofts, 1952), 184.

10 Moore does, however, disavow an equation of objective value with intrinsic value in“The Conception of Intrinsic Value,” in G. E. Moore, Philosophical Studies (London: Routledgeand Kegan Paul, 1922), 255.

11 Thomas Nagel, “The Many in the One,” The New Republic (February 27, 2006), 32. Nageldescribes intrinsic values as those values that are “not reducible to their value for anyone.”Thomas Nagel, The View from Nowhere (New York: Oxford University Press, 1986), 153;emphasis his. For closely related claims, see ibid., 154–55, 162; and Thomas Nagel, ThePossibility of Altruism (Princeton, NJ: Princeton University Press, 1970), 96.

12 Russell Shafer-Landau, Moral Realism (New York: Oxford University Press, 2003), 8. Forsimilar statements of moral objectivity understood in terms of independence of people’sbeliefs, see Peter Railton, “Moral Realism,” Philosophical Review 95 (April 1986): 164; andDavid Brink, Moral Realism and the Foundations of Ethics (New York: Cambridge UniversityPress, 1989), 20. For a brief examination of intrinsicism in the work of another contemporaryethicist, see my “Comment on George Sher, ‘Perfectionism: A Theory’ —Chapter 9 of Sher’sBeyond Neutrality,” paper delivered at a conference on “Concepts and Objectivity” held atthe University of Pittsburgh, September 22–24, 2006, esp. pp. 8–9.

Notice that Rand’s thesis does not demand that all advocates of objective value beintrinsicists. Pointing out that many are intrinsicists usefully clears brush that typicallyobscures the alternatives in the debate.

130 TARA SMITH

consciousness and external reality. Moral objectivity is a function of therelationship between subject and object. It is neither located in nor aproduct of either, to the exclusion of the other. While one might be able tofind elements of the intrinsicist and subjectivist schools with which Randwould agree, it is crucial to understand that such elements represent onlypartial truths, in her view, that need to be supplemented by the fullexplanation of the dynamic that gives rise to values.13

Let us turn, then, to Rand’s account of what this third alternative tosubjectivism and intrinsicism is.

III. Objectivity Without Intrinsicism: Rand’s View

To understand Rand’s account of the objectivity of morality, we need tounderstand some of the framework within which it is set. Thus, I will firstsketch Rand’s broader view of morality and then zero in on the featuresthat distinguish its objectivity.

A. The nature and origin of value

As I have explained in greater depth elsewhere, the function of ethics,in Rand’s view, is to enable human beings to flourish.14 Ethics is “a codeof values to guide man’s choices and actions —the choices and actions thatdetermine the course and the purpose of his life,” she writes.15 A value,

13 Jorge Gracia offers a broadly similar perspective to Rand’s in his “The OntologicalStatus of Value,” The Modern Schoolman 53 (May 1976): 393–97. Note that what Rand dubs“intrinsicism” does not refer only to the explicit assertion of the existence of intrinsic values.While advocates of intrinsic value certainly qualify as intrinsicists, so does anyone whosetheory, in essence, divorces claims about value and morality from things’ effects on humanbeings. For an argument against the existence of intrinsic value in particular, see my “Intrin-sic Value: Look-Say Ethics,” The Journal of Value Inquiry 32 (1998): 539–53. For more onRand’s understanding of the intrinsicist conception of value, see Darryl Wright’s essay inthis volume.

Rand’s statement that the intrinsicist believes that “the good is inherent in certain thingsor actions as such, regardless of their context and consequences” should not be taken toimply that intrinsicism precludes consequentialism. A consequentialist could be an intrinsicistabout which states of affairs (which types of consequences) are valuable and are thus to bepromoted. Bear in mind that consequentialism is a view about what makes actions right orwrong, rather than about which things are good or bad. Intrinsicism is a view about valueor goodness. Rand is not maintaining that intrinsicists hold that “context and consequences”are irrelevant to rightness or wrongness. Rather, the context of her statement makes clearthat she is speaking of views about what things are good (and good as ends, rather thangood as means). She is observing the nonrelational character of value, on the intrinsicistview, rather than making claims about the consequentialist character of rightness. WhileRand does refer to “actions” as well as to “things,” one could hold that an action hasintrinsic value if one believes that it is good (as opposed to: right) independently of itsconsequences. Thanks to Michael Huemer for prompting me to clarify this.

14 See Smith, Viable Values, esp. chaps. 4 and 5.15 Ayn Rand, “The Objectivist Ethics,” in Rand, The Virtue of Selfishness (New York: Signet,

1961), 13.


in turn, is “that which one acts to gain and/or keep.” 16 Values range frommaterial objects to relationships, conditions, positions, qualities of char-acter, psychological dispositions —things such as coats, cars, friendship,freedom, a career, knowledge, perseverance, initiative. Ethics’ concern isto identify the most basic, abstract values necessary to steer a person tothe achievement of more concrete values. The reason for having ethics isthat its guidance is a necessity of survival.17

Each living organism’s life depends on goal-directed action that satis-fies the organism’s needs. However simple or sophisticated the life-formin question —be it a sycamore, a silkworm, a hawk, or a human being —itmust engage in certain activities in order to obtain and process the fuelthat its sustenance depends on. Life is a process that consists of self-generated, self-sustaining action. Every organism must do certain thingsin order to maintain that process.

For human beings, unlike lower life-forms, much of the requisite actionmust be deliberately chosen. Consequently, we must know how to chooseand what to choose. This creates our need for morality. A moral codeidentifies the kinds of ends that a person should seek (values) and thekinds of actions that he should take to secure values (virtues).18 (It doesso, again, at a fairly abstract level, with concretes to be filled in as objec-tively validated by a particular individual’s context.)

The phenomenon of value in the normative sense —of things’ beinggood or bad, of certain ends as worthwhile and others not —is intelligibleonly against this background.19 Nothing can be valuable to inanimateobjects. All sorts of things can happen to them, of course, many of whichmay be good or bad for particular people who have reason to care aboutthe objects in question. (It is bad for me when my car’s front tire goes flat.)But it is only relative to goals and purposes —which are the exclusiveprovince of living organisms —that events can be good or bad. Nothingcan be good or bad for an entity unless something is at stake for it; it muststand to lose or gain something. While numerous layers of goals or pur-poses may be needed to understand how something is a net gain or lossin a given case, the fundamental alternative that gives rise to any “gain-

16 Ibid., 16. I will use terms such as “gain,” “acquire,” “obtain,” and “secure” as short-hands to refer to gaining and/or keeping.

17 Throughout, I will be speaking of “life,” “survival,” and “flourishing” interchangeably,such that “life” and “survival” refer not to the minimal clinging of a heartbeat, but to a lifethat is led in an optimal, life-furthering manner. For a detailed explanation of the relation-ship between survival and flourishing and of the justification of my usage, see Smith, ViableValues, chap. 5; and Smith, Ayn Rand’s Normative Ethics, 28–33. See also Rand, “The Objec-tivist Ethics,” 25–27; and Peikoff, Objectivism, 219–20.

18 I explain Rand’s understanding of the nature of virtues and their relationship to moralprinciples in Ayn Rand’s Normative Ethics, 48–52.

19 I will sometimes use “value” in a neutral sense to refer to things that a person does infact seek, whether with or without good reason, and sometimes in the stricter, normativesense to refer to those things that a person objectively should seek. It should be clear fromthe context which sense I am using.

132 TARA SMITH

ing” or “losing” is the alternative between life and death. Impact on anorganism’s living provides the basis by which we can objectively distin-guish good from bad.20 On Rand’s view, accordingly, for any humanbeing who seeks to live, those things that enhance his life are good; thosethings that jeopardize or detract from his life are bad. “Good” meansgood for the person in question, and “bad” means bad for the person.21

As this quick introduction indicates, Rand’s moral theory is naturalistic.It is facts of man’s nature that give rise to value and set the terms of moral-ity because it is these facts that create our need to treat certain things as valu-able and certain things as threats or impediments to values. Just asresearchers in medicine seek to learn the sorts of substances and actionsthat are most conducive to peak physical flourishing for human beings,Rand maintains that the role of ethicists is to identify the sorts of ends andactions (albeit, at an abstract level) that are most conducive to all-encompassing human flourishing, material and spiritual.22 A value is anend that contributes to a person’s flourishing. The basic normative guid-ance that emerges from this foundation is egoistic: a person should act inthe ways conducive to his well-being. (In this respect, Rand’s view is unlikethe views of most naturalists, but egoism is not my focus in this essay.)23

It is important to emphasize that a person should act in this way if heseeks to live. Unlike many advocates of the objectivity of morality, Randdoes not regard the commands of morality as a given; we do not bear anyunchosen duties. “Reality confronts a man with a great many ‘musts,’”she writes, “but all of them are conditional. The formula of realistic neces-sity is: ‘you must, if . . . if you want to achieve a certain goal.’” Thisapplies to morality as much as to anything else. If a person wishes tolive, he should follow the prescriptions of a rational moral code that tellshim how.24

20 Peter Railton offers a similar argument about goodness and badness being intelligibleonly to things for whom other things’ differing effects could matter. See Railton, “Facts andValues,” Philosophical Topics 14, no. 2 (Fall 1986): 9.

21 For much further explanation, see Smith, Viable Values, 83–97, and Smith, Ayn Rand’sNormative Ethics, 19–25.

22 Rand understands the spiritual as that which pertains to one’s consciousness (“TheObjectivist Ethics,” 35). Obviously, medicine is not indifferent to psychological well-being.Whereas medicine is concerned with dimensions of human functioning over which a personhas no direct control as well as dimensions over which he does, however, ethics is ultimatelyconcerned exclusively with the latter.

23 Henceforth, I will confine my discussion to values for humans and leave aside valuesfor plants and animals, unless the context indicates otherwise.

24 Ayn Rand, “Causality versus Duty,” in Rand, Philosophy: Who Needs It (New York:Bobbs-Merrill, 1982), 118–19. For fuller explanation of this important aspect of Rand’s thought,see all of that essay as well as Peikoff, Objectivism, 244–45; and Smith, Viable Values, 101–3,105–11. Douglas Rasmussen’s recent discussion of Rand’s account of the foundation ofmorality, I think, gives insufficient weight to this aspect of her view and thereby delivers anoverly naturalistic rendering of it. See Douglas B. Rasmussen, “Regarding Choice and theFoundation of Morality: Reflections on Rand’s Ethics,” The Journal of Ayn Rand Studies 7,no. 2 (Spring 2006): 309–28; see esp. 318–19.


Values, then, are conditional —as is all of the derivative moral instruc-tion that rests on the identification of values. Values are based on factsthat are not of man’s making —his nature and the needs of his survival —aswell as on a fact that is of his making: his embrace of his life, his com-mitment to living.

All of this indicates the basic manner in which value is objective, inRand’s view. Ethics serves a practical mission; things are good or badaccording to whether they help to accomplish that mission. Life is thestandard of value.25 To appreciate the difference between objective valueand intrinsic value, however, we need to turn to her more exact accountof what the objectivity of values consists in.

B. The relational character of objective value

Rand writes:

The objective theory holds that the good is neither an attribute of“things in themselves” nor of man’s emotional states, but an evalua-tion of the facts of reality by man’s consciousness according to arational standard of value. (Rational, in this context, means: derivedfrom the facts of reality and validated by a process of reason.) Theobjective theory holds that the good is an aspect of reality in relation toman —and that it must be discovered, not invented, by man. Funda-mental to an objective theory of values is the question: Of value towhom and for what? An objective theory . . . does not permit theseparation of “value” from “purpose,” of the good from beneficia-ries, and of man’s actions from reason.26

Notice that Rand characterizes the good as both “an evaluation” (by arational standard) and “an aspect of reality.” In his elaboration of Rand’sview, Leonard Peikoff makes value’s two-dimensional nature equally plain:“The good designates facts —the requirements of survival —as identifiedconceptually, and then evaluated by human consciousness in accordancewith a rational standard of value (life).” 27

One might well wonder: how can it be both? But there is no greatpuzzle here. A distinction from Christine Korsgaard may be helpful.Korsgaard has observed that the term “values” can refer to at least threedifferent things: valuable objects themselves (such as operas, flowers,friendships); the properties in virtue of which we deem certain things

25 For more on this, see Rand, “The Objectivist Ethics,” 13–27; Peikoff, Objectivism, 206–20;and Smith, Viable Values, 83–95.

26 Rand, “What Is Capitalism?” 22; emphasis in original.27 Peikoff, Objectivism, 243; emphasis added. See a similar statement on p. 397, which also

stresses that the evaluation must be rational, i.e., determined according to the facts of reality.

134 TARA SMITH

valuable (in flowers, for example, their appealing colors and scents); andthe valuableness that we attribute to the objects in virtue of those prop-erties.28 In Rand’s usage, a value is a thing (such as a flower or anopera —Korsgaard’s first sense), insofar as it possesses the properties(Korsgaard’s second sense) that render it rationally to be pursued by aparticular person (Korsgaard’s third sense). (There is actually a little moreto Rand’s view, but we’ll get to that shortly.) In their statements present-ing the nature of objective value, Rand and Peikoff are simply indicatingthe different aspects of a value that can be important to stress in differentcontexts.

The main substantive message is that objective value is relational, inRand’s view. Value neither exists nor is created entirely in our heads norentirely in external reality. Values are not reducible to psychological states,as in subjectivism; nor are they independent of them, as in intrinsicism.Both external and internal conditions are necessary for values. To be avalue, the thing in question must, in fact, serve a person’s life, and thatperson must seek his life. It is important to understand each of these.29

First, a value must in fact be beneficial to a person; it must bring himsome net gain. Given Rand’s argument that the phenomenon of value isintelligible only in the context of impact on a life, no freestanding value“in” various external things could exist. Nothing is valuable apart fromits bearing on a life. What a particular thing’s impact is, of course, cansometimes be difficult to judge. Moreover, the magnitude of impact canrange from the life-changing to the quickly forgotten. Winning a ten-dollar bet is not typically as valuable as gaining a job promotion; discov-ering a restaurant you like is not typically as valuable as discovering apassion for a particular field of work. Nonetheless, any objective value, beit modest or momentous, must offer some positive contribution to theperson’s flourishing. And, contrary to subjectivism, whether a given thingdoes so is a matter of fact rather than of anyone’s beliefs or wishes. Peoplecan be as mistaken about what is good or bad for them —about the effectsthat various things carry on their well-being —as about anything else.

I should caution, however, against reading this feature of objectivevalue in an unrealistically rigid way. Rand’s contention is not that thedesignation of something as an objective value for a particular personrequires infallible advance knowledge of things’ ultimate impact on ourlives. If it is truly, objectively rational for Blake, in his context of knowl-edge, to believe that item x or activity y would be a net benefit for him buthe turns out to have been mistaken about this, it remains the case that it

28 Christine Korsgaard, “The Dependence of Value on Humanity,” in Joseph Raz et al., ThePractice of Value, ed. R. Jay Wallace (New York: Oxford University Press, 2003), 67.

29 Moral naturalists such as Philippa Foot similarly see value as relational to humanwell-being. Their accounts of the necessary ingredients of that relationship are not alwaysthe same as Rand’s, however. For succinct statements of the subjectivist and intrinsicistviews, see Peikoff, Objectivism, 245–46.


was objectively valuable from his perspective. When discussing the objec-tivity of value, it is important to always bear in mind the different per-spectives of the relevant individual himself at the time of choice or actionand that of a more fully informed or hindsighted observer. This is not aconcession to relativism, but a simple recognition of the fact that anyassessment of something as objectively valuable is necessarily someone’sassessment and that humans are fallible, non-omniscient beings. Conso-nant with this, we can legitimately evaluate the same object or event(Blake’s having that house, taking that medicine, accepting that job) inopposing ways from the two perspectives, so long as we are clear aboutwhich perspective is being employed. There is no contradiction in simul-taneously holding that a particular job was not objectively valuable forBlake insofar as it did not deliver the expected net benefit and that the jobwas objectively valuable for Blake insofar as it was objectively rational forhim, at the time he accepted it, to expect that it would deliver this benefit.Error is not proof of a failure to be objective.30

Another way of viewing this might seem more natural. We shoulddistinguish between assessments of whether a person is being objective inreaching a conclusion about a thing’s value and assessments of whetherthe thing in question is in fact beneficial to him. Judgment of the latterrequires the more fully informed perspective that is not always available.Accordingly, when I say that an objective value must be beneficial to therelevant person, on Rand’s view, I mean that it must be beneficial as faras he has objective reason to expect.31 The objective character of this beliefshould become clearer as we proceed.

It is not enough that something stands to benefit a person, however, forit to be an objective value to him. Further, he must seek his life. (This isthe second critical element of the value relation that I noted above.) More-over and closely related, he must identify the thing’s beneficial impact onthat end.

Unless a person chooses to embrace his life —to maintain it, by seekingto achieve his well-being —all the facts in the world about various things’effects on his life will not matter. They will carry no value significance.Those causal relationships themselves will remain, of course (protein willstill be nutritious, certain activities will still be dangerous while manyothers will still be life-enhancing), but it is only in the context of the questfor life that values arise. It is a person’s desire to live that converts clinical,impersonal facts about things’ effects into values: things that he shouldact to gain or keep (because he seeks his life and because of the nature of

30 This is similar to the case of a jury that proceeds with scrupulous objectivity in reachinga verdict but later discovers that, for reasons that reveal no deficiency of objectivity on itspart, its verdict was factually incorrect. (The jury convicted an innocent man, for instance.)

31 I will leave aside here the further details of what this involves, but certainly it isincompatible with objectivity for a person to be casually indifferent to pertinent knowledgethat is readily available.

136 TARA SMITH

the things in question and their impact on that end). In an essay, Randobserves that “material objects as such have neither value nor disvalue;they acquire value-significance only in regard to a living being —particularly, in regard to serving or hindering man’s goals.” 32 Goals,however, are things that human beings deliberately adopt.33

I noted above that a thing’s being an objective value for a person alsorequires his recognition of the thing’s potential benefit to him. This rec-ognition is important because a value is a thing to go after; it is the objectof action (“that which one acts to gain and/or keep,” quoted in Sec-tion III.A above). Human beings must act in order to acquire many of thethings necessary for our existence. We need to engage in life-sustainingactivities; we need to secure life-sustaining ends. While some of the req-uisite actions occur automatically thanks to our genetic coding (digestion,circulation, etc.), many other requisite actions are volitional and must bechosen from among alternatives. Consequently, a person must know whichthings to choose. A person will not act for the sake of x unless he thinksx worth acting for, among the things that he has reason to seek. What thismeans is that no beneficial thing can function as a value for a personwithout his having identified it as beneficial. Nothing can function in theway that some things must function, in other words (namely, as prods tolife-sustaining action), unless they are seen as worth the effort.34 That willonly be the case if the person wants to live and sees a particular end as ameans of furthering that larger end. When a person does recognize some-thing’s potential benefit for him, his attitude, in effect, is: “Yes, okay, thisis likely to get me some of the sort of thing I am seeking.” If he fails torecognize a thing’s potential benefit, however, while the object in ques-tion may continue to carry some of its effects on his existence (the anti-biotic will still heal a wound; milk will still nourish), it will not be a valueto him.

One might object that I am begging the question here by assumingRand’s definition of value as something that a person acts to acquire. Ifthat is what a value is, then of course the recognition of a benefit will benecessary for the existence of a value. But isn’t hers an odd definition ofvalue?

The validity of Rand’s definition rests in the fact that a value is morethan a benefit. While it is useful to know which things stand to benefit

32 Ayn Rand, “From the Horse’s Mouth,” in Rand, Philosophy: Who Needs It, 96. AlthoughRand was speaking of material values in the passage cited, I am aware of no evidencesuggesting that she would restrict this claim to apply only to material values. Moreover, ina workshop with students, Rand observes that “that which satisfies a need is not necessarilya value . . .” (“The Objectivist Workshop —#4, Ethics and Politics,” transcribed by Ben Bayer,Ayn Rand Archives, Ayn Rand Institute, 11).

33 While we do refer to the goal-directed actions of a person’s bodily organs, whenspeaking of the goals of a human being (as opposed to the goals of one of his body parts),“goals” refers to objects of deliberate choice.

34 Since human action is volitional, the prods in question are not mechanical.


human beings, it is useful to know this because human beings must takeaction in order to acquire many of the goods that our lives depend upon.While, again, certain of the necessary actions occur automatically, it isbecause numerous equally vital activities do not so conveniently tran-spire that we need to identify the further types of actions that will enableus to achieve life-furthering ends and flourish. For this reason, we mustidentify not simply benefits, but values —that is, ends whose attainmentwould be beneficial but whose attainment is possible only if we takeappropriate action. The rewards of a particular friendship or career, forinstance, are benefits that depend in part on a person’s taking certainactions. The rewards of a desired commodity, for that matter, such as acoat or a car or a house, also typically depend on a person’s actionsenabling him to acquire and maintain the item.

I should thus refine my claim above that a value is “more than abenefit.” More exactly, a value is a particular type of benefit. Many thingsbenefit human beings without any deliberate cooperation on our parts(sunlight, rainfall, episodes of fortunate coincidence that help a person tobe hired for a job when he is one of several who are equally qualified, forinstance). Identifying benefits is not sufficient to sustain us, however.Among all potential benefits, we further need to identify those whosebenefit depends on or can be enhanced by our action, so that we cansubsequently take the relevant action. We need, in other words, to dis-tinguish benefits from the subset that constitute values, things we mustact to acquire. Rand’s definition, I think, is thus vindicated.35

The larger point vis-à-vis the relational character of objective value,again, is that because flourishing requires deliberate action of certainsorts, a value is not simply something that stands to benefit the personwho embraces his life. It is also something that he recognizes as such.36

We now, I hope, have a fuller grasp of the relationship that constitutesobjective value, in Rand’s view. For something to be a value is for it to bea benefit to a particular person who wishes to live and who identifies itsbenefit. The relationship involved is to a person’s consciousness as well asto his existential condition. As Allan Gotthelf and Gregory Salmieri havesummarized it, for Rand, values “are formed by a consciousness in accor-dance with the facts of reality. To be a value something must be identifiedby an agent as furthering his life. The identification is, in Rand’s termi-nology, man-made, as is the choice to live that gives it meaning. But therelationship between the value and the agent’s life is metaphysically given,

35 For more on the relationship between values and benefits, see Smith, Viable Values,84–85. On the related issue of the inability to impose or force a value on another person, seeRand, “What Is Capitalism?” 23.

36 A benefit, in contrast, does not require a person’s embrace of his life or his recognitionof the beneficial thing’s salutary effects. Certain things are beneficial for a person insofar asthey are life-preservative. This holds, regardless of the person’s attitude toward his life.Whether something strengthens a person’s prospects of living is quite independent ofwhether he wishes to live.

138 TARA SMITH

as is the need to identify this relationship conceptually.” 37 The objectivityof values thus reflects both facts that are independent of a person’s mentalstates and facts that are not.

C. The broader character of objective morality

We can now step back from the details of this relationship to appreciateRand’s broader image of the objectivity of morality. One significant impli-cation of the relationship I have described is that objectivity pertainsprimarily to people, rather than to things “out there” in the world. Objec-tivity, at bottom, is a normative rather than a descriptive concept. The twoare intimately connected, in Rand’s view; any valid normative claim mustbe grounded in an accurate description of the relevant facts. Indeed, thepractical purpose of morality makes facts all-important to the determi-nation of what a person should and shouldn’t do. The immediate point,however, is that objectivity refers primarily to a manner of using one’smind. In any sphere of knowledge, objectivity concerns the method thatguides a person’s thinking about an issue and the basis on which hereaches conclusions. Whatever the subject matter —biology, economics,history, cooking, snorkeling, brick-laying —an objective truth is one thathas been reached and can be validated by adherence to all and only therelevant evidence, governed strictly by logical inferences therefrom. WhatRand stresses is that the same holds in the realm of morality. A moralclaim is objective when it can be validated by objective methods.

Rand allows that objectivity is not entirely methodological. She distin-guishes two senses of “objective”:

Objectivity is both a metaphysical and an epistemological concept. Itpertains to the relationship of consciousness to existence. Metaphys-ically, it is the recognition of the fact that reality exists independentof any perceiver’s consciousness. Epistemologically, it is the recog-nition of the fact that a perceiver’s (man’s) consciousness must acquireknowledge of reality by certain means (reason) in accordance withcertain rules (logic). This means that although reality is immutableand, in any given context, only one answer is true, the truth is notautomatically available to a human consciousness and can be obtainedonly by a certain mental process which is required of every man whoseeks knowledge. . . . Metaphysically, the only authority is reality;

37 Allan Gotthelf and Gregory Salmieri, “Ayn Rand,” Dictionary of Modern American Phi-losophers, ed. John R. Shook (Bristol, UK: Thoemmes Continuum, 2005), 1998. For Rand’sdiscussion of the importance of appreciating the difference between the man-made and themetaphysical, see Ayn Rand, “The Metaphysical Versus the Man-Made,” in Rand, Philoso-phy: Who Needs It, 28–41.


epistemologically —one’s own mind. The first is the ultimate arbiterof the second.38

What Rand finds undeniable in the metaphysical sense of moral objec-tivity is that some things are such that a person should seek them. Doingso would advance his life. Moreover, human beings are capable of know-ing of these things. When we do know of them, however, it is because wehave identified them as values on objective grounds. To say that moralityis objective is thus largely a comment on how we attain knowledge ofwhat is good. Strictly, facts are not objective. They are not the sort of thingthat could be. It is only human beings who can be objective or non-objective in how we think about facts and in the claims that we makeabout them. Statements asserting facts can become objective when theyresult from the objective method.39

Given my claim that a moral claim is objective when it can be validatedby objective methods, one might wonder about the order of dependencehere. Rand’s distinction between metaphysical objectivity and epistemo-logical objectivity is widely accepted, but which, in her view, is pri-mary?40 Am I saying that, for Rand, epistemological objectivity determinesmetaphysical objectivity?41

No. Recall her claim that reality is “the ultimate arbiter” of epistemol-ogy. This certainly means that epistemological objectivity does not deter-mine reality. The relationships involved are somewhat intricate, however,in part because the question itself can be taken to refer to either episte-mological or metaphysical primacy.

The answer is best understood, I think, by considering: What is itthat leads human beings to devise the concept of “objectivity” in the firstplace? What leads us to differentiate valid from invalid means of gain-

38 Ayn Rand, “Who Is the Final Authority in Ethics?” in Rand, The Voice of Reason, ed.Leonard Peikoff (New York: New American Library, 1988), 18. Notice that both senses of“objective” were reflected in the statements from Rand and Peikoff that I quoted earlier,claiming that the good designates facts, conceptually identified and evaluated by a rationalstandard.

39 Accordingly, Rand does not believe that the phenomenon of objective value arises,strictly, prior to the conceptual level of consciousness. The term “objective,” Peikoff stresses,applies “only to values chosen by man. The automatic values that govern internal bodilyfunctions or the behavior of plants and animals are not the product of a conceptual process.Such values, therefore, are outside the terminology of ‘objective,’ ‘intrinsic,’ or ‘subjective.’ ”Peikoff, Objectivism, 243. Further, Rand remarks in the workshop cited in note 32 that theclaim that values are objective reflects a relationship in reality as identified by a humanconsciousness.

40 See Brian Leiter’s characterization of the distinction between metaphysical and episte-mic objectivity, for instance, in his “Introduction,” in Objectivity in Law and Morals, ed. BrianLeiter (Cambridge: Cambridge University Press, 2001), 1–2. I also discuss this distinctionbriefly in my “ ‘Social’ Objectivity and the Objectivity of Values,” in Science, Values, andObjectivity, ed. Peter Machamer and Gereon Walters (Pittsburgh, PA: University of Pitts-burgh Press, 2004), 143–71; see esp. 150–56.

41 Thanks to Scott MacDonald for prodding me to clarify this.

140 TARA SMITH

ing knowledge? Answer: the recognition of certain facts about humanconsciousness —in particular, its fallibility and its means of gaining knowl-edge. What we are responding to, in conceiving of the objective and thenon-objective, is the independent character of reality (external reality aswell as the pertinent features of the nature of the human mind). Yet thereason that we need the concept “objective” stems most directly from thefact that we control our consciousnesses —what we think, whether wethink, how we think. Human beings make choices that determine thecourse of our thought; we do not automatically make rational choices orthink in ways that reach accurate conclusions about reality. For this rea-son, we find it necessary to distinguish reality-tracking methods of usingour minds (objective) from non–reality-tracking methods (non-objective).

On the question of which type of objectivity is primary, then: Whatqualifies as epistemologically objective is grounded in and must be faith-ful to the nature of reality. In that respect, one might say that metaphys-ical objectivity is primary. Yet given that the concept of objectivity arisesin order to govern human beings’ thinking about reality, in an equallyimportant sense, one might say that epistemological objectivity is pri-mary. While reality “comes first” insofar as it dictates the need for theconcept of objectivity as well as what constitutes (epistemological) objec-tivity, it is significant that we coin the concept of “metaphysical objectiv-ity” only after realizing the need for objectivity in our method of thinking(that is, after recognizing the value of distinguishing the epistemologi-cally objective from the epistemologically non-objective). Indeed, speak-ing very strictly, the claim represented by a thing’s purported “metaphysicalobjectivity” (namely, that a given thing’s nature is as it is independentlyof us) is not properly characterized as a “metaphysically objective” fact orexistent.42 Facts are simply facts. The concept of objectivity arises becausewe come to realize the need to discipline our thinking about reality, butreality is what it is —facts are what they are —quite independently of thatthinking.

To put this last thought in a slightly different way: Strictly, the questionof objectivity does not arise for reality. Reality cannot be either objective ornon-objective. (Any purportedly “non-objective reality” is not truly real-ity.) The question of objectivity and the alternatives of being objective ornon-objective arise for us, rather, in distinguishing ways of thinking aboutreality. Then, once we have developed that epistemological concept ofobjectivity, we can ascribe the concept in an indirect way to reality itself,calling certain things “objectively” so or “metaphysically objective.” Butthe concept of metaphysical objectivity is derivative from the prior con-cept (epistemologically prior) of epistemological objectivity.

42 The parenthetical is a somewhat crude statement of the claim of metaphysical objec-tivity and leaves aside, for present purposes, certain exceptional sorts of “things” (such ascertain of our own mental states).


Thus, on reflection, even the statement that in one sense, metaphysicalobjectivity is prior seems not quite right. For it is not that “metaphysicalobjectivity” is prior to epistemological objectivity. Rather, reality isprior. Reality is what sets the terms of objectivity, what determines whatcounts as objective —epistemologically objective as well as metaphysi-cally objective. Our fallible means of gaining knowledge of reality leadsus to distinguish between objective and non-objective methods of thought(epistemological objectivity), and we subsequently apply that distinctionindirectly to reality itself, via the concept of metaphysical objectivity.

The upshot, for our purposes in differentiating Rand’s understandingof objective value from intrinsicism, is this. In Rand’s view, the assertionof objective values is not a report on the ontological status of certainpreexisting good things woven into the fabric of the universe. The searchfor objective values is not akin to a treasure hunt, driven by the question:Where is the stuff? References to objectivity in the moral realm are, to aconsiderable extent, prescriptive. They instruct a person not so much inwhat to think, as in how to think —how to use his mind in order to answermoral questions and to evaluate competing answers.43 The essence of theobjective method, again, is deliberate, logical adherence to reality, to rel-evant facts.44 Thus, the moral objectivity that Rand affirms consists in thefact that moral truths —whether assertions of moral goodness, badness, orderivative moral claims —can be validated by employment of this method.And, to tie this back to her view of the overall function of morality, thereason that a person should be objective in thinking about moral mattersis that objectivity offers his only means of identifying the facts that heneeds to identify, in order to chart a course of action that can result in hisflourishing. Reality —or a person’s successful navigation of reality —demands objectivity.

A second noteworthy implication of Rand’s conception of moral objec-tivity is her denial of a thesis often associated with moral objectivism, thebelief that objective values form a single set, identical for everyone. In herview, not all values are universal or equally applicable to everyone. Manythings, to be sure, will be valuable to all human beings in virtue of ourcommon human nature. Protein, exercise, and hygiene, for instance, aswell as certain spiritual values such as friendship, art, education, andpurpose, will be values to everyone who seeks to live.45 At the same time,the exact means by which a person achieves such universal values canvary, within limits. It is morally optional whether a person satisfies his

43 Rand once described ethics as “a science devoted to the discovery of the proper meth-ods of living one’s life,” which suggests that ethics is not primarily concerned with iden-tifying the appropriate objects of our pursuits, objects that antecedently possess intrinsicvalue. Ayn Rand, Introduction to Objectivist Epistemology, expanded second edition, ed. Leon-ard Peikoff and Harry Binswanger (New York: Penguin, 1990), 36.

44 See Peikoff, Objectivism, 116–21; and Darryl Wright’s essay in this volume.45 I do not mean that everyone is aware of the value of these things or that their value is

self-evident.

142 TARA SMITH

needs for art and relaxation through jazz and crossword puzzles or throughtheater and pinochle. It is morally optional whether he obtains his exer-cise through running or walking or playing basketball. The boundaries ofoptional values must be respected, in order for the more specific “values”to be genuine. Not any form of recreation will pass muster, since someare antithetical to human flourishing (binge drinking or Russian rouletteare obvious examples). The parameters defining the permissible range arethemselves objective insofar as they are grounded in the natural require-ments of human life.

Rand is not saying, then, that honesty may legitimately mean one thingfor Ben and another for Jerry. The basic principles by which a person canachieve the values that serve his objective well-being are uniform for allhuman beings. She is acknowledging, however, that many particular thingscan play objectively life-enhancing roles in some individuals’ lives with-out doing so in others’. Correlatively, moral objectivity does not entail asingle, exhaustive set of values that are identical for all persons.46

IV. The Distinctive Character of Rand’s Moral Objectivism

All of this presents the core of Rand’s understanding of moral objec-tivity. My suggestion at the outset was that this perspective could facili-tate progress in the long-standing debate over moral objectivity. Let meunderscore, therefore, how hers is a distinctive view.

Like the subjectivists, Rand maintains that value is relational. Unlike inthe subjectivist view, however, it is not relational simply to individuals’beliefs or attitudes. Rather, value is a function of a thing’s relationship toa person’s life, which encompasses his existential status as well as hisconsciousness. Matters of independent fact are indispensable to objectivevalue. In order to advance our lives, human beings must respect certainfacts that are not of our choosing. (Human beings need sleep, for exam-ple; therefore, I should value sleep. Unprotected intercourse is risky; there-fore, to avoid pregnancy and disease, I should value contraception. Thatcolleague is unreliable; I should not depend on him. That sister-in-lawtends to be insightful; I should speak with her more often.) The personaware of objective value is recognizing that the object in question hascertain qualities, rather than creating that fact.

In this, obviously, Rand sides with the traditional defenders of moralobjectivity. She differs from these intrinsicist “objectivists,” however, inso-far as they treat moral values as metaphysical existents. They conceive ofobjective values as givens of the universe, akin to rocks and minerals,prefabricated and contained in nature. The intrinsicists, in Rand’s view,reify an abstraction. They treat the product of objective thought —the

46 For more on optional values, see Peikoff, Objectivism, 323–24; and Smith, Viable Values,99–101.


identification of a particular thing as standing in a life-furthering rela-tionship to an individual —as if it were a freestanding entity.

It is important, in this vein, not to mistake Rand for a “relationalintrinsicist.” Her contention is not that life is inherently valuable and thatwhatever bolsters life is thereby a value,47 nor that certain things areinherently valuable but must be linked with a particular beneficiary inorder to constitute “working” or full-fledged values. On her view, moredistinctively, value does not exist for a person until and unless he seeksto live and appreciates the positive role that the would-be value can playin his life. Prior to that, all that exists are potential values. Because valueis not free-floating within certain specially endowed things, on her view,the relevant relationship is not one of simply finding some of this golddust and directing it to the benefit of a particular individual. What isnecessary is not awareness of preexisting value, but identification of thepositive relationship in which a given thing stands to one’s life.48

Rand holds that moral objectivity, like all objectivity, is essentially amatter of method. The intrinsicist, by contrast, believes that no particularmethod for discerning value is to be followed. Instead, you “just know”which things are good and right. According to Ross, “We have no moredirect way of access to the facts about rightness and goodness and aboutwhat things are right or good, than by thinking about them; the moralconvictions of thoughtful and well-educated people are the data of ethics,just as sense perceptions are the data of a natural science.” 49

This, however, is mysticism. It is utterly arbitrary. To the rational ques-tion of how one knows that a particular thing is valuable, the response “Itjust is” or “Some of us can tell” is a reassertion rather than an answer. Andit is every bit as subjective as subjectivism.

Ross provides an especially stark disavowal of method. Many intrinsicistsare not intuitionists, as he is, of course, and many are not nearly asforthright or as self-aware about their underlying methodology. What issignificant to all intrinsicism, however, is the fact that, because no intrin-sic value can be found or shown “in” things (as intrinsicists themselvesoften admit), assertions of intrinsicist moral truths inevitably reduce toprojections of the non-objective opinions of the speaker. These are non-objective insofar as we have no means of objectively validating assertionsof value that are put forward as, at bottom, independent of rational foun-dations. The discipline imposed by evidence and logic is precisely what

47 Rand explicitly rejects this notion, as we saw in the passage cited earlier that insistedon the conditional character of value.

48 Rasmussen, I think, in the piece cited above (in note 24) fails to appreciate that the sheerfact that a thing stands in a beneficial relationship to a person does not render it a value.Such a relationship is part of what is necessary for objective value, on Rand’s account, butnot all.

49 Ross, “What Makes Right Actions Right?” 192. Along the same lines, Moore appeals to“the sober judgment of reflective persons.” G. E. Moore, Principia Ethica (New York: Cam-bridge University Press, 1903), 94.

144 TARA SMITH

the intrinsicists set aside. Whether one posits a Form of the Good in arealm apart from sensible experience from which all other goodness andrightness derive (as in Plato), or one proclaims a less grandly titled “good-ness” that inheres in all good things (as in Ross, Moore, et al.), or oneasserts moral obligations that bear no necessary relation to the obligatedperson’s beliefs or commitments (as in Shafer-Landau and a train ofothers), the known absence of rational demonstration of the professedmoral qualities means that they are professed as independent of the stan-dards of rationality. The intrinsicist version of objectivity, in other words,shorn of its misleading labeling, collapses into subjectivism.

A recent defense of an intrinsicist form of intuitionism against chargesof subjectivism offers a telling example.50 Michael Huemer contends thatan intuitionist does, contrary to critics, have a means of confirming anintuition’s truth: by asking other people whether they share that intuitionas well as by “seeing whether it coheres with other intuitive beliefs” ofone’s own.51 What apparently escapes Huemer is that these tests arethemselves subjective. Whether one checks one’s intuitions against thebeliefs of many subjects or of one subject (oneself ), subjects remain deter-minative of moral truth. Subjectivism is again in command when Huemerresponds to alleged caricatures of intuitionism in part by employing thetest of whether disputed examples are controversial.52 Degree of contro-versy, he fails to notice, is simply an index of popularity; it reflects theviews of a large quantity of subjects. Thus, this intrinsicist attempt toavoid subjectivism also fails. To regard what most people think of anygiven case as the arbiter of objective truth is, at bottom, merely anotherform of subjectivism.53

Certainly, Huemer does not mean to be a subjectivist; he is explicitlyresisting it. Huemer’s view collapses into a subjectivist one, however, inso-far as it renders the subjective contents of individuals’ minds determina-tive of objective value. Its effect, in other words, is the equivalent ofsubjectivism. If I understand his position correctly, Huemer would notmaintain that people’s intuitions create moral reality. Rather, our intu-

50 Bear in mind that intuitionism per se is not a direct marker of intrinsicism. What issalient to certain intuitionists’ also being intrinsicists is the kind of value that they believeis intuited. Also recall that intrinsicism encompasses a wider stable of thinkers than thosewho openly assert the existence of intrinsic values as such. See note 13 above. I examine therelationship between intuitionism and belief in intrinsic value in Viable Values, 66–71, 77; seealso 20–28 on intuitionism.

51 Huemer, Ethical Intuitionism, 109.52 Ibid., 133–34; see also 21 for related discussion.53 In offering this last defense, Huemer is contesting a portrait of Samuel Clarke and using

degree of controversy as the test of whether Clarke meant a specific moral claim to beinfallible. Thus, Huemer’s more precise contention is that degree of controversy is thebarometer of a claim’s fallibility rather than directly of its truth. Nonetheless, even if it isonly allegedly infallible truths that are purported to be known through intuition (either byClarke or by Huemer), as long as the defense of the existence of such truths is made byappeal to the level of controversy that these claims engender, Huemer is employing sub-jective props for claims that he asserts to be objectively valid.


itions are our only means of access to such reality. For this reason, mycharge of subjectivism may initially appear off-target. Huemer’s accountnonetheless cedes intuitions too much power to forestall subjectivism.For on his view, the alleged objectivity of a moral truth depends entirelyon people’s beliefs, rather than on the nature of the object. Because hisaccount of objectivity includes no requirement that people’s intuitionsbe grounded in the actual nature of the object in question, the accountends up, in practical application, being no different from subjectivism.What is objectively so, in morality? What a consensus of thinkers agreesis so.

This is but one example of a wider phenomenon. Rand’s contention,in short, is that the dominant traditional image of moral objectivity,intrinsicism, is false to both the metaphysics and the epistemology ofvalue. It assigns values an ontological status they do not actually pos-sess and breeds a mental passivity (“open your eyes and see value”)that is directly contrary to the thoughtful, rational evaluation of rela-tionships that human life requires. Intrinsicism invents value wherenone exists and fails to provide the type of guidance that a personneeds in order to identify values (namely, a rational, evidence-anchoredassessment of whether a given thing will in fact further his life andthus should be pursued).

While the subjectivists are right to preserve a role for the subject in theirexplanation of moral truth, then, and the intrinsicists are right to insist onthe significance of the object, neither camp accurately captures the role ofits favored feature, and both miss the crucial interaction between the two.Rand argues for recognition of the role of both subject and object. Objec-tivity in ethics is attained neither through revelation of the intrinsic prop-erty of goodness nor through the subject’s creation of goodness, but througha rational procedure of evaluation that is strictly governed by the methodof objectivity. Rand’s theory identifies the process by which a subject canrationally identify objects as genuine values.

Rand’s conception of moral objectivity does not, accordingly, purportto reflect a “view from nowhere.” Any view originates somewhere. Thequestion that a person faces, in aspiring to moral objectivity, is not how toescape his vantage point, either literally or figuratively, but how to makehis view conform with reality. What is the nature of this thing that I amconsidering? And what sort of impact is it most likely to exert on my life?These are the principal questions that a person must address.

The basis for evaluating answers to these questions is rooted in theobject, although the evaluation, as well as the person’s embrace of his life,are also necessary for the facts in question and the benefits they stand tooffer to be objectively valuable to him. Thus, to the perennial challenge toassertions of objective truth in ethics —“Who decides?” —Rand’s responseis: it is not a matter of decision. She explains: “Nature does not decide —itmerely is; man does not decide, in issues of knowledge, he merely observes

146 TARA SMITH

that which is.” 54 On her theory, again, objective value reflects recognitionof mind-independent facts coupled with a person’s seeking his life andcorrelatively seeking the specific values that propel his life.

V. Conclusion

The long-standing debate over the objectivity of morality has beenhampered, in Rand’s view and in mine, by a skewed portrait of thealternatives. The dominant understanding of objective value is actuallyintrinsicist, reflecting the belief that certain things are good or right “in,by, and of” themselves. While the appeal of firm moral answers that arenot contingent on people’s ever-shifting opinions is understandable, whatis missing from most assertions of objective moral truth is appreciation ofthe significant role of the subject, the person to whom and for whomanything can be valuable. Moreover, because assertions of intrinsic valueultimately collapse into subjectivism, the prevalent conception of objec-tivism in ethics does not offer a genuine alternative to subjectivism.

Rand’s view, I think, helps to refine our understanding of what objec-tivity in ethics actually is. To claim objectivity for any particular assertionof value is to make a claim about the nature of the mental processes thatled to that conclusion. It is not the assertion of some sort of particles ofreality that exist as values, independently of us, but a statement of ourability to learn of the positive or negative impact that different thingshave on our existence. Claims about the value of specific things are claimsabout factual matters and, as such, are true or false. The positing of theobjective value of particular things thus tells a person not where to lookto find values, but how to use his mind in order to determine whethervarious things are values. One of the significant strengths of this theoryis that its account of morality’s objectivity is entirely of a piece with thebasic nature of objective knowledge in any field. Morality’s objectivity isno different in kind, qua objective, from medicine’s objectivity.

I suggested at the outset that the philosophical roots of popular trepi-dation about morality’s objectivity were essentially twofold: misunder-standings of the nature of moral objectivity, which I have examined here,and misguided views of the function of morality. Ultimately, the twocannot be severed. Rand’s commitment to objectivity in ethics rests on herbelief that the function of ethics is a practical one that can only be servedby adherence to an objective course. This obviously raises even largerquestions about the nature and function of morality. Examination of thesequestions is vital, however, not simply for reaching a verdict on Rand’sview, but for progress in the objectivism-subjectivism standoff. For how-

54 Rand, “Who Is the Final Authority in Ethics?” 18; emphasis in original. She also observesthat the question’s presupposition that it is a matter of decision itself reflects subjectivism(ibid., 19).


ever deficient some of people’s more superficial reasons for resistance tomoral objectivity may be, people are right to be spooked by assertions ofobjective value that are, in essence, arbitrary. And this is what intrinsicistversions of moral objectivity offer. To escape the arbitrariness in subjec-tivism as well as in intrinsicist “objectivism,” we need to recognize thatwhat answers to moral questions depend on (which is the central bone ofcontention between objectivists and subjectivists) itself depends on whatanswers to moral questions are for. It is only by also confronting this, Ithink, that real progress can be made.

Philosophy, University of Texas at Austin

148 TARA SMITH

EVALUATIVE CONCEPTS AND OBJECTIVE VALUES:RAND ON MORAL OBJECTIVITY*

By Darryl F. Wright

I. Introduction

Those familiar with Ayn Rand’s ethical writings may know that shediscusses issues in metaethics and that she defended the objectivity ofmorality during the heyday of early noncognitivism.1 But neither hermetaethics, in general, nor her views on moral objectivity, in particular,have received wide study.2 In this essay, my aim is to elucidate someaspects of her thought in these areas. That elucidation, I hope, will alsoexhibit her account of the objectivity of morality as innovative and plau-sible, though for reasons of space I will not mount a direct defense of theaccount here, but will focus on interpretive issues. There are two dimen-sions on which, I believe, Rand’s account has something new to offercurrent discussions: first, in its conception of the way in which moralvalues serve a biologically based human need; second, in its conjunctionof two ideas that have seemed hard to reconcile without appealing to aquestionable nonnaturalism about values —the idea that moral conceptsand judgments have an essentially practical or action-guiding function,and the idea that moral judgment can be objective in fundamentally thesame way as science. I emphasize both of these dimensions in whatfollows.

Rand’s account of moral objectivity depends on features of her generalepistemological views, views that are also richer and more extensive thanis usually recognized.3 Specifically, the account of moral objectivity depends

* This essay has benefited from comments I received on previous versions of some of thismaterial from Allan Gotthelf, Fred Miller, Ellen Frankel Paul, Peter Railton, Connie Rosati,Greg Salmieri, Geoff Sayre-McCord, and audiences at the University of Pittsburgh and atBowling Green State University. Much of the final work was completed with the support ofthe Social Philosophy and Policy Center at Bowling Green State University, through theCenter’s Visiting Scholars program. I am most grateful to the Center’s directors and staff forproviding a comfortable and supportive work environment during my time there in theSpring 2007 semester.

1 See Ayn Rand, “The Objectivist Ethics,” in Rand, The Virtue of Selfishness: A New Conceptof Egoism (New York: Signet, n.d.), 13–39.

2 The only book-length treatment is Tara Smith, Viable Values: A Study of Life as the Root andReward of Morality (Lanham, MD: Rowman and Littlefield, 2000). My account of Rand’smetaethics in this essay differs significantly from Smith’s (and some of the issues I will takeup lie beyond the scope of Smith’s book).

3 These can be found mainly in Ayn Rand, Introduction to Objectivist Epistemology, expandedsecond edition (New York: Meridian, 1990).


on general views of hers about the nature of objective judgment. Further,the general norms or standards of objectivity that Rand puts forwardfurnish a useful template for understanding her approach to norms inother areas, including in morality. Thus, before proceeding to her meta-ethical views, I will explore some issues in her epistemology. That meansI will aim to cover quite a bit of terrain and, as a result, will discussRand’s views in perhaps regrettably broad terms. But given that Rand’sdefense of moral objectivity is so largely unfamiliar, a wide-angle shot ofit may well be the best way to begin.

The next section introduces some terminology for thinking about ques-tions of objectivity. Sections III and IV then focus on Rand’s general viewsof the canons of objective judgment, including, in Section IV, a principalobjection that her views may elicit, and that might seem to prevent heraccount of moral objectivity from even getting off the ground. Sections Vand VI take up her account of moral objectivity. Section VII is a conclusion.

II. Conceptions of Objectivity—Some Preliminaries

In a helpful recent discussion, Brian Leiter contrasts two kinds of claimsabout objectivity. Claims about metaphysical objectivity concern whether(or in what respect or degree) “things are what they are independent ofhow we take them to be.” 4 Objectivity in this sense comes to mind-independence. If I say that the facts in some domain are what they areregardless of anyone’s beliefs about them, I am attributing metaphysicalobjectivity to those facts.5 By contrast, claims about epistemic objectivityconcern whether (or to what extent) “the cognitive processes and mech-anisms by which we form beliefs about the world [are] constituted in sucha way that they at least tend toward the production of accurate represen-tations of how things are.” 6 As this formulation makes clear, a cognitiveprocess can qualify as objective even if it errs, provided that the error isnot systemic —provided that, over time and all else equal, its continuationwould tend to be self-correcting. The candidates for objectivity in thisepistemic sense are not reality or a set of facts, but our methods for gettingat the facts. By extension, one might characterize as “epistemically objec-tive” those conclusions —the claims or bodies of theory —that emergefrom epistemically objective methods. Although Leiter is not explicit aboutthis, unlike “metaphysical objectivity,” “epistemic objectivity” is a nor-

4 Brian Leiter, “Introduction,” in Objectivity in Law and Morals, ed. Brian Leiter (Cam-bridge: Cambridge University Press, 2001), 2; emphasis in original.

5 Metaphysical objectivity is not the same as what Rand calls the “metaphysically given.”She contrasts the latter with the man-made. Thus, institutions of government are not meta-physically given, for instance, but once in place are metaphysically objective: they exist andare as they are whether or not a given person is aware of them, and whatever a given personmight believe about them. See Ayn Rand, “The Metaphysical Versus the Man-Made,” inRand, Philosophy: Who Needs It (New York: Signet, 1984), 23–34.

6 Leiter, “Introduction,” 1.

150 DARRYL F. WRIGHT

mative concept, a concept by means of which we evaluate cognitive pro-cesses and their products. Epistemic objectivity in a process or product ofcognition is a form of cognitive success, and failures of epistemic objec-tivity constitute cognitive deficiencies, at least in those areas where wetake it that objectivity might have been attained. In contrast, metaphys-ically objective facts are not on that account “successful” facts (whatcould this mean?) —they’re just facts.

If one understands issues involving objectivity in these terms, then, forinstance, the “objectivity of the natural sciences” would consist roughlyin the epistemic objectivity of the methods of the natural sciences —theaptness of those methods for accurately representing metaphysically objec-tive natural facts. The “objectivity of the social sciences” would consistroughly in the epistemic objectivity of the methods of the social sciences —the aptness of those methods for accurately representing metaphysicallyobjective social facts (which, of course, in a broader sense of “natural,”can also be considered natural, even if they are not reducible to the sortsof facts that the “natural sciences” study). In a parallel fashion, we mightthink of moral objectivity as consisting in the epistemic objectivity ofcertain methods of moral reasoning, where this is understood as theaptness of these methods (whatever they may turn out to be) for accu-rately representing metaphysically objective moral facts. Moral facts, inturn, might be seen either as a species of natural or social fact, or as a suigeneris order of nonnatural, nonsocial facts —facts that in some way tran-scend or sit apart from the facts investigated in the natural and socialsciences. But some defenders of moral objectivity would balk at thischaracterization of it. Some defenses of moral objectivity are domain spe-cific: they deny that ethics attains objectivity by sharing those featureswhich are widely held to make science objective, such as its accuratelyrepresenting a mind-independent world.7 According to domain-specificaccounts, moral objectivity need not be modeled on scientific objectivity;it might, for example, require only that ethical claims be capable of inter-subjective justification or that ethics have room for a distinction betweenbeing right and seeming right, even if objectivity in science would requiremore (or other) than the satisfaction of these criteria.8

As I will try to show, Rand views moral objectivity as continuous with,but not isomorphic to, objectivity in the natural sciences. Thus, the con-ception of moral objectivity envisioned in the last paragraph, which wasmodeled on Leiter’s conception of epistemic objectivity, is not quite broadenough to encompass Rand’s conception of moral objectivity. What it

7 On the issue of domain specificity, see ibid., 2.8 For examples of domain-specific approaches, see, e.g., John McDowell, Mind and World

(Cambridge, MA: Harvard University Press, 1996); and Christine Korsgaard, “The ReasonsWe Can Share: An Attack on the Distinction between Agent-Relative and Agent-NeutralValues,” in Ellen Frankel Paul, Fred D. Miller, Jr., and Jeffrey Paul, eds., Altruism (Cam-bridge: Cambridge University Press, 1993), 24–51.

EVALUATIVE CONCEPTS AND OBJECTIVE VALUES 151

omits is the distinctively evaluative nature of moral reasoning, which, forher, is not reasoning directed at a new set of facts —morally evaluativeones —but rather reasoning of a new kind —evaluative rather than inves-tigational and explanatory —about (some of ) the same sorts of facts thatconcern us in other departments of thought.9 At the same time, the can-ons of objectivity that apply to such reasoning, in Rand’s view, are offundamentally the same nature as those that apply to scientific and otherforms of reasoning. In this sense, Rand’s conception of moral objectivityremains non-domain-specific. As a preliminary formulation, then, we mightsay that, for her, moral objectivity comes to the epistemic objectivity ofcertain methods of moral reasoning, where this is understood as theaptness of these methods for accurately evaluating what one morally shoulddo, in view of certain metaphysically objective natural facts. This formu-lation will require modification later; in addition, we will see shortly thatLeiter’s characterization of epistemic objectivity —though, in the respectjust indicated, too narrow for Rand to accept —also turns out to be, in adifferent respect, too broad for her to accept. For the time being, however,I will leave both this preliminary conception of moral objectivity andLeiter’s conception of epistemic objectivity in place as rough orientationsto the terrain ahead. (Hereafter, references to objectivity are to epistemicobjectivity unless otherwise specified.)

I turn now to the epistemic foundations of Rand’s metaethics, and thento the metaethics itself.

III. Rand on Objectivity

Rand rejects what might be called the self-transcendence model ofepistemic objectivity. It has been common to view objectivity as demand-ing the transcending of obstacles that human nature throws in the way ofcognition. Plato expresses this view insofar as he sees sense perception asa cognitive obstacle, a form of cognitive distortion.10 More recently, thissort of view seems implicit in Thomas Nagel’s conception of objectivity asthe effort to see the world as if from no particular vantage point.11 Nageldoes not quite say, but seems to assume, that the view from where we arecould not fail to be a “subjective” one. Though he lacks Plato’s contemptfor subjectivity thus conceived, he thinks we properly aspire to somethingmore. In Rand’s view, however, the ideal of a self-transcendent grasp ofthe world amounts to an aspiration to a magical kind of cognition, inwhich reality impresses itself on our minds by no identifiable means, and,like Potter Stewart spotting pornography, we know truth when we see it

9 For this point, see Rand’s discussion of moral evaluation in “What Is Capitalism?” inAyn Rand, Capitalism: The Unknown Ideal (New York: Signet, 1967), 22; and in Rand, TheRomantic Manifesto, revised edition (New York: Signet, 1975), 18.

10 For example, in the Phaedo, 64e–65d.11 Thomas Nagel, The View from Nowhere (New York: Oxford, 1986).


but can never manage to say how we recognized it.12 But if anything lacksobjectivity, according to Rand, it is the sort of unaccountable truth claimmade by someone who purports to know something but cannot say how.There is no need to idealize such invocations of cognitive magic, shethinks, if we reject the idea that our active involvement in a cognitiveprocess must distort its results. For then we can regard objectivity not asa process of getting around the obstacles that human nature throws in theway of the needed magical deliverances, but rather as a process of exploit-ing the cognitive resources that human nature puts at our disposal, in justthose ways in which they are (biologically and psychologically) suited tobeing exploited. Broadly speaking, objectivity will be a matter of adher-ing to cognitive methods tailored to our distinctively human cognitiveneeds and capacities. In this section, I will elaborate this broad charac-terization, with attention to one central demand of objectivity (as Randconceives it) that may raise both general concerns and concerns about theprospects for moral objectivity. Let us begin, however, by exploring Rand’sviews about the source and structure of norms of objectivity, for thesenorms will serve as a useful reference point in working through her viewsabout the source and structure of moral norms.

In one respect, Rand’s conception of epistemic objectivity is narrowerthan the conception that Leiter’s formulation (discussed in Section II)suggests. Whereas Leiter’s formulation makes no distinction between theautomatic and the volitional aspects of cognition —between, for example,the way in which our sense organs function and the way in which wereason —in Rand’s use of the term only the volitional aspects of cognitioncan be assessed for epistemic objectivity. As she views it, epistemic objec-tivity is not merely normative but prescriptive: it concerns, as we mightput it, not the mechanisms and processes of cognition in general butspecifically its methodology, that is, those component processes that are atleast potentially under the subject’s conscious control. One reason for thisnarrowing is that Rand considers skeptical doubts about the reliability ofthe automatic aspects of cognition self-undermining, whereas the exis-tence of a sphere of cognitive autonomy or volition, within which thesubject can regulate his or her own cognitive activities, does not under-mine worries about reliability, but is precisely what generates such wor-ries.13 Although a person’s sensory capacities or memory can, of course,become impaired, and thereby impair cognition, specifically epistemic fail-ures, in Rand’s view, are failures within the sphere of our cognitive auton-omy. Some of these, presumably, are all but unavoidable: we overlookthings, we hit the wrong key on the calculator, and so forth. But the

12 For the point about Rand, see Introduction to Objectivist Epistemology, 79–80. For JusticePotter Stewart’s statement that he knew hard-core pornography when he saw it, eventhough he might never succeed in defining what it was, see his concurrence in Jacobellis v.Ohio, 378 U.S. 184 (1964).

13 Rand, Introduction to Objectivist Epistemology, 78–79.


central threat to the reliability of cognition in its autonomous aspects lieselsewhere, namely, in the lack of adequate cognitive methods (either gen-eral methods or methods of specialized inquiry). We need canons ofepistemic objectivity in order to remedy this deficiency.

Perception, for Rand, is an aspect of the life-functions of certain livingorganisms and is thus subject to biological norms. Our perceptual capac-ities can function well or poorly in a biological sense. But there are nocognitive norms for perception, in her view; perception fulfills its cogni-tive function just in case it functions well in a biological sense (e.g., we donot need a methodology for seeing material objects lying before us, thoughwe need one for understanding what we see).14 Conceptual thought,however, is subject to cognitive as well as biological norms. Our concep-tual capacities can be in biologically good order and yet not deliver cog-nition. Rand’s explanation for this is that, beyond the stage at which weform our earliest, perceptually-based concepts (such as those for perceiv-able entities like chairs or dogs), our conceptual capacities do not functionautomatically, nor have we any built-in grasp of how to utilize themproperly, since such a grasp would itself have to be held in conceptualform and thus result from the right use of those very capacities. There isa distinction to be drawn, then, between the right and wrong use of ourconceptual capacities that is not reducible to the distinction between theirbeing healthy or impaired (biologically excellent or defective). Norms forthe right use of our conceptual capacities fall into two broad categories, inRand’s view: norms for concept-formation and norms for conceptual judg-ment (roughly, for epistemic justification, i.e., for certifying a judgment aswarranted). Her conception of epistemology is broader than prevailingconceptions in that it encompasses the delineation of both types of norms.

But what is the standard of “right use” here? Rand says that “the rulesof cognition must be derived from the nature of existence and the nature,the identity, of [man’s] cognitive faculty.” 15 Cognitive norms, as she viewsthem, are teleological norms; their telos is cognition, knowing. We formu-late them by asking how we must use our cognitive faculties, given theirnature and the nature of reality, if we are to attain knowledge.16 Settling

14 We may also need instruments for seeing them, of course (a microscope, a telescope,corrective lenses), and (in some cases) a methodology for using these instruments. ButRand’s point is that we need no methodology for using our eyes.

15 Ibid., 82.16 The content of Rand’s norms of objectivity depends on the end or telos to be reached

and on the nature of the means at our disposal for reaching that end —the nature of a humanbeing’s cognitive capacities. But in a sense it is not just the means (the norms of objectivity)but the end they serve (knowledge) whose specification would have to depend on us. Forhow exactly are we to think of the end of acquiring “knowledge”? What counts as successhere? Rand seems to think that epistemology assumes this end, and perhaps helps specifyits satisfaction-conditions precisely. But she does not seem to view the delineation of the endas a wholly epistemological matter. Fundamentally, it would seem that this has to be amatter for ethics, which prescribes the basic ends of human action. Since, as we will seebelow, it is our needs as a certain kind of living organism that set the standards for moral


this requires settling certain issues in metaphysics, as well as settlingcertain questions about the nature of our cognitive capacities. On themetaphysical side, Rand defends particularism; she holds that, apart fromour minds, only particulars exist —that there are no abstractions or uni-versals inherent in reality. Her particularism is promiscuous: it admitsnot only particular entities but also particular property- and relation-instantiations (the roundness of this particular ball, or the nutritiousnessof this particular apple). Conceptual thought, then, does not discloseabstract features of reality but serves as “a cognitive device of man’s typeof consciousness,” a device that we need because “the range of what mancan hold in the focus of his conscious awareness at any given moment, islimited.” 17 According to Rand, “[c]onceptualization is a method of expand-ing man’s consciousness by reducing the number of its content’s units,”and thus “concepts represent condensations of knowledge, which make furtherstudy and the division of cognitive labor possible.” 18 Broadly speaking,the normative standards for concept-formation must consider which pro-cedures of conceptual classification are cognitively productive; as Randputs it, “[t]he requirements of cognition determine the objective criteria ofconceptualization.” 19 Not all classifications will serve equally well; at thesame time, there are bound to be options, between which we must simplychoose.20

Classification is therefore not a subjective process, according to Rand —there are standards by which to evaluate our classificatory preferences —but neither are conceptual classifications correct because they reflect amind-independent conceptual order, intrinsic in reality as such. Rather,correct conceptual classifications are, as she puts it, objective —correct asevaluated by a nonarbitrary standard whose grounds lie in the functionalnature of the process of conceptualization, that is to say, in what conceptsdo for us and in the cognitive need they fill. More broadly, objectivity, asRand conceives it, is a matter of adherence to cognitive norms —normsgoverning our cognitive methods —derived in the manner indicated above.She contrasts her conception of objectivity with two archetypical views ofconcepts and conceptual thought that have been influential historically(and continue to have influence). What she calls intrinsicism sees conceptual-ity as a feature of reality, rather than of our methods of cognition, and seesconceptual thought as a fundamentally passive process of receiving con-ceptual content. Intrinsicism denies that the process of receiving concep-

evaluation, on Rand’s view, it would seem to be those needs that should most fundamen-tally determine the content of the end of acquiring knowledge. The end might be seen asthat of attaining the full range of conceptual knowledge that human beings need in order tolive fully human lives (an end whose attainment would, of course, require cooperation anda division of cognitive labor).

17 Ibid., 67, 63.18 Ibid., 64, 65; emphasis in original.19 Ibid., 72.20 Ibid., 73–74.


tual content has any distinctive nature —asserting that it is receptivity andnothing more —in order to preserve its cognitive character, that is, itsstanding as a source of knowledge about a mind-independent world.What she calls subjectivism sees conceptualization as an active process buttakes the objects of conceptual thought to be constituted by our concep-tual framework. Subjectivism denies that conceptual thought yields knowl-edge of a mind-independent world, in order to give due recognition to itsactive character, its dependence on our activity and on the nature of ourcognitive equipment. In contrast, Rand’s conception of objectivity seeksto reconcile the active character of conceptual thought with its cognitivecharacter, by jettisoning the assumption that cognitive success requiresself-transcendence.

What, then, does objectivity look like in more detail, according to Rand?One of its central demands, she argues, is the systematic integration ofour thinking into a unified whole. Integration functions to ensure theconsistency of our thinking; to ensure that conclusions reached in onearea are not cast in doubt by those reached in other areas, and to modifythem when they are cast in doubt.21 But integration presupposes materialto work on, and objectivity’s other central demand concerns the sourceand initial justification of such material. This demand grows out of acertain kind of epistemic foundationalism that Rand embraces, which issummarized in the following passage from her writings:

Concepts are the products of a mental process that integrates andorganizes the evidence provided by man’s senses. . . . Man’s sensesare his only direct cognitive contact with reality and, therefore, hisonly source of information. Without sensory evidence, there can be noconcepts; without concepts, there can be no language; without lan-guage, there can be no knowledge and no science.22

Objectivity, for Rand, depends on respecting the foundationalism impliedhere, both in concept-formation and in judgment. In regard to concepts,“one must be able to retrace the specific (logical, not chronological) stepsby which they were formed, and one must be able to demonstrate theirconnection to their base in perceptual reality.” 23 Finding these connec-tions is not a simple matter of finding referents for every concept, orevidence for every judgment, in perceptual reality; Rand is explicit thatmany concepts (perhaps most) either lack perceptual referents altogetheror have referents that can be perceptually identified only by the inter-

21 Ayn Rand, For the New Intellectual (New York: Signet, n.d.), 125–26; Leonard Peikoff,Objectivism: The Philosophy of Ayn Rand (New York: Meridian, 1993), 118–28.

22 Ayn Rand, “Kant Versus Sullivan,” in Rand, Philosophy: Who Needs It, 90; emphasis inoriginal.

23 Rand, Introduction to Objectivist Epistemology, 51.


position of other concepts (she gives the example of “furniture,” whoseinstances can be perceived but are recognized as furniture by being rec-ognized as tables, chairs, and so forth).24 Thus, different concepts will, inher view, be “based” in perception in different ways. In the formation ofpsychological concepts, she argues that the “basing” involves a recogni-tion of intentionality —that, for instance, to see is to see something.25 Inother cases, she argues, the “basing” may advert to ways in which aconcept facilitates theoretical generalization and explanation, or math-ematical computation, bearing on our ability to comprehend (and act in)perceptual reality.26

Rand, then, can allow quite intricate ways of basing concept-formationand judgment on perceptual evidence. If such “basing” is central to objec-tivity, then exploring her conception of moral objectivity must involveexploring how her “basing” requirements apply to moral evaluation. (Awider treatment of her views would also consider the role of the demandfor integration; however, I set this aside for another time.) But beforegetting to that, let us now take up an important objection to the “basing”requirements themselves.

IV. Rand, McDowell, and the “Myth of the Given”

I can present the objection by briefly rehearsing some arguments ofJohn McDowell’s. The “basing” requirements, it might be objected, arejust Rand’s resurrection of the infamous “Myth of the Given.” The “Mythof the Given” —or, as McDowell calls it, the “dualism of [conceptual]scheme and Given” 27 —is the view that the epistemic justification of anempirical belief is, in the last analysis, a matter of its being warranted byexperience, where experience is seen as the raw material for concept-formation but as being, in itself, conceptually unstructured. How do thosewho view epistemic justification in this way view the concepts employedin empirical thought? McDowell suggests that they must take the viewthat the content of an empirical concept is fixed by the justificatory rela-tions that hold between bits of the Given and predications of the concept,which in turn are fixed by what bits of the Given the user of the conceptinitially forms the concept to refer to. The same conception of empiricalwarrant, then, plays into the dualism’s account of the epistemic justifica-tion of empirical beliefs and its account of the content of empirical con-cepts. In McDowell’s view, however, the dualism’s conception of empiricalwarrant rests on a confusion.

24 Ibid., chapter 3 and passim.25 Ibid., 29–30.26 Ibid., 148–49, 304–6, and chapter 7, passim.27 McDowell, Mind and World, 4. Below, I will refer to the dualism of scheme and Given

as simply “the dualism.”


Unlike coherentists, such as Donald Davidson, McDowell does notpropose to do without the idea of empirical warrant, for this idea, hethinks, is presupposed in the idea of an empirical belief. We cannot rejectit, therefore, without undermining the intelligibility of this latter idea.28

Where the dualism goes wrong, McDowell suggests, is not in invokingthe idea of empirical warrant but in supposing that there are justificatoryrelations spanning the boundary between the “conceptual sphere” andthe raw, conceptually unstructured material of experience, and thereforethat, as McDowell puts it, “the space of reasons is . . . more extensive thanthe space of concepts.” 29 In McDowell’s view, however, the apparentplausibility of the dualism’s expansive view of the space of reasons dependson a failure to see that “a bare presence cannot be a ground for any-thing[,]” 30 because “we cannot really understand the relations in virtueof which a judgment is warranted except as relations within the spaceof concepts: relations such as implication or probabilification [and soforth]. . . .” 31 A “bare presence” is all an experience could be, on thedualism’s conception of experience. But relations bearing on epistemicjustification are relations between pieces of propositional content. Thisseems clear if we think about what sort of thing justifications or reasonsare: considerations that give rational support or, as Stephen Darwall callsthem (in discussing practical reasons), dicta —statements in a language orthe propositions expressed in such statements.32 Bare presences are pre-cisely not dicta and thus are not the type of thing that could figure as aterm in a justificatory relation. According to McDowell, therefore, thedualism’s attempted expansion of the space of reasons fails.

McDowell argues that it is possible to retain the idea of empiricalwarrant while jettisoning this way of construing what such warrantamounts to. We can do this if we allow that experience itself comes“equipped with conceptual content.” 33 He elaborates:

In a particular experience in which one is not misled, what one takesin is that things are thus and so. That things are thus and so is the contentof the experience, and it can also be the content of a judgment: itbecomes the content of a judgment if the subject decides to take thejudgment at face value. So it is conceptual content. But that things arethus and so is also, if one is not misled, an aspect of the layout of theworld: it is how things are.34

28 Ibid., 14–18, esp. 18.29 Ibid., 6.30 Ibid., 19.31 Ibid., 7. One judgment “probabilifies” another if the truth of the first would make the

truth of the second more probable.32 See Stephen L. Darwall, Impartial Reason (Ithaca, NY: Cornell University Press, 1983), 31.33 McDowell, Mind and World, 25.34 Ibid., 26; emphasis in original.


McDowell’s view is not idealism, because although it claims that theworld is nothing over and above the totality of “thinkable contents,” itdoes not claim that the world is nothing over and above thinking.35 Theworld is independent of our thinking, in his view, and empirical justifi-cation demands that we square our beliefs with an independent reality,since it demands that we square our beliefs with experience. But his viewdoes share, with coherentism —and in opposition to the dualism —thepremise that epistemic justification is always a matter of how one piece ofpropositional content stands with respect to others. In that sense, asMcDowell points out, his conception of justification is Neurathian: it seesjustification as proceeding always from within our going conceptual schemeand framework of beliefs.36

The “dualism of scheme and Given,” as McDowell presents it, is basedon an internalist conception of epistemic justification, according to whichwhat makes it the case that a belief is justified is a rationale for the beliefthat could, in principle, be the believer’s reason for holding it, and that,if it were, would justify the act of believing. McDowell shares that internal-ism and, therefore, does not challenge it. What he challenges is the dual-ism’s conception of what could qualify as a rationale, specifically, theview that experiential inputs from outside the conceptual sphere could bepart of such a rationale. Now, like the dualism, an externalist conceptionof justification could make room for justificatory linkages that extendoutside the conceptual sphere. But since externalist accounts of epistemicjustification do not construe justifications as reasons for belief, such a viewwould not be vulnerable to the charge that it had illicitly extended thespace of reasons beyond the conceptual sphere. In order to reject anexternalist view that incorporated justificatory linkages extending out-side the conceptual sphere, McDowell would have to attack not the appealto extra-conceptual justificatory linkages, but the underlying externalistpicture of epistemic justification.37 I will not pursue the issue of external-ism further here, however, because I believe that Rand’s views of episte-mic justification are largely internalist, and that where they diverge from

35 Ibid., 28.36 Ibid., 52–53. McDowell’s explicit reference to Neurathian reflection comes in his dis-

cussion of practical thought (see ibid., 81), but the earlier passage I cite here seems to makea similar point in criticizing Gareth Evans’s conception of “non-conceptual content.”

Otto Neurath (1882–1945) was an Austrian philosopher of science and Vienna Circlemember who likened scientific progress to the attempt to rebuild a boat while at sea. Just asthe boat could only be rebuilt plank by plank —not all at once —since those aboard wouldneed somewhere to stand while they worked, so our beliefs about the world can only berevised in stages, since it is only by holding some part of our worldview fixed that we gaingrounds for challenging and improving other parts. “Neurathian reflection” refers to suchstaged belief revision and stands opposed to the project of finding external justificatorygrounds for our worldview as a whole.

37 Two forms of externalism need to be kept apart in this context. One is externalism aboutjustification, which is the position I discuss in the text. The other is externalism aboutknowledge, which denies that justification is a necessary condition of knowledge.


standard forms of internalism, it is not just by pulling in externalist ele-ments. Instead, I want to suggest a fairly modest defense of Rand’s epis-temology against the charge that it buys into a myth about given-ness.Other, more aggressive lines of defense may be available to Rand, but thisone is modest in that it concedes a good deal of the framework fromwhich the “myth” charge derives —but, crucially, not all of that frame-work. Besides being modest, it also strikes me as being correct.

The dualism maintains that conceptually unstructured experience con-stitutes the ultimate rational ground for epistemically justified empiricalbeliefs. If experience is to play this role, it must be capable of figuring intoa person’s reasons for believing something: it must be possible for aperson to hold a belief because he sees (extra-conceptual) experience aswarranting it. In order for that to be possible, however, it would seemnecessary for the contrast between the Given and our conceptual schemeand beliefs to be an object of some form of direct acquaintance. For,otherwise, it would seem impossible to appeal to elements of the one inorder to launch (or, looked at from the opposite direction, to terminate) ajustification of elements of the other. Suppose, then, that you acquire thebelief, “The house key is on the kitchen table,” upon going into thekitchen and looking on the table. How is an appeal to the Given toproceed? What you see before you are things that you immediately rec-ognize as falling under various concepts —the concepts “table,” “key,”and so forth. Is it that, in some sense, you have the experiential elementsand the conceptual overlay or structuring both in view at once? If so, thenthis contrast must present itself even to someone whose capacity to rec-ognize justifications is as limited and undeveloped as could be, since thejustifications at issue here are those which are supposed to supply therational ground of all the more sophisticated empirical justifications aperson might later come to recognize. It seems dubious, however, thatany such contrast presents itself to direct acquaintance, for it is not asthough the Given and the conceptual overlay sit next to one another, liketwo contrasting patches of color on a fabric. Whatever distinction there isto be made between them is a fairly advanced and abstract one.

It seems unlikely that Rand would want to claim that we are directlyacquainted with the contrast between scheme and Given. The type ofacquaintance involved here would have to be partly introspective, since(according to the dualism) a conceptual scheme is something constructedby our minds. But Rand regards introspection as a relatively difficult andhighly fallible process. Moreover, she thinks that, as a psychological mat-ter, many people remain unaware of the distinction between perceptualinput and the conceptual processing of such input; as a result, they “treat. . . the first-level abstractions, the concepts of [perceivable] physical exis-tents, as if they were percepts. . . .” 38 Yet surely even such people are able



to grasp elementary empirical justifications. It is hard to see how Randcould think that people could remain unaware of the distinction betweenscheme and Given if she thought that it were one that we could bedirectly acquainted with at a relatively early level of cognitive develop-ment. Indeed, according to Rand, we see the world through our concep-tual scheme; our conceptual scheme is part of our means of cognition andnot, except in special cases, the object of our awareness. On this view, onewould expect the scheme/Given distinction to be more elusive thanevident.

If Rand does not think we have the distinction between scheme andGiven in view all at once, is it possible that she thinks we learn to makea kind of Gestalt shift, from a conceptually structured experience of theworld to one that is all Given and no conceptual structuring? That alsoseems very unlikely. Consider what, by her lights, would have to beinvolved in our being able to make that kind of Gestalt shift. Rand thinksthat to remain on the purely perceptual level (her term for the stage of ourcognitive development that precedes our having any explicit concepts) isto be without the subject-object distinction; this is the position that shethinks animals and very young children are in.39 It is also the position inwhich she thinks adults remain in experiencing the “emotional meaning”of music —a position in which we cannot tell what is objectively attrib-utable to the music and what depends on us.40 But is it conceivable thatwe could lose (and regain) the subject-object distinction in adult percep-tual experience by an act of will? That seems flatly inconceivable, and Ican see no reason to ascribe such a position to Rand. What seems true,and what it seems more likely that she herself believes, is that the devel-opment of our conceptual capacities irrevocably transforms adult percep-tual experience, and, specifically, transforms its form. Phenomenologically,there seems to be no psychological distance at all between seeing some-thing and seeing it as a table, once one has the concept “table”; even whenwe cannot quite make out exactly what something is (something off in thedistance, say) we see it, for example, as something up there on the hillside.41

To my mind, the view that makes most sense, and that it also makesmost sense to ascribe to Rand, is that —rather than something that we aredirectly acquainted with —the distinction between the conceptually unstruc-tured Given and our conceptual scheme is a theoretical one. That is not,of course, to impugn it. I do not have any very precise criterion in mindfor what counts as a theoretical concept. But what I mean to suggest is

39 Rand, The Romantic Manifesto, 55.40 Ibid.41 A further point, which I do not have space to develop here, is that Rand views human

perceptual experience as implicitly conceptual from the very first. See, for instance, Introduc-tion to Objectivist Epistemology, 5–7. Her idea that there are implicit concepts requires analy-sis, but that she sees the need to describe perception in such terms suggests how far she isfrom the dualism of scheme and Given that McDowell attacks.


that the idea that, before we begin acquiring any concepts, we have aconceptually unstructured form of perceptual awareness, is one that wearrive at not by having or recalling perceptual experiences matching thisdescription, but, rather, in the course of building up a theoretical accountof how we think. It is something that we infer, after having built up afairly sophisticated conceptual framework for raising and addressing ques-tions about human cognition. If this is correct, then the idea of the Givenhas an epistemic status analogous to that which Rand assigns to theconcept of a sensation. Percepts, she says, are automatic, neurologicalintegrations of sensations. But “the knowledge of sensations as compo-nents of percepts is not direct, it is acquired by man much later; it is ascientific, conceptual discovery.” 42

To treat the distinction between scheme and Given as a theoretical oneis to eschew the idea that epistemic justification has inputs from outsidethe conceptual sphere. If this is how Rand views the scheme/Given dis-tinction, then she can agree with McDowell that what forms the startingpoint (or the terminus) of epistemic justifications is not something outsidethe conceptual sphere, but just some such conceptually structured bit ofperceptual content as that the house key (or maybe only a key) is on thekitchen table. McDowell says that perceptual experience is “equipped withconceptual content.” 43 Rand, perhaps, would not say just this —it de-pends on how we take the notoriously ambiguous term “content.” SinceMcDowell regards experience as passive,44 to say that it has conceptualcontent is to locate the source of its conceptuality outside of our minds, onthe side of the object rather than of the subject. If those are the onlypossibilities, then we must locate conceptuality in what we experience inorder to reject the view that there are conceptually unstructured experi-ential inputs to justification. But, for Rand, awareness always involves notonly a subject and an object but a form, which depends on both. Randdenies that conceptual structuring is inherent in what we perceive. Thisdenial, however, need not commit her to recognizing conceptually unstruc-tured justificatory inputs, for she can maintain, and I believe does main-tain, that adult perceptual experience has an ineliminably conceptual (orconceptualized) form.

But can Rand exclude conceptually unstructured inputs from her accountof epistemic justification consistently with her theory of concept-formation?Her theory holds that perceptual input is the conceptually unstructuredraw material for concept-formation or, more precisely, for the formationof our most basic concepts; other concepts, she thinks, are formed byconceptualizing already conceptualized material —by engaging in “abstrac-tion from abstractions,” as she calls it.45 If McDowell is right, however,

42 Rand, Introduction to Objectivist Epistemology, 5.43 McDowell, Mind and World, 25.44 Ibid., 10–13.45 Rand, Introduction to Objectivist Epistemology, 19.


abstractionist theories of concepts like Rand’s must explain the content ofbasic empirical concepts in terms of the justificatory relations betweenbits of the Given and predications of these concepts.46 In that case, Rand’stheory of concepts would commit her to the very conception of justifica-tion that I have suggested that she can (and should) avoid.

It is significant, I think, that in Rand’s view, early-stage concept-formation is not a deliberate process. The acquisition of language obvi-ously depends on guidance from current speakers, but the psychologicalside of the process, the side that allows language to be used with under-standing, unfolds more or less automatically, she believes.47 In the initialstages of conceptual development, we do not seek out a conceptual per-spective but fall into one. What must this seem like to the subject? If Randis correct that concept-formation is a process of classifying things accord-ing to observed similarities, but that the nature of this process is nottransparent to the subject from the outset, then it is plausible to suggestthat in the early going, we find ourselves confronted by the impression ofa wholly external sort of normativity —the impression that certain thingsjust are all of a kind and just do belong together under the same conceptor term. On the picture I am suggesting, we do not (initially) see ourselvesas having put them together but as recognizing them for what they are;matters of classification do not seem to depend on us —they seem todepend entirely on the way things are in the world.

I am suggesting, in effect, that McDowell’s description of how percep-tual experience works is a correct description of how it seems to workpre-reflectively. That is, pre-reflectively, perception seems to be “saddledwith conceptual content” right from the start, where the “start” is thepoint at which we begin learning language. But I also want to suggest thatthis account of how things pre-reflectively seem can cohere with Rand’saccount of how things actually are. According to Rand, classificatorynormativity is not wholly external to our minds, nor is conceptual contentthere for the perceptually experiencing whether or not we ever experienceit. Its being right to classify certain things together as “chairs” or “keys”or “dolphins” depends both on mind-independent facts about those thingsand on facts about our means of cognition, such as that certain ways ofclassifying (but not others) facilitate the expansion of knowledge throughgeneralization. Classifications are not right in themselves —intrinsicallyright —but right relative to the goals of cognition. Similarly, “conceptual-ity” is not there in the world apart from our experience but rather is aproperty of our developed perspective on the world. But even assumingRand is right in all this, as I think she is, nothing in what she says aboutconcepts requires us to conclude that the nature and mechanics of con-

46 McDowell, Mind and World, 19.47 Rand, Introduction to Objectivist Epistemology, 150–51; see also ibid., 144–45, and Ayn

Rand, “The Comprachicos,” in Rand, Return of the Primitive: The Anti-Industrial Revolution,ed. Peter Schwartz (New York: Meridian, 1999), 54–55.


cepts and concept-formation are transparent to the subject all along. Whatis, psychologically, a process of mental sorting can seem initially to be aprocess of coming to know a presorted world. It is, I suggest, this appar-ent presortedness and the normativity it entails that anchors our initialconceptual classifications. There need be no thought to the effect that thesebits of conceptually unstructured experience justify the predication of,and thereby help to constitute the content of, this concept.

V. Evaluative Objectivity and the Structure of Norms

The model of objectivity as self-transcendence suggests that we findout what is (objectively) morally good by detaching moral reflection fromthe context of our own aims and interests —by reflecting on our aims andinterests, impartially considered as those of one person among others,rather than through those aims and interests. Because Rand rejects thegeneral self-transcendence model of objectivity, she also rejects this modelof moral objectivity. What, then, does she put in its place? In general,moral objectivity will involve forming and deploying moral conceptsaccording to the general canons of objectivity adumbrated in Section III.In considering these canons of objectivity, I have focused on the basingrequirements, and I will continue that focus in taking up Rand’s viewsabout moral objectivity. Besides being integrated with one another andwith other knowledge, Rand holds, moral concepts and judgmentsmust conform to the basing requirements in order to attain objectivity andthus earn their place in our conceptual schemes, theories, and practicalactivities.48

In the last section, I suggested one fairly modest way of defending thebasing requirements from the charge that they themselves are based in amyth about the given-ness of conceptually unstructured perceptual data.Even if Rand is not vulnerable to this objection, though, one might won-der how the basing requirements can come into play for moral concepts.On the face of things, fitting moral concepts to the basing requirements —giving them the right sort of connection to their “base in perceptualreality” 49 —seems fairly daunting. For it is not even clear where in theperceptually given environment we could find a basis for these concepts.If we set aside our own responses, and consider only what is availableperceptually, it seems clear that moral concepts do not refer to anythingthat we can directly observe in perception, although of course they mayapply to such things. This seems to be agreed on all sides. G. E. Moore, forinstance, thought that goodness could be predicated of things that we

48 This is implied, for instance, by her view that objective moral judgment must employa standard derived from mind-independent facts of reality. See Rand, “What Is Capitalism?”22.



perceive, but insisted that it could not itself be perceived.50 His answer tothe question of how we gain moral knowledge sometimes seems to appealto a quasi-perceptual process that delivers abstract knowledge, withoutthe involvement of the sense organs.51 But Rand would consider theappeal to such a process just another invocation of the magical.

What if we bring our own responses back into the picture? In Rand’sview, it is through our own pleasure-pain and emotional responses thatwe first become aware of normative issues.52 In that respect, the role ofthese responses in the formation of normative (but not yet moral) con-cepts bears a certain resemblance to the role of perception in the forma-tion of cognitive concepts. Through these responses, the world strikes usas normatively valenced (that is, as containing things that are good and bad,things that are worthy of being sought or avoided). The philosophicquestion these responses raise is how to understand them. Physical plea-sure and pain, Rand claims, are the forms in which we experience theimmediate beneficial and harmful effects of their sources on our physicalwell-being.53 They are sources of normative information, albeit limited intheir reach. But Rand denies that they supply an adequate foundation formoral concepts, as opposed to a rudimentary generic concept of one’sgood.54 Emotional responses she views as the affective form in which weexperience our own (sometimes implicit) evaluations (thus fear, she holds,is the form in which we experience the judgment that something is to beavoided on account of its dangers). Emotions cannot be treated as evenlimited purveyors of normative information, she argues, except when wecan be sure of the evaluations that underlie them.55 The world strikes usthrough our emotions as being normatively valenced, because we makeimplicit evaluations before we realize that we are doing so, and before weare in a position to reflect critically on the concepts and standards bymeans of which we do so. Unlike our early conceptual framework forperceptual judgment, Rand regards our early evaluative framework asdemanding more cognitive work to structure appropriately and, accord-ingly, as far more error-prone.56 Whereas our early perceptual concepts

50 G. E. Moore, Principia Ethica (Cambridge: Cambridge University Press, 1903), 38–39, 41.51 Ibid., 16–17.52 Rand, “The Objectivist Ethics,” 18; Rand, Introduction to Objectivist Epistemology, 228–29.53 Rand, “The Objectivist Ethics,” 18–21.54 This is clear both from her rejection of hedonistic standards of moral judgment in “The

Objectivist Ethics” (see 32–33) and from her comments during a workshop on her ethicaltheory attended by a group of university faculty in the late 1960s, a recording and transcriptof which are part of the collections of the Ayn Rand Archives at the Ayn Rand Institute(Irvine, California). She is asked whether moral evaluation is reducible to a process ofgauging the long-range consequences of alternative choices on the basis of the amount ofphysical pleasure and pain that they would be likely to yield. Her answer is that suchcalculations could not give rise to the idea of morality. The workshop was one of a seriescovering both ethics and epistemology. The relevant material occurs on pages 11–14 of thetranscript of Workshop #4.

55 Rand, “The Objectivist Ethics,” 30–31.56 See, e.g., ibid., 31.


are a basis from which to move forward, whatever minor corrections theymay require, an implicit, emotionally articulated evaluative frameworkmay demand substantial reflection and reconstruction.

It is not apparent, then, how perception or anything analogous to per-ception in its concrete immediacy —such as pleasure/pain responses oremotional responses —could provide a foundation for moral concepts andjudgments. In bringing out Rand’s views about this, I will first argue thatshe considers the teleological structure exhibited by her norms of objec-tivity (and by other cognitive norms that she would endorse) to be char-acteristic of all norms, including moral norms. That argument will occupythe remainder of this section. In the next section, I will explore her con-ception of (teleological) moral norms in greater depth. There I will focuson how she thinks the telos of moral norms should be established, andthen on how moral norms thus conceived stand with respect to the basingrequirements.

The teleological structure of cognitive norms is not atypical, in Rand’sview, but is characteristic of all genuine norms, including moral norms.When something is evaluated, Rand maintains, it is always appropriate toask for the standard of value; and objective evaluation, she holds, isalways a matter of assessing something by an appropriately chosen stan-dard.57 In this sense, she views properly formed normative concepts asteleological concepts. For something to be good (for instance) is for it tosatisfy the relevant standard of value, which in some way specifies whatthe evaluated object should be or attain. This, Rand thinks, provides aclue to how normative discoveries, beyond those enabled by physicalpleasure and pain responses, are made. One finds out what is good byapplying the relevant standard of value to the case being judged. Randcalls this process teleological measurement.58 As long as the standard andits criteria of application are sufficiently determinate, questions aboutwhat is teleologically good will have objective answers, answers that canin principle be discovered, not magically, but by the application of aspecific method of discovery. This may seem like a pretty big “as long as,”though. How optimistic can we be about the prospects for determinate-ness? One point that Rand makes is that what teleological measurementaims for are ordinal rather than cardinal judgments —judgments of betterand worse that facilitate choice.59 That relieves some of the pressure fordeterminateness or, more precisely perhaps, gives us an understanding ofwhat determinateness in teleological measurement should mean that offersgrounds for optimism about attaining it. Further, sometimes, of course,making standards more determinate is just a matter of getting clearerabout what we want —say, from a set of student papers. But the more

57 See, e.g., Ayn Rand, “The Shanghai Gesture, Part II,” The Ayn Rand Letter 1, no. 14 (April10, 1972): 62–63; and Rand, “What Is Capitalism?” 22.

58 Rand, Introduction to Objectivist Epistemology, 32–34.59 Ibid., 33.


sway we give to what we want, the more we open up another source ofworry about objectivity —a worry not about lack of determinateness inour standards but about arbitrariness.

How would Rand address these worries? If we are interested in tele-ological evaluation, the first question is really where to get nonarbitrarystandards for it. Determinateness is no virtue in standards picked out ofthin air; it would hardly reassure one’s students to tell them that one willassign grades to papers according to a rule based on the order in whichthey end up stacked in one’s mailbox on submission. What does Rand sayabout the provenance of teleological standards? She says at one point thatthe standards for evaluating inanimate objects are “determined by theparticular purpose for which one evaluates them.” 60 On its face, this mayseem like an endorsement of a fairly extreme relativism about standards,a relativism that reinforces the worry about arbitrariness. But the relativ-istic reading of her point is hasty, for she does not say that we get to pickthe purpose for which we evaluate things, or get to pick it arbitrarily.Further, elsewhere she says that in evaluating human activities in (to usea quaint, early American term) the useful arts, such as medicine or man-ufacturing, the correct standards depend on the constitutive aims of theseactivities.61 Evaluation in the useful arts is relative to purposes, but it isnot up to the evaluator to decide arbitrarily what purposes a manufac-turing technique is to be judged by, so that one might, for example,legitimately say that a good agricultural technique was one that wastedtime. More generally, when we judge something “good of its kind” —good farming, good automotive manufacturing, a good pen, a good read-ing lamp —it is not up to the evaluator to set the standards, and evenwhere those standards require significant interpretation (raising our otherworry, about determinateness), we must hold our interpretations answer-able to something beyond themselves. We might call this sort of evalua-tion “functional,” meaning evaluation with respect to the function (or inthe case of activities like farming, the aim) that a kind of thing has andthat helps constitute its identity as that kind of thing.

It seems to me that the statement above that raised concern aboutrelativism can just as easily be read to embrace functional evaluation forinanimate objects. Rand does apply this form of evaluation to one categoryof inanimate objects, namely, works of art. The purposes by reference towhich art is to be evaluated as art, she argues, are not those that we mighthappen to bring to the judgment of an artwork but, rather, depend on thepsychological function that the creation and experience of art has in ourlives.62 Further, as we have seen, she applies the same functional, teleo-logical form of evaluation to another category of human artifacts that are

60 Ibid., 51.61 Ayn Rand, “Who Is the Final Authority in Ethics?” in Leonard Peikoff, ed., The Voice of

Reason: Essays in Objectivist Thought (New York: Meridian, 1990), 18.62 Rand, The Romantic Manifesto, 42; see also chapters 1–3, passim.


of central concern within her thought, namely, to concepts themselves.Concepts, in Rand’s view, are functional devices whose purpose is toexpand the range of human cognition (and action) beyond the perceptu-ally given environment —to enable us to think about and act with a viewtoward that which is not perceptually present (including importantly, inthe case of action, desired future states of affairs). We (epistemically)justify a given concept, in her view, by showing that it is needed for theseends. Thus, for Rand, the epistemic assessment of concepts involves aprocess of “teleological measurement” broadly similar to that employedin the other forms of assessment mentioned above. Indeed, it is suchfunctional teleological assessment that underlies Rand’s basing require-ments for concepts. Concepts, in her view, need to be connected in certainways to perceptual data because their cognitive function is to enable us toorganize and make inferences from perceptual data. (I am admittedlycondensing and oversimplifying a good deal here.)

I think that Rand considers all objective evaluation to be both teleolog-ical and (more specifically) functional; that, at least, is how she treatsevery case of evaluation that I am aware of her mentioning. It is ofparticular interest to note, here, that this is the approach she takes tomoral evaluation. It is in a discussion of the objectivity of moral evalua-tion that Rand refers to teleological evaluation in the useful arts. Sheanalogizes moral evaluation to teleological evaluation in the useful arts,suggesting that it, too, is a form of teleological evaluation.63 The questionof what is the morally right way to live a human life is no less amenableto objective resolution, she maintains, than the question of what is theright way to cure an illness or build an automobile.64 In some sense,despite clear differences, the questions are analogous. Why is this? In ageneral way, the actions of any living thing can be evaluated teleologi-cally by reference to the requirements of its particular form of life, sinceany life-process must be carried on in a certain way, appropriate to thenature of the particular living thing in question, in order to be sustained.Now Rand’s analogy to medicine and manufacturing refers specifically tothe moral evaluation of human actions, not to the generic biological eval-uation of the actions of living things. But Rand’s aim is to assimilatemoral evaluation to biological evaluation, by arguing that moral valuesare needed in the life of a human being. Thus, elsewhere, she treats moralvalues teleologically. The proper standard for assessing moral values, sheargues, is to be drawn from “the function of moral values in humanlife.” 65 If moral values have a biologically related function, then in prin-ciple they are amenable to objective, functional assessment analogous tothe functional assessment of medical or manufacturing techniques and tothe functional assessment of artifacts.

63 Rand, “Final Authority,” 18.64 Ibid., 18–19.65 Rand, The Virtue of Selfishness, x.


But if this is Rand’s aim, then her analogizing moral evaluation tofunctional, teleological evaluation in the useful arts may seem peculiar. Ineffect, that analogy treats moral values as analogues of medical or man-ufacturing techniques. One can see some of the point of this analogyeasily enough: elsewhere, Rand describes ethics as a science of method,and in another place she analogizes it to engineering.66 The moral valuespromulgated by an ethical theory function as methods for attaining acertain end, in her view, and in this respect they are analogous to medicaland manufacturing techniques. But the biological basis that Rand seeksfor them —their grounding in the needs of human beings as a certain kindof living thing —may seem to render the functionality of moral values (ifRand is correct in ascribing a kind of biological functionality to them)quite different from the technical functionality of medical and manufac-turing techniques, but similar to the functionality of the parts and systemsof living things. The analogy to medicine and manufacturing highlightsthe artifactual nature of moral values and virtues —their status as humancontrivances. That status may seem to stand in tension with the biologicalstatus that Rand also wants to claim for them. Moral values, on Rand’sconception of them, seem to be both artifacts of human culture —ways ofconducting ourselves that we devise —and subject to biologically basednorms of assessment. It may seem problematic to accord moral valuesboth statuses, since the teleology of human artifacts seems to be driven bythe purposes for which we create them, whereas biological teleology isindependently given.

But surely this dichotomy is too stark. In the first place, it may not betrue that the functions of artifacts are always as closely tied to the inten-tions of their designers as the foregoing comments imply. Although atight connection is often assumed,67 there are arguments in the literatureon artifact functions suggesting that artifact “functions” (in at least onesense of this multifaceted term) may depend not only on the intentionalcauses of the design and production of artifacts but on the intentionalcauses of their continued reproduction, and that the bearing of the lattercan sometimes be to pull function and designer’s intent apart from oneanother.68 Leaving that aside, the creation of artifacts (and thus the func-tions intended by their creators) is clearly driven in significant part byrecognized biological needs. Sometimes these needs are clearly graspedand artifact functionality is precisely targeted; other times, perhaps, therelevant needs are grasped only inchoately and we grope around. That,I believe, is the situation Rand finds in regard to moral values: that an

66 Rand, Introduction to Objectivist Epistemology, 36; Rand, The Romantic Manifesto, 169.67 See, for example, Paul Sheldon Davies, Norms of Nature (Cambridge, MA: MIT Press,

2001), chapter 1.68 For an account of artifact functions according to which this is the case, see Beth Preston,

“Why Is a Wing Like a Spoon? A Pluralist Theory of Function,” Journal of Philosophy 95(1998): 215–54.


inchoate awareness of our need for them has driven the creation of moralcodes and theories, but that we have groped around without a clear viewof this need’s nature and source.69 In any event, though, there is no incon-sistency in the supposition that although we have a biologically based needto shape our moral characters in certain ways, we must identify and respondto this need by our own wits and choices, through the delineation and adop-tion of a code of moral values.70 The question we will need to raise is nothow moral values can be both artifactual and biologically based, but ratherwhat makes the biologically based need for moral values —if such a needcan be established —decisive for their evaluation. What makes a standardof moral value based on such a criterion objectively correct, by Rand’s lights?I take this up toward the end of the next section. First, though, I turn to asketch of Rand’s argument that moral values do have a biologically basedfunction —that we need them in order to live.

VI. Moral Objectivity and the “Function of Moral Values”

In a discussion of the role of art in moral development, Rand gives thefollowing general characterization of the normative function of morality:

[M]orality is a normative science —i.e., a science that projects a value-goal to be achieved by a series of steps, of choices —and it cannot bepracticed without a clear vision of the goal, without a concretizedimage of the ideal to be reached. If man is to gain and keep a moralstature, he needs an image of the ideal, from the first thinking day ofhis life to the last.71

What the realization of moral values “gains and keeps” for a person, inRand’s view, is what she calls a moral stature. One’s stature, of course, isone’s standing, usually one’s height (physical stature) or one’s standing ina profession (professional stature) or in some other endeavor. Some dic-tionaries link these nonphysical senses of the term with reputation gainedby virtue of achievement or growth within a field, or simply with achieve-ment or growth itself. It is clear from her fiction and nonfiction writingsthat moral stature, in Rand’s sense, is a function of a person’s own choices

69 Rand does not think moral values are unique in this regard. She thinks the same is trueof both art and (perhaps to a lesser extent) government, and her arguments concerning theproper standards for evaluating art works and institutions of government run parallel to herargument concerning the proper standard of moral value. See Rand, The Romantic Manifesto,42; and Ayn Rand, “The Nature of Government,” in Rand, The Virtue of Selfishness, 125–34.

70 Indeed, it is precisely this dual status that, for Rand, gives rise to issues of objectivity inregard to moral values. It is the combination of dependence on choice and choice-independentgrounds of assessment that generates both a need and a source for a standard of moralvalue, and the corresponding need and possibility of making epistemically objectiveevaluations.

71 Ayn Rand, “Art and Moral Treason,” in Rand, The Romantic Manifesto, 147.


and does not depend on reputation, but it does seem to depend on achieve-ment and growth. To develop a moral stature is to become a certain kindof person, a person distinguished by the sort of character he or she hasformed. And although Rand does not ascribe to art a primarily didacticfunction, she does see art as an important source for “learn[ing] theconcept of moral values and of a moral character in whose image [one]will shape [one’s] soul.” 72

How, though, can moral character have a biological function for us? Inoutlining Rand’s argument here, I will focus on three of the charactervirtues that she emphasizes in her discussions of substantive moral theory:rationality, productiveness, and pride.73 Each of these deserves muchfuller treatment than I can give it here, but what I will say should sufficeto indicate the central idea behind Rand’s claim that moral values have abiological function. Roughly, Rand understands rationality as a commit-ment to using and developing one’s intellectual capacities to their fullest,facing facts and refraining from magical or wishful thinking, and notacting on unconsidered desires or emotions. She understands productive-ness as involving “the consciously chosen pursuit of a productive career,in any line of rational endeavor, great or modest, on any level of ability.It is not the degree of a man’s ability nor the scale of his work that isethically relevant here, but the fullest and most purposeful use of hismind.” 74 Pride, for Rand, centrally involves “moral ambitiousness,” pri-

72 Ibid., 146.73 Although, in some places, Rand distinguishes between moral values and moral virtues,

there is a wider sense of the term “moral values” in her writings that includes the moralvirtues. In this essay, I employ the wider sense, since the distinction between virtues and (inthe narrower sense) values has no bearing on the arguments I make. In the narrow sense,moral values are the ends attained by moral virtue. Thus, for example, Rand designates“rationality” as a moral virtue and “reason” as the corresponding moral value, since onevalues reason by developing and practicing the virtue of rationality. A person of good moralcharacter values reason (values his mind) and nourishes and protects that value throughrationality. However, in a larger sense, rationality is itself a moral value, for Rand, since indeliberately practicing rationality one makes rationality one’s end. The wide sense of “moralvalues” is in play, for instance, in Rand’s discussions of the role of art in moral developmentin her book The Romantic Manifesto, and in her discussion of moral objectivity in her essay“What Is Capitalism?” See Rand, The Romantic Manifesto, 146; and Rand, “What Is Capital-ism?” 21–22.

74 Rand, “The Objectivist Ethics,” 29. The qualifiers “productive” and “rational” are clearlyintended to constrain what forms of activity can satisfy the requirements of this virtue, sothat, for instance, earning a living as a drug dealer would presumably not count. AlthoughI cannot argue the point here, it seems to me that Rand’s criteria for the rationality andproductiveness of a “line of endeavor” would primarily involve consideration of whetherone’s activity contributes to the fulfillment of a genuine human need (including needs forpleasure and recreation, and “spiritual” or psychological needs such as those to which thefine arts respond) —whether a need of one’s own or, through the division of labor, of others.Much further work is needed here to unpack Rand’s conception of productiveness, but thisat least indicates the direction in which I think Rand would argue in order to excludeactivities such as drug dealing or organized crime from the scope of the virtue of produc-tiveness. For further discussion, see Rand, For the New Intellectual, 130; and Tara Smith, AynRand’s Normative Ethics: The Virtuous Egoist (Cambridge: Cambridge University Press, 2006),chapter 8.


marily in the form of a commitment to the consistent, uncompromiseddevelopment of rationality and productiveness; it is what might be calleda second-order virtue that one maintains by always according first-ordervirtues such as rationality and productiveness primacy over other con-siderations in deliberating about what to do.75

Rand outlines her view of the role of moral values in human life asfollows:

Just as man’s physical survival depends on his own effort, so does hispsychological survival. Man faces two corollary, interdependent fieldsof action in which a constant exercise of choice and a constant cre-ative process are demanded of him: in the world around him and inhis own soul (by “soul,” I mean his consciousness). Just as he has toproduce the material values he needs to sustain his life, so he hasto acquire the values of character that enable him to sustain it, andthat make his life worth living.76

According to this passage, moral values (“values of character”) have twobroad functions. They enable one to sustain one’s life, and they makeone’s life worth living. Developing and maintaining moral virtue involves“a constant exercise of choice and a constant creative process” in a per-son’s own soul. The passage claims, first, that this inner process is anenabling condition of the “constant exercise of choice and [the] constantcreative process” that is required of a person “in the world around him”to “produce the material values he needs to sustain his life.” That is,rationality and productiveness, in Rand’s view, are needed in order toproduce the material values that our survival depends on, and they qual-ify on that account as human virtues.77 Her discussion of these virtues

75 A number of Rand’s moral virtues have this second-order status, not only pride but alsointegrity and justice, both of which “operate” on and thus presuppose other, first-ordervirtues. Integrity involves the refusal to sacrifice one’s convictions, including one’s moralconvictions, to the wishes or opinions of others; justice involves (among other components)judging other people by objective moral standards. See Rand’s discussion of these virtues in“The Objectivist Ethics,” 27–30; and For the New Intellectual, 128–31. See also Peikoff, Objec-tivism, chapter 8.

76 Ayn Rand, “The Goal of My Writing,” in Rand, The Romantic Manifesto, 169.77 This claim raises some obvious questions about free-riding: for instance, about whether,

if rationality and productiveness qualify as human virtues on this account, a person has areason to be virtuous rather than to free-ride on the virtue of others. Although these issuesare largely outside the scope of the present essay, the discussion in the text of Rand’sconception of psychological survival is relevant to how she would address them. Reasons ofpsychological survival militate against free-riding and, in her view, undermine the appear-ance that free-riding could be beneficial to a person. I intend to address these matters in anarticle in preparation on John McDowell’s criticisms of what he calls “neo-Humean natu-ralism” in ethics. For McDowell’s argument about how naturalistic virtue ethics should andshould not deal with free-rider problems, see his “Two Sorts of Naturalism,” in RosalindHursthouse, Gavin Lawrence, and Warren Quinn, eds., Virtues and Reasons: Philippa Foot andMoral Theory, Essays in Honour of Philippa Foot (Oxford: Clarendon Press; New York: OxfordUniversity Press, 1995), 149–79.


makes it clear that what is at issue is long-range survival, “throughout thewhole of [a human] lifespan.” 78 But Rand also sees moral values ashaving a psychological function, which is more complex and, in a sense,more important than their function of enabling the production of materialvalues.

What does Rand mean by “psychological survival”? For any livingorganism with some form of consciousness, its consciousness will be thesource of its actions. The actions of a conscious being will be governed bya system of motivation and reward that depends on its psychologicalcapacities. The rewards will be that for the sake of which it acts —the endor goal of action —as well as the incentive for further action toward moreof the same kinds of rewards. In nonhuman animals, of course, the pur-suit of such reward-incentives, although purposive, is instinctual ratherthan intentional. In humans, it is intentional and runs through our con-ceptual capacities —our capacities for thought. In a generic (not strictlymoral) and descriptive (not normative or evaluative) sense, the reward-incentives sought in purposive action are values sought by human beingsand other species that engage in purposive action. From a psychologicalperspective, the pursuit and attainment of such reward-incentives —ofvalues —is what the “life” of any living organism possessing conscious-ness consists in. Psychologically, for such organisms, the process of pur-suing values just is the process of living. Rand notes that there is nothingbiologically discontinuous about the purposive action of conscious organ-isms, in contrast with the nonconscious, merely vegetative action of plants,for instance. For in an even broader sense of the term “value,” the life ofa plant is also made up of value-pursuit. Although plants do not actpurposively, they act goal-directly in a more general sense. Thus, we saythat plants turn their leaves in order to expose them to sunlight. In Rand’sview, the nutrients that plants draw from the soil and the sunlight towardwhich they turn their leaves qualify as plant-values, even though they arenot objects of conscious or purposive pursuit.79 Like the reward-incentivesdriving the actions of conscious species, they are the ends of the plant’saction and (a central element in) the cause of its further action towardmore of the same kinds of ends. Thus, they are values in the broadestsense of “that which one acts to gain and/or keep.” 80

Fundamentally, the survival of a plant consists in the maintenance,primarily through its own activities, of its capacity for (nonconscious,nonpurposive) value-pursuit. Something analogous is true for other kindsof living things: the survival of a human being or an animal consists,fundamentally, in the maintenance, primarily through its own activities,of its capacity for conscious, purposive value-pursuit —its capacity to

78 Rand, “The Objectivist Ethics,” 26.79 Ibid., 16–17 and 17n.80 Ibid., 16.


pursue reward-incentives. But Rand sees a central difference betweenthe kind of self-initiated action required to maintain an animal’s capacityto value and the kind of self-initiated action required to maintain ahuman being’s. Self-maintenance, for an animal, consists in physical self-maintenance. Although animals can experience externally inflicted psy-chological damage (e.g., through psychologically taxing laboratoryexperiments), absent such damage the maintenance of an animal’s capac-ity to value follows on the maintenance of its physical capacities; there isno such thing as psychological self-maintenance in the case of nonhumananimals; an animal’s psychological capacities remain intact provided itobtains the nutrition, shelter, and so forth that it needs physically (and,again, absent external sources of psychological damage). But, for Rand, ahuman being’s self-maintenance of his capacity to value is primarily amatter not of physical but of psychological self-maintenance. Even absentexternal psychological dangers (such as psychological abuse by otherpeople), a human being, in Rand’s view, must take certain kinds of psy-chological action —that is, must make certain kinds of choices —in order tomaintain his or her capacity to value.

What does human psychological self-maintenance —the maintenanceof the capacity to value —involve, according to Rand? Return to the pointthat our human incentive-reward system runs through our conceptualcapacities. Every kind of conscious organism has a range of psychologicalrewards that its nature “fits” it to receive and enjoy. There is overlap, ofcourse; a human being and a dog can both be incentivized by the smell ofa juicy steak and rewarded by its taste. But Rand maintains that humanbeings have a distinctive need for rewards whose attainment and enjoy-ment calls upon the volitional exercise of our conceptual capacities, and,further, that we are psychologically debilitated by the absence of suchrewards; without them, we experience the process of carrying on our livesas unhappy, as a burden, as a source of anxiety, as motivationally drain-ing.81 Let us call the rewards at issue here conceptual pleasures. In morefamiliar terms, the need for conceptual pleasure comes to the need tohave one’s life mean something or, as Rand puts it in the foregoingpassage, the need for it to be worth living. I take it that what Rand meansby “making one’s life worth living” is not primarily making it a life that,considered third-personally, is worthy of being lived, but rather makingit a life that the person living it experiences as a life worth living. For, aswe will see below, Rand contends that the only normative standards thatapply to living one’s life are standards derived from the nature andrequirements of the human life-process itself. To find one’s life worthliving, then, must be to experience the process of living —the activitiesthat define and give substance to one’s life —as intrinsically motivating,as a source of pleasure and fulfillment.

81 Ayn Rand, “Our Cultural Value Deprivation,” in Peikoff, ed., The Voice of Reason, 103–4.


How, then, does Rand view the sources of conceptual pleasure? I do nothave the space here to cover her views in detail, for on her account thereare several important sources of such pleasure that need to be considered,including art, human relationships, sex, and certain forms of recreation, allof which importantly depend on our intellectual capabilities and are inac-cessible (tout court or at least in the forms in which we experience them)to creatures lacking these capabilities.82 But since my aim in this section isto sketch her account of the function of moral values, let me comment onthe sources of conceptual pleasure that she sees as being most directlyrelated to morality. Rand holds that the two primary sources of conceptualpleasure for a human being are precisely the “thinking and productivework” that make up the core of moral virtue.83 These are our fundamentalsources of meaning in life and of a person’s sense of identity; further,because of the ways in which the other major sources of conceptual plea-sure noted above depend on one’s sense of identity, Rand sees the acces-sibility of these other sources of conceptual pleasure as depending on one’shaving fitted oneself for the primary sources. Fitting oneself for these pri-mary sources of conceptual pleasure involves developing and sustainingthe self-chosen commitment to the virtues of rationality and productive-ness described above. It is by making oneself into a rational, productiveperson that one accesses the forms of conceptual pleasure on which themaintenance of a human being’s capacity to value depends. The lack of thesevirtues leads to what Rand calls “value-deprivation”:

The choice to think or not is volitional. If an individual’s choice ispredominantly negative, the result is his self-arrested mental devel-opment, a self-made cognitive malnutrition, a stagnant, eroded, impov-erished, anxiety-ridden inner life.84

What explains the connection Rand makes between rationality, produc-tiveness, and other moral virtues, and conceptual pleasure? The moralvirtues, for her, are self-chosen postures of mind —ways of functioningmentally in one’s intellectual and practical lives. What moral virtue servesto do is to bring oneself fully into being as a human organism —to com-plete the human life-process which, unlike that of other species, does not

82 The inclusion of art here tells us something important about the passage most recentlyquoted in the text. Rand describes the two fields of human action that she mentions —innerand outer —as “interdependent.” The field of outer action centrally involves the productionof material values, and it might be supposed that this field of action concerns itself exclu-sively with the provision of our physical survival. But works of art are material values thatanswer to psychological rather than physical needs. This, it seems to me, is a good exampleof the sort of interdependence that Rand sees between the inner and outer fields of action.Although I mention one other sort of interdependence in the text below, I defer fullerdiscussion of these interconnections to another time.

83 Rand, “The Objectivist Ethics,” 25.84 Rand, “Our Cultural Value Deprivation,” 102.


unfold automatically in our case. There is a real sense in which a personwho does not utilize his mind and his capacities for productive actionfully or consistently is not wholly engaged in the process of living abiologically human life. The psychological depletion that Rand claims sucha person would feel is the psychological correlate of the biological incom-pleteness of such a life. By contrast, the psychological fulfillment that sheclaims comes with the full use of our capacities (at least absent externalmisfortune) is the correlate of the biological completeness of the life-process. Of course, it is primarily for the psychological rewards of livingthat we do want to live; merely soldiering on as a physical organism hasno independent value for us. Thus, the bearing of moral virtue on ourlong-range physical well-being can assume importance for us only oncewe have some reason to care about our long-range physical well-being.On Rand’s view, it is the direct psychological rewards of moral virtue —the forms of meaning or conceptual pleasure that we access through theexercise of our rational and productive capacities —that lead us to careand thus that make a person’s life “worth living.”

I have only sketched a view that, of course, needs considerable elabo-ration and argument. My intent has been simply to outline the kind ofbiological functionality that Rand ascribes to moral values. According toRand, rationality and productiveness enable the production of the mate-rial values that human beings need in order to survive; but they alsoenable the various forms of conceptual pleasure that our psychologicalsurvival depends on. Further, the requirements of our psychological sur-vival are as much an aspect of our biological nature as a certain kind ofliving thing as the requirements of our physical survival are. In bothrespects, then, these aspects of human character have a biologically basedfunction that provides a reference point for evaluating them as moralvirtues (and for evaluating their opposites as vices). With this sketch inplace, let us turn to the connection of this framework for moral evaluationwith Rand’s canons of epistemic objectivity.

In asking about the function of moral values, Rand assumes at least aloose shared understanding (among us as inquirers) of what moral valuesare, but not necessarily any shared agreement as to what the correct moralcode is (if any). There needs to be some kind of shared understanding ofwhat it is whose function is in question, or else we cannot even pose thequestion intelligibly. But if this shared understanding is to allow for vari-ations among people’s substantive moral views, pending the derivationof an objective standard for making moral evaluations, then our looseshared understanding must itself be a loose functional (rather than sub-stantive) one. Rand’s point of departure is the thought that moral valuesfunction to shape a person’s character and thereby to shape a person’slife. The question is then whether something with this narrow functionhas any wider function from which a standard of assessment might bedrawn and, if so, what this wider function is. (Any code that we can


adopt will function to shape our characters, but the question is whetherwe ought to shape them in a particular way, whether shaping our charactersalso has a function.) Rand’s claim is that if moral values have no biolog-ical function —if there is no way in which we need them in virtue of ourhuman nature —then there will be no objective standards available forassessing them. This is why she argues, in her essay “The ObjectivistEthics,” that the derivation of an objective code of morality depends on ananswer to the question of whether and why human beings need moralvalues.85 It is the biological function of moral values, Rand claims, thatproperly sets the standard for what is to be counted as an objectivelycorrect moral value to hold. Rationality and productiveness, for instance,are to be evaluated as objectively correct moral values —as genuinevirtues —on the grounds of their contribution to our physical and psy-chological survival. Correspondingly, irrationality and slothfulness are tobe evaluated as vices by the same standard.

But what anchors moral values to this particular teleology, to this par-ticular standard of assessment? In effect, Rand is proposing a reconstruc-tion86 of the concept of a “moral value,” on which a moral value is

(1) a value in the generic sense of “that which one acts to gainand/or keep”;

(2) functional in the narrow, (relatively) uncontroversial sense of serv-ing to shape the character of the person who espouses it; and

(3) functional in the wider sense of playing a certain kind of role infurthering the physical and psychological survival of the personwho espouses it.

On this reconstruction of the concept, something that was either notcharacterized by (1) or not characterized by (2) could not qualify as amoral value at all. (To that extent, the proposed reconstruction substan-tially accords with ordinary usage and therefore cannot be accused ofchanging the subject.) Something that was characterized by (1) and (2) butnot by (3) would qualify as a moral value but would be evaluated nega-tively, as a deficient or improper moral value (as morally bad). Somethingthat was characterized by (1), (2), and (3) would be evaluated positively,as a proper or genuine moral value (as morally good). What we need toknow is how Rand would justify this reconstruction of the concept of a“moral value,” in terms of her canons of epistemic objectivity for conceptsand, particularly, in relation to the basing requirements.

85 Rand, “The Objectivist Ethics,” 13–14.86 I draw the idea of viewing Rand’s project as one of conceptual reconstruction from

Harry Binswanger, “Life-Based Teleology and the Foundations of Ethics,” The Monist 75(1992): 84–103.


We have said that, for Rand, moral concepts do not have the kind ofgrounding in perception that our concepts for observables like tables andtrees have, or even the extensively (and theoretically) mediated ground-ing that concepts in the natural sciences have. On her view, moral con-cepts need to be linked to perception in a different way from these others,corresponding to the difference in the nature of moral thinking noted inSection II above. Moral thinking serves to evaluate a certain class ofnatural facts —namely, certain human choices —rather than to disclose aset of moral facts. If Rand is correct, moral thinking, when objective,evaluates those choices in relation to the needs of the agent’s physical andpsychological survival. Now this way of viewing moral evaluation can bejustified by reference to Rand’s general view of evaluation as teleologicalmeasurement. That would make the connection of her standard of moralvalue, and of the corresponding conception of “moral values,” to thebasing requirements dependent on the connection of her general concep-tion of evaluation (as teleological measurement) to those requirements.We can evaluate other things besides moral choices, and to have the ideaof evaluating moral choices, we need to have the general idea of “eval-uating” things, of gauging or measuring their value. Here, of course, wemean not gauging whether or not they are things we do in fact “act togain and/or keep,” but rather gauging what to act to gain and keep (andwhat not to). What, then, does Rand think is the “base in perceptualreality” of the idea of gauging what to act to gain and keep?

It may be helpful to bring out her view on this issue through a contrastwith a line of argument that Christine Korsgaard sets out in taking up avery similar issue. Korsgaard argues that it is not the case that we havenormative concepts because we have “spotted some normative entities, asit were wafting by”;87 rather, we have normative concepts because wehave normative problems: we have choices to make and action to decideupon, and normative concepts are primarily tools for choosing and act-ing, rather than tools for investigating and theorizing about the world.Rand, I believe, would agree with Korsgaard about this. Rand maintainsthat a being that could only observe the world but could not respond toit in any way could form no normative concepts; it would not notice anynormative entities wafting by (for there are none to be noticed), norwould it have any normative problems due to its utter lack of involve-ment in the world it observed.88

Korsgaard argues that we have normative concepts in virtue of ourcapacity to reflect on our actions.89 It is our capacity to reflect that gen-

87 Christine Korsgaard, The Sources of Normativity (Cambridge: Cambridge UniversityPress, 1996), 44.

88 This is part of the point of a thought experiment she proposes involving an “immortal,indestructible robot” incapable of being affected or changed. The robot, she argues, couldhave no values and no sense of good or evil. See Rand, “The Objectivist Ethics,” 16.

89 Korsgaard, The Sources of Normativity, 92–93.


erates the normative problems that we fashion normative concepts inorder to help us sort through and resolve. Thus, I can step back from whatI am about to do —say, stop working for the day —and reflect on whetheror not that is a good thing for me to do, whether or not it might be betterto go on working a while longer. But this account of the source of ournormative concepts is open to the objection that it treats reflection assomething sui generis and unaccountable (and, indeed, Korsgaard arguesprecisely that since it is reflection that generates normative problems,reflection must supply its own internal standards for resolving theseproblems).90 Rand, I think, would agree with Korsgaard that we havenormative concepts because we have normative problems, but she wouldlocate the source of these problems differently. Rand locates the source ofnormative problems not in reflection itself, but in that which prompts,and occasions a need for, reflection —for we do not reflect spontaneouslybut in recognition of some need (or apparent need) to do so. If I hesitatebefore stopping work, and reflect on whether I should, it is presumablymy sense that something of consequence hangs on whether or not I stopthat prompts me to reflect on whether it might be better to work a bitmore instead.

Rand’s view is that the idea of gauging what to gain and keep —theidea of practical deliberation —is grounded in facts about our nature asliving things. A living organism faces an alternative between life anddeath the outcome of which depends, in substantial part, on its ownactions. The maintenance of a living organism’s life depends on what itacts to gain and keep, what values it pursues. The concepts and judg-ments through which these biological facts are expressed themselves sat-isfy the basing requirements if the biological sciences do so. But thisbiological context furnishes a standard for gauging what is good for aliving thing; the good for a living thing will be that which is required byits nature for sustaining its life. Nonhuman organisms, of course, have nocapacity or need to deliberate about their actions; it is only from ourvantage point in studying and tending to them that questions about theirgood get raised. According to Rand, however, it is this same generalbiological context that supplies the basis, in our own case, for the idea ofdeliberating about what to act for. The “base in perceptual reality” for theidea of deliberation, of gauging what to gain and keep, is the alternativesthat a human being faces between physical and psychological survival ordestruction, which are grasped in the first instance as alternatives betweenboth physical and conceptual pleasure and pain. This means that there isa “base in perceptual reality” for (but only for) a conception of evaluationas teleological measurement. Outside the context of human survivalneeds, Rand maintains, there is no basis for a concept of human practicalgood —good to be sought through action. That is, there is no concept of

90 Ibid., 93.


the good that can satisfy the basing requirements. Thus, the question ofwhat is morally good must also, in Rand’s view, be resolved according towhat moral values a human being’s physical and psychological survivaldepends on.

VII. Conclusion

This broad-brush account of Rand’s metaethics has been aimed at elu-cidating how, in her view, moral values are connected to our survival, andhow, for her, this connection secures the epistemic objectivity of moralevaluation. Certainly there is much more to be said on both of thesefronts, even to elucidate her views fully, let alone to defend them ade-quately. But I want to close by indicating the sense in which, for Rand,morality is essentially practical or action-guiding.

Rand takes the needs of one’s psychological and physical survival to bethe fundamental source of all one’s reasons for action. Life or death is thefundamental alternative each of us faces, and it is only in the face of thatalternative that there is any need or context for practical deliberation andthe evaluation of options concerning what to seek. The relativization ofpractical reasons to life-based needs means that for Rand there will be nopractical reasons directing us to live. The use of our distinctive capacitiescan be a source of meaning and pleasure that make life a value to us; butthat is just to say that we come to value our lives by living them, since touse our capacities is precisely to throw ourselves into the process of livinga human life. There is no standpoint outside all the activities that make upour lives from which we can take up the question of whether or not tolive, according to Rand. Living is in this sense an end in itself, on herview.

Rand says that the source of practical reasons lies in the choice to live.There is such a thing as letting the needs of one’s survival, particularly ofone’s psychological survival, go —such a thing as not taking pains toattend to them. Practical deliberation, in Rand’s view, responds to thegeneralized commitment to bring oneself fully into being as a humanorganism —the generalized commitment to set in motion the pattern ofchoices that is required to sustain one psychologically and physically.91

That generalized commitment supplies a motive for filling in the details,that is, for finding out and doing what is required. Thus, practical rea-

91 If my interpretation of Rand is correct, this generalized commitment is equivalent towhat Rand calls the “choice to live.” But it bears an interesting resemblance to the virtue ofpride described in Section VI of the text. Pride, as Rand understands it, might be viewed asa more fine-grained, articulated form of that same comprehensive commitment to one’s life.The choice to live, on my interpretation, would amount to the commitment to develop andmaintain the kind of moral character one needs in order to live. Pride would be the com-mitment to live up to a (more) specific vision of what kind of moral character that is. But itmight also be said that the choice to live amounts to a rudimentary form of pride —ageneralized form of the commitment that characterizes someone with the full virtue.


soning is largely constitutive reasoning —finding out what living a lifegeared to the needs of our physical and psychological survival consists in.What it most fundamentally consists in, according to Rand, is rationalityand productiveness, for the reasons discussed in Section VI, as well as thepractice of the other moral virtues that she defends on similar grounds.But this means, in her view, that a person who has any practical reasonsat all will have (overriding) moral reasons. There will still be no uncon-ditional rational requirements, requirements that could bind a personwhatever his choices and values. The essential practicality of moralitywill lie not in its commanding us unconditionally but in our never havingreasons to act immorally.

Philosophy, Harvey Mudd College


ARISTOTELIAN CONSTRUCTIVISM*

By Mark LeBar

I. Introduction

Constructivism about practical judgments, as I understand it, is thenotion that our true normative judgments represent a normative reality,while denying that that reality is independent of our exercise of moraland practical judgment. John Rawls put constructivism on the map forpolitical theory with his monumental work A Theory of Justice, whichargued that criteria for the justice of basic social institutions were consti-tuted (or constructed ) through deliberation conducted under certain highlyspecified conditions. Rawls credited Immanuel Kant for the basic struc-ture of the approach, and, more recently, Christine Korsgaard and othershave followed Rawls in developing more general constructivist accountsof moral and practical truths.

Indeed, the Kantian strain of practical constructivism (through Rawls,Korsgaard, and others) has been so influential that it is tempting to iden-tify the constructivist approach in practical domains with the Kantiandevelopment of the outlook.1 In this essay, I will explore a somewhatdifferent variety of practical constructivism, what I call “Aristotelian con-structivism” (hereafter “AC”). My aim is to establish conceptual space forthis form of constructivism by indicating in what ways AC agrees with itsKantian counterparts and in what ways it differs. I shall claim that AC is,in one sense, more faithful to the constructivist enterprise than the Kant-ian varieties, in that its understanding of both the establishment of prac-tical truth and the vindication of the theory itself is constructivist.

The essay proceeds as follows. In Sections II and III, I survey the Kant-ian constructivist enterprise, considering both Kant’s own view as con-structivist (in Section II), and Korsgaard’s contemporary developments ofKantian constructivism (in Section III). In Section IV, I sketch AC in light

* For helpful discussion and comments on earlier drafts of this paper, I thank Julia Annas,Alyssa Bernstein, Chris Freiman, Nathaniel Goldberg, Chris Gowans, Tom Hill, Mike Huemer,Paul Hurley, Simon Kirchin, Dan Layman, Tibor Machan, Eric Mack, Fred Miller, JamesPetrik, Dan Russell, Valerie Tiberius, Mark Timmons, Elizabeth Willott, David Wong, TadZawidzki, the other contributors to this volume, and its editors.

1 There are other kinds of accounts, such as ideal observer theories, which may alsoplausibly be made out as “constructivist”; however, they typically do not lay claim to thelabel, and I shall have little to say about them. One which does self-identify as constructivistis the account in Ronald Milo’s essay “Contractarian Constructivism,” Journal of Philosophy92 (1995): 181–204. Like Rawls, Milo is concerned only with “social norms,” not with acomprehensive account of practical truth, and I shall not address his view.


of the nature of constructivism discernable from the Kantian approach.But this sketch lays the groundwork for a challenge to AC that can beexpressed as a dilemma: either AC is implicitly not ultimately construc-tivist, or, if it is, it can be so only by giving up any plausible claim toobjectivity and opening itself to an objectionable relativism. Sections Vand VI argue that AC is constructivist “all the way down,” and Section VIIcontends that this does not eventuate in an objectionable relativism, butthat AC can vindicate a claim to a plausible and attractive form of objec-tivity in ethical judgments.2

A preliminary note about terminology: Some theorists identify construc-tivism with a sort of proceduralism about practical rationality.3 This, Ibelieve, is a mistake. I follow what I take to be Rawls’s conception of whatis distinctive about constructivism: a commitment to the metaphysicalposteriority of practical truths to our apprehensions of them, as I shallexplain in Section II. In this light, identifying constructivism with proce-duralism amounts to taking a species for a genus. Though no substantiveissues hang on the terminology, I believe the understanding of construc-tivism I employ here makes it easier to see how an Aristotelian construc-tivism can address some of the concerns about constructivism in ethics.

II. Kant’s Constructivism

I take Kant’s view to be constructivist in at least the following sense:there are truths about what we have reason to do (or how we have reasonto act), and those truths are themselves established by the exercise ofpractical reason. They do not, that is, exist as objects of moral cognitionprior to the activity of practical rationality in us as agents. Kant himself

2 The significance for constructivism of Rawls’s work —both in political theory and ininterpreting Kantian themes for the purposes of that work —cannot be overstated. Here,however, I will mostly set aside his deployment of constructivism for the defense of justiceas fairness, for several reasons. The most important is that Rawls’s positive view makes noclaims to be a “comprehensive moral doctrine.” Rawls sets out his theory of justice not onlyas more limited than a complete moral theory, but with a focus on the basic structure ofsociety which falls short even of specifying a full social ideal; see John Rawls, A Theory ofJustice (Cambridge, MA: Harvard University Press, 1971, 1999), sec. 2, esp. p. 9. Elsewhere,Rawls explicitly distinguishes the aims and content of that view from “comprehensivedoctrines,” including, specifically, Kant’s “moral constructivism”; see Rawls, Political Liber-alism (New York: Columbia University Press, 1993), 12ff., 99ff., 125. My aim, by way ofcontrast, is to consider constructivism as a more comprehensive doctrine about the natureand content of normative practical truth.

3 This identification is suggested by Stephen Darwall, Allan Gibbard, and Peter Railton,in “Toward Fin de Siècle Ethics: Some Trends,” in Darwall, Gibbard, and Railton, eds., MoralDiscourse and Practice (New York: Oxford University Press, 1997), 12–15. See also SharonStreet, “Constructivism about Reasons” (manuscript). This is also suggested by ChristineKorsgaard’s contrast between “substantive realism” and “procedural realism”; see ChristineKorsgaard, The Sources of Normativity (Cambridge: Cambridge University Press, 1996), 35.However, Korsgaard drops this terminology in later work, and I believe her considered viewis that it is most useful to understand constructivism as a thesis about ontological priority,as I suggest here.

ARISTOTELIAN CONSTRUCTIVISM 183

makes just this priority claim for the nature of good and evil and theirconceptual dependence on the moral law, which is the form of purepractical reason:

[T]he concept of good and evil is not defined prior to the moral law,to which, it would seem, the former would have to serve as founda-tion; rather the concept of good and evil must be defined after and bymeans of the law.4

The contrasting view would be one in which the order of explanationwould be reversed. For Kant, this would mean that the moral law wouldbe understood as undertaking the realization of good conceived as some-how prior to and independent of that law; as Rawls frames the contrast,it would involve an order of moral facts prior to and independent of ourinvestigation of that order, which it is the aim of moral cognition to detectand represent accurately (a view he refers to as “rational intuitionism”).5

Berys Gaut calls this contrasting view the “recognitional model” of prac-tical reason, and says that it construes practical reason as “the capacity torecognize and be motivated by what has objective value.” 6

Constructivism, as I understand it, may be seen as one form of non-skeptical negation of recognitionalism, as a competing theory about theontological priority of practical truth and our recognition of it. “Practicalphilosophy,” Korsgaard says, “is not a matter of finding knowledge toapply in practice. It is rather the use of reason to solve practical prob-lems.” 7 In speaking of “finding knowledge,” she refers to the recognitional-ist conception of the role of practical reason which constructivists reject.The model of the normative world and our relation to it is, on therecognitionalist conception, very like the model of the empirical worldand our relation to it. We gain empirical knowledge when we are impingedupon by elements of that world which exist prior to our apprehension ofthem. In a similar way, the recognitionalist sees us as being impinged

4 Immanuel Kant, Critique of Practical Reason, trans. Lewis White Beck (New York: Macmillan,1956), Ak. pp. 62–63. (“Ak.” refers to the pagination of the standard Prussian Academyedition.)

5 John Rawls, “Kantian Constructivism in Moral Theory,” in John Rawls: Collected Papers,ed. Samuel Freeman (Cambridge, MA: Harvard University Press, 1999), 343; see also Rawls,“Themes in Kant’s Moral Philosophy,” in ibid., 510.

6 Berys Gaut, “The Structure of Practical Reason,” in Ethics and Practical Reason, ed. Gar-rett Cullity and Berys Gaut (Oxford: Clarendon Press, 1997), 183.

7 Christine Korsgaard, “Realism and Constructivism in Twentieth-Century Moral Philos-ophy,” in Journal of Philosophical Research, APA Centenary Supplement (Charlottesville, VA:Philosophy Documentation Center, 2003), 115. In this work, Korsgaard refers to what I amcalling “recognitionalism” as “realism”; this is somewhat confusing because in The Sourcesof Normativity she refers to it as “substantive realism.” However, my concern is the questionof order of explanation between recognitionalists and constructivists, both of whom thinkpractical judgments bear representational content, so I will not concern myself further withthese taxonomic questions.

184 MARK LEBAR

upon in some fashion by elements of a moral or normative order —again,elements which exist prior to our apprehension of them. By way of con-trast, on the constructivist conception our relation to those elements isessentially active, and in some fashion our activity explains the existenceof the normative truths themselves.

Kant’s primary motivation for constructivism, as I understand him,arises from his conviction that our autonomy as moral agents can be pre-served only if we give to ourselves the laws governing our conduct.Autonomy, as Kant makes it out, is the capacity we have for the “deter-mining grounds” of our wills to be self-imposed by the exercise of ourreason, rather than imposed on us by natural law, as is the case withnonrational animals.8 Such autonomous willing is necessary, he main-tains, for the very possibility of value in the world. The only uncondi-tional good —and the condition of all the good that there is —is goodwilling.9 And good willing is just willing that is determined by reason —bythe application of practical rationality to the principles upon which wechoose to act.

From this framework follows the nature and extent of Kant’s construc-tivism.10 True claims about how we ought to conduct ourselves are madetrue in virtue of a certain property of the principles (or maxims) on whichwe act. That property is, in its most perspicuous formulation, the prop-erty of being suitable for willing as universal law. Kant’s idea is, first, thatto see ourselves as acting —as being the causes of the actions and out-comes we will to bring about in the world —our willings and our actionsmust be connected in a law-like (that is, uniform) way. We must, that is,act according to laws. For our willings to have moral worth, however, thelaws upon which we act must be rational, not merely causal.11 Moralworth depends on willing in ways that are determined by reason, and thismeans that the principles upon which we intend to act —our maxims —can be morally valuable only when they pass a certain test of reason. Thetest of reason Kant proposes is the most basic and indispensable of ratio-nal tests: the bare avoidance of contradiction. Kant says of this test that itis needed “in general for the possibility of any employment of reason. . . .

8 Immanuel Kant, Groundwork for the Metaphysics of Morals, trans. Allen Wood (NewHaven, CT: Yale University Press, 2002), Ak. pp. 440, 452.

9 Ibid., Ak. p. 396.10 In addition to Korsgaard’s well-known readings of Kant as constructivist, Andrews

Reath offers a carefully-worked-out interpretation in Reath, “Legislating the Moral Law,”Noûs 28 (1994): 435–64; and Thomas Hill offers a more limited constructivist conception inHill, “Hypothetical Consent in Kantian Constructivism,” Social Philosophy and Policy 18, no. 2(2001): 300–329. It is certainly not beyond controversy that Kant should be read as construc-tivist. For an argument that he should not, see Patrick Kain, “Self-Legislation in Kant’sMoral Philosophy,” Archiv für Geschichte der Philosophie 86 (2004): 257–306. Barbara Hermanalso argues for limits on the extent of “creation” in Kant’s constructivism: see Herman,“Justification and Objectivity: Comments on Rawls and Allison,” in Kant’s TranscendentalDeductions, ed. E. Förster (Palo Alto, CA: Stanford University Press, 1989), 131–41.

11 Kant, Groundwork, Ak. p. 412.


[It is] the condition of having any reason at all.” 12 Impermissible maxims,he tells us, are those that cannot be willed as universal because doing soinvolves the agent in a contradiction of one sort or another. Thus, thecrucial property that establishes the truth of claims about how we oughtto act is that of surviving a certain process of rational scrutiny —a processthat it is incumbent upon us, as moral agents, to undertake in determin-ing how to act. In this sense, the truth-makers for true normative judg-ments are constructed by a process we engage in as rational agents. Thejudgments are true in virtue of their surviving this process, and since theprocess is one in which our practical rationality is the effective agency,we can be said to have constructed that truth.13

The contrast to this sort of constructivism —an archetypical recognitionalview —might be represented by a proposal we might take from Plato’sRepublic.14 Suppose there is a Form of the Good —an idea that exists in aworld of ideas, and which is the essence of what is common to all thingsgood; their resemblance to (or “participation in”) this Form is the reasonwe call all these divers things good. On this picture, when a practicaljudgment is true, what would make it true is X’s standing in a certainrelation with the Form of the Good.15 Our task as rational agents wouldbe to recognize and respond to that Form as best we could; this is justwhat rational agency would consist in.

Why does Kant reject recognitionalism and see moral truth as con-structed instead? The best place to look for an answer to this question isKant’s response to the only viable alternatives he considers to the auton-omy approach: theological voluntarism, and what he calls “the ontolog-ical concept of perfection.” 16 The latter amounts to one way we mightconstrue the Platonic proposal. Kant’s objection is that the requisite con-cept of perfection is uselessly indeterminate:

[The ontological concept of perfection] is, no doubt, empty andindefinite, and consequently useless for finding in the boundless

12 Kant, Critique of Practical Reason, Ak. p. 120.13 Some care is required here. The content of the moral law holds not only for all human

beings, but for rational beings generally, and necessarily so (Kant, Groundwork, Ak. pp. 389,408, 411–12). So the sense in which we might see ourselves as constructing practical andmoral truth can be only that we are “legislating” this universal law for ourselves: we areimposing it, making it authoritative, determining our wills by it, or (as Kant puts it) author-ing not the law itself but its “obligation” (Kant, The Metaphysics of Morals, trans. Mary J.Gregor [Cambridge: Cambridge University Press, 1997], Ak. p. 227). I owe this way ofthinking about the limitations on autonomy to Patrick Kain.

14 I make no claim that this proposal is Plato’s own considered view; I offer reason forthinking otherwise in Section IV below.

15 Cf. Plato, Republic 517c.16 Kant, Groundwork, Ak. p. 443. Rawls offers a somewhat different account of Kant’s

reasons for rejecting “rational intuitionism” in Rawls, “Themes in Kant’s Moral Philoso-phy,” 520. “Theological voluntarism” makes value depend on divine will. Kant rejects thatapproach, and I will not attend to it.

186 MARK LEBAR

field of possible reality the greatest amount suitable for us; moreover,in attempting to distinguish specifically the reality of which we arenow speaking from every other, it inevitably tends to turn in a circleand cannot avoid tacitly presupposing the morality which it is toexplain. . . .17

Clearly, Kant believes that a focus on a rationally intuited moral idealmust “turn in a circle” —it must inevitably lead back to the very processhe has been engaged in, namely, attempting to identify what principlecould govern an “absolutely good will.” But why?

Kant might have in mind something like the following.18 In earlierwork, Kant considered how our representations come to “agree with”their objects.19 He thinks there are only two ways this is possible: ourintellect is either passive or active with respect to its objects, and we canconstrue the issue as taking the form of a dilemma. First, consider anyrepresentation we might have of this “ontological perfection.” It makesno sense to suppose it could be the result of our being acted upon by thisobject (viz. that we are passively receptive to it). Kant sees no way tounderstand such receptivity outside the causal order. He identifies passiv-ity with the “world of sense” and activity with the “intellectual world”;20

rationality just is a form of activity, so the proposal that we could pas-sively experience a rational intuition of such an object is unintelligible.21

And the involvement of our sensible natures would entail that the result-ing principles of the will must be heteronomous and thus useless as asource of moral authority.

Suppose, then, that we take our intellect to be active in the apprehensionof this object (the Good). What we would need to understand, then,would be the content of the representation such an active intellect couldhave, as an active intellect. But that is precisely the question Kant has beenattempting to answer in the second section of the Groundwork. In otherwords, if we take the only promising branch of the dilemma, we end upjust where Kant had brought us before considering the objection that therecould be a recognitional alternative to his construal of the good will andautonomy. This, then, is the “circle” Kant is referring to. The alternative

17 Kant, Groundwork, Ak. p. 443.18 Here I am grateful to James Petrik for helpful discussion.19 See his letter to Marcus Herz, reprinted in Lewis White Beck, ed., Kant: Selections (New

York: Macmillan, 1988), 81–33 (Ak. X:130–31).20 Kant, Groundwork, Ak. p. 451.21 Kant says that Plato “assumed a previous intuition of Divinity as the primary source of

the pure concepts of the understanding,” which makes no sense: “the deus ex machina in thedetermination of the origin and validity of our knowledge is the greatest absurdity onecould hit upon and has —besides its deceptive circle in the series of inferences from ourhuman perceptions —also the additional disadvantage that it provokes all sorts of fancyideas and every pious and speculative sort of brainstorm” (Kant, letter to Herz, 131). WhileKant is referring here to theoretical rather than practical knowledge, I take the proposal thatknowledge of the Good could give us reason to act to be an unholy synthesis of the two.


is vacuous because it fails to supply the content of such a representationfrom an analysis of what the active, practical intelligence must be like.Since the activity of the intellect must somehow be essential to the contentof the representation itself, constructivism is inescapable.

Now, if Kant’s theory succeeds, it yields a constructivist understandingof normative, practical truths —truths about what we ought to do. In animportant sense, however, Kant’s constructivism does not go “all the waydown”: the overall justificatory structure of his account is foundational,resting on considerations that are not themselves constructed.22 The argu-ment for the authority of the moral law in the Groundwork is driven, first,by reflection on common-sense intuitions about moral value. Kant seeshimself as developing the concept of the good will, “just as it dwellsalready in the naturally healthy understanding, which does not need tobe taught but rather only to be enlightened.”23 Kant thus takes our rec-ognition of the value of the good will as something like a datum, not itselfan article of construction. Further, the reality of the instantiation of thisconcept in us —the fact that it is not merely a “chimera” —is shown byKant’s argument in Section III of the Groundwork that we must see our-selves as under the “idea of freedom,” and hence as subject to the morallaw.24 In the Critique of Practical Reason, Kant famously reverses the direc-tion of this argument and holds that, since we know that we are subjectto the moral law, we know that we are free: “[T]he moral law is given, asan apodictically certain fact, as it were, of pure reason, a fact of which weare a priori conscious.” 25 Thus, while Kant is a constructivist about nor-mative truth, his argument for the account of that truth seems to dependon cognitions that do not seem to have as their objects constructions, butrather antecedent normative facts about our nature and the nature ofvalue. As Rawls suggests,26 this constructivism is built upon principlesthat are not themselves constructed. This structure, to which I shall developa contrast below, is also characteristic of Kant’s constructivist legacy, inKorsgaard’s neo-Kantianism, to which I now turn.

22 Larry Krasnoff is, so far as I know, the only person to have explicitly distinguishedthese structural features of Kant’s theory. See Krasnoff, “How Kantian Is Constructivism?”Kant-Studien 90 (1999): 385–409. However, Hill also argues that Kant’s constructivism islimited in ways similar to those I will suggest, in Hill, “Hypothetical Consent in KantianConstructivism.”

23 Kant, Groundwork, Ak. p. 397.24 Ibid., Ak. pp. 445, 448.25 Kant, Critique of Practical Reason, Ak. p. 47.26 Rawls maintained that the facts pertaining to just basic social institutions and the

persons that realize them were a subset of all the moral facts there are, and denied that theproposal that constructivism could go “all the way down” was even intelligible: “Thus, wedon’t say that the conceptions of persons and society are constructed. It is unclear what thatcould mean. . . . We should not say that the moral facts are constructed, since the idea ofconstructing the facts seems odd and may be incoherent. . . .” (Rawls, “Themes in Kant’sMoral Philosophy,” 514, 516; reiterated in Political Liberalism, 104, 121–22). See also PoliticalLiberalism, 108: “The conceptions of society and person as ideas of reason are not, certainly,constructed any more than the [procedural] principles of practical reason are constructed.”

188 MARK LEBAR

III. Korsgaard’s Kantian Constructivism

In her early constructivist work in The Sources of Normativity, Korsgaardunderstood the essential activity of our practical reason in constructingmoral truth as a function of the development of practical identities.27 Theseidentities are “descriptions under which you value yourself . . . find yourlife to be worth living and your actions to be worth undertaking.” 28 Wesee ourselves in certain ways (as child, teacher, lawyer, neighbor, Lutheran,and so on) that shape our conduct, and from these ways of seeingourselves —from these “practical identities” —flow our obligations andreasons for action. But the fact that we are so constituted that we needreasons to act, and find such reasons in the various practical identitieseach of us acts from and is obligated by, gives us a further practicalidentity, to which reflection on the nature and source of obligation leadsus. This is our identity as moral beings, our “moral identity” or “human-ity,” which Korsgaard understands as tantamount to seeing ourselves ascitizens of the Kingdom of Ends on Kant’s conception.29 Only the recog-nition of this identity can afford us the sort of reflective success that,Korsgaard argues, our quest for the source of normativity demands.

As Korsgaard acknowledges, however, this argument alone does notdeliver all of what morality involves; in particular, it does not deliver theright story about our obligations to others. For that, an additional bit ofargument is required, one which denies that it is possible to have reasonsonly to value one’s own humanity without similarly valuing the human-ity of others. Korsgaard chides various forms of “neo-Hobbesian” and“neo-Kantian” theories for attempting to begin with “private” reasonsand then argue to conclusions that we are committed, by some sort ofrational inference, to recognize the humanity of others as a source of“public” reasons.30 Instead, she argues —drawing a parallel to LudwigWittgenstein’s argument that no private meanings are possible —that rea-sons must be shareable to be reasons.31 Because they are shareable, if myhumanity is to be reason-giving for me, it must be reason-giving for youas well, and of course, on the same grounds, your humanity must bereason-giving to me. Our most essential practical identity establishes some-thing like the Categorical Imperative understood under Kant’s Formulaof Humanity.

Thus, our activity as moral agents consists in the reflection upon andadoption of these identities, and in our acting on the reasons that emanate

27 Korsgaard, The Sources of Normativity, lectures 3, 4, and 9.28 Ibid., 101.29 Ibid., 115. Korsgaard also thinks her argument for the grounding work performed by

the moral identity is “a fancy new model” of Kant’s Formula of Humanity (which requiresus to treat the “humanity” in ourselves and others always as an end, never merely as ameans; ibid., 122). Since Kant thinks his formulations of the Categorical Imperative areequivalent, it is hardly surprising that Korsgaard sees elements of both in her conception.

30 Ibid., 133.31 Ibid., 135.


from them. However, we have these obligations only in virtue of the iden-tities we endorse through reflection.32 In considering an objection from G. A.Cohen, Korsgaard makes clear that it is our activity in endorsing that doesthe normative work. What, Cohen asks, is an “idealized Mafioso” obli-gated to do, given that he takes his identity to be that of a loyal soldier forthe family, with a code of strength and honor? If he is asked to murder byhis paterfamilias, is he thereby obligated to do so?33 Korsgaard bites the bul-let in her reply. The Mafioso’s obligation is real, she says: “And this isbecause it is the endorsement, not the explanations and arguments that pro-vide the material for the endorsement, that does the normative work.” 34

This is an expression of Korsgaard’s constructivism: if she were to hold thatit is the “explanations and arguments” that do the work, she would be opento the suggestion that, at bottom, her picture is one of agents discoveringthrough reflection normative facts that are there antecedent to their inquiry.And that is the model she rejects. It is only the constructive activity of moralagents that brings into existence normative truth.

But, surely, having to maintain that the Mafioso has any obligation tocarry out his repugnant assignments is a high price to pay to maintainconstructivism, and Korsgaard appears to have thought better of it. Herlater constructivist work (in particular, her “Locke Lectures”) makes littleuse of the notions of practical identity or endorsement. The emphasisinstead is on self-constitution. Her argument (drawing on threads recog-nizable from The Sources of Normativity) is that our human predicament issuch that we cannot help but act; however, action itself is constituted bycertain rational norms. “Principles of practical reason,” she argues, “areprinciples of the unification of agency.” 35 The option for us is not whetherto will badly or well, relative to Kantian imperatives; it is whether or notto constitute ourselves as agents at all.

Korsgaard argues here (as earlier) that particularistic willing —that is,willing that is intended to bind just in a particular case —is impossible.36

Either we identify ourselves with the “principle of choice” in deciding toact, or we are not agents at all —we are merely theaters in which a conflictof causal impulses plays out. But particularistic willing, she argues, amountsto collapsing the distinction between these two alternatives. A particular-istic will would be one that simply identified with whatever impulse wasdecisive in a particular case, and thus really ceased to be a will at all. Soagency requires willing in ways that are not particularistic. On Korsgaard’sview, that entails willing universally; and thus, to constitute ourselves asagents, we must govern ourselves by the Categorical Imperative.

32 Ibid., 252.33 Ibid., 183.34 Ibid., 257.35 Christine Korsgaard, “Locke Lectures,” 1.3.4 (http://www.people.fas.harvard.edu/

;korsgaar/#Publications).36 Korsgaard, “Locke Lectures,” 2.5.2.

190 MARK LEBAR

Like Kant, Korsgaard emphasizes that principles of practical rationalityare principles governing activity, not bits of knowledge to be appre-hended. In his essay “What the Tortoise Said to Achilles,” Lewis Carrollmade the point that there is no reducing rules of inference to mere prem-ises; rules of inference govern something we do, and cannot be takenwithout remainder as truths that themselves form only premises for fur-ther inferences.37 Korsgaard deploys just this argument to hold that stan-dards of practical rationality cannot obligate us to act. To think of standardsin this way —as things that can be objects of knowledge, and carry withthem obligations —is to make just the mistake poor Achilles made. Wemust instead see these principles as constituting agency; to be an agentjust is to be governed by principles of practical rationality, in precisely thesame way that to have a mind just is to be governed by rules of infer-ence.38 Construction, as activity, is what norms of practical rationality andmorality are about. Practical truth is constructed, not discovered, becauseit is activity in accordance with the norms of practical rationality, whichare themselves constitutive of agency.

In Korsgaard’s work (both earlier and later), as in Kant’s, we can seeboth constructivism and foundations for the enterprise of construction. InThe Sources of Normativity, practical agency is characterized (at times, orwhen conflicts in motivation arise) by a reflective search for a “source ofnormativity” —an answer to the question of why I must do what I wouldrather not do. That search can be terminated only by recognizing that thefact that I am searching for such a reason reflects something importantabout the kind of creature I am, namely, a creature who craves and actsupon reasons. That recognition is a recognition of my “moral identity,”and that is the source of normativity. As in Kant, the justificatory questends with the discovery of a more fundamental fact, though in this casea fact about the sort of creature that could seek justification in the firstplace. In Korsgaard’s “Locke Lectures,” the argument is more nearlyKant’s own, though here the grounding argument rests on what is con-stitutive of agency —namely, governing how one acts by the CategoricalImperative. Once I recognize that I am an agent, and that I act, I recognizethat doing so is impossible without doing so on principles, and indeed onprinciples that are (so it is claimed) universally binding. In both versionsof the story, the truth-makers for normative practical claims are the objectsof construction rather than recognition: they do not exist except through

37 Lewis Carroll, “What the Tortoise Said to Achilles,” Mind 4 (1895): 278–80. Achilles triesto persuade the Tortoise (in effect) that, given P and P r Q, he must conclude Q. Tortoiseasks why, and Achilles responds that it is logically necessary that he do so. Tortoise then askswhy a skeptic must accept the inference, and Achilles unwisely suggests that it is becauseof the truth of the proposition that P, Pr Q, and (P & Pr Q)r Q. When Tortoise asks whythat inference must be accepted, Achilles even more unwisely suggests that that is due to thetruth of a further proposition, and an infinite regress has begun. The moral: no number ofpremises can substitute for a rule of inference.

38 Korsgaard, “Locke Lectures,” 2.4.1ff.


our rational agency. As in Kant, however, the processes of constructioninvolve rationally mandated procedures of deliberation that are success-fully concluded only upon recognizing some fact that gives us reason toacknowledge the truths in question. With that framework as our back-ground, I now turn to sketching a form of constructivism that is, by wayof contrast, constructivist “all the way down.”

IV. Aristotelian Constructivism

Aristotelian constructivism (AC), as I shall call it, shares with Kant’sand Korsgaard’s versions of constructivism a comprehensiveness of ambi-tion: it claims to give a general account of the nature of what is true aboutwhat we have reason to do that sees such truth as constructed rather thanrecognized. It differs from its Kantian cousins in the nature and justifi-cation of the method of construction which constitutes the truth-makersfor normative practical claims.

Kant’s view is constructivist in holding that truths about what weought to do are constructed through a procedure applied to the maximsor principles upon which we propose to act. Korsgaard amends this storyto maintain that the procedure of maxim-testing constitutes us as agents,unifies the diverse parts of our “souls,” and thus constructs the “forms”of our selves as the particular persons we are.

In contrast, on AC the enterprise of construction involves not merelyprocedures but substantive normative judgments. Truth about what wehave reason to do is established in light of the aim of living well, wherewhat counts as living well is an object of construction, though not theupshot of any formal procedure. Part of the Aristotelian legacy drawnupon here is the idea in Aristotle (as in Plato and later Hellenistic phi-losophies) that eudaimonia (happiness) is the ultimate end for us as humanbeings, both in the psychological sense that it is, in fact, what people wantfor themselves, and in the normative sense that it is what we have great-est reason to seek in living and acting. As in Kantian constructivism, theheart of AC is the denial that the truth about how to live and act is outthere somehow, waiting for us to recognize and act on it, even in thesubstantive judgments that are incorporated as part of the enterprise ofconstruction.

The role of such judgments has led some to see Aristotle as exemplify-ing a recognitionalist view. Gaut maintains that Aristotle begins with “anindependent account of the conditions under which actions are good, and[derives] from this an account of practical rationality.” 39 At some points,Aristotle’s language does invite a recognitionalist reading —in particular,

39 Gaut, “The Structure of Practical Reason,” 13. See also Ralph Wedgwood, “PracticalReasoning as Figuring Out What Is Best: Against Constructivism,” Topoi 21 (2002): 139–52.

192 MARK LEBAR

his description of practical wisdom as a kind of perception.40 Aristotle’smany claims that virtuous action aims at the fine or noble (to kalon) are alsonot incompatible with such a reading. Though I shall challenge these tex-tual claims briefly in what follows, my main argument to the contrary drawson what I take to be the broader thrust of Aristotle’s view.41

The constructivism of Kantian value theory is closely related to itsemphasis on the conditional nature of the value of everything but thegood will. Constructivism flows from this thesis about value, since theidea just is that value is established through acts of rational willing. Now,less well-known is the widespread endorsement of a similar but distinctconditionality thesis in ancient eudaimonism, and this is the place tobegin grasping the constructivism in it. Plato says that in themselves“both ‘good’ and ‘bad’ things are valueless,” 42 and that all the thingscommonly called good are “possessions of great value to the just andpious, but . . . to the unjust they are a curse.” 43 Aristotle also is explicitthat their goodness is conditional upon their deployment in accord withvirtue: “to the noble and good man things profitable are also noble; but tothe many the profitable and the noble do not coincide, for things abso-lutely good are not good for them as they are for the good man.” 44 Evenstarker is the Stoic doctrine that things besides virtue are themselvesnever good; they are at best “indifferents” that may be “preferred” as theobjects of wise or virtuous “selection.” 45 On all these accounts, valuedepends on virtuous agency, and this conditionality of value leads toconstructivism in these accounts just as it does in Kant’s.46

40 Aristotle, Nicomachean Ethics (EN ) VI.8.1142a25–32. Quotations are from the Ross/Urmson translation, in The Complete Works of Aristotle, ed. Jonathan Barnes (Princeton, NJ:Princeton University Press, 1984).

41 Still, I do not claim to be offering an interpretation of Aristotle; instead, I believe theview is one naturally congruent with major themes in his ethics.

42 Plato, Euthydemus 281d.43 Plato, Laws II.661b; see also Gorgias 470e. It is not easy to reconcile the metaethical

outlook expressed in these passages with the sort of robust realism that the theory of theForms in, e.g., the Republic is taken to exhibit. But the fact that these passages are drawnfrom both what are taken to be some of Plato’s earliest dialogues, and what is without adoubt his last, suggests that it is an outlook from which he likely never departed.

44 Aristotle, Eudemian Ethics VII.15.1249a10–12. Should a constructivist worry about thenotion of “absolute good” here? No. Elsewhere (EN V.1.1129b2–7), Aristotle holds thatthe goodness of such goods is conditional, and (EN III.4.1113a32) that the good man is the“norm and measure” of good.

45 Cf. Cicero, De Finibus III.53–55. The Stoic value theory on which nothing but virtue canproperly be called “good” is the clearest case of an account of conditional value in theancient world. However, as the text indicates, I think different versions of the notion thatvalue is conditional (and thus that it is not some nonrelational property of things, waitingfor us to recognize it) are found in Plato and Aristotle as well; and, in fact, Aristotle’s viewoffers the best overall understanding of the tacit value theory at work in the other accounts.I thank Fred Miller for pressing this point.

46 In this way, the ancient accounts are vulnerable to the objections Gaut offers againstKant’s conditionality thesis just as Kant’s account is (Gaut, “The Structure of PracticalReason,” 165–170). I do not believe these arguments are telling against conditionality ineither case, but I do not have space to address these questions here.


According to AC, value or goodness is constructed in the followingsense: the reasons we have for undertaking particular actions, realizingparticular aims, etc., are not a matter of recognizing or responding to anysorts of normative properties to be found in the objects of our actions. Inoutline, the idea is that the goodness of such things is conditional upontheir contribution to a good human life. The goodness of (say) a cold beeris not anything one recognizes and responds to. Instead, one recognizesand responds to the beer’s natural properties (these, of course, are notconstructed), and to the extent that, in virtue of its natural properties, itcontributes to one’s living well, only in that way does it have value —onlyin that sense is it good. In turn, the goodness of a good life is cruciallydependent on the exercise of practical rationality in the right way (that is,on practical wisdom, or phronêsis). This is true both because good livesrequire the direction of practical wisdom and because only successfulpractical rationality can determine which lives are genuinely good. Con-versely, the criterion for success in practical rationality —practical wisdom —just is the construction of a good life. Neither eudaimonia nor practicalwisdom can be specified without essential reference to the other.

Now this story is patently circular, so there is at least a threat of vacuityhere. But that threat is empty, for two reasons. First, the circularity at issueis not conceptual, but metaphysical. The claim is not that our concepts ofeither living well or practical wisdom are mutually dependent. Instead,the claims of AC are substantive claims about what in the world theseconcepts should rightly be seen as picking out. In each case, what ispicked out makes essential reference to the other.

Second, what is picked out is something about which we can learn inother ways. For example, the considerable work in recent years on theo-ries of well-being can contribute to our understanding of what makeshuman lives good ones, as can common-sense judgments about successesand failures in living.47 The eudaimonist insight into that work is thatsuch lives cannot be successful except when they recognize the appro-priate place for practical wisdom. And as for practical wisdom, there hasbeen no shortage in human history of reflection on what courses of humanconduct are wise. Here eudaimonism gives point and focus to that reflec-tion, and can add what we learn about thinking about virtue more gen-erally. These examples are hardly exhaustive, but I trust their point isclear. While we can fully understand neither what a good life is nor whatpractical wisdom consists in independently, the way they are interdepen-dent does not leave us with a vacuous circularity. This is not to say thatthere isn’t work to be done in expanding the interdependence between

47 The history of Greek ethics is itself a lively debate over exactly what lives are goodlives, and why. I think it is useful to see the enterprise of answering these questions as oneof establishing reflective equilibrium between judgments about particular cases and abstractprinciples attempting to unify and harmonize those judgments.

194 MARK LEBAR

eudaimonia and practical wisdom, but there is no reason to assume at theoutset that that task is impossible.

In contrast to the structure of Kantian constructivism, the constructiv-ism in AC goes “all the way down,” in that the fundamental normativeelements at work in the account are themselves products of construction.Now this does not mean that everything is constructed; as Rawls pointsout, every construction needs materials.48 For AC, those materials are thenatural facts that bear on normative judgments and properties. Thesefacts are not themselves constructed, but are straightforwardly objects ofrecognition. Thus, the sense in which AC’s constructivism goes “all theway down” refers to the nature and content of its fundamental normativeelements —eudaimonia and practical wisdom —and indicates that AC hasno normative “primitive” or base notion on which the normative struc-ture of the account is founded.49 I will explore this picture by explaininghow AC addresses a problem any constructivist view must face. Theproblem takes the form of a dilemma.50 One horn of the dilemma is aconcern that the construction cannot succeed without appeal to somenormative standard that is not itself constructed (so that constructivism isreally just window-dressing on recognitionalism). The other horn is thecharge that, in avoiding the first horn, constructivism becomes arbitraryor unacceptably relativistic: in repudiating any independent standard ofevaluation, it loses the footing necessary to establish the kinds of norma-tive judgments we want to make against noxious practices and beliefs.51

I contend with this second horn in Section VII of this essay; in the nexttwo sections, I undertake a two-part explanation of how AC avoids thefirst horn, and in doing so try to explain further how its fundamentalnormative elements are objects of construction. I take up the constructionof eudaimonia in Section V, and of practical wisdom in Section VI, withthe aim of showing that both elements are constructed, not recognized.

V. Eudaimonia

The ancients agreed that a good human life must be the life of a goodspecimen of our kind, and that what is distinctive of our kind is the

48 Rawls, “Themes in Kant’s Moral Philosophy,” 514; see also Rawls, Political Liberalism,103–4.

49 I thank David Wong for pressing for clarity on this point.50 Various Euthyphro-style dilemmas have been posed as problems for constructivism; for

two recent examples, see Russ Shafer-Landau, Moral Realism (Oxford: Clarendon Press,2003), 42; and Mark Timmons, “The Limits of Moral Constructivism,” Ratio 16 (2003): 400ff.The version I set out here differs somewhat from both these formulations, but I think it isthe natural concern given the species of constructivism I offer.

51 We might understand Cohen’s Mafioso objection to Korsgaard to take this form:Korsgaard’s practical-identities account is impaled on this second horn in virtue of its“content-neutral” emphasis on reflective endorsement. Cohen presses his objection in theseterms in Korsgaard, The Sources of Normativity, 184.


capacity for reflective practical rationality. Our practical reasoning cantake as its object any number of things, including its own operation: wecan reason not only about what actions to perform, but about how weshould reason. Aristotle, like Kant and others, is struck by the fact thatwe have this capacity; and it is distinctive of us that we characteristicallylive via its exercise. One implication of this fact is that because we live thisway, doing so well establishes normative standards for action for us. Thatis the basic point of Aristotle’s notorious ergon argument in NicomacheanEthics I.7: while we share nutritive, perceptual, and motor capacities withother living beings, what is distinctive about us is that we reason abouthow to think and act. So living well as the kind of things we are involvesreasoning well, and that is why successful practical rationality —practicalwisdom —plays such a crucial role in living well as a human being.

Nonetheless, while the attention to practical rationality is somethingAristotle shares with Kant, Aristotelian constructivism departs from theKantian variety in attending to the deep and complex connection betweenthis rational capacity and our animal natures, and in maintaining in lightof this complexity that there is no hope for a merely formal or proceduralaccount of success in practical rationality. I return to the latter point in thenext section; at present, I want to attend to the significance of this con-nection for living well. AC couples the attention to reflective rationalitywith the recognition that we are organisms and share a biological naturewith other animals and plants. We are ourselves animals, with the fullsuite of nutritive, perceptual, appetitive, and affective systems whichsuffice for life in simpler organisms. However, in addition we have reflec-tive rationality, as the ergon argument observes, and it is no overstatementto say that this changes everything.

One important effect of this change is captured nicely by Kant, whosays:

Freedom of the power of choice has the characteristic, entirely pecu-liar to it, that it cannot be determined to action through any incentiveexcept so far as the human being has incorporated it into his maxim. . . .52

Kant’s point, with which the ancients agree, is that our capacity for reflec-tive rationality opens up a critical distance between the conative states weexperience as animals and our eventual response in action.53 This effectsan important transformation in these conative states: they call into play

52 Immanuel Kant, “Religion within the Boundaries of Mere Reason,” in Religion andRational Theology, trans. Allen Wood and George di Giovanni (Cambridge: Cambridge Uni-versity Press, 1996), Ak. p. 24; emphasis in original. Henry Allison calls this Kant’s “Incor-poration Thesis,” and claims that it “underlies virtually everything that Kant has to sayabout rational agency.” See Henry Allison, Kant’s Theory of Freedom (Cambridge: CambridgeUniversity Press, 1990), 40.

53 Cf. Korsgaard, The Sources of Normativity, 93, 223; Korsgaard, “Locke Lectures,” 4.2.4.

196 MARK LEBAR

practical rationality, and can for the first time become reasons for action.John McDowell indicates how this critical distance is manifest in theAristotelian picture in writing of our need to be trained in the deploy-ment of practical reason:

Moral education does not merely rechannel one’s natural motiva-tional impulses, with the acquisition of reason making no differenceexcept that one becomes self-consciously aware of the operation ofthese impulses. In imparting logos, moral education enables one tostep back from any motivational impulse one finds oneself subject to,and question its rational credentials.54

It is not simply that what would otherwise be merely sources of motiva-tion now have simply become reasons; it is that action has become possible,as an alternative to our being just another link in the extended causalorder —patients of prior causes, channeling them to whatever effect nat-ural law dictates. By our nature, then, we are committed to the necessityof some determination of what sorts of lives we will choose to lead by thedeployment of our capacities of reflective rationality: how we live and actbecomes a matter of choice in a way it is not for other creatures.

A second implication of our having this capacity, emphasized far moreby the ancients than in the Kantian tradition, is the transformative effectof this capacity on the rest of our animal natures —on our passions anddesires in particular. McDowell calls this rational transformation a sort of“second naturalism,” and it is a theme running throughout the ancients.55

Aristotle insists that “habituation” —in large part, a training and educa-tion of the passions —is essential for virtue. Less well-known, perhaps, arePlato’s ways of making the same point. In the Laws, Plato emphasizeswhat we might call the “plasticity” of the objects of our pleasure andpain —the great degree to which we can learn to experience pleasure orpain in consequence of achieving certain aims:

I maintain that the earliest sensations that a child feels in infancy areof pleasure and pain, and this is the route by which virtue and vicefirst enter the soul. . . . I call “education” the initial acquisition ofvirtue by the child, when the feelings of pleasure and affection, painand hatred, that well up in his soul are channeled in the right courses

54 John McDowell, “Two Sorts of Naturalism,” in Virtues and Reasons, ed. RosalindHursthouse, Gavin Lawrence, and Warren Quinn (Oxford: Clarendon Press, 1995), 170.

55 This form of “naturalism” differs from that developed by some Aristotelian“naturalists” —such as Philippa Foot, Natural Goodness (Oxford: Oxford University Press,2001), and, to a lesser degree, Rosalind Hursthouse, On Virtue Ethics (Oxford: Oxford Uni-versity Press, 1999), esp. chap. 9 —in recognizing that our rationality makes an essentialcontribution to the normative properties of anything that is good. Cf. McDowell, “Two Sortsof Naturalism,” 166ff.


before he can understand the reason why. Then when he does under-stand, his reason and his emotions agree in telling him that he hasbeen properly trained by inculcation of appropriate habits. Virtue isthis general concord of reason and emotion.56

Just because our experiences of pleasure and pain are plastic in this way,moral education is critical for us. Before our capacity to reflect upon andrecognize reasons, we need the benefit of wise guidance to “channel” ouraffective and conative responses to the right sorts of things. But it is thetask of practical rationality to determine what these right sorts of thingsare, and to establish the canons for moral education. The point is notextirpation of the motivating effects of these other aspects of our animal-ity, nor indeed to bend them to the service of an idealized rationality.Instead, the point is (as so nicely illustrated by the central analogy of theRepublic) the thriving of the integrated whole —the good life of a creaturethat is both essentially animal and essentially rational.

This emphasis marks the significant difference between Aristotelianand Kantian forms of constructivism. Whereas, in the latter view, thefocus is the nature and exercise of pure practical reason, “fully cleansedof everything that might be in any way empirical and belong to anthro-pology,” 57 the focus in the former view is a life: a process (or, better,activity) played out over a span of history, with a concrete beginning, aconcrete ending, and an unfolding of successive stages of life in between.We not only have these lives but we live and shape them, and the goal ofliving well is both informed by the fact that we have the capacity forreflective rationality and the object of practical rationality as we exerciseit. Indeed, Aristotle goes so far as to suggest that our selves are constructedthrough the activity of living our lives deliberatively:

We exist by virtue of activity (i.e. by living and acting), and . . . thehandiwork [ergon] is in a sense, the producer in activity; he loves hishandiwork, therefore, because he loves existence. And this is rootedin the nature of things, for what he is in potentiality, his handiworkmanifests in activity.58

A good human life is the life of one of a certain kind of animal, in whichpractical reason molds and shapes the natural drives, motives, and pas-sions, and in so doing creates a self. This is what renders Aristotle’s useof the locution of “perception” (aesthesis) for wise judgment congruentwith the constructivist view I am proposing here: he says the perception

56 Plato, Laws II.653a–b (Saunders translation), in Complete Works of Plato, ed. John Cooper(Indianapolis, IN: Hackett, 1997).

57 Kant, Groundwork, Ak. p. 389.58 Aristotle, EN IX.7.

198 MARK LEBAR

in practical judgment is of “another sort” (allos eidos) than perception ofobjects.59 He may be suggesting that this capacity for perception is some-thing we are active in shaping; the reflective work of practical wisdom inhabituation is a way of coming to see the patterns that are constitutive ofliving well. Once shaped, the capacities thus formed are capable of detect-ing, of responding to, the descriptive features of conditions and circum-stances that warrant action, and in this sense the capacities engage inperception. What responds is not merely an intellect but a self —the selfthat is itself the product of construction, through habituation guided bypractical wisdom.60 The result is a life that such a creature can judge to begood in light of the criteria we can best identify for evaluating the kindsof lives of which we are capable. The normative success of that life, itsgoodness, is a construction of the effective exercise of that very capacity forpractical rationality. Both what counts as a good life and the goodness ofsuch a life depend on the exercise of our practical rationality. There are nofacts about living well as a human being apart from the wise judgmentsof human agents about living so.

The constructivist credentials of AC’s notion of eudaimonia thus cometo rest on its notion of practical wisdom. But the canons for success inpractical rationality are themselves objects of construction on AC. Practi-cal wisdom ( phronêsis), in Aristotle’s own view and in AC, just is practicalrationality, exercised in such a way as to live well:

Now it is thought to be a mark of the man of practical wisdom to beable to deliberate well about what is good and expedient for himself,not in some particular respect . . . but about what sorts of thingconduce to the good life in general.61

Thus, the standard of success in practical rationality just is eudaimonia —asubstantive, not merely procedural, standard.

This is a curious picture, in more ways than one. How exactly is eudai-monia supposed to afford this standard? The challenge for AC is to resistthe idea that the norms for success in either living or practical reasoningcan be understood in recognitional terms, as normative facts that holdindependently of the very processes by which we come to know them.Gaut, for example, maintains that only by understanding effective prac-tical rationality without “ineliminable reference to evaluative content”can a constructivist account avoid collapse into recognitionalism.62 Butthe eudaimonist account of success in practical rationality manifestly

59 Aristotle, EN VI.8.1142a30.60 For more on the idea of construction of self in this way, see my “Eudaimonist Auton-

omy,” American Philosophical Quarterly 42 (2005): 171–84.61 Aristotle, EN VI.5.62 Gaut, “The Structure of Practical Reason,” 177–78.


refers to such content (viz. eudaimonia). If Gaut is right, then, the threatof the first horn of the dilemma described above (at the end of Section IV)is especially acute just when we turn to practical wisdom to explicate thenotion of eudaimonia. In the next section, I hope to show how that hornmay be averted, and, in Section VII, I hope to show that we are notthereby plunged into relativism.

VI. The Structure of Practical Rationality

The challenge Gaut poses is the worry that, in its account of successfulpractical rationality, so-called Aristotelian constructivism is just a novelface put on an ultimately recognitionalist view. Many (like Gaut, andperhaps Kant) assume that the only way to avoid this objection is todeploy a purely formal or procedural conception of successful practicalrationality. In other words, the canons for successful exercise of practicalrationality apply purely to its form or procedure. Constructivism, then,would as a procedural approach be contrasted with recognitionalistapproaches that take practical reason to be aimed at some substantivegoal, such as the good, and to be assessable in light of that goal.63

This contrast is useful enough when it is Kantian constructivism that isunder scrutiny. However, this reflects more on the idiosyncrasies of theKantian conception of practical rationality than on its constructivism. Thehallmark of the Kantian conception of practical rationality is that sub-stantive constraints on practical rationality can be derived from correctlyspecified formal requirements. From the bare idea of a categoricalimperative —which Kant construes as the demand to avoid various typesof contradiction in willing —Kant thinks we can derive the substantiverational requirements which constitute the moral law. This twist to theKantian conception of practical rationality is much of its genius, but it isthe real target of Gaut’s charge. If the Kantian account fails (as Hegel andothers have argued) in its attempt to extract substance from form, then,just as Gaut suggests, it must covertly rely on principles that are notconstructed.

Moreover, even if it does not rely on such principles, its structure (as wehave seen) situates the justification for the procedure (and the method ofconstruction of practical truth) in substantive (and nonconstructed) a pri-ori concepts of the understanding. This is just to say, again, that on theKantian approach neither the proceduralism nor the constructivism goesall the way down. Korsgaard argues that Kant’s philosophy generally(that is, both his practical and his theoretical philosophy) should be under-stood as a search for the “unconditioned” —the stopper of a justificatoryregress that itself stands in need of no justification:

63 Cf. ibid., 163; Wedgwood, “Practical Reasoning,” 139–41; and Street, “Constructivismabout Reasons.”

200 MARK LEBAR

The claim that reason seeks the unconditioned is not based on ananalysis of the abstract concept of reason. It is more a claim about theplight of self-conscious beings who because we are self-consciousneed reasons to believe and to act. When we go looking for thosereasons we find ourselves —via a form of regress argument that isperfectly natural to any rational being —on a road that leads to theunconditioned, a road that threatens to have no satisfactory stoppingpoint.64

The threat is empty, on the Kantian conception, because we can arrive atthe unconditioned. In Kant, it is the value of the good will and ourfreedom or subjection to the moral law. In Korsgaard, it is the nature ofagency, or the experience of valuing, itself. These are each in their ownway regress-stoppers, and they afford the justificatory foundation for theenterprises of construction of normative truths that are the projects ofthese various Kantian theories. The foundation provides the warrant foraccepting the output of procedures which, when appropriately executed,produce the normative results we seek, and in that way makes the worldsafe for constructivism. The foundations themselves are not the objects ofconstruction, and it is this foundational structure which invites Gaut’schallenge. Insofar as such a foundation is necessary, the worry that thereis an incipient recognitionalism at work gains traction.

By way of contrast, the overall justificatory structure of AC is not foun-dational but coherentist. There is no normative foundation on which thewarrant for other normative claims rests; instead, the normative claims ofAC draw their justification from their membership in an overall structurethat both fits with and explains our substantive normative and theoreticaljudgments.65 The normative crux of the view rests on judgments about

64 Christine Korsgaard, “Motivation, Metaphysics, and the Value of Self,” Ethics 109 (1998):61.

65 On Scott MacDonald’s reading of Aquinas (in his essay “Foundations in Aquinas’sEthics,” elsewhere in this volume), Thomistic eudaimonism takes a “thin foundationalist”form. As I understand MacDonald’s reading, Aquinas is committed to the view that we havehappiness (or eudaimonia) given to us as an ultimate end by nature (in virtue of our desirefor it), but in only a formal and indeterminate form; and it is the task of practical reasoning(deliberation) to arrive at determinate content for that end. As I read the ancients, they agreethat in general we have such a desire, but they offer no claim that it is not, at least inprinciple, capable of being rejected as reason-giving, just as any other desire is. That is to say,we can step back even from the desire for happiness and ask whether or not we have reasonto try to satisfy it (as well as to make it determinate). Because of the reflexive nature of ourrational capacities, there is no brute motivational state that can escape this sort of justifica-tory scrutiny. In this sense, even Aquinas’s “thin” foundations are too thick for the sort ofcoherentism I am envisioning here.

One might imagine an even thinner foundationalism, in which, rather than a desire forhappiness, one had a sort of intuition as to its reason-giving nature —an intuition that wasnevertheless corrigible and defeasible, as I have insisted, but which carried a sort of primafacie justificatory force nonetheless —which then features (alone or with other intuitions) ina justificatory system for normative practical judgments. Such a foundationalism is getting


concrete, particular good human lives and about the effective exercise ofpractical rationality. In a fashion similar to the way Rawls claims we builda conception of political justice from the “materials” of antecedent con-ceptions of citizen and society, on AC truth about how to act is built on thematerials afforded by normative conceptions of a good human life andeffective practical rationality. However, unlike the Rawlsian picture, onthis picture these conceptions are themselves part of the objects of con-struction, and (unlike the Kantian picture) there is no fundamental nor-mative notion to stop a justificatory regress. In no case do we begin withnormative facts that await our recognition, prior to and independent ofour cognition of them. True claims about eudaimonia and practical wis-dom are not established prior to the judgments of wise human agents thatthey are true.66 In this sense, on AC, there is no normative truth that is notthe product of construction.67

This approach requires that AC not invoke a merely formal or proce-dural conception of practical rationality, but a substantive one. AC main-tains that practical rationality is effective when it not only satisfies theprocedural requirements commonly applied to practical reasoning, butalso delivers substantively correct judgments about how to live well. Thejudgment that I ought to X is true just insofar as it shapes my activity intoliving a good life. But we have already seen that the content of eudaimo-nia is itself constructed. Thus, the short response to Gaut’s challenge is theobservation that, while on AC practical reason does have a substantivestandard, that standard —what we take the good life to be —is itself anobject of construction.

vanishingly less distant from the coherentism I espouse, and I am unsure what rests oninsisting on categorizing either view one way or another. The key point, in my view, is thein-principle defeasibility and corrigibility of any motivational state or apprehension thatsome course of conduct is right, or of any judgment that something is good. I am uncertainwhat concerns (apart from epistemological ones) might lie beyond that point. I thank MikeHuemer for discussion of this issue.

66 In this sense, though it is true that multiple agents, equipped with the same knowledgeand equally wise, would all arrive at the same judgments for particular cases, the explana-tion for this is not that they are each discovering an antecedent fact about what practicalrationality requires. Instead, that fact just consists in the fact that practically wise agentswould see that course of action as the thing to do. There would be no such fact exceptthrough the deliberations of practically wise agents. I thank Christopher Gowans for press-ing this point.

67 Tom Hill has suggested (in private conversation) that the regress-stoppers in Kant’saccount, anyway, might well be seen as the objects of construction. Perhaps a similar casemight be made for Rawls’s and Korsgaard’s accounts too. It is striking, however, that innone of these theories is there any account of that constructive process. Certainly it cannotbe the same sort of constructive procedure that yields normative truth on those accounts.However, my aim here is less to attack the various Kantian positions than to show how oneinviting reading of their structure lays their constructivism open to Gaut’s objection. If theirstructure is, implicitly, more like the structure of the Aristotelian picture I give here, all thebetter for them. The point remains: the coherentist structure is the sort of structure thatenables this particular recognitionalist objection to be avoided, and it is central to theAristotelian picture.

202 MARK LEBAR

By way of a somewhat fuller sketch: on AC, successful practical ratio-nality involves both successful reasoning (in the sense of marking theproper inferential relations among beliefs and other attitudes) and mak-ing substantively right choices about how best to live. First, AC can endorsethe procedural requirements on practical reasoning accepted not only byKant but (for example) by formal theories of rational choice. Any formalor procedural constraint on practical reasoning which we take, on reflec-tion, to be a genuine canon of effective rationality is one that will countas such a canon on AC. So AC can embrace canons not only of deductiveinference (e.g., consistency) but also of less formal inferential patterns:coherence more broadly construed, canons of successful explanation, andthe like. The warrant for accepting these canons is itself the product of thevery faculties to which the canons apply (what else could do the war-ranting?). In this sense, AC requires that practical wisdom “bear its ownsurvey” (to put the point in David Hume’s way).68

The sticky part of the story, obviously, is what it says about the sub-stantive requirements on practical wisdom. These may not be identifiedmerely as the results of the operation of some purely procedural test. Itmay be helpful to begin by thinking of AC as endorsing something like anObjective List of things that normally belong in a good human life. It isimportant to note both the ways in which this picture is accurate, and theways in which it differs from a recognitionalist way of conceiving sucha list.

The picture is accurate in according at least prima facie warrant to justthose judgments we make about people who do not take the items typ-ically on such a list (e.g., their own health) to be reason-giving; we takesuch people to be irrational or at least open to rational criticism. Suchjudgments are part of everyday life and the common-sense exercise ofwisdom, and AC endorses such judgments, at least as a starting-point. Itis difficult to avoid judgments that health, friendship, knowledge, and soon —the items usually found on an Objective List —typically have such aplace in good human lives, and AC embraces them.69 But it differs fromrecognitionalism in its account of how things come to be on that list.Gaut maintains that the proper way to understand such a list is on therecognitional model: the contents of the list itself are there to be recog-

68 David Hume, A Treatise of Human Nature (Oxford: Clarendon Press, 1978), 620.69 This way of putting things breezes past a considerable complication for eudaimonism,

taking as it does a “formally egoistic” form: Can the view give the right account of the waythat other persons afford us reasons for acting? There are, I believe, two distinct sub-questionshere, one pertaining to (roughly) the welfare or well-being of others (perhaps: the demandsof beneficence), and the other pertaining to something like their rights and the correspond-ing obligations of respect we owe them. Both are important issues, but I cannot take themup here. I address the latter question explicitly in “Virtue Ethics and Deontic Constraints”(manuscript). On the former question, see Julia Annas, “The Good Life and the Good Livesof Others,” Social Philosophy and Policy 9, no. 2 (1992): 133–48. I thank Tibor Machan forpressing this point.


nized, prior to our conception of practical rationality, and thus can affordus a canon for assessing which actions are rational and which are not.70

On AC, however, membership on that list is itself the upshot of effectivepractical rationality. We can, and do, reflect on putative judgments aboutthe constituents of a good life, and it is success in such reflection thatwarrants our substantive judgments about living well. In this sense, full(as opposed to prima facie) warrant for something’s being on that list isdelivered only by broad reflection on living well. A claim that somethingbelongs on that list is established by vindicating to the judgment of prac-tical wisdom that it is an element in a good human life; hence, somethingis on that list just because it is judged to be. This is the reverse of the orderof explanation the recognitionalist offers.

Aristotle himself holds that virtuous action (or wise judgment) alwaysaims at the “fine” or “noble” —to kalon.71 The Greek idea here includesboth the ethical force of “noble” and the aesthetic element of being “fine”(or “beautiful”). As Aristotle deploys the idea, both of these elements, aspredicates of ethical action, track the degree to which right action is fittingor appropriate (to prepon).72 We make judgments of fittingness concerningboth our own lives (How do I strike a balance between time spent onwork and time spent with my family?) and the lives of others (Does thisepisode of plagiarism deserve failure of the course, or a report to Uni-versity Judiciaries?). We make these judgments on large-grained ques-tions (Am I a good fit for this job?) and on more finely-grained issues(Should I wear a tie to this party?). We make them in theorizing (“This isthe better explanation”) and in practice (“That was uncalled for”). Sincejudgments of fittingness must be highly contextually dependent, theycannot be codified, and consequently cannot be seen as produced orwarranted by any purely procedural rationality. They are substantive —this just is fitting or appropriate, and that just is not —but they are nottasks of recognition of some property of appropriateness in their objects.Instead, they are products of the constructive task of piecing together theright contributory elements to living a good life. That is the major workof practical wisdom on the Aristotelian conception.

This proposal is quite sketchy, but I hope it suggests the picture I havein mind. It is an outline of a different conception of practical rationalitythan the procedural accounts often on offer. Whereas on the Kantian story,substantive practical and moral truth is supposed to emerge from a pureprocedure, the Aristotelian story countenances and embraces substantivejudgments and commitments to ends we must have to be practicallyrational. This is the reply to the concern that AC has shoved recognitional-

70 Gaut, “The Structure of Practical Reason,” 183.71 See, e.g., Aristotle, EN III.7.1115b12–13; III.8.1117a8; III.11.1119b15–16; and many others.

This is a dominant theme in Aristotle’s account of virtue and virtuous action.72 See Kelly Rogers, “Aristotle’s Conception of To Kalon,” Ancient Philosophy 13 (1993):

355–71.

204 MARK LEBAR

ism out the front door only to usher it in at the back. I have explained thatthis is not the case, that even the normative foundations (such as they are)of Aristotelian constructivism are themselves products of construction.But that invites just the opposite form of concern: that constructivismallows, or perhaps even just is, a noxious form of relativism. To thatconcern I now turn.

VII. Objectivity, Subjectivity, and Relativism

Suppose AC is, in fact, constructivist “all the way down.” That runs theaccount straight into the objection that it has so little structure that it canbe used to justify any number of sets of moral principles, each groundedin some agent- or culture-relative set of assumptions and judgments, withno basis for critical engagement among such constructed sets.73 The chal-lenge is to show that AC can vindicate a claim to sufficient objectivity tolay these concerns aside.

This is not a novel objection to neo-Aristotelian theories, and the basisfor concern is not entirely the constructivism offered in the present ac-count. The view that the judgment of the wise and virtuous agent —thephronimos —is even epistemically indispensable in picking out virtuous atti-tudes and actions is itself sufficient to raise many of the same questionsabout the objective credentials of the virtue approach. Much of the virtueethics literature on (for example) the impossibility of codification of moraljudgments, the impossibility of moral rules, and the like, is aimed atmeeting this challenge.74 AC goes beyond the epistemic claims to makesuch judgments constitutive of what is required by successful practical

73 What this horn of the dilemma looks like in the formulations of Shafer-Landau andTimmons is not clear. Both think the dilemma is between characterizing the constraints onconstruction in moral terms and eschewing the use of such terms. On the latter horn (whatTimmons calls “thin” characterizations), Timmons takes the problem to be indeterminacyamong the principles constructed, while Shafer-Landau is concerned that the resultingconstruction may not even be recognizable as a set of moral principles. On the former horn(what Timmons calls “thick” characterizations), Shafer-Landau holds that the approach isno longer constructivist, since there are moral principles at work prior to the construction,while Timmons worries about “conceptual chauvinism” and “relativism,” in that differentsets of moral assumptions will yield different constructed outputs. Both of these formula-tions seem to me to assume a foundationalist structure to the enterprise of construction, soneither quite fits the structure of the account I have set out here. I hope I have met Shafer-Landau’s argument that the approach has the problems of the “thick” characterization; thisleaves the concern that the result is no longer recognizable as moral. As I observe in the text,this is a problem for any ancient ethical account, and fully meeting this charge is possibleonly in a successful defense of a normative, eudaimonist, virtue ethic. Yet the success of sucha broad normative theory would set Timmons’s charge of indeterminacy to rest, only to runinto his worry about relativism; a similar concern is formulated by Cullity and Gaut (“Intro-duction,” in Cullity and Gaut, eds., Ethics and Practical Reason, 16). My argument here isdirected at the concerns about relativism, though I hope it speaks to Shafer-Landau’s wor-ries as well.

74 In particular, Rosalind Hursthouse considers these points at length in On Virtue Ethics;see esp. chaps. 1, 8–11.


rationality, so in some respects its problem is more grave.75 My strategywill be to consider what forms of objectivity we might want, to suggestthat the desire for some of these forms is unwarranted, and to argue thatAC provides a form of objectivity worth having.

Bernard Williams has argued that questions about objectivity in ethicaldiscourse boil down to what we will say about ethical disagreement —inparticular, the possibility or necessity of at least one party being in error.76

One way to understand the present concern is in this light. For example,there are contrasting beliefs about the moral permissibility of female gen-ital mutilation (“female circumcision”); part of what we want by way ofobjectivity is a basis for claiming that those who think this practice is per-missible have wrong or wicked attitudes, not just different ones, and fordefending the thought that such a claim is not merely “chauvinistic,” 77 sothat when we disagree with proponents of such a practice we are not merelytalking past them. Given the patent objective purport of claims about bothmorality and practical rationality more generally, Williams is surely rightthat providing such a basis is a necessary condition for a plausible moraltheory. It might be, of course, that no such objective standard can be vin-dicated, but it will certainly count as an indictment against the objectivityof an approach that it cannot even propose such a standard.

Here, Kantian constructivism has a card to play that AC does not.Consider Kant’s aspiration to provide a moral philosophy applicable toall rational beings, considered purely as such. If successful, this aspirationarguably could establish an objectivity drawn from rational necessity.Moral laws, Kant says, “should be taken from pure reason”: we are re-quired, “since moral laws are to be valid for every rational being ingeneral, to derive them from the universal concept of a rational being ingeneral. . . .” 78 If he is right, moral and (in general) practically rationalrequirements could lay claim to an objectivity which anything known tobe true a priori —because necessarily true —must have. Kant virtuallyequates the objective with what remains after abstraction of everythingempirical,79 and at the same time holds that this amounts to the universalvalidity of practical laws.

75 Nevertheless, it is Aristotle’s problem as well: “Excellence, then, is a state concernedwith choice . . . this being determined by reason and in the way in which the man of practicalwisdom would determine it” (EN II.6.1107a1–2).

76 Bernard Williams, “Saint-Just’s Illusion,” in Williams, Making Sense of Humanity (Cam-bridge: Cambridge University Press, 1995), 145. This passage is central to Rosalind Hurst-house’s discussion of objectivity in virtue theory (On Virtue Ethics, chap. 11), to which I amdeeply indebted.

77 Timmons calls “chauvinistic” those versions of relativism which maintain that moralconcepts are such that “where two individuals or groups really do seem to be thinking oruttering contradictory judgments employing those terms and concepts, the judgments inquestion are not really contradictory at all” (Timmons, “The Limits of Moral Constructiv-ism,” 406).

78 Kant, Groundwork, Ak. pp. 411–12.79 Ibid., Ak. p. 427.

206 MARK LEBAR

If we think ethical objectivity requires necessity, then Kant’s approachmay be the only game in town, whether it succeeds or not. However,surely it is odd to think that the most basic truths about how we ought tolive our lives must be norms shared with creatures as different from us asother rational beings —considered purely as such —could be. Could itreally be that our practical requirements are the same as those would befor creatures that were not mortal? That the principles governing ourinteractions with others are the same as those that would govern crea-tures who did not live in characteristic scarcity? Who were not vulnerableto each other? Who reproduced asexually or not viviparously? Thoughcontingent, it is hard to see how the facts about how we differ from suchcreatures could not be deeply significant for the content of our practicalprinciples.80 As McDowell says, the thought that such contingencies neednot be significant is a “consoling myth” in coming to grips with our needfor objectivity,81 and there is a history of worries that a priori principlescannot generate the necessary traction for contingent circumstances.

In contrast, if the contingency of our natures and circumstances neednot preclude the objectivity of prescriptions for action, then we are inposition to seek objectivity in something about our natures, contingentthough they may be. Like its Kantian cousins, the Aristotelian approachfocuses on our natures, not as conative subjects, but as rational agents;however, unlike the Kantian approach, it focuses on that rationality, notas abstracted from, but rather as immanent in, our animal natures.82 Wecan look to find objectivity in what is called the “publicity of reasons” —the notion that reasons are open to a public discipline, that they are notconstrained merely by private and individual attitudes about them.83 Oursusceptibility to reasons, and (in turn) their publicity, grounds the hope ofidentifying a standard of correctness in the exercise of practical rational-ity, and thus a plausible form of objectivity in AC.

This way of establishing a standard has its roots deep in Aristotle. The“mark to which the man who possesses reason looks,” he says, is rightreason (orthos logos).84 Unlike Kantian constructivism, AC does not main-

80 An excellent exploration of some of these contingencies, and their bearing on ourethical life and practice, may be found in Rosalind Hursthouse, Beginning Lives (Oxford:Blackwell, 1987), 247–59.

81 John McDowell, “Virtue and Reason,” Monist 62 (1979): 339. See also his “Non-Cognitivism and Rule-Following,” in Steven H. Holtzman and Christopher M. Leich, eds.,Wittgenstein: To Follow a Rule (London: Routledge and Kegan Paul, 1981), 155, where hecriticizes the thought that, if rationality is present, it should be recognizable from an “exter-nal standpoint.”

82 The fact that we have reflective practical rationality, and characteristically live bydeploying it, is a natural fact about us, and thus is unobjectionable as one of the “materials”to be used in a constructivism that goes “all the way down,” as I claim. I thank Dan Laymanfor raising this issue.

83 Korsgaard has explored this idea at great length in a number of places, including TheSources of Normativity, lecture 4. I shall indicate the ways in which I understand this “pub-licity” differently than she does.

84 Aristotle, EN VI.1.1138b25.


tain that this mark can be purely procedural: practical reason must satisfysubstantive constraints as well. But because those constraints are subjectto the discipline of reason, the standards in question are objective. Thenormative nature of the enterprise of giving and exchanging reasons fortaking a given action or way of living as being conducive to living wellboth assumes and establishes the standard of correctness for such claims.

On AC, these reasons will depend on truths about living good humanlives; reasons for engaging in conduct of one sort or another will comedown to claims about the conduciveness of doing so to living well —claims which are subject to an interpersonal discipline. This disciplinewill be partly procedural: as I have said, AC accepts just those canons forinferences that will be endorsed by any plausible view of rationality. Butthe justification for particular claims about reasons will depend in addi-tion on substantive judgments as to which putative reasons really arereasons, and which are not. In a crucial sense, these candidate claims aresubject to a form of public vindication.

Now, Korsgaard offers a conception of the publicity of reasons in herdevelopment of Kantian constructivism, but the picture of the publicity ofreasons on the Aristotelian approach differs from Korsgaard’s. Her aim inestablishing her Kantian account of publicity is to argue that there are no“private reasons” —reasons of the sort that would provide a toehold forrational egoism and, in consequence, a basis for asking for a justificationfor claims about reasons to respect others. Korsgaard maintains that thisapproach to reasons is wrongheaded. Because reasons are public by theirvery nature, she claims, they must be shareable.85 Since this feature ofreasons is supposed to show that rational egoism is impossible, the per-tinent sense in which reasons are shareable —hence public —must requirethat they are in fact shared.86

The sense in which reasons are shareable and thus public (and thussuitable to ground claims to objectivity) on AC differs from Korsgaard’ssense in two ways. First, AC distinguishes between reason-tokens andreason-types. Our reasons fall into indefinitely many types: reasons forprudence, reasons for concern, reasons to come in out of the rain, and soon. Reason-tokens are the particular reasons that fall into these types. Onetype of reason, for example, may be the reason not to drink to excess. Buteach of us has a different reason-token in that type. Your reason for notdrinking to excess is that it is bad for your health, whereas mine is thatdoing so is bad for my health. In a recognizable sense, our reasons aresimilar, though their bases are distinct, and they are reasons for different

85 Korsgaard, The Sources of Normativity, 135.86 However, at some points, Korsgaard indicates that by saying that reasons are shareable

she means merely to deny the claim that they cannot be shared (e.g., The Sources of Norma-tivity, 135, 141). I argue that this ambiguity is fatal to her account of reasons for respectingothers in my essay “Korsgaard, Wittgenstein, and the Mafioso,” Southern Journal of Philos-ophy 39 (2001): 261–71.

208 MARK LEBAR

conduct by different people. The similarity captures the type, while thedifferences are reflected in the tokens.

Korsgaard does not draw this distinction, and her claims are aboutreason-tokens; but for AC the publicity of reasons means they are share-able in the sense that their types are in fact shared. On a eudaimonist view,each of us has most reason to make our own life go well. These reason-tokens —reasons for making a particular, contingently-situated humanlife go well —are not necessarily shared, and thus are not shareable in thesense Korsgaard requires. It is not a requirement on your having reasonto make your life go well that I must see myself too as having reason tomake your life go well, as Korsgaard maintains. Your reasons are, indeed,shareable in the ordinary and unremarkable sense that I can, in variousways, come to have reason to make your life go well (and thus to takeyour reasons to be mine as well), and you may well have reason toreciprocate. AC maintains, then, that reason-tokens can be shared, butwhat makes them reasons is their membership in reason-types that areshared; most fundamentally, the relevant type is reasons to live well, butthere are indefinitely many other reason-types as well.87

What makes a putative reason-token a token of a shared reason-type?The second difference from Korsgaard’s view is AC’s requirement on suchtypes that they be intelligible.88 It is difficult to get a precise grip on justwhat this standard comes to, but that it is a condition on reasons cannotbe doubted. Failure of intelligibility is a reason to reject putative rea-sons.89 Elizabeth Anscombe cites the unintelligibility of the reply “To getmy camera” to the question “Why are you going upstairs?” — in the face ofthe explicit acknowledgment that the camera is in the basement —as anexample of this sort of failure.90 Anscombe’s diagnosis is that such a caserepresents a failure correctly to pick out an intention as a reason for action,and the conclusion she draws as a result is that “[a] man’s intention in

87 Access to shared reason-types and processes of reasoning is the counterpart in thisview, I believe, to the way that Kant thinks we can grasp the nature of practical rationalityin a completely abstract way, so as to arrive at a “universal concept of a rational being ingeneral” (Kant, Groundwork, Ak. p. 412). On what I am calling the Aristotelian approach, wehave no confidence that we can apprehend what such rationality might consist in. The onlyrationality we know is embodied and shaped by a variety of contingencies. We can distin-guish it in kind from our experience of the causal order, but we have no grasp on itindependently of our cognition within that order. I thank Tom Hill for pressing this point.

88 Cf. McDowell: “[I]t is our common human nature that limits what we can find intel-ligible in the way of theses about how human beings should conduct their lives. . . .” JohnMcDowell, “The Role of Eudaimonia in Aristotle’s Ethics,” in Essays on Aristotle’s Ethics, ed.Amélie Rorty (Berkeley: University of California Press, 1980), 371.

89 Intelligibility cannot plausibly be thought of as a sufficient condition on reasons. We canmake perfect sense of a wide range of things people (including ourselves) do out of errorsin judging what they have reason to do, or even in the absence of reasons at all. RosalindHursthouse characterizes actions of the latter sort in her essay, “Arational Actions,” Journalof Philosophy 88 (1991): 57–68. The point I would make here is that it is in virtue of theintelligibility of such actions that we classify them as actions (as the expression of intentionsof agents like ourselves) at all.

90 G. E. M. Anscombe, Intention (Ithaca, NY: Cornell University Press, 1957), 36.


acting is not so private and interior a thing that he has absolute authorityin saying what it is —as he has absolute authority in saying what hedreamt.” 91 While our dreams are private, the reasons to which our inten-tions respond are not. These two elements of AC’s picture of the publicityof reasons are connected. Part of what makes intelligible the reason-tokens that we each derive from our interest in living well is the fact thatthe type is shared. It is just because I can so readily grasp the reason-givingforce of my own eudaimonia that I can appreciate the reason-giving forcefor you of yours.

AC agrees with Korsgaard that the notion that reasons are public isessential to the claim that there are standards of correctness for practicalreason and thus to the case that judgments about what one has reason todo (or not to do) may claim a form of objectivity.92 AC disagrees withKorsgaard in holding that no purely procedural set of canons for suchjudgments will suffice, and that substantive judgments —in particular,judgments about what we have reason to do and seek in living a good lifeas a human being —are necessarily part of those standards.

This is not to deny that there are deep disagreements on such judg-ments. Mark Timmons argues that a serious problem for moral construc-tivist views is presented by cases of “moral symmetry.” An example is acase, offered by Hilary Putnam, of deep evaluative disagreement with hiscolleague Robert Nozick. Here is Putnam’s account of the disagreement:

One of my colleagues is a well-known advocate of the view that allgovernment spending on “welfare” is morally impermissible. On hisview, even the public school system is morally wrong. If the publicschool system were abolished, along with the compulsory educationlaw (which, I believe, he also regards as an impermissible govern-ment interference with individual liberty), then the poorer familiescould not afford to send their children to school and would opt forletting the children grow up illiterate; but this, on his view, is aproblem to be solved by private charity. If people would not becharitable enough to prevent mass illiteracy (or mass starvation ofold people, etc.) that is very bad, but it does not legitimize govern-ment action.

In my view, his fundamental premises —the absoluteness of theright to property, for example —are counterintuitive and not sup-

91 Ibid. (emphasis in the original).92 From the standpoint of any individual agent, the publicity of reasons entails that he or

she may be mistaken about considerations taken to be reasons. That, in turn, opens thepossibility that the acquisition of wisdom can be a procedure of discovery for particularagents. The constructive enterprise establishes norms governing eudaimonia and practicalwisdom, as a matter of a shared and public enterprise of the exercise of reflective practicalrationality. But from the first-person standpoint of a particular deliberating agent, thisenterprise of construction and the discovery of reasons are indistinguishable. I thank EricMack for pressing this point.

210 MARK LEBAR

ported by sufficient argument. On his view I am in the grip of a“paternalistic” philosophy which he regards as insensitive to indi-vidual rights.93

Such disagreements are, of course, not unfamiliar. Putnam uses this exam-ple to remark on what sorts of attitudes toward one’s interlocutor occurin such disagreements. In this case, he says, the disagreement results in anattitude of contempt: neither for Nozick’s mind nor for his character norfor him as a person, but for a “certain complex of emotions and judg-ments” in him.94 This contempt (which Putnam takes to be reciprocal)involves a judgment that the other is, in a certain way, failing to besensitive to reasons —not just any reasons, apparently, but reasons towhich it is a kind of moral failing to be insensitive.

The point Timmons draws from the case is quite different: it is a text-book example of the problem of moral symmetry for constructivism. Thefeatures he takes the case to exemplify, which make it problematic, are thefollowing:

(1) [T]wo individuals or groups are engaged in a moral disagreementover some issue or case, (2) because they are placed in circumstancesthat are as similar as possible, their disagreement is not explicableowing to [differences in application of universal principles], (3) theindividuals or groups are plausibly interpreted as making no factualerrors in relation to their moral judgment about the case at hand, (4)their respective moral outlooks enjoy wide reflective equilibrium,and hence (5) their respective moral views on the topic in questionare stable.95

The stability of these views under this sort of wide equilibrium, Timmonsbelieves, compels constructivists either to see the interlocutors as simplytalking past each other (in only apparent disagreement) or to construeboth views as true for the respective sensibilities involved; either way,constructivism is committed to an objectionable relativism.

However, I think AC makes perfect sense of disagreements of this sort,and accounts for just the attitudes Putnam cites as flowing from them. IfAC is right, there is no reason to accept that the Putnam-Nozick caseexemplifies Timmons’s features (3), (4), or (5). Cast in a eudaimonist light,part of the disagreement is due to the differences between Putnam’s andNozick’s conceptions of what human well-being involves, and in partic-ular their conceptions of the roles of governmental as opposed to other

93 Hilary Putnam, Reason, Truth, and History (New York: Cambridge University Press,1981), 164 (emphasis in original).

94 Ibid., 165.95 Timmons, “The Limits of Moral Constructivism,” 412–13.


social institutions for the provision of education.96 But the facts —both thepurely empirical facts, and the way these facts fit with other facts inconstituting good lives and good relations between persons as parts ofthose lives —are exceedingly complex. These facts involve substantivejudgments about what we have reason to do, and AC need not accept thateither party has gotten these judgments right.

Furthermore, the way in which AC takes reasons to be public meansthat the mere fact of this sort of disagreement must result in a degree ofdisequilibrium. Neither Nozick nor Putnam would be justified in seeinghis own position as beyond reasonable doubt; by hypothesis, each isperfectly aware that there is reasonable doubt about his position. Thatdoes not mean that Nozick and Putnam may hold their positions onlytentatively; given their best assessments of things, they see the conclu-sions they have reached as irresistible. Still, a considerable degree ofepistemic humility is appropriate. Just as Anscombe suggested, claimsabout what one has most reason to do are always corrigible. If AC is true,then in the dispute between Putnam and Nozick, at most one of them isright; perhaps both are wrong and there is some third position that rep-resents the facts about how human beings may best organize themselvesso that they educate their young in a way most conducive to humanflourishing. The present point is that both Putnam and Nozick see thosefacts as objective, and the contempt they feel for the pattern of responsesin the other is a reflection of that presumption of objectivity. Those atti-tudes are perfectly consistent with the claims of AC. Moreover, AC canexplain why the disagreement is so deep: it is just because each sees thereasons for his own view as so deeply tied to what is necessary for humanwelfare that each has the attitude he does toward the other’s insensitivity.But since on AC there is no reason to think that either Putnam or Nozickis in possession of the facts as practical wisdom would see them, AC is notcommitted to the pernicious relativism that Timmons deplores. AC nei-ther sees them as talking past each other, nor maintains that each may befully vindicated by his own lights. Instead, it holds that there is a standardof correctness to be applied —the truth about what conduces to goodhuman lives —and that that standard can be vindicated only by the bestexercise of reason and judgment of which we are capable.

VIII. Conclusion

The promise of constructivism about practical rationality is the hopeit holds of avoiding both the metaphysical and epistemological prob-lems of recognitionalism and the lack of fidelity to the experience ofresponding to reasons which is characteristic of views that doubt thatthere is practical truth, or that our practical judgments aspire to track a

96 Presumably, they do not disagree over the role of education in a good human life.

212 MARK LEBAR

normative reality. The Kantian version of the constructivist enterpriselives or dies on its ability to precipitate substantive reasons for actionfrom a conception of rationality as procedural and suitable for rationalbeings considered purely as such. The viability of Aristotelian construc-tivism, in contrast, depends on there being truth about how to live well,and on that truth’s being something that is constructed by human judg-ments about what a good life is. Such truth (if it exists) would be the mosthard-won of human intellectual and moral accomplishments, and my aimhere has not been to set it out. Instead, in the framework of practical truthconstructed from a substantive conception of successful practical ratio-nality, directed at the goal of living good human lives, I hope to havesketched an approach appropriate for understanding objectively wisepractical rationality and objectively good human lives.

Philosophy, Ohio University


MORAL CONSTRUCTION AS A TASK:SOURCES AND LIMITS

By Thomas E. Hill, Jr.

I. Preliminary Remarks, Aims, and Preview

Many different questions have been debated under the rubrics of“objectivity” and “relativity” of morals. My plan is to begin by distin-guishing several of these questions as background discussion. Then,after a few remarks about the earlier questions, I sketch a broadlyKantian position on the last two of these questions: First, how, if atall, can we derive, justify, or support specific moral principles andjudgments from more fundamental moral standards and values? Sec-ond, how, if at all, can (alleged) fundamental standards be defended?Famously, Kant attempted to provide nonskeptical answers to both ques-tions. His explicit answers and arguments have often been criticized,and there are probably few today who accept them all without quali-fication. Many contemporary philosophers, however, have been influ-enced by his ideas about the nature and structure of moral thinking,his attempt to articulate a supreme moral standard, and his strategy fordefending moral principles. My sketch here of a broadly Kantian per-spective represents my attempt to develop and extend some of theseideas beyond Kant’s texts.1 The influence of John Rawls’s theory ofjustice may also be evident.

To preview, the main features of my broadly Kantian position on ourtwo main questions are these. First, Kant’s later formulations of the Cat-egorical Imperative provide important elements of a deliberative proce-dure that can be used to guide and constrain our thinking about morespecific moral principles in a way that helps to correct for mere personal

1 By calling a perspective “broadly Kantian” I mean to acknowledge that it draws heavilyfrom Kant’s ethical writings but does not follow Kant’s texts in all respects. The broadlyKantian perspective that I sketch here and elsewhere modifies and supplements Kant’stheory in various ways, one of which is its attempt to disassociate normative aspects of thetheory from Kant’s theory of noumenal freedom (as this is usually understood). I give afuller (though still incomplete) description of the broadly Kantian perspective I have inmind in other essays, especially Thomas E. Hill, Jr., Respect, Pluralism, and Justice: KantianPerspectives (Oxford: Oxford University Press, 2000), chaps. 2, 4, and 8; Hill, Human Welfareand Moral Worth: Kantian Perspectives (Oxford: Oxford University Press, 2002), chap. 3; andHill, “Assessing Moral Rules: Utilitarian and Kantian Perspectives,” Normativity, Philosoph-ical Issues (A Supplement to Noûs), vol. 15, edited by Ernest Sosa and Enrique Villanueva(2005): 158–78.


preference and parochial bias.2 The procedure is far less than a determi-nate decision procedure or strict “derivation” of particular moral conclu-sions, but it promises a way of supporting judgments that initially mayhave appeared more doubtful. Second, Kant has two strategies for defend-ing the ideas in his various formulations of the Categorical Imperative.The first has merit, at least when modified and presented more modestlyas akin in some respects to Rawls’s method of reflective equilibrium. Thebackground of this strategy, too frequently ignored, is Kant’s principledrejection of previous errors in foundational moral theory. Also, thoughnot a major focus here, Kant’s second line of defense, at least in its finalmodified form, may still serve its purpose in practical, rather than meta-physical, reflection and dialogue. If what is questioned is the rationalauthority of the basic standards identified by the previous method, thebest we can do, and often all that is needed, may be to appeal to deeppervasive commitments and consistency in the practice and serious think-ing of each rational moral agent.

Words such as “construction,” “objectivity,” and “relativity” are at bestblunt instruments for discussion, and by now they have been used in somany ways they may have outlived their usefulness. Nevertheless, toconnect my discussion with some recent literature, I propose that thebroadly Kantian perspective sketched here leaves us with two ways inwhich moral construction is an ongoing task. First, and most obviously,defining the broadly Kantian deliberative perspective is an incomplete work inprogress —an attempt to articulate for practical purposes an appropriateframework for deliberating about the interpretation and limits of specificmoral principles (or “rules”). This construction project starts from theassumption that Kant’s own efforts to articulate a comprehensive suprememoral principle were inadequate for the purpose but contain important,promising elements. The aim would be to put these Kantian elementstogether, modifying and supplementing them as needed in the light ofarguments, to build an account of an appropriate deliberative perspectiveor stance for addressing more specific moral issues. The challenge wouldbe to identify and defend certain more basic values as constituting theframework for deliberation about further, often more controversial issues,for example, the interpretation and limits of rules about lying, preemptivekilling, and torture. To continue the metaphor of construction, this sub-stantive deliberation about more specific principles can be seen as a second,more practical construction project. In any broadly Kantian view, contro-versies about the interpretation of (and exceptions to) specific moral rulescannot be resolved by science, metaphysics, theology, or appeal to anyauthority based merely on power, tradition, or convention. Instead, such

2 Immanuel Kant, Groundwork for the Metaphysics of Morals, trans. Arnulf Zweig, ed.Thomas E. Hill, Jr., and Arnulf Zweig (Oxford: Oxford University Press, 2002), 222–37[4:421–38]. Bracketed numbers in all citations of Kant’s works refer to volume and pagenumbers in the standard Prussian Academy edition.

MORAL CONSTRUCTION AS A TASK 215

principles and their limits must, in a sense, be “constructed” in reflectionfrom an appropriate moral point of view by the moral agents who mustbe subject to them.3

My remaining discussion will be divided into the following three mainsections: In Section II, as background, I distinguish several different ques-tions that arise in debates about moral relativism and objectivity. Then, inSection III, I sketch my broadly Kantian perspective for assessing specificmoral principles. Finally, in Section IV, I describe briefly Kant’s groundsfor rejecting previous moral theories and his two strategies for defendinga proposed deliberative perspective.

II. Different Questions Regarding Relativismand Objectivity

Some worry and others happily accept that moral principles and judg-ments may be “relative” and “subjective,” as opposed to “absolute” and“objective.” These terms, unfortunately, have been understood in so manydifferent ways that they are now more often confusing than helpful towardunderstanding deep philosophical concerns.4 At least their use in eachcontext needs to be explained and supplemented with discussion in otherterms. So I shall try to distinguish several questions without relying on the

3 The idea that moral principles can be constructed is a metaphor that can be illuminatingin some respects but is also liable to be misleading. For example, the metaphor is intendedto suggest that in order to reach reasonable conclusions about what moral principles requireand what exceptions they allow, what is needed for practical purposes is not acute “per-ception” of natural or nonnatural metaphysical facts but rather reflection and discussionabout what standards we have most reason to endorse in the light of certain basic moralvalues, norms of rationality, and relevant facts about the human condition and the morelocal conditions in which the principles are to be applied. The metaphor of constructionwould be misleading if it suggested that we may construct whatever principles we pleaseas if we were children given a pile of miscellaneous materials with which to build somethingbut without any given blueprint, purpose, or standard for construction. When I suggest thatwe may think of substantive principles (for example, about lying, stealing, and killing) asconstructed (or to be constructed) from a broadly Kantian deliberative perspective, I do notmean to imply that they are (or are to be) constructed from morally neutral or value-freeassumptions or from values that are themselves constructed in the same sense. In saying, inaddition, that we may think of articulating the broadly Kantian deliberative perspective as a taskfor construction, I mean to suggest only that it remains an unfinished task in moral theory towork out and express as clearly as possible a description of the most basic constraints andguiding values that are reasonable, and appropriately called (broadly) “Kantian,” as aframework for assessing more substantive and derivative moral principles. When under-stood this way, construction of substantive principles need not be redundant or viciouslycircular, and construction of the deliberative basic perspective does not necessarily presup-pose any particular metaethical theory. A more radical and controversial claim would bethat all moral values and principles, including those treated as basic in a broadly Kantiandeliberative perspective, are constructed in some further sense.

4 Philosophers have often noted important distinctions among the questions relevant todiscussions of relativism and objectivity. An excellent example is Richard Brandt, EthicalTheory (Englewood Cliffs, NJ: Prentice Hall, 1959), chap. 11.

216 THOMAS E. HILL, JR.

terms “relative,” “absolute,” “subjective,” and “objective,” even though Iwill use some of these terms as temporary labels for ideas that I explain.

A. Linguistic relativity and related deeper disputes

Some words are incomplete without explicit or implicit reference tosomething further. An obvious example is “relevant.” Back in the 1960s,students often clamored for new courses that were “relevant” but withoutspecifying what they were supposed to be relevant to. Some no doubt hadthe general idea that new courses should be relevant to their interests instopping the war in Vietnam, reforming or destroying the existing gov-ernment, living untraditional lifestyles, etc., but others seemed to thinkthat the word “relevant” by itself expressed a complete idea. More com-monly, people often talk about various things being important withoutsaying to whom or for what they are significant. No doubt many couldpause, if asked, and fill in the assumed or neglected qualification, but toooften conversations proceed as if no further comment is required, somethings just being important period, not for any specifiable purpose andnot to you, or to me, or even to all human beings, to all rational beings,to all living things, etc. In these cases, we naturally wonder what thestandards for evaluating such claims are and whether, after all, the claimsmake sense.

Thomas Hobbes says that “good” is always “relative to the person,”usually relative to the person that uses the word but relative to the sov-ereign authority in a state when the sovereign makes a ruling about whatis to be regarded good.5 At least part of the point seems to be that theword “good,” like “relative” and “important,” is incomplete by itself, andto make sense needs to be supplemented, for example, by specifying to orfor whom, or from whose perspective, something is alleged to be good.We might add a variety of other possible supplements, such as “good for. . . (a purpose),” “a good . . . (kind of thing),” or “good from the point ofview of . . . (a group, project, or institution with shared aims, etc.).” Sim-ilar things might be said of “valuable,” “desirable,” “ideal,” and othervalue terms (including negative terms such as “bad,” “undesirable,”“repugnant,” etc.). If we want a label, we might use linguistic relativity forthis idea that familiar value terms, such as “good,” are incomplete with-out further specification.

Underlying any dispute about the linguistic point, however, is likely tobe a deeper disagreement in moral theory or metaphysics. Plato held thatthere is a Form of Goodness that exists independently of existing earthlybeings, their purposes, and their natural kinds. Hobbes made his linguis-

5 Thomas Hobbes, Leviathan, ed. Edwin Curley (Indianapolis, IN: Hackett Publishing Co.,1994), part I, chap. 6, pp. 28–29. This famous passage in Hobbes is open to several differentinterpretations. My concern here is with the idea I draw from Hobbes, which may or maynot be exactly what Hobbes meant.


tic point in a way that also expressed his antipathy to a Platonic concep-tion of goodness. Unlike Plato, he thought that we call things good becausewe desire them, not that we desire them because (independently) they aregood. Like Plato, G. E. Moore also implied that various things are good inthemselves, or good period, as opposed to good for something, good tosomeone, good of a kind, etc. His underlying point was not merely lin-guistic but a controversial metaphysical thesis about the existence of good-ness as a real nonrelational, nonnatural property.6 Disputes about this gobeyond linguistics and more properly fall under later questions aboutbasic standards and justification.

B. Causal dependence on culture and history

When discussing whether morality is “relative,” one often hears theclaim that moral beliefs are determined by cultural and historical factorsthat vary widely. This is an empirical thesis about the causes of moralbeliefs. Its relevance to disputes about the truth or validity of universalmoral principles is less clear than is often supposed. There are two mainpoints to consider. First, if true, the thesis would suggest (but does notentail) that there are no moral principles that are universally believed.Although it is not necessary, it is perhaps natural to expect that variablecauses will produce variable effects. So the causal dependence claim issignificant partly because it can be taken to support a “relativist” positionon the next issue on my list, that is, whether or not all moral beliefs varywith cultural and historical conditions. The second and more importantpoint to consider, however, is that the causal dependence of moral beliefson variable cultural and historical factors is thought to undermine moraltheories committed to other sources of moral belief —intuition, con-science, innate reason, divine commands, or universally shared senti-ments. It is a fallacy to argue that a belief must be false because it wascaused by something other than evidence for its truth. Hence, tracing thecauses of moral beliefs to variable cultural conditions unrelated to moraltheorists’ alleged sources does not prove that the normative principles ofthese theorists are false, but it tends to undermine moral theories thatdepend on alternative causal stories to explain how they know theirprinciples are true or valid.

C. Diversity of many or all moral beliefs

As I just noted, sometimes the “relativist” position is an empiricalthesis that what is accepted as morally right or good varies widely overtimes and places. A radical version would hold that there are no univer-sally shared moral beliefs or values, and for every moral belief and value

6 G. E. Moore, Principia Ethica (Cambridge: Cambridge University Press, 1903, 1959).


accepted in one culture there are incompatible beliefs and values in someother culture. That many particular moral beliefs diverge (for example,beliefs regarding sex, property, or the conduct of war) is undeniable, butit remains controversial whether there are some general moral valuesunderlying the diversity of particular moral beliefs accepted in differentcultures. This is an empirical question that cannot be easily settled. (Acomplication is that to assess evidence one must first take a positionabout what beliefs and norms count as moral.) When there is culturalvariance in beliefs about facts such as the shape of the earth, there areways to resolve the issue, showing that certain beliefs are true (the earthis round) and others mistaken (the earth is flat). Mere diversity of beliefdoes not entail that there is no fact of the matter. The claim that all moralbeliefs vary with cultural and historical conditions, then, does not entailthat there are no true or valid universal moral beliefs. It does, however,raise doubts about moral theories committed to certain explanations ofmoral beliefs that would lead us to expect almost universal agreementabout moral issues. If, for example, the theorists tell us that learningmoral truths is like perception of colors or “seeing” simple mathematicalpropositions to be “self-evident,” then they will be hard-pressed to explainpervasive and persistent diversity of moral beliefs.

D. Normative dependence of specific moral rules and particularjudgments on context

Some believe that there are fairly specific, substantive moral rules thatare true or valid in all human conditions. Many understand the TenCommandments this way. In a notorious late essay, Immanuel Kant arguesthat telling lies is always wrong, and even some who espouse “virtueethics” hold that acts of certain kinds (such as torture or intentionallykilling innocent persons) are always wrong.7 By contrast, act-utilitariansmaintain that specific conclusions about right and wrong depend onwhatever will have (or is expected to have) the best consequences in thecontext. They typically affirm this act-utilitarian standard as true or validbut grant that the rightness or wrongness of specific acts can vary withthe circumstances. For example, according to the act-utilitarian standard,malicious self-serving lies tend to be wrong because they cause moreharm than good, though it would not have been wrong to tell a lie to aNazi at the door to save the Jewish family hiding in one’s attic. Even

7 For Kant’s extreme position on lying, see “On a Supposed Right to Lie from Philan-thropy,” in Immanuel Kant, Practical Philosophy, trans. and ed. Mary Gregor (Cambridge:Cambridge University Press, 1996), 605–15 [8:423–30]. Elizabeth Anscombe’s view is expressedin her classic essay “Modern Moral Philosophy,” Philosophy 33 (1958): 1–19. Some otherphilosophers associated with virtue ethics (for example, Philippa Foot) have reportedlyendorsed this view in discussions.


deontologists such as W. D. Ross hold that, although there are true uni-versal principles of prima facie duty, particular judgments about one’s“actual duty” in a situation depend also on the variable facts of the caseat hand.8 Contemporary moral “particularists” (for example, JonathanDancy) go even further, maintaining that correct moral judgments aredetermined by the particular facts of each case independently of moralprinciples.9 Dancy does not deny that particular moral judgments are true,“objective,” and supported by reasons, but only that there are valid anduseful moral generalizations. All of these theorists hold that there are trueor valid particular moral judgments, as determined either by universalmoral principles or by good reasons in a context. Their primary disagree-ment is about the scope of intermediate-level moral rules or principles —for example, those forbidding lying, promise-breaking, torturing prisoners,and engaging in extramarital sex.10 Are these valid rules or true gener-alizations for everyone, or only within certain cultures, or only for certaingroups or individuals, or not at all?

E. Questions about the scope and authority of basic standards anddeliberative procedures

Those who hold that specific moral rules and judgments are true orfalse, or valid or invalid, depending on their context typically believe thatthere are more basic, general standards or procedures that explain thisvariability of specific rules and judgments. Both act-utilitarians and rule-utilitarians affirm such basic standards, and so do deontologists such as

8 W. D. Ross, The Right and the Good (Oxford: Clarendon Press, 1930), chap. 2. Ross heldthat it is the principles of prima facie duty together with the facts of the particular case athand that determine one’s actual duty, but because none of those principles is absolute andthey cannot be ranked with respect to one another, one’s actual duty can vary with evenslight changes in circumstance. To say that one has a prima facie duty to keep one’s promisein a particular case is, in effect, to say that the fact that one promised is a pro tanto ordefeasible moral reason to keep the promise or, in other words, a reason for doing what waspromised that is sufficient in the absence of weightier countervailing moral reasons but notnecessarily decisive in all contexts.

9 Jonathan Dancy, Ethics without Principles (Oxford: Oxford University Press, 2004),chap. 1.

10 Principles can be more or less general and comprehensive. In Kant’s moral theory, thevarious formulations of the Categorical Imperative are supposed to express the suprememoral principle, which is the most general principle relevant to all questions about what wemorally ought to do. What I call “intermediate-level principles” (or sometimes just “specificprinciples”) are not as general and comprehensive, applying only to certain kinds of often-recurring cases, for example, cases involving lies, theft, various sexual practices, developingone’s talents, giving aid to the needy, etc. These are intermediate in generality between theCategorical Imperative and extremely specific rules and judgments concerned with a nar-rowly restricted range of cases (e.g., the special responsibilities of doctors and social workersregarding patients with Alzheimer’s disease). Kant outlines a system of intermediate-levelprinciples for law and individual ethics in Immanuel Kant, The Metaphysics of Morals, trans.and ed. Mary Gregor (Cambridge: Cambridge University Press, 1996).


Ross and Kant.11 The standards or procedures are commonly regarded astrue or valid for all competent adults. They are conceived as the ultimatetest of what is moral, and for many they are also (or thereby) an ultimaterational authority. Opposing theorists (and some who are opposed to alltheory) deny that there are universally valid basic moral standards ordeliberative procedures that can guide and constrain our assessment ofmore particular moral rules and judgments (or make some true and oth-ers false).12 There are different ideas, of course, about what it means forbasic standards to be true, valid, and rational.

F. The possibility of proof, justification, or defense of basic standards

Many who accept that there are true or valid basic moral standardsoffer arguments of various kinds in support of their position, but somesimply pronounce their standards to be self-evident. Kant and John StuartMill are two prominent moral philosophers who took seriously the tasksof explaining a sort of “proof” or defense that they believed possible andof offering arguments to serve the purpose. There are different views, ofcourse, about what would constitute a proof, justification, or defense, andmany theorists are skeptical about whether these are possible in any form.What constitutes a proof, justification, or defense depends heavily onhow a given moral theory construes moral judgments —as stating naturalfacts, expressing the speaker’s attitudes, ascribing real nonnatural prop-erties to things, applying necessary rational principles, and so on.

There are, no doubt, other questions and further distinctions in dis-putes about moral relativity and objectivity, but my aim here has beenonly to set the stage for discussion. The primary focus of the rest of thisessay will be on the last two issues —more specifically, on a broadlyKantian position on whether there are general and basic moral standards thatdetermine which more specific moral principles and judgments are jus-tifiable and how, if at all, the more basic standards can be defended. These are

11 The act-utilitarian standard (roughly) is that one ought to do whatever brings about, oris expected to bring about, the most good (e.g., pleasure or well-being) in the long run. Thestandards of rule-utilitarianism are more complex to state because of various necessaryqualifications, but the main idea is that one ought to conform to a set of rules (a code) theinternalization of which by most people would bring about the most good. See Brad Hooker,Ideal Code and Real World (Oxford: Clarendon Press, 2000), 1–2 and 33. Ross’s basic standardsare his principles of prima facie duty concerning fidelity, gratitude, reparation, justice,self-improvement, non-injury, and beneficence. See note 8. Kant’s basic standards or delib-erative procedures are expressed in the several formulations of the Categorical Imperative.See note 2.

12 The opposing theorists include not only nihilists (who dismiss morality entirely) butalso subtle expressivists (who deny literal truth and validity to moral principles but do notdeny their importance) and particularists (who affirm particular moral truths, possibly evensimilar ones in different cultures, but deny that there are basic general standards). Anotherpossible view would be that there are general standards that are basic and true or validwithin different historical or cultural contexts but no basic moral standards that are true orvalid for all contexts.


difficult questions because they require not only interpreting and extend-ing Kant’s ideas but also discussing them in the terms of our contempo-rary debates in moral theory.

Before turning to the main two questions, however, here are a few briefcomments on my broadly Kantian perspective on the other questions —comments which may explain why those are not my focus of concern. Thequestions in subsections B and C above are entirely matters for empiricalinvestigation. Kant had opinions about them but no relevant expertise,and the main features of his moral theory, arguably, do not depend on hisopinions on such empirical questions.13 The main question in subsectionA presumably belongs to traditional metaethics because it concerns themeaning, or proper use, of the word “good” and related terms. Kant’sposition on this matter could easily be misunderstood. He repeatedly writes

13 The relevance of various empirical facts to Kant’s basic normative claims is not obvi-ous. Kant evidently believed that virtually every mature and competent moral agent, at leastevery one in relatively “advanced” or “enlightened” parts of the world (and history), haspractical reason and a conscience that acknowledges “the moral law,” no matter how badlyhe or she may behave. He also thought that reason can be latent and conscience dulled, andthat “evil” people intentionally elect to subordinate their moral predispositions. See Kant,The Metaphysics of Morals, 160–61 [6:400–401] and 189–91 [6:438–40], and Kant, Religion withinthe Boundaries of Mere Reason, ed. Allen Wood and George di Giovanni (Cambridge: Cam-bridge University Press), 179–80 [6:184–87] and 45–65 [6:19–44]. Kant’s deplorable anthro-pological remarks about peoples of other “races” at least suggest that, in his opinion, eventhe capacity for morality may be more limited than his inspiring words about humandignity would seem to imply. See Bernard Boxill and Thomas E. Hill, Jr., “Kant and Race,”in Race and Racism, ed. Bernard Boxill (Oxford: Oxford University Press, 2001), 448–71. Itremains a difficult question to what extent Kant’s main theses and arguments depend on hisempirical beliefs and prejudices. Groundwork I starts with assertions about common rationalknowledge of morality and proceeds to analyze their presuppositions. If Kant was mistakenabout how wide the community that shared the “common” beliefs is, this would limit theinterest in his main argument, making it applicable only to those who share the commonbeliefs. This, however, would not necessarily refute or render useless the sort of conditional(“analytical” style) argument that he admitted the argument in Groundwork I to be. Similarly,Groundwork II argues that our common conception of duty has certain striking presuppo-sitions, but Kant concedes rather dramatically that this does not entail even that “we” whohave that conception are really under moral obligations —and, all the more, the argumentclearly does not entail that human beings who lack the conception are under moral obli-gations. In Groundwork III, and the Critique of Practical Reason, Kant tries to convince readersof the stronger claim that all agents with practical reason are subject to the moral require-ments prescribed by the Categorical Imperative, but he is committed to the idea that thesearguments cannot rest on empirical claims (for example, the claim that, as a matter of fact,all cognitively competent human beings acknowledge these moral requirements or woulddo so after being given rational arguments). See sections I and II of Kant’s Groundwork and,for an interpretation of the main arguments of those sections, see my “Analysis of Argu-ments,” in Immanuel Kant, Groundwork for the Metaphysics of Morals, trans. Arnulf Zweig anded. Thomas E. Hill, Jr., and Arnulf Zweig (Oxford: Oxford University Press, 2002), 195–245[4:393–445] and 109–14. See also Immanuel Kant, Critique of Practical Reason, trans. and ed.Mary Gregor (Cambridge: Cambridge University Press, 1997), 28–64 [5:30–50]. If Kant’sconception of duty and a good will were limited to eighteenth-century Prussians, the inter-est in his arguments would be severely limited in scope, but there is apparently a wideraudience that shares Kant’s conception of duty, or at least the belief that a moral duty iswhat one has sufficient reason to do and not just because it serves one’s interests. Whether,despite objections, Kant’s arguments have any merit remains a controversial question, noteasily settled by reference to anthropological facts and still subject to doubts of other kinds.


of moral acts as “good in themselves” and thus does not buy into Hobbes’sthesis that “good” is always “relative to the person” (more specifically, tothe desires of the speaker or the command of the sovereign). In Kant’s view,however, there is nothing, not even what he calls “good in itself,” that is“good” in a practical sense independent of its relation to actual or possiblerational agents and their reasons for intending, willing, and acting. Kantdoes not believe in intrinsic values in the metaphysical sense that, for exam-ple, is central in G. E. Moore’s intuitionist ethics.

Regarding the issue discussed in subsection D, Kant famously held thatthere are some specific moral principles that hold for all people at alltimes. His notorious example was an absolute prohibition on (serious)lying, but it is noteworthy that even famous advocates of “virtue ethics”(for example, Elizabeth Anscombe) apparently accept some completelyinflexible universal principles about other matters, such as the intentionalkilling of innocent people.14 On my broadly Kantian view, when (if ever)principles regarding lying and killing innocents are subject to exceptionsmust be determined by reflection from a more basic deliberative perspec-tive. What this is and whether it is defensible —the questions in subsec-tions E and F —are the topics for my remaining discussion.

III. A Broadly Kantian Perspective for AssessingSpecific Moral Principles

Here I can only give a brief summary of relevant points in the Kantianperspective that I have in mind. Kantian moral theory belongs to a ratio-nalistic tradition that maintains that the most fundamental moral valuesand standards are accessible through the use of practical reason, appli-cable to all rational human beings, and contrary to reason to dismiss orviolate.15 These fundamental values and standards are supposed to beexpressed in Kant’s several formulations of the Categorical Imperativeand accompanying explanations.16 Kant regarded these basic standardsas presupposed by, and even implicit in, ordinary moral deliberation andjudgment. He illustrates the guiding function of two versions of the Cat-egorical Imperative by applying them directly to four examples that he

14 See note 7 above.15 In Kantian theory, as I understand it, “values” are not entities or facts independent of

what rational agents would care about (or “value”) in various conditions. To say thathumanity is a basic Kantian value, then, is not to claim that it has a perceptible quality thatexplains and justifies the idea that we have a rational requirement or good reasons to respectit. Saying that humanity is a basic value implies that it should be respected; i.e., we have arational requirement or good reasons to respect it, because these are just different ways ofsaying the same thing. The Kantian view, so construed, is an instance of what T. M. Scanloncalls the “buck-passing” view of value —that is, value-claims do not state the justifying basisof reason-claims but rather express the idea that there are good reasons for taking certain (notyet specified) attitudes and actions toward the thing. T. M. Scanlon, What We Owe to EachOther (Cambridge, MA: Harvard University Press, 1998), 95–100.

16 Kant, Groundwork for the Metaphysics of Morals, 222–37 [4:421–38].


sketches. In his later, more thorough treatment of practical problems,however, he appeals to basic standards together with general facts aboutthe human condition to identify and (sometimes) argue for a system ofgeneral principles about property, marriage, legal authority, punishment,lying, self-neglect, helping others, gratitude, respect, and friendship.17

This is a system of principles of mid-level generality, standing betweenthe most comprehensive supreme moral principle (expressed in the Cat-egorical Imperative) and moral judgments in particular contexts. Someprinciples state juridical duties, which can and ought to be enforced bystate authorities. Other principles state directly ethical duties, which requireus to adopt maxims of various sorts to promote the ends of self-perfectionand the happiness of others. The strict perfect duties have the form “Never. . .” or “Always . . . ,” but the imperfect duty to help others leaves room forjudgment and choice regarding when, how, and to what extent to help.

According to Kantian theory, then, there is a structure that can guide usto correct moral judgments in particular cases. The ultimate standard, a prin-ciple or procedure expressed as the Categorical Imperative, is taken to bea comprehensive, fundamental, and even necessary requirement of prac-tical reason —and so the standard for our reasoning about what we oughtto do. Applying this standard to pervasive human conditions, Kant attemptsto identify (and justify relative to the standard) a system of mid-level prin-ciples of several kinds intended to be applicable independently of varia-tions in local conditions. The principles of perfect duty must never entailcontradictory prescriptions, but the “grounds of duty” from which they aredrawn may be in tension.18 All the more, imperfect duties may be in ten-sion, for promoting one moral end may limit the ways that one can pursueanother. The system is supposed to be structured and consistent, but itsimperfect duties allow considerable room for choice, and all duties requireinterpretation and judgment appropriate to the context.19

When explaining the basic Kantian standard for identifying more spe-cific moral principles, most contemporary reconstructions and develop-ments of Kant’s moral theory concentrate on interpretations of eitherKant’s formula of universal law or his formula of humanity as an end initself.20 Skeptical of these (or any formulas) as case-by-case moral deci-

17 Kant, The Metaphysics of Morals, parts I and II, 1–170 [6:203–413].18 Ibid., 16–17 [6:224–25].19 Ibid., 153 [6:390].20 The formula of universal law states: “Act only on that maxim by which you can at the

same time will that it should become a universal law.” Kant, Groundwork for the Metaphysicsof Morals, 222 [4:421]. A variation that Kant uses in discussing his examples is: “Act asthough the maxim of your action were to become by your will a universal law of nature.”Kant, Groundwork, 222 [4:421]. The formula of humanity states: “Act in such a way that youtreat humanity, whether in your own person or in any other person, always at the same timeas an end, never as a means.” Kant, Groundwork, 230 [4:429]. Distinguished scholars haveinterpreted these formulas in different ways, many of which I review in my essay “KantianNormative Ethics,” in David Copp, ed., The Oxford Handbook of Ethical Theory (Oxford:Oxford University Press, 2006), 480–514.


sion procedures, I have suggested that a Kantian framework for assessingmoral rules (or mid-level principles) might be constructed from Kant’sidea of the perspective of a lawmaking member of a community of ratio-nal, autonomous, and mutually respectful persons (what Kant calls aReich der Zwecke or “kingdom of ends”).21

In section II of the Groundwork for the Metaphysics of Morals, Kant’s attemptto articulate a basic moral standard unfolds in a progressive argumentthrough the earlier formulas (of universal law and humanity as an end)to the idea of rational persons with autonomy legislating universal lawswithout basing them on contingent personal desires. Loosely paraphrased,the argument proceeds from “Conform to universal law” to “Respect andplace the highest value on humanity in each person,” followed by anexplanation that this requires, among other things, respecting each personas a rational autonomous legislator of universal laws. This idea, in turn,is further developed in a formula of the Categorical Imperative that Kantsuggests is his most complete characterization of moral requirements, theprinciple that says that we should conform to the laws of a possible Reichder Zwecke, or ideal commonwealth in which all rational autonomouspersons, “abstracting from personal differences,” legislate the moral lawsthat unite and bind them.22 Although Kant still preferred his famousformula of universal law as the best guide for judgment in particularcases, he obviously regarded this later principle as an inspiring concep-tion of the moral point of view, “closer to intuition” and “feeling.” 23

The main elements of Kant’s idea of moral legislation are the following.The legislators are rational, seeking to legislate a consistent and coherentset of “laws” (moral principles), making no law without sufficient reasonor in ignorance of relevant facts. They are committed to finding andrespecting laws that apply to all and respect human dignity in all, and so,in legislating, they must rely on reasons that any moral agent can acknowl-edge and take into account apart from special interests. Their “auton-omy” also implies freedom from various distorting and morally irrelevantinfluences and a genuine acknowledgment that the rational standard

21 See note 1 for references to my fuller discussions of this Kantian perspective. Accordingto David Falk in conversation, the frequent translation of Reich der Zwecke as “kingdom ofends” is due to H. J. Paton’s insistence on using the term in his translation of Kant’sGroundwork despite Falk’s warning that the translation is misleading. H. J. Paton’s transla-tion is Groundwork of the Metaphysic of Morals (New York: Harper and Row, 1964). A Reich(literally) is a state or commonwealth, a governing legal system that may or may not involvea king or monarch. There is a possible analogy with biblical talk of a “kingdom of God” thatmay have influenced Kant but is not essential to his idea. When Kant describes the “king-dom of ends” (Reich der Zwecke) in the Groundwork, it has a “head” with features onlypossible for a divine being, but the head has no function apparently but to endorse the samerational principles as every other member. The primary difference is that the “head,” whois not conceived as a political sovereign with independent legislative authority, lacks sen-suous temptations to act contrary to reason and so is not “subject to” or “bound by” therational principles. That is, they are not moral “imperatives” for the divine being.

22 Kant, Groundwork for the Metaphysics of Morals, 231–37 [4:431–37].23 Ibid., 237 [4:436–37].


behind particular moral reasons is not merely rational prudence orefficiency in the pursuit of one’s own personal ends. Obviously, andfor good reason, certain moral values, assumed to be fundamental, arebuilt into this conception of moral legislation. The aim is not to give abootstrapping argument for basic moral values but rather to constructfrom these an intuitively appealing and inspiring conception of themoral point of view that should frame our thinking about more spe-cific moral requirements.

A more modest application of Kant’s idea, akin to Rawls’s initial theoryof justice, would make use of the model of “legislation” from an idealperspective adapted for a more specific range of practical problems. Forexample, we might use this approach in addressing the question whetherthe prohibition of torture is open to exception. Given certain basic moralvalues, the presumption against torture (as well as against murder, rape,fraud, and manipulation) is rather obvious. The difficult questions areabout cases in which moral values conflict. Granting that some particularhard cases may have no resolution, we can still ask whether it is morallyimportant to maintain shared public principles about such matters, inde-pendently of law, and what exceptions, if any, these principles shouldallow. The broadly Kantian idea is that it may be helpful to treat suchquestions in two stages —articulating the relevant standpoint for jointdeliberation (in effect, the rules of debate) and then reviewing argumentsfor proposed principles.

From the Kantian perspective, the most fundamental values that are toguide and constrain deliberation about rules are drawn from the idea ofhuman dignity, or humanity as an end in itself. In Kantian theory, as Iunderstand it, saying that something is valuable is equivalent to sayinghow we ought to regard and treat the thing that is said to have value. Wecan interpret the idea of human dignity, then, as a cluster of prescriptionsabout how to regard and treat human beings. In an earlier essay, I sum-marized these prescriptions as follows:24

In sum, they are, negatively, not to regard persons as having a merelycommensurable value and not to violate principles that rational auton-omous persons would endorse, all things considered; and, positively,to exercise, preserve, develop, and honor rational nature in all per-sons, to prioritize the needs for rational autonomous living overmore trivial concerns, and to count the happiness of others (or, strictly,their permissible ends) as significant in our deliberations. The pro-posed Kantian perspective regards the second negative prescriptionas an inflexible constraint on action: never violate the principlesthat, in one’s best judgment, rational autonomous deliberators would

24 Hill, “Assessing Moral Rules: Utilitarian and Kantian Perspectives,” 168–69; emphasisin original.


endorse, if guided and constrained by the idea of human dignity andmindful of the complexities of the real world for which the principlesare intended. The first negative prescription (“never regard personshas having a merely commensurable value”) and the remaining pos-itive prescriptions are also non-optional demands, but they directlyconcern our attitudes when we deliberate and act.25

Obviously much more needs to be said, but it may be helpful to con-clude my sketch with a few general remarks about a significant differencebetween the Kantian perspective and some theories that are in otherrespects similar. For example, the Kantian framework is meant to reflectwhat deliberation from a moral point of view about the issue at handwould be like, not necessarily what rational self-interested agents wouldchoose in a state of nature or what “mutually disinterested” rationalagents would choose under a “veil of ignorance.” 26 Because those whodeliberate from the broadly Kantian perspective are directly moved andconstrained by basic moral considerations, the Kantian perspective,as construed here, resembles Locke’s social contract theory more than

25 Of course, they will rule out certain acts as immoral and show that others are morallyrequired, but they do so by demanding that our deliberations, choices, and intentions arecompatible with the prescribed attitudes. It is not as though one could muster up a moti-vationally inert and non-reason-giving “inner state of mind,” such as a mere wish or “purethought,” and then anything one does will be permitted.

26 As Fred Miller has noted in discussion, theories that aim to derive specific principlesfrom a “self-interested” or any other deliberative point of view will not be able to reach anysubstantive conclusions unless they explain the interests or values that motivate the partiestaking the deliberative (or contractual) perspective. “Self-interest” can be conceived in dif-ferent ways, for example, as having the most pleasure possible, as realizing one’s mostimportant personal ends, or as Aristotelian eudaimonia. In Hobbes’s state of nature, thosewho make the social contract are conceived as moved by their various desires, especially thepredominant desire to survive, and their desires are always for some “good” for themselves.In Rawls’s “original position,” the parties are exclusively motivated to secure, within theconstraints of the choice situation, the most favorable mix of “primary goods” (e.g., wealth,income, opportunities, powers, and self-respect) for themselves. They are mutually disin-terested for purposes of deliberation under a veil of ignorance that bars them from knowingtheir special talents and circumstances, but they “represent” themselves as actual persons inthe real world who may have deep religious convictions, family ties, and even altruisticdispositions. Both Rawls and Hobbes (on my reading) conceive of the parties in the relevantdeliberative position (Rawls’s original position and Hobbes’s state of nature) as not moti-vated by explicitly moral concerns. The choice situations are deliberately defined to ensurethat we can draw moral conclusions from arguments that the parties in those choice situa-tions would endorse certain principles, but the parties themselves are not engaged in moralreflection. There are some theoretical advantages in conceiving of the parties’ motivations asnonmoral, but, following Kant, my broadly Kantian deliberative perspective neverthelessconceives of the “legislators” of specific moral principles as motivated in part by stipulatedbasic moral values (for example, humanity as an end in itself ). This approach, of course,invites questions about how the stipulated values can be defended and whether, given thatmoral values are stipulated, arguments for specific principles are circular or redundant.My brief discussion here suggests Kantian responses, but these questions obviously requiremore attention.


Hobbes’s.27 In this respect, the Kantian perspective also appears to differfrom T. M. Scanlon’s contractualism, for Scanlon stipulates that in deter-mining what principles can reasonably be rejected we must restrict our-selves to comparing the weight of reasons that different agents have fromtheir perspective.28 The reasons for endorsing or rejecting a principle thatmust be weighed are not moral reasons, even though the outcome ofScanlon’s contractualist reasoning is supposed to be judgments aboutwhat is (or is not) morally wrong. My broadly Kantian approach, then, isjust one of several ways of thinking about morality that attempt, first, todefine an appropriate point of view for deliberation and debate and then,second, to use this perspective to derive specific principles about how todeal with various recurring practical problems. My approach differs fromvarious contract-oriented theories, however, in explicitly calling for reli-ance on basic moral values in deliberating about specific principles.

IV. Strategies for the Defense of theDeliberative Perspective

To sketch Kant’s strategies, we must review three main points: a back-ground of critical arguments and then two different lines of defense forKant’s own perspective. The background (sketched in subsection A below)consists of objections to previous moral theories. Then the first positivestrategy (subsection B) is an attempt to show “analytically” that the basicvalues in Kant’s deliberative perspective are presupposed in conceptsessential to our ordinary understanding of morality. The second positive

27 John Locke’s political philosophy employs the idea of reasonable agreements or con-tracts that were, or would be, made by persons who take for granted certain fundamentalmoral norms (the “laws of nature”) as requirements of reason. Locke then tries to justifyvarious specific principles regarding government, revolution, etc., by arguing that theseprinciples would be endorsed by contracting parties who have personal interests but arealso committed to these laws of nature (as moral guides and constraints). Thomas Hobbes,by contrast, tries to justify principles regarding government, law, etc., by arguing thatrational persons in a state of nature would have sufficient prudential grounds, apart fromany independent moral commitments, to accept those principles to be enforced by a sovereignpower. Interpretations differ, but on my reading Locke, like Kant, is explicitly arguing frombasic general moral values to more specific moral principles. Hobbes is apparently moreconcerned to avoid potentially controversial assumptions about moral values and instead tojustify specific principles as rationally prudent to accept and follow if one is assured thatothers will. Among those principles are Hobbes’s “other laws of nature,” many of whichresemble familiar moral norms, but Hobbes apparently regards these as rationally bindingonly because following them serves to enhance one’s power to survive and satisfy one’sdesires. See John Locke, Second Treatise of Government, in Classics of Moral and Political Theory,ed. Michael L. Morgan (Indianapolis, IN: Hackett Publishing Co.), esp. chaps. 1–9; andThomas Hobbes, Leviathan, part I, esp. chaps. XIV and XV. We should note, however, that Ihave been comparing a Kantian moral perspective with the political theories of Locke andHobbes. A comparison of the latter with Kant’s political and legal theory is complicatedbecause the relation between the Categorical Imperative and Kant’s principles regardinglaw is unclear and controversial.

28 Scanlon, What We Owe to Each Other, esp. chaps. 2, 4, and 5.


strategy (subsection C) is simply to appeal to each reader’s moral con-sciousness, which Kant thought was a virtually inescapable recognition ofbeing under moral obligation (as Kant conceived this).29 Apart from thedetails of Kant’s own exposition, arguably these general background argu-ments and strategies are potentially relevant in assessing contemporarymoral theories, including the broadly Kantian perspective sketched here.

A. The background: Critique of previous moral theories

The background consists of negative arguments to reject other (non-Kantian) views about the nature and source of morality. These argumentsare most explicit in Kant’s Groundwork for the Metaphysics of Morals andCritique of Practical Reason.30 Kant categorizes all previous moral theoriesinto four types, which we may label “ethical egoism,” “moral sense theo-ries,” “divine command theories,” and “perfectionism.” Having arguedthat moral concepts presuppose that “autonomy of the will” is the fun-damental source of morality, Kant charges that all the previous moraltheories are mistaken because they fail to acknowledge this fact. Here, Itake it, Kant is engaging in what is now considered “metaethics,” for hemaintains that his “analytical” argument in Groundwork II shows that thetheories in question fail to understand common moral concepts and, fur-ther, that they leave us without any resources to argue (as Kant attemptsin Groundwork III) for the rational authority of basic moral standards.Kant alleges that each type of moral theory is open to special objections,but the common error of each theory is to suppose that morality is some-thing other than following principles that moral agents with practicalreason and autonomy would give (or “legislate”) to themselves as con-straining their social relations and personal choices. We do not need toaccept the full Kantian idea of autonomy to appreciate his critique ofother moral theories.

Roughly, the main negative points, reworked and modified for con-temporary discussions, are these: (1) Some moral theories attempt toderive the content and rational authority of moral principles from theidea that, fundamentally, each person should (or has sufficient reason to)

29 Those who were looking for a logical demonstration or empirical confirmation ofKant’s basic value claims are likely to find this appeal to their moral consciousness initiallydisappointing and perhaps not even worth calling a “defense.” It also falls short of Kant’sambitious project in Groundwork III to “establish” that anyone who acts for reasons does so“under the idea of freedom” and is therefore committed to the Categorical Imperative as asupreme standard. However, given a background of arguments against other moral theo-ries, the attempt to get us to see, in our own reflections, that we deeply and persistentlyacknowledge certain basic moral values as rational might suffice, for practical purposes, inresponse to certain kinds of theory-driven doubts about whether acting morally makessense.

30 See Kant, Groundwork for the Metaphysics of Morals, 242–45 [4:441–44], and Kant, Critiqueof Practical Reason, 32–36 [5:35–41].


do whatever most effectively promotes his or her interests. This approachis inadequate, Kant argues, because it is a constitutive conception of moralrequirements that they prescribe what we must do independently ofwhether they advance our own interests. Thus, even if only the pursuit ofself-interest is rational, the requirements that would be prescribed by thismethod would not be moral requirements. Kant classifies egoistic theoriesas “empirical,” apparently because their advocates typically base them onpsychological egoism31 and because they make specific moral require-ments vary widely with each person’s empirical circumstances. Kant’smain critical point, however, is that moral requirements cannot be derivedfrom egoistic principles of any kind, rational or otherwise.

Further, (2) Kant also rejects moral sense theories, presumably in partfor reasons that comparable theories tend to be rejected now. Such theo-ries attempt to derive moral conclusions from patterns of moral feelingsthat human beings typically have in recurrent situations —for example,impartial spectators’ feelings of disapproval of cruelty. The core objectionis that judgments about what we ought to do are normative judgmentsabout what we have good reasons to do, and these do not follow (withoutfurther premises) from facts about human sentiments, no matter howuniversal these may be. The point for contemporary discussions would bethat empirical psychology alone cannot establish moral claims about whatwe ought to do, even though it can undermine ill-informed particularmoral judgments and cast doubt on our capacity to live by certain allegedmoral standards. Assuming (with Kant and others) that moral claimspurport to give us good reasons to act, empirical studies by themselvescannot determine which moral claims, if any, are true or valid.

Turning to theories that Kant classifies as nonempirical or rational theo-ries, (3) he rejects theories that attempt to derive moral requirements fromtheology on the ground that either their derivations are circular or thederivative commands are not moral. In other words, either the divinesource is defined as giving only rational moral commands or else its com-mands are nonmoral because the authority comes from divine power, notreason. This objection also poses a challenge for any contemporary moraltheory that rests the ultimate reasons for being moral on actual com-mands, secular or ecclesiastical. Similar considerations might apply to theassumption that the “force” of tradition, custom, or common opinion(alone) is the basis of moral requirements.

Finally, (4) Kant rejects forms of perfectionism that base the rationalauthority of moral requirements on their alleged promotion of the meta-physical value of “perfection,” an idea prominent in the work of Leibnizand Christian Wolff. This idea, Kant suggests, is too indeterminate to be

31 Psychological egoism, as I understand it, is the view that, as a matter of fact, everyperson is always ultimately motivated by the desire to advance his or her own interests. Itcan take different forms, depending on how “one’s own interests” is interpreted and howdeep or near the surface self-interested desires are taken to be.


action-guiding unless we interpret “perfection” in a morally loaded waythat would make derivations of moral requirements from it circular orquestion-begging. Again, details aside, we can draw a message relevantin moral theory today. Metaphysical properties, whether labeled “perfec-tion” or “intrinsic value” or whatever, are supposed to be features of theworld as it is, not necessarily as it ought to be. What we have reason to docannot be determined by such facts alone. Although we are not concernedwith the sort of perfectionism that Kant explicitly rejected, his objectionmay be relevant to current theories that rest too easy with “intuitions”about which facts are practical “reasons” (or “considerations that favor”choices), in the absence of any general account of what constitutes reasonsof various kinds.

Returning now to my review of Kantian strategies for defending allegedbasic moral standards, the first main point (which provides backgroundfor more positive arguments) has been that certain other strategies areinadequate for the defense of Kantian basic standards and, equally impor-tant, that those strategies cannot by themselves refute Kantian proposalsabout what the rational moral standards are. This is a strong negativemetaethical claim that, perhaps for good reasons, most contemporaryKantians do not wish to debate at length, but it is a background idea that,if correct, makes it easier to move on to positive considerations.

B. Argument that the perspective is presupposedin common moral judgments

Kant’s main attempts to defend his conception of the basic moral stan-dard are presented in the first two chapters of the Groundwork, where heargues that concepts deeply embedded in common morality presupposethe supreme principle expressed in his versions of the Categorical Imper-ative. His aim in these chapters is not to “establish” his basic principle orstandard as rational but only to identify formulations of that standardthat express its essential features and can serve to guide moral delibera-tion. If successful, these arguments taken by themselves would providean important but limited “defense” of Kant’s basic deliberative perspec-tive as reflecting the indispensable core of our common morality, ratherthan mere custom or personal preference.

Kant’s arguments do not start from a survey of moral opinions aboutparticular matters, such as lying and suicide, but from relatively formalideas about duty and a good will. An essential feature of these ideas,supposedly pervasive in ordinary moral thinking, is that being good anddoing what one ought is, among other things, a matter of respectingconstraints that we have sufficient reason to follow and not just becausedoing so advances our interests. This could be so, Kant argues, only ifthere is a basic rational standard, distinct from instrumental and pruden-tial rationality, which can support more particular moral judgments. More-


over, Kant argues that the content of the basic standard(s) is expressiblein different ways, which are supposed to be fundamentally the same.Most abstractly expressed, the standard is that we must (rationally) choosein a way that respects “universal laws.” As eventually interpreted, thestandard is to respect the more specific principles of conduct that wouldbe “legislated” by rational agents who have “autonomy of the will” andrespect humanity in each person. The form of Kant’s argument is condi-tional, and so is his conclusion. That is, essential concepts of commonmorality presuppose his basic deliberative perspective as the standard fromwhich genuine moral obligations derive. This defense is distinct from hislater attempt, in Groundwork III, to show that common morality is notmistaken in taking its requirements to be rational to follow.

Let us set aside details and doubts about Kant’s actual argument (andmy brief summary) and focus instead on the kind of positive strategyKant uses to defend his attempt to articulate a basic moral standard.Kant’s strategy is similar in some respects to Rawls’s method of reflectiveequilibrium, but there are also important differences.32 Some differencesin their methods stem from differences in their ambitions and hopes formoral theory, but others reflect Rawls’s greater respect for the empiricalrealities which influence and provide the context for moral reasoning.These differences understandably make Rawls’s method more acceptableto contemporary philosophers, but the ways in which Rawls’s methodremains similar to Kant’s are also instructive.

To set the background very briefly, Rawls draws from widely accepted,relatively uncontroversial assumptions about basic moral and politicalvalues to develop (or “construct”?) an appropriate idealized point of view(“the original position”) for assessing competing conceptions of justice.33

He proposes two general principles of justice and then argues that thesewould be preferred to certain alternatives from the perspective of theoriginal position. Then Rawls argues that applications of the principles, inseveral stages, more or less match the “considered judgments of compe-tent judges” in due reflection. When arguments from the appropriatedeliberative standpoint (“the original position”) as this was initially definedlead to conflicts with our “considered judgments,” we need to reflectfurther on the definition and arguments for the original position, its appli-cation, and our own pretheoretical considered judgments. The task ofseeking reflective equilibrium is to revise the arguments, refine the originalposition, or abandon our initial moral judgments, until we can find nomore acceptable way to bring our various convictions into a coherentwhole. Like Kant, Rawls argues from what he takes to be deep andpervasive ordinary moral ideas to support an account of an appropriate

32 John Rawls, A Theory of Justice, Revised Edition (Cambridge, MA: Harvard UniversityPress, 1999), 18–19, 42–45, and 507–8.

33 Ibid., esp. part I, 10–19.


deliberative perspective for assessing more specific principles, and thenhe illustrates the application of the deliberative perspective by applyingit to several practical problems.34

Significant differences, of course, should be noted. Most obviously,Kant’s ambition was to present a comprehensive moral standard, whereasRawls aims to construct a deliberative perspective for a more limitedproject. That is, even in A Theory of Justice, Rawls constructs his “originalposition” primarily to assess competing, historically prominent concep-tions of justice as they apply to the basic structure of well-ordered, coop-erative, closed societies —in particular, with regard to how the basicstructure affects the distribution of primary social goods. Later, in PoliticalLiberalism, he restricts his project further to “political conceptions” ofjustice that reflect background assumptions common in Western liberaldemocracies.35 A significant feature of Rawls’s arguments for his delib-erative procedure for assessing principles is that, unlike Kant, Rawls doesnot try to make his case by drawing out the implications of a few, rela-tively formal and allegedly essential features of common morality (suchas the idea of duty and the value of a good will). Rawls rightly acknowl-edges more places where empirical facts are relevant to his arguments,and he is obviously more aware of the possibility that revisions may benecessary.

Important similarities remain, however —and for good reasons. BothKant and Rawls see moral theory as a project with a structure that dis-tinguishes basic deliberative procedures from derivative principles, andboth distinguish arguments for the deliberative procedure from argu-ments from the procedure to derivative principles. Although both rely onassumptions about common moral opinions, neither treats these opinionsas self-evident or infallible. Despite his repeated warnings about not bas-ing moral theory on examples, Kant (like Rawls) draws out the (sup-posed) implications of his principles as some sort of confirmation ofthem.36 Neither Kant nor Rawls addresses his theory to radical moralskeptics or pretends to have arguments sufficient to show that living amoral life is necessarily in one’s interest. Finally, both Kant and Rawlsseem to realize that, in the end, defense of their projects and results

34 Rawls takes up various practical problems in part II of A Theory of Justice, for example,toleration of the intolerant (190–94), governmental institutions to promote just distributionof wealth and income (242–51), justice between generations (251–58), and civil disobedience(319–31).

35 John Rawls, Political Liberalism (New York: Columbia University Press, 1993).36 Kant says: “[M]y claims about this central question [the nature and justification of the

supreme principle of morality] would be greatly clarified by seeing the application of thatsupreme principle to the whole system, and they would be strongly confirmed by the ade-quacy the principle would manifest throughout.” Kant, Groundwork, 193–94 [4:492]; empha-sis added. He nevertheless strongly condemns any attempt to derive fundamental moralprinciples from examples. See, for example, Groundwork, 208–14 [4:406–12]. The Metaphysicsof Morals contains his most systematic attempt to present and support a system of intermediate-level principles.


depends on a different sort of appeal to each reader —the subject to whichI turn next.

C. Appeal to each rational agent’s awareness or acceptanceof the basic standards

A final strategy for defending an account of basic moral deliberativestandards is suggested by Kant’s attempts to address the question that heso dramatically poses at the end of Groundwork II. That is, even if anaccount is demonstratively reflective of common moral ideas, is it reallyrationally necessary to accept it? Might it not be that morality’s claim to berational is just an illusion? Kant tried in Groundwork III to meet the chal-lenge by arguing that whenever we act for reasons, we cannot help butsee ourselves as free in a sense that ultimately commits us to the Cat-egorical Imperative.37 To meet the objection that he may have argued ina circle, he adds that in the theoretical use of reason (for example, inscientific investigations) we must take ourselves to be following reasonindependently of causal determination, and this again commits us to theCategorical Imperative.38

Later, in his Critique of Practical Reason, Kant reverses his argument,apparently conceding the inadequacy of the early arguments. He thenproposes that consciousness of the moral law is “a fact of reason,” ade-quate for practical purposes.39 Scholars debate what he meant, but appar-ently at least part of the idea is that for us, human beings who takeourselves to be rational moral agents, no further argument is neededbecause, try as we will to deny and escape moral requirements, we can-not. That is, although self-deception, special pleading, and moral dis-agreements are all too common, in serious honest reflection we cannotconsistently regard ourselves as rationally entitled systematically to ignorethe claims of others as not (rationally) binding on us. The rational author-ity that supports such claims, Kant suggests, is inevitably recognized asthe authority of our own reason, however much we may wish to deny it.This is not presented as a demonstration, or a claim to perceive self-evident rational truths, or an appeal to moral feelings. As I understandKant’s point, it is an appeal to individuals, in their own first-personreflections, to acknowledge that they are under moral obligations or atleast cannot help but conceive of themselves in this way. Arguably, despiteKant’s own views, such reflections should not be used to justify assumingthat others are as committed to morality as we are —assuming this, forexample, when we are inclined to justify harsh moralistic treatment ofcriminals without regard to their actual background and attitudes. Never-

37 Kant, Groundwork for the Metaphysics of Morals, 245–48 [4:444–48].38 Ibid., 248–57 [4:448–58].39 Kant, Critique of Practical Reason, 28–64 [5:30–50].


theless, Kant apparently thought his appeal to each person’s moral con-sciousness a relevant kind of response to certain kinds of doubts aboutmorality, given the fact (argued earlier) that contrary claims about ratio-nal choice could not be proved by science, theology, metaphysics, orsocial practice.

Kant’s expectations for the practical success of this appeal to each per-son’s moral consciousness were higher, I suspect, than the expectationsmost of us share. This is especially so if the person’s moral consciousnessis supposed to acknowledge Kant’s particular statements of the moral lawand associated ideas about freedom of will. But at least when our aim andcontext is practical, rather than an attempt to establish a universal truthof metaphysics or abstract moral theory, we may find that the best we cando, and all that is needed, is to appeal to the commitments of each rationalmoral agent. If agents profess to deliberate and debate in terms of sharedreasons, reasons about which others outside their circle should care, thenthey may be able to see, on due reflection, that their moral and reason-based commitments exceed their wishes to escape moral obligation. Andperhaps a more modest, reasonable contemporary account of moral obli-gations (compared to some ideas in Kant’s actual texts) can facilitate thisprocess, and arguably it can be a (practically) rational process. Rawls’sappeals to each reader to consider, revise, and endorse his theory ofjustice is similar and, in my view, remains worthy of serious consideration.

V. Conclusion

In this wide-ranging essay, I began by distinguishing different ques-tions that are often confused in discussions of moral relativism and objec-tivity, and then I attempted to sketch the responses that a broadly Kantiantheory might give to two of these questions. The first of these questionswas: How (if at all) can we derive, justify, or support specific moralprinciples and judgments from more fundamental moral standards andvalues? In response, I briefly sketched a broadly Kantian deliberativeperspective designed for this purpose, drawing from Kant’s later formu-lations of the Categorical Imperative. The second question was: How (ifat all) can fundamental standards such as those found in the Kantianperspective be defended? In response, I briefly described two of Kant’sstrategies for defending his fundamental standards and the importantbackground of arguments against previous moral theories.

The Kantian deliberative perspective, like rule-utilitarianism, sets usthe task of finding or constructing a set of specific moral principles (or“rules”) that satisfy basic values and constraints but also take into accountrelevant facts about the human condition and the circumstances in whichthe principles are to apply. Unlike rule-utilitarianism, however, the Kant-ian perspective essentially includes values associated with human dignity


that shape Kantian deliberations about the set of specific principles to beadopted. Kant’s strategies for defending the basic values inherent in thedeliberative perspective are, first, to try to show that these values arepresupposed in common moral thought and practice and, second, toappeal directly to the moral consciousness of the would-be skeptic. Inneither strategy does Kant argue for his standards from (allegedly) moreself-evident premises or appeal to intuitive “perception” of an indepen-dent moral order. Instead, he analyzes the presuppositions of commonmoral concepts and appeals to each person to acknowledge his or herown deep and persistent practical commitment to the basic standards asrationally authoritative. How far such arguments can succeed in the worldtoday is an open question. Kant’s own confidence in them was a kind ofmorally based faith not refutable by empirical studies, but it is hard toshare that confidence in today’s world.

Philosophy, University of North Carolina at Chapel Hill


CONSTRUCTING NORMATIVE OBJECTIVITY IN ETHICS

By David B. Wong

I. The Problem of Moral Normativity

A prominent problem for all naturalistic theories of morality has beento account for the apparent normative force of a moral demand. By “nor-mative force,” I mean the scope of proper application of that demand,along with the conditions for its proper application. To characterize thenormative force of a moral demand is to characterize to whom it properlyapplies and what conditions must be fulfilled for it to be properly applied.In this essay, the crucial issues about normative force are whether and inwhat manner a moral demand can properly apply to an agent regardlessof whether that agent has inclinations or desires that would be served byconforming to the demand. In ordinary moral discourse, it seems thatmoral demands can apply regardless of the agent’s possessing the rele-vant inclinations. To judge that U.S. officials have violated a moral dutyby imprisoning foreign nationals without charges is not necessarily torefer to any inclination those officials have that would be satisfied bydoing their duty. Indeed, one may judge so while assuming they have nosuch inclination. We talk as if the duty applies regardless of the existenceof such a motivation.

That moral demands are generally regarded as inescapable or non-hypothetical in this sense has generally dictated two opposing responses:on the one side, attempts to validate the apparent inescapability of moraldemands; on the other side, attempts to show that the appearance corre-sponds to a deep and pervasive error on the part of moral-language users.My response generally falls into the first category and is based on recog-nition of the role of morality and, more generally, the role of substantivepractical reason in shaping basic human motivations. By “substantivepractical reason,” I mean to suggest that the apparatus for reasoningabout what to do includes not simply rules of inference for passing frompremises to practical conclusions about what to do, but also an array ofreasons for agents to act in certain ways, where these reasons are situa-tional features that weigh in favor of agents’ acting in certain ways. Myconception of practical reason contrasts with the instrumentalist construalthat has been dominant within naturalistic approaches to morality: theconception of practical reason as incapable of dictating ultimate ends butrather a formal faculty for guiding the transitions from basic, exogenous(relative to reason) motivations to nonbasic motivations and ultimately toDOI: 10.1017/S0265052508080096© 2008 Social Philosophy & Policy Foundation. Printed in the USA. 237

actions. On the instrumentalist conception, normative demands on anagent ultimately derive from some motivation the agent already has.

There is a connection between the way that the apparent inescapabilityof moral demands poses an explanatory problem for naturalistic approachesto morality and the dominant construal of practical reason. I subscribe toa naturalistic approach distinguished by three features: first, it seeks tounderstand the role of morality in human life by drawing on the humansciences; second, it rejects any a priori method for yielding substantivetruths shielded from questioning informed by empirical evidence; andthird, it refuses to take moral judgments, even ones we are stronglyinclined to believe as true, as corresponding to some sui generis portion ofthe universe, existing independently of human needs, feelings, and will.Right away, problems arise from such a naturalistic perspective when welook at the third feature alongside the apparent inescapability of moraldemands. What kind of normative content can apply to agents regardlessof their inclinations if there is no moral reality existing independently ofsuch inclinations? The dominant instrumentalist conception of practicalreason would seem to eliminate the possibility that even if there were anysuch normative content it could necessarily provide everyone reasons tobehave morally. Rather, the existence of reasons contingently depends onthe particular inclinations and will of the agent.

A classic example of the conclusion a naturalist might draw from thisdilemma is Philippa Foot’s 1972 essay “Morality as a System of Hypo-thetical Imperatives.” She accepts that whether one has reasons to con-form to a moral demand depends on whether one has the appropriatedesires or inclinations that would be served by conforming to the demand.On Foot’s instrumentalist interpretation of moral reasons, Immanuel Kantis simply wrong to hold that, regardless of the content of one’s desires, itis irrational to disregard one’s moral duty. It is true, she grants, that moraldemands are nonhypothetical in applying to people whether or not theyhave any desires that would be served by being moral. Foot, however, hasanother way of explaining how moral demands are nonhypothetical. Theyare nonhypothetical in the same way that the demands of etiquette are.1

The demands of polite behavior apply even to those who have no inter-ests served by being polite. Rude behavior remains subject to criticismfrom the standpoint of the institution of etiquette. To say of someone thathe is rude does not mean, according to Foot, that he has a reason to bepolite or that he is irrational to be rude.

The same is to be said of morality, except that morality is taught withmore emphasis. In general, it matters more to others that one should be

1 Philippa Foot, “Morality as a System of Hypothetical Imperatives,” Philosophical Review81 (1972): 305–16. In her later work, however, Foot takes a very different position on theforce of moral demands. See her “Does Moral Subjectivism Rest on a Mistake?” OxfordJournal of Legal Studies 15 (1995): 1–14; and her Natural Goodness (Oxford: Clarendon Press,2001), chap. 4.

238 DAVID B. WONG

moral than that one should be polite, and this accounts for the feeling ofinescapability we generally attach to moral demands. Kant, in Foot’sview, wrongly infers from this kind of inescapability that the moral law isa universally and necessarily binding law of practical reason. We can seethe error in this kind of inference by recognizing that, in terms of scopeand conditions for application, etiquette possesses the same kind of inescap-ability as morality. But no one supposes that there is a law of etiquettethat is a binding law of practical reason. Thus, Foot preserves a kind ofinescapability for moral demands by divorcing it from the having ofreasons to conform to them.

Richard Joyce, in contrast, argues that moral judgments purport to havea normative grip directly on the agent. To grant that a moral demandapplies to an agent and yet to deny that she necessarily has a reason tosatisfy it sounds very odd. This is to deprive moral demands of thepractical clout or “oomph” they purport to have, to invoke Joyce’s felic-itous technical term.2 Joyce correctly observes that Foot could have refor-mulated her position to admit the existence of reasons to be moral that donot depend for their existence on the agent’s having the right kind ofdesires. These are reasons offered by morality as an institution, such that,given its rules, for example, everyone has a reason not to harm innocentpeople. However, Joyce argues that this kind of reason does not supplythe practical clout that we seem to attribute to moral demands in every-day moral discourse.

Consider the parallel case of etiquette. Everyone has reasons to observethe rules of polite behavior in the sense that the rules of the institution ofetiquette require the various forms of polite behavior. Someone who caresnot at all for the institution of etiquette can disregard the reasons it offersto wield one’s dinner knife in the polite manner. The fact that reasons ofetiquette lack practical clout may be an acceptable result, but it is notacceptable in the case of morality because we seem to expect moral rea-sons to have this clout.3

Moral reasons seem to be genuine in an institution-transcending sense,Joyce observes, such that they seem to yield genuine deliberative consid-erations for those to whom they are ascribed. But in the end, Joyce sug-gests, the appearance may be an illusion. Since Joyce does not see hownonhypothetical reasons could transcend institutions and the motivationsof agents to whom they are ascribed (and I think it is fair to say that hisversion of a naturalistic perspective has something to do with this), he isprepared to attribute a fundamental error to moral-language users. Inmaking moral demands on each other, we presuppose a kind of transcen-dent normativity that does not exist: there are no nonhypothetical reasonsto be moral that are external to or transcend those specified by human

2 Richard Joyce, The Evolution of Morality (Cambridge, MA: MIT Press, 2006), 202–3.3 Ibid., 194.

CONSTRUCTING NORMATIVE OBJECTIVITY 239

institutions and the existing motivations of agents. It should be noted,however, that Joyce’s version of error theory, unlike J. L. Mackie’s version,does not end with a definite dismissal of practical clout, but rather withthe expression of pessimism that it will ever be vindicated. Unlike Mackie,who flatly denies that moral reasons have the kind of practical clout thatis attributed to them in everyday moral discourse, Joyce simply doubtsthat we will ever find a way of explaining how moral reasons could havesuch clout.4 His position is closer to a kind of moral agnosticism, ratherthan being a straightforward embrace of Mackie’s moral atheism.5

But why do we presuppose, in everyday moral discourse, that moralreasons do have practical clout? Joyce’s explanation begins with the ideathat moral beliefs promote helping, cooperative behavior and regulateinterpersonal relationships in such a way as to avoid great harm as wellas produce significant benefit for all concerned. He points to recent devel-opments in evolutionary theory that could explain why helping and coop-erative behaviors could have a genetic basis. Joyce’s further hypothesis isthat a moral sense, one of the functions of which is to produce guilt whenone fails to be helpful or cooperative, could have an innate basis. Thismoral sense presupposes the concept of moral reasons that transcendinstitutions and apply independently of the individual’s motivations. Moralconscience, conceived in this way, might provide the extra “oomph” weneed to be helpful and cooperative even in circumstances where being somight seem to go against prudence. In the long run, being a steadycooperative partner could enhance one’s reproductive fitness, suggestsJoyce, expanding upon a theme taken from Robert Frank.6 The practicalclout of moral reasons, in other words, may be an adaptive illusion.

How compelling is Joyce’s response to Foot? Joyce is correct, it seemsto me, in viewing Foot’s account of the inescapability of moral demandsas deflating in its effect. On Foot’s account, moral reasons could apply toall agents, but the sense in which they could apply seems purely formal.Yet there is something unsatisfying about Joyce’s explanation of why wehave a concept of transcendent moral reasons. According to him, themistaken belief that such reasons have a normative grip on us indepen-dently of our existing motivations serves the function of nudging us

4 Ibid., 223.5 J. L. Mackie, in Ethics: Inventing Right and Wrong (Harmondsworth: Penguin, 1977),

holds that it is a presupposition of moral discourse that objective values exist. However,such values cannot exist, he argues, because they are supposed to have a “categoricallyimperative element” that is “objectively valid” (ibid., 29). He means that objective values aresupposed to give anyone reasons to behave in certain ways —reasons that are not contingenton having inclinations that would be served by so behaving. Therefore, Mackie’s position isthat moral demands purport to give us noncontingent reasons to behave in certain ways, butsuch reasons cannot exist. Joyce’s position is more that no one has given a plausible way ofexplaining how such reasons could exist. He stops short of declaring that there is no suchway.

6 Joyce, The Evolution of Morality, 121. See also Robert Frank, Passions within Reason: TheStrategic Role of the Emotions (New York: Norton, 1988).

240 DAVID B. WONG

toward cooperation when prudence might move us the other way. Butwhy should the belief in transcendent reasons have such a motivatingeffect on us? Joyce says that guilt over the thought of acting against suchreasons might nudge us toward acting according to them, but why feelguilt if we can make no sense of their normative grip? Either the beliefitself is sufficient to motivate us, which does not fit very well with Joyce’snaturalism,7 or it is simply the guilt that motivates us, but then the moti-vating effect of guilt suggests there is some prior motivation to act onbehalf of others in the first place.

On two crucial points, Joyce is right. First, we do treat moral demandsas if they are not contingent on their serving the agent’s existing desires(on this, he and I agree with Foot). Second, we do treat moral demands asif they provide reasons for action (on this, he and I disagree with Foot). Itsounds odd to say that Joe morally ought to stop cheating on his taxesand that he has no reason to do so. It sounds downright contradictory tomy ears to say that Joe ought to stop cheating and that he has no moralreason to do so. Ought-to-do’s and duties imply reasons for those whoought to and who have duties.

The next crucial question is, “What kind of reason is implied by moralought-to-do’s and duties?” The reason is clearly a justifying reason. Withinthe moral perspective, to say that Joe has a moral reason to stop cheatingon his taxes is to say that there is something about his situation thatjustifies his ceasing to cheat: for example, the fact that he would be ben-efiting from services funded by other taxpayers but not contributing tothese services. Perhaps this is what Joyce means when he says that every-one is supplied reasons by the institution of morality. Joyce thinks thatsomething more than a justifying reason from the moral perspective isneeded for moral demands to have the practical clout ordinarily attrib-uted to them. Otherwise, a reason that justifies an action from the pointof view of etiquette could have as much of a normative grip on the agentas a moral reason. As to the nature of what more is needed, Joyce seemsto have in mind something that could license saying to the agent, “You

7 Naturalists generally tend to accept David Hume’s thesis on the motivational inertnessof reason: a belief alone cannot move an agent; it can do so only in conjunction with somemotivation that has the structure of desire. On Elizabeth Anscombe’s way of distinguishingdesire from belief, according to which desire aims to make the world fit with it rather thanaiming to fit the world, see Anscombe, Intention (Oxford: Blackwell, 1957). I think themotivational inertness thesis should not be an a priori feature of naturalism, but rathershould be accepted on the grounds that it is part of more fruitful explanations of humanaction. Antonio Damasio’s work on what goes wrong with the practical reason of certainbrain-damaged patients is a start at vindicating the empirical fruitfulness of Hume’s thesis.Damasio’s patients are distinguished by the normality of their theoretical reasoning, evenabout practical matters, and by the abnormal inability to act on reasoning, even in mattersof prudential reasoning about which they are able to arrive at sound judgments in theabstract. The damage is to the parts of their brains that process emotions, and Damasio’sthesis is that practical reasoning, in order to be effective, needs to have various actionscenarios marked by bodily processes that correspond to emotion. See Antonio Damasio,Descartes’ Error (New York: Avon Books, 1994), 50.


are making a mistake if you don’t take this reason into account in yourpractical deliberation.” And the kind of mistake is not merely going againstthe norms of an institution, whether the institution is that of etiquette ormorality.

The intuition behind Joyce’s objection to Foot is that reasons addressedto an agent, if they have legitimate normative force, must go deeper thanbeing based in an institution that claims some kind of authority over theagent. They must go deeper, I think, in three ways. First, the normativeforce is conceived as rooted in something that is not something purelyexternal to the individual. It is not just, “They judge me to be mistaken ifI do not conform to this demand.” Second, the root is not the same asprudential normativity. The normative force is not derived, for example,from concern about being punished by disapproving others. Third, thereis a practical point to whatever normative force moral reasons possess,and a practical point that concerns the agent. Otherwise it’s just “brow-beating,” to use Bernard Williams’s term.8 That is, it must be possible forthe agent to act on the basis of recognizing them as moral reasons.

We are now in a better position to understand why Joyce thinks he hasto resort to error theory. Moral reasons purport to have a deeper grip onthe agent than seems explicable on the basis of an instrumentalist theoryof reasoning. It seems as if we do believe in transcendent, nonhypotheti-cal reasons that apply to the individual externally to his motivationalsystem, yet the instrumentalist theory, on a naturalistic view, seems to bethe clearest theory we have of how reasons have a grip on us. We must beerring in conceiving moral reasons to be nonhypothetical, and Joyce islooking for an explanation of why that error is committed. He suggeststhat the error has a use, and if it is not instrumental to the individual’smotivation, perhaps it is instrumental to the “purposes” of natural selection.

This explanation does not work well, as I have argued. I will offer analternative explanation that affirms the existence of reasons to be moralthat can get a normative grip on the individual’s motivations, even whenthe reasons cannot be explained in terms of what serves the individual’sdesires and inclinations. I will argue that there is a good naturalisticexplanation for their existence and of how they could come to motivateindividuals whose existing interests are not served by acting according tothem. To see our way clear to such an explanation, we must let go of oneassumption that has been dominating the discussion among philosopherssuch as Foot and Joyce: the assumption that the apparatus of practicalrationality exercises normative force only on the individual who is alreadyconstituted as an agent and an individual self. In contrast to this assump-tion, I would argue that the learning of moral reasons must be recognized

8 Bernard Williams says this about the idea of applying a reason requiring an individualto act even though that reason is “external” to the individual’s subjective motivational set.See Bernard Williams, “Internal and External Reasons,” in Williams, Moral Luck (Cambridge:Cambridge University Press, 1981), 101–13.

242 DAVID B. WONG

as going into the very constitution of the agency and selfhood of theperson. That is, moral reasons go deeper by influencing the very structureof motivation and by enabling practical deliberation itself. This mightsound constructivist and Kantian. To the first, I plead a hearty “guilty.” Tothe second, I plead guilty if there can be a naturalistic and relativisticKantianism. So that might be “not guilty.”

Section II lays the groundwork for an explanation of the existence ofnonhypothetical reasons to be moral by explaining how moral normsplay a crucial role in making cooperation possible by shaping and reinforc-ing the diverse and potentially conflicting array of human motivations sothat they are better suited for cooperative life. Section III sets out the waythat moral norms are related to moral reasons, and also sets out how aparticular set of moral reasons gets established as part of the morality ofa group. Section IV explains how moral reasons so characterized help toshape motivation by becoming “embedded” in motivational propensities.Section V argues that given the role that moral reasons play in shapinghuman motivation and guiding practical deliberation, moral reasons gointo the construction of persons, that is, into the construction of who theyare as persons. This last theme leads to a naturalistic explanation of howmoral reasons can have a deeper normative grip on the individual thancan be explained by an instrumental conception of practical reason. Moralreasons do not simply answer to the desires and inclinations that make uswho we are; they help to shape and are embedded in the desires andinclinations that make us who we are, and they can be transcendentreasons in that sense. The dominance of the instrumental conception ofreason makes it harder to see that moral reasons are not simply based onindependently existing desires and inclinations. Section VI attempts toundermine this dominance by arguing that the concept of a self thatextends over time is a concept constructed to meet the demands of socialcooperation. Prudential reasons are reasons to act on behalf of the self asit extends into the future and are often taken to constitute the paradigmof reasons that are instrumentally based on desires and inclinations of theindividual. But such reasons, like moral reasons, are constructed to pro-mote human cooperation and go into shaping the individual’s motiva-tions, and are not merely based on the individual’s motivations.

II. Some Scene Setting: The Cooperative Nature of HumanLife and the Role of Moral Norms in Making This Possible

Let me begin by sketching a picture of the diverse array of motivationalpropensities that are plausibly considered to be innate to human beings.9

9 Most but not all of the rest of this section draws from ideas expressed in chapter 2 of myNatural Moralities: A Defense of Pluralistic Relativism (New York: Oxford University Press,2006).


Alongside their instincts for self-preservation, members of the humanspecies (or many of them)10 developed capacities of care for kin, a will-ingness to engage in mutually beneficial practices of cooperation withothers if they show a similar willingness, a willingness to punish thosewho violate the agreements and norms that make cooperative practicespossible (even when the expenditure of resources to punish cannot bejustified on the grounds of pure self-interest), and some degree of altru-istic concern for nonrelated others. Human beings developed all thesecapacities because they were fitness enhancing in an inclusive sense, aconclusion that much of the latest work in evolutionary theory sup-ports.11 While human beings certainly evolved as strongly self-interestedcreatures, they also evolved motivational capacities that allowed them toform cooperative bonds with each other. Human beings, then, were selectedto have a diverse array of innate psychological tendencies that can poten-tially come into conflict with each other if they do not have ways ofregulating and tempering the expression of these tendencies. Humannature can be profoundly ambivalent in this way.

Cultural norms, I suggest, play a large part in this structuring of moti-vation. The human capacity to regulate the self through cultural normscoevolved with biological traits. The long period of the Pleistocene duringwhich human beings evolved social instincts overlapped considerablywith the period in which people began living in social groups with cul-tural institutions. If culture was a partner in this biological evolution, thenit is plausible to hypothesize that some of our biological traits, as anthro-pologists Robert Boyd and Peter Richerson have suggested, might pre-pare us to regulate ourselves through culture: for example, the dispositionto follow the majority or to emulate the most successful members of one’sgroup. (The latter strategy requires people to be selective imitators and tohave at least an inkling of what good solutions to common problemsare.)12 Such traits could have conferred an evolutionary advantage on

10 The reason for this qualification will surface later in this essay.11 On the selection mechanism for altruism toward kin, see W. D. Hamilton, “The Genet-

ical Evolution of Social Behavior,” Journal of Theoretical Biology 7 (1964): 1–52. On the mech-anism for selecting a willingness to cooperate with others if they show willingness tocooperate, see Robert Trivers, “The Evolution of Reciprocal Altruism,” Quarterly Review ofBiology 46 (1971): 35–56. For a theory of “group selection” as the mechanism behind concernfor non-kin, see Elliott Sober and David Sloan Wilson, Unto Others: The Evolution and Psy-chology of Unselfish Behavior (Cambridge, MA: Harvard University Press, 1998); and foranother theory emphasizing the role of sexual selection in altruism, see Geoffrey Miller, TheMating Mind (New York: Anchor Books, 2000). For evidence supporting the existence ofnon-self-interested willingness to punish, and to reward others who cooperate, see HerbertGintis, Game Theory Evolving (Princeton, NJ: Princeton University Press, 2000).

12 Robert Boyd and Peter J. Richerson, Culture and the Evolutionary Process (Chicago: Uni-versity of Chicago Press, 1985); Boyd and Richerson, Not by Genes Alone: How CultureTransformed Human Evolution (Chicago: University of Chicago Press, 2005). In the latter book,they define “culture” as “information capable of affecting individuals’ behavior that theyacquire from other members of their species through teaching, imitation, and other forms ofsocial transmission.” By “information,” they mean “any kind of mental state, conscious or

244 DAVID B. WONG

members of a group by enabling them to adopt satisfactory solutions toproblems that were worked out by other members. Individuals do notneed to “reinvent the wheel” on their own but can instead follow culturalnorms established over some period of time, provided that the group’senvironment changes slowly enough so that the solutions embodied inthose norms remain satisfactory.

Moral norms, on my view, culturally evolved to promote beneficialsocial cooperation, not simply through requiring behavior that is coop-erative and considerate of the interests of others, but also through encour-aging, strengthening, and directing the sorts of feelings and desires thatmake people promising partners in social cooperation.13 A virtue of thisfunctional conception of moral norms is that it helps to organize andsystematize many of the most central moral beliefs that appear acrosscultures and historical periods: beliefs that specify the conditions forpermissibly killing or conducting aggression against other human beings;beliefs about the right to assign and distribute the basic resources neededto sustain life; and beliefs that require reciprocation of good for good.There is a lot of variation in how these beliefs are filled in with specificcontent and in the nature of the particular restrictions and distributions,but a common end these beliefs serve is the regulation and promotion ofsocial cooperation.

Some prominent functionalist accounts of morality claim that the pri-mary purpose for which morality is invented is to counteract the destruc-tive effects of self-interest (e.g., Thomas Hobbes) or the limitations on oursympathies for others (e.g., G. J. Warnock, J. L. Mackie).14 I favor a morecomplex functional picture, under which the profound ambivalence ofhuman nature is managed in a variety of ways and not just throughconstraint of potentially destructive self-interest. Moral norms need totake into account the strength of self-interest in order to accommodate

not, that is acquired or modified by social learning, and affects behavior” (Not by GenesAlone, 5). I agree with Dan Sperber and Nicholas Claudière (“Defining and ExplainingCulture,” Biology and Philosophy 22 [2007]) in their observation that Boyd and Richerson’suse of “information” is too broad in one sense —culture is better taken to include only widelydistributed information —and too narrow in another sense —the relevant kind of informationcan be implemented not only in the form of mental representations but also in the form ofbehaviors, artifacts, and institutions.

13 This does not mean that directly facilitating social cooperation is the only function ofmorality. Some moral norms take the form of character ideals and conceptions of the goodlife specifying what is worthwhile for the individual to become and to pursue. This intra-personal function of morality comprehends what has been called the “ethical,” as opposedto what might be called the “narrowly moral.” Morality in the broader sense used herecomprehends the ethical. This part of morality helps human beings to structure their livestogether in a larger sense, i.e., not just for the sake of coordination with each other but alsofor the sake of coordination within themselves.

14 See Thomas Hobbes, Leviathan, part 2, chaps. 13–17; G. J. Warnock, The Object of Morality(London: Methuen, 1971); and J. L. Mackie, Ethics: Inventing Right and Wrong (London:Penguin, 1977), chap. 5.


that motivation and to encourage its integration with motivations thatmore directly lead to acting on behalf of others. Effective moralities,then, do not merely restrain actions from self-interest or encourage thedevelopment of opposing motivations, though they do these things.Effective moralities provide outlets for the expression of self-interestthat can be consistent with the expression of other-regarding motiva-tions. Self-interested motivation can clearly have undermining effectson social cooperation when it motivates noncooperation and aggressionagainst others. However, in the right circumstances self-interest cansupport, rather than oppose, other-regarding motivations.15 Arrange-ments that generate some self-interested return to other-regarding behav-ior can create an “ecological niche,” Jane Mansbridge suggests, thathelps to sustain that behavior. By making that behavior less costly,these arrangements can increase the degree to which individuals feelthey can afford to indulge their concerns for others.16 Rather than say-ing that an effective morality should always constrain self-concern andreinforce other-concern, I would suggest that it should often attempt toaccomplish a productive balance or reconciliation between those typesof concern.

Norms of reciprocity that require a return of good for good receivedplay a crucial role in such reconciliation. The need to reconcile self- andother-concern appears first in family relationships. Across widely differ-ent cultures, there are duties to respect and to honor parents and otherswhose roles involve raising and nurturing the young. Performance ofsuch duties constitutes a kind of return of good for good, though what isreturned, of course, is not always the same kind of good as what wasoriginally given. Sometimes the return is similar to the original good, asin the case of children’s care of aged parents. Most other times, however,the return is a good that is fitting to the nature of one’s relationship tothose who have cared and nurtured: obedience and receptiveness to whatis taught, for example. Perfectly selfless parents might not need suchreinforcement, but profoundly ambivalent beings might not be able to dowithout it. Note, however, that the benefit of being reciprocated for an actof helping need not be greater than the cost of helping in order to reinforcethat initial act of helping. It need merely generate enough return so thathelpers feel they can afford to indulge whatever other-directed concernsthey have.

Furthermore, we humans cooperate on a scale that is far larger than kingroups and the small groups in which it is feasible to directly reciprocate

15 See Gintis, Game Theory Evolving.16 Mansbridge has suggested that while other-regarding motivations such as those stem-

ming from empathy do exist in most individuals, they do not have infinite value. HerbertGintis’s portrait of Homo reciprocans suggests the same qualification. If the costs of benefitingothers are very high, many will simply decline to pay. See Jane Mansbridge, “On theRelation of Altruism and Self-Interest,” in Mansbridge, Beyond Self-Interest (Chicago: Uni-versity of Chicago, 1990), 133–43.

246 DAVID B. WONG

to those who have conferred benefits upon us.17 Moral norms play a keypart in the shaping of motivations appropriate to such widely ranging coop-eration. Moral norms, for example, might encourage the favorable eval-uation of those who engage in helping and cooperative behavior, regardlessof whether one has directly benefited from that behavior. In what has beencalled “indirect reciprocity,” others become favorably disposed through rep-utation toward the helper as a potential cooperative partner.18 A moralitycan help this kind of reinforcement via indirect reciprocity through plac-ing value on trustworthiness. Learning who is trustworthy is learning toassess others for their reliability as potential cooperative partners, and com-munication of judgments of trustworthiness helps to sort the cooperatorsfrom those free-riders who move from group to group to avoid identifi-cation and punishment. Gossip may have a moral function after all, andthe negative aspect of that function might be at least as important as thepositive aspect. There is evidence that cooperators tend to punish free-riders by ceasing to cooperate with them even if noncooperation costs moreto the cooperators than cooperation with free-riders.19

Moral norms direct the manner of expression of self- and other-concernso that they are more compatible. Such norms reinforce and strengthenother-concern. The performance of these functions is crucial for making ago of human cooperative life. In the next section, I explain how moralnorms are related to moral reasons.

III. What Moral Reasons Are: Their Relation to MoralNorms and Their Place in Moralities

Reasons are those considerations weighing in favor of or against anagent’s doing something.20 They are structured as three-place relations

17 Explanations of altruism toward non-kin that rely solely on the idea of natural selectionover genetic variations rely on rather special conditions being set in place. For example, CharlesDarwin (in The Descent of Man) thought that natural selection sometimes operates on groupsand not just on individuals, so that in the case of human beings, a tribe with members willingto sacrifice for other members will prevail in competition with other tribes with no such mem-bers, or will do well in adverse natural circumstances, and will therefore gradually predom-inate among the human species. This explanation, however, depends on groups’ preservingthe genetic differences between them as the ones with greater proportions of altruists win outover the ones with lesser proportions. Members of a winning group who intermarried withmembers of a losing group would undermine this special condition, but this seems to havebeen a common occurrence. In contrast, the tendency toward conformity with cultural normsmight preserve intergroup variation and allow groups with prosocial cultural norms to winout over others and continue adherence to those norms even in the face of intermarriage withmembers of other groups. See Peter J. Richerson, Robert T. Boyd, and Joseph Henrich, “Cul-tural Evolution of Human Cooperation,” in Genetic and Cultural Evolution of Cooperation, ed.Peter Hammerstein (Cambridge, MA: MIT Press, 2003), 368–69.

18 Richard Alexander, The Biology of Moral Systems (New York: Aldine de Gruyter, 1987).19 See Gintis, Game Theory Evolving. Robert Trivers, in “The Evolution of Reciprocal Altru-

ism,” identified a crucial role for “moralistic aggression” (negative reactions to perceivedviolations of reciprocity) in helping to reduce the incidence of free-riding. However, it isGintis who correctly points out that in many instances there is an altruistic element to thewillingness to retaliate against free-riders.

20 This section summarizes ideas expressed in chapter 2 of my Natural Moralities.


between an agent A, an action X, and a feature F in the agent’s situationthat weighs in favor of A’s doing X. For example, A may have a reason tohelp B in virtue of B’s being in imminent danger of being harmed and A’sbeing able to help with no risk and low personal cost to herself. As Iindicated in Section I, a reason defined in this way is a justifying reason.F is the feature that purportedly justifies A’s doing X. A justifying reasonis not necessarily one that would motivate A to do X. It is not necessarilya motivating reason, but I will explain in Section IV how a justifyingmoral reason can become motivating.

We may think of reasons in general as providing the basic vocabularyfor thinking and talking about what to do. Judgments as to what a personmorally ought or ought not to do or as to what is morally right or wrongfor that person to do can be construed as judgments as to what relevantmoral reasons require for that person. To think about what we morallyought to do, we need to identify the moral reasons relevant in the givensituation, and in cases where more than one situational feature is relevantand the different features weigh in favor of different and incompatibleactions, we need to judge what the balance of reasons requires.

Moral norms guide our thinking about what reasons require. Some-times they simply identify and articulate kinds of reasons, such as reasonsto do what one has promised or agreed to do, or reasons to tell the truthin communicative contexts where there is a standard expectation of truth-telling. Sometimes norms provide guidance in thinking about what the bal-ance of reasons requires, as in a situation where there are reasons againstkilling other human beings and reasons permitting one to save one’s ownlife. The relevant norm permits killing in self-defense. Sometimes a normmarks the relative centrality of a value to a morality, such as the norm requir-ing respect for individual liberty and autonomy or the norm that empha-sizes the realization of one’s humanity in relationship to others.

The content of reasons (e.g., which situational features go into theidentification of what we have moral reason to do) and of norms (e.g.,how to prioritize in cases of conflicts between reasons) is constrained bymorality’s function of promoting and sustaining cooperation. That is, agenuine moral reason is such that acting on it under the appropriatecircumstances contributes to this function of morality. (The contributionthat acting on any single reason might make to morality’s functions mighthave to be assessed according to the ways in which it works with otherreasons recognized within that morality.) Receiving the help of anotherperson, for example, would constitute a reason to reciprocate. The moral-ities of particular societies yield a diverse and rich set of moral reasonsthat operate on much more specific levels (e.g., the forms of help thatwould create duties to reciprocate and the appropriate forms of recipro-cation would be specified more concretely), but to have a chance at beinggenuine moral reasons, the reasons specified within these moralities mustsatisfy the general constraints.

248 DAVID B. WONG

How does a set of specific moral reasons get established as part of themorality of a given group? I approach this question as a question aboutthe concepts of moral reasons and how they acquire the reference theyhave.21 Much recent work on the nature of concepts has undermined the“classical” model that posits necessary and sufficient conditions for theirapplication. One of the proposed replacements for the classical model isprototype theory, according to which concepts include features possessedby their instances, the features embodying the average or most typicalinstances. To take a frequently used example in the prototype literature,the concept of dog includes features making up a kind of composite“everydog” (has four legs, a tail, emits barking sounds), and an object thatis a candidate for falling under that concept is more likely to qualify themore it resembles the composite typical dog.22 Another proposed replace-ment for the classical model is exemplar theory, which holds that havingconcepts involves the ability to call up particular instances that serve asthe standards of comparison for candidate instances. Having the conceptof a dog involves the ability to call up from memory particular dogs onehas encountered, and one compares dog candidates to the closest exem-plars to see if one gets a close enough match.

Concepts need not be limited to one structure. Indeed, some conceptsmight involve the acquisition of prototypes that are constructed on thebasis of exemplars. A child might acquire her prototype of everydog onthe basis of encounters with particular dogs she knew while growingup.23 She might call up the dog prototype to categorize most dogs sheencounters, but if she were to encounter a difficult case, she might recallan atypical dog exemplar that most closely resembles the present animal.Concepts of moral reasons seem to exhibit this kind of versatility. Con-sider the concept of a reason to help another person. In acquiring thisconcept, we might have had certain concrete situations identified as exem-plars of a reason to help: a parent demonstrates for us what is to be donewhen a sibling falls down and hurts himself; the experience of an exchange

21 Much recent discussion of truth and reference has employed the language of conceptswhere the language of meanings has previously held sway. I prefer the language of conceptsbecause it has become the language for expressing and defending alternatives to the “clas-sical” model which holds that a term (or the concept expressed by it) refers by virtue of aset of necessary and sufficient conditions embodied in the concept. Furthermore, much ofthe empirical evidence undermining the classical model has come from studies in cognitiveand developmental psychology of ways that people categorize things, and these studies arecouched in the language of concepts. See Eleanor Rosch and Carolyn B. Mervis, “FamilyResemblances: Studies in the Internal Structure of Categories,” Cognitive Psychology 7 (1975):573–605.

22 See Jesse Prinz, Furnishing the Mind: Concepts and Their Perceptual Basis (Cambridge,MA: Bradford Books, 2003), 51–72.

23 See Andy Clark, “Connectionism, Moral Cognition, and Collaborative Problem Solv-ing,” in Mind and Morals: Essays on Ethics and Cognitive Science, ed. Larry May, MarilynFriedman, and Andy Clark (Cambridge, MA: Bradford Books, 1998), 109–13. Jesse Prinz’s“proxytype” model of concepts is supremely eclectic in incorporating prototype, exemplar,and other models of concepts. See Prinz, Furnishing the Mind, chap. 6.


of gifts or favors is in many cultures an occasion exemplifying a reason toreciprocate. Alternatively, people can construct prototypes of such rea-sons that generalize over these exemplar cases, giving rise to a conceptualrepresentative of a “typical” reason, say, to help or to reciprocate. Onmany occasions of classification, we might call up a prototype if thecurrent situation seems typical, but in novel or borderline cases, we mightcall up from memory the closest exemplar and try to determine whetherthere is a close enough match. Peter Singer’s argument that the relativelyaffluent have a strong duty to help famine victims drew much of itspower from his analogy to the duty to save a drowning child if doing somerely required wading into a shallow pool and ruining one’s clothes.24

Saving the life of a child at very low cost to oneself may be an exemplarof what we have very strong reason to do (or very similar to such anexemplar), and Singer’s argument is that helping famine victims at com-paratively low cost to oneself is a very close match to the case of savingthe child.

Recall that moral norms specify the situational features that go into theconstitution of reasons (e.g., the situational feature of having receivedhelp is specified in a moral norm as a reason to reciprocate in somemanner) and also specify the priorities among reasons in case these rea-sons require incompatible actions in a given situation (conflict may occurbecause a situation contains several morally relevant features that call forincompatible actions). People specify reasons and decide on priorities,and hence construct these norms, as they develop their cooperative lifetogether and as they reflect on how well or how badly that life goes. Thenorms that emerge and get accepted within a group establish the truthconditions for moral judgments made by its members, since these judg-ments are equivalent to judgments about what there is reason to do orwhat the balance of reasons requires in a given situation. Therefore, themoral truth is constructed by human beings as they construct the normsthat specify what there is moral reason to do, but the task of constructionis subject to constraints on adequate moralities that spring from humannature and the functions of morality.

I have already implied one such constraint: a morality that lackednorms requiring reciprocation for help received, for example, would be awoefully inadequate morality, given the functions of morality and giventhe way human beings are. Another constraint follows from the way thatmorality works through a large degree of voluntary acceptance of itsnorms. If conformity to its norms and reasons depended solely on thethreat of force or coercion, the costs would detract greatly from the ben-efits of cooperation. It makes sense that human beings evolved a systemfor regulating and promoting cooperation that governs in this way. A

24 Peter Singer, “Famine, Affluence, and Morality,” Philosophy and Public Affairs 1 (1972):229–43.

250 DAVID B. WONG

further step in the evolution of morality also makes sense for creatureswho explain and justify their actions to one another: voluntary acceptanceof moral norms came to be seen as based on their justifiability to thosegoverned by them. Hence, another constraint on moralities is that justi-fication for following the norms and reasons of an adequate moralitycannot crucially depend on falsehoods. Norms that permit the subordi-nation of the interests of some members to those of others often violatethis constraint because they are rationalized on the basis of false presup-positions about the supposedly inferior capacities of the subordinated forrational thought and self-control.

I have identified a couple of the ways in which the set of norms thatgroups actually construct could fail to satisfy the constraints on morali-ties. A group could fail to include norms that are necessary for the pro-motion of cooperation. The justification the group gives to some normscould be false or inadequate. When members of a group judge what thebalance of moral reasons requires in a given situation, they invoke thenorms the group has constructed that specify what moral reasons thereare and the kinds of priorities that obtain among them. They invoke thesenorms on the assumption that they are adequate and meet whateverconstraints apply to them, and because this assumption can be wrong,there is always room for criticism and debate over the group’s currentnorms and over the truth about what the balance of moral reasons requires.

In this section, I have specified the structure of moral reasons and therole of moral norms in a group’s specification of what moral reasons thereare and the priorities among them. I have given examples of constraintsthat apply to a set of moral norms that stem from the role of morality inpromoting cooperation and from features of human nature. Morality per-forms this role through shaping human motivation. In the next section, Idescribe how moral reasons help to shape motivation.

IV. How Moral Reasons Help to Shape Motivation

The crucial concepts in my explanation of how moral reasons functionin shaping motivation are those of motivational propensities, the inten-tional objects of such propensities, the embedding of moral reasons withinthe intentional objects, and the channeling of propensities through restruc-turing of intentional objects. Motivational propensities are functional statesgrounding dispositions to act or to feel under certain circumstances. Theycan take the form of a felt urge toward an intentional object, as in thirstingafter a drink of cool water (where the object is the action of drinking thewater), but it is not necessary for the intentional object to serve as aphenomenological content of a propositional attitude. That is, it need not bean object of awareness for the agent. When we specify the intentionalobject of a motivational propensity, we are specifying the motivational


direction of the relevant dispositions to act and the way they can form anintelligible bundle of action and feeling tendencies. Human motivationalpropensities are rooted in the imperatives of our biological being, but thebasic directions of those propensities can be refined, rechanneled andreshaped. One of the ways in which these directions get changed is throughour learning what reasons there are for doing things. The mere learningof such reasons does not assure any motivational propensity to act accord-ingly. The recognition of reasons must get embedded in the propensity.25

To support the possibility that the embedding of reasons can serve thepurpose of channeling motivational propensities, let me point to the phe-nomenon of moods that become emotions. One way to distinguish moodsfrom emotions is to say that moods are free-floating. They have no inten-tional object in the way that emotions have —no “aboutness.” One isafraid of something, but one can be anxious about nothing in particular.Free-floating anxiety can easily turn into specific fears. Consider the caseof a man who suffered from chronic feelings of anxiety. After the birth ofhis first child, he became frequently concerned about his child’s safety, tothe point of worrying that his child might one day climb onto the garageroof, fall off, and hurt himself on the stone bench below. This motivatedthe man eventually to hire workmen to break up the bench with sledge-hammers and cart away the rubble.26 This case concerns an excessive fear,even a pathological one (though some of us who are parents may admitto taking only slightly less ridiculous precautions for the sake of ourchildren), but it dramatically illustrates how moods acquire intentionalobjects. Certainly, the man had reason to be concerned about the safety ofhis child: e.g., the child’s vulnerability to harm and lack of awareness orjudgment about what constituted dangerous play. This case also illus-trates how reasons can get embedded in motivational propensities. Thatpart of an emotion that is independent of the intentional object (the free-floating anxiety) can provide the motivational energy behind thoughts(the possibility of one’s child falling onto a stone bench) that serve asreasons for action (having the bench broken up), and it thereby is chan-neled into that specific course of action.

The aforementioned case of the anxious man illustrates the mutabilityof our emotional lives. It has been remarked that flexibility of response ischaracteristic of creatures with emotions and that such flexibility con-

25 The theme of reasons getting embedded in motivational propensities is expressed inchapter 7 of my Natural Moralities, and in my essay “Moral Reasons: Internal and External,”Philosophy and Phenomenological Research 72 (2006). The characterization of embedding in thepresent essay, however, is revised from that previously published material, and is connectedto empirical examples that I have not previously discussed. The application of this theme tothe debate between the Footian and Joycean views of moral inescapability is not discussedin that previous work.

26 This case is described by Gerald L. Clore and Andrew Ortony in “Cognition in Emo-tion: Always, Sometimes, or Never?” in Cognitive Neuroscience of Emotion, ed. Richard D.Lane and Lynn Nadel (Oxford: Oxford University Press, 2000), 27.

252 DAVID B. WONG

ferred some evolutionary advantage on them. If you have a dog, as Ido, who is very interested in chasing rabbits, you might have experi-enced the reaction of rabbits upon spotting a potential predator approach-ing from a distance. A common initial response is to freeze while bloodflows to the muscles in preparation for fleeing. Whether the rabbitcontinues to freeze or runs away seems to depend in part on the rab-bit’s noticing whether the predator has zeroed in and is ready to pounce.Human beings have far greater flexibility of response in analogoussituations. A member of a racial minority walks down the street of anunfamiliar part of town. As he notices the looks from those he isapproaching, there might be a similar preparation for action, but a lotwill depend on how he interprets the looks —as curious in a friendlyway, as hostile, as getting ready to deliver an assaultive remark, and soon. As psychologists Gerald L. Clore and Andrew Ortony observe, “thereseem to be two fingers on the emotional trigger: one controlled byearly perceptual processes that identify stimuli with emotional valueand activate preparation for action, and a second controlled by cogni-tive processes that verify the stimulus, situate it in its context, andappraise its value.” 27 They observe that increased flexibility is con-nected to an expanded capacity for subjective experience, which “reg-isters the urgency of a situation, provides information, and allowsprocessing priorities to be revised.” People “can entertain alternativecourses of action and sample how they would feel about different out-comes,” and in order to do this, “they must be aware of the stimulusthat occasions the processing.” 28 Thus, if our passerby registers looksof sullen but silent hostility, he may simply think to himself, “So what’snew?” and walk on by without a further glance. He may considerthrowing a hard stare back but may feel that the possible escalatedoutcome of such an action would not be worth his own emotionaldisturbance, much less the potential bodily harm. A further bit of com-plexity in this example is that our passerby can, in some sense, decidehow angry he’s going to get. He can decide how he will react to thesituation and knows in advance that certain ways of reacting will esca-late whatever feelings exist on the other side. He knows that he willreact to such feelings, thereby raising the stakes on both sides.

Let me clarify. I am not claiming that emotions simply are cognitions ofa certain sort but rather that cognition can enter into complex relation-ships with affective and conative components of emotions, producingnew emotions, changing existing ones, and channeling the motivationaldirections of those emotions. The interactivity of cognition and emotionallows for an adaptive flexibility of response. This flexibility and capacityfor a greater self-regulation of one’s emotions characterizes human nature.

27 Ibid., 41.28 Ibid.


It is not only a capacity exercised by individuals over themselves, but alsoa capacity exercised by groups over their members.

Here is an example of a use of moral reasons to shape motivation. In1A7 of the Mencius, an account is given of a conversation between Menciusand King Xuan, the ruler of a Chinese state. Mencius is attempting topersuade the king to adopt the Confucian dao or way of ruling by becom-ing a true king to his people, namely a king who tends to their welfarerather than launching his state into wars of territorial expansion, therebyover-taxing and drafting his subjects into his army. The king wonderswhether he really has the stuff to be this kind of king Mencius is describ-ing, and Mencius replies that he thinks so, asking whether the followingstory he has heard about the king is true. The story is that the king sawan ox being led to slaughter for a ritual sacrifice. The king decided tospare the ox and substituted a lamb (that he did not see) for the ritualsacrifice. The king acknowledges that this story is true, and thinking backon that occasion, recalls that it was the look in the ox’s eyes, like that ofan innocent man being led to execution, that led him to substitute thelamb. Mencius then comments that this story demonstrates the king’scapability to become a true king, and that all he has to do is to extend thesort of compassion he showed the ox to his own people.29

The conversation between Mencius and King Xuan portrays, in myview, an attempt by Mencius to embed a reason to care for one’s peoplein the king’s motivations. The king had shown a prior capacity for com-passion in sparing the ox (even if he substituted the lamb, but that isanother aspect of the story I cannot linger on), and Mencius is seizing onthe king’s recall of that past event in pressing for another expression ofthat capacity, but now directed toward his people. The king is no moralexemplar, but he is likely to have already learned that the vulnerability ofhis people to suffering (much of it potentially inflicted by him) is a jus-tifying reason from the standpoint of morality for him to take certainactions and to refrain from others.

What would constitute that reason’s being embedded in the motiva-tional propensities that constitute the king’s compassion? The king wouldsee the potential suffering of his people as a reason to act, and thatrecognition would become associated with so acting, such that alleviatingor preventing the suffering of his people would now be part of the inten-tional object of his compassionate motivational propensities. This doesnot mean that the king’s recognition of his people’s suffering wouldinvariably result in the appropriate action, but if it did not, it would meananother set of appropriate reactions. One sort of reaction would be toquestion himself. He would question whether he had failed to do what he

29 For a translation of this passage, see Readings in Classical Chinese Philosophy, ed. PhilipJ. Ivanhoe and Bryan Van Norden (New York: Seven Bridges Press, 2001), 116. The versionof the text available today is dated from the second century c.e.

254 DAVID B. WONG

had reason to do. Other appropriate reactions would be remorse andactions such as apology or an attempt to compensate those he had failed.Therefore, the embedding of a reason in motivational propensities meansnot only that the agent recognizes the reason, but also that recognition ofthe reason guides the exercise of the propensity in a way it had not beforethe embedding took place. And in cases where recognition of the reasonfails to result in exercise of the propensity, embedding may still showitself in appropriate emotions (such as remorse) or actions (such as apol-ogy or attempted compensation).

How is such embedding accomplished? The answer to this must ulti-mately rely on a great deal of empirical study. Here are some ideas to test.Perhaps reminding the king of a previous exercise of his compassion,prompting him to recall it so that he relives and refeels that occasion,while reminding him that he has reason to alleviate or prevent the suf-fering of his people, is an effective way of trying to embed that reason inhis compassion. Perhaps that is why dry philosophy lectures in norma-tive ethics are so ineffective at bringing about genuine moral change intheir audiences, and why films, narratives, and poetry can be more effec-tive. The embedding of a reason is more likely if the propensity is acti-vated at the time that the reason is being recognized by the agent. Whenthe propensity involves complex emotion of the sort that can interact withnew cognition, we have the possibility of transformation of emotion asthe propensity is channeled.

Much of the most interesting work in the cognitive neuroscience of emo-tion and cognition is still very speculative. Allow me to speculate, how-ever, on the neural mechanisms of embedding, using Antonio Damasio’stheory of somatic markers.30 Somatic markers are bodily reactions to amental image and correspond to emotions. They ultimately derive from ourbiological drives, though they can become greatly modified through expe-rience. A negative marker, attached to the mental image of a possible actionand/or its outcome, can lead one immediately, before any cost-benefit analy-sis, to reject an option, roughly in the way one might recoil from the sightof a snake. Somatic markers can get connected to predicted future out-comes of certain scenarios. In deciding whether to alienate a friend for thesake of financial gain, one might picture the look on his face when he findsout what you’ve done, for example. Positive markers are beacons of incen-tives; that is, they highlight a certain response option. One of Damasio’shypotheses is that when different somatic markers are juxtaposed to dif-ferent combinations of images, they modify the way the brain handles thoseimages, and thus operate as a bias. The bias might allocate attentionalenhancement differently to each component image, the consequence beingthe automated assigning of varied degrees of attention to varied contents.

30 Damasio, Descartes’ Error, 173–89. Damasio is a behavioral neurologist and neuro-scientist at the University of Southern California.


This hypothesis rings true to me phenomenologically and also as anexplanation of how others attach greater weight to certain reasons ratherthan others. Damasio’s picture provides a way of making sense of theoperation of Aristotelian habituation, to which John McDowell referswhen he talks about the acquisition of a “second nature” in the process ofsocialization.31 Consider the example of the negative marking of the sce-nario of a friend’s face when he finds out you’ve betrayed him for thesake of financial gain. The negative marking biases your attention towardthat scenario in considering all the consequences of the betrayal, perhapseclipsing alternative scenarios of what you would do with the money. Itmight be so strong a negative marker that it exhibits the phenomenon thatMcDowell mentions as characteristic of a virtuous person’s deliberations:a moral reason “silencing” other nonmoral considerations that, in othercontexts, might weigh significantly in practical deliberation. The thoughtof betraying a friend “silences” other considerations in altogether divert-ing your attention from the prospect of financial gain that normally weighsin favor of a course of action.

Damasio’s somatic marker theory might be part of the explanation ofwhat happens when moral reasons become embedded in motivationalpropensities. The situational features designated by reasons get somaticmarkers attached to them. If Mencius succeeds with King Xuan, the somaticmarker that was attached to looking into the eyes of the terrified ox getsattached to images of his people suffering from the wars he embarks hisstate upon.

My suggestion in previous sections has been that human beings evolvedto guide themselves through cultural norms that specify reasons andestablish priorities among reasons. In this section, I have given an accountof the role that moral reasons play in the shaping of motivations. In moraleducation and socialization, these reasons become embedded in motiva-tional propensities and render individuals more suitable for the forms ofcooperative activity that characterize human life. By shaping motiva-tional propensities that can become central to who we are, reasons canenter into the kinds of persons we are.

V. The Challenge of Accounting for Normative Objectivity:How Reasons Go into the Construction

of Motivations and Persons

Let me return to the problem of making sense of nonhypothetical moralrequirements that are external to individuals’ existing motivations. Suchrequirements purport to have a deeper grip on the agent than seems expli-cable on the basis of an instrumentalist theory of reasoning, according towhich these requirements must concern actions that serve some existing

31 John McDowell, Mind and World (Cambridge, MA: Harvard University Press, 1996), 84.

256 DAVID B. WONG

motivations of the agent. Philippa Foot responds by construing the require-ments as social demands that individuals only have reason to satisfy whenit would serve some of their motivations to do so. Richard Joyce correctlysees that these demands purport to have a deeper normative grip on indi-viduals, but he does not see how individuals necessarily would have rea-sons to satisfy these demands, independently of their existing motivations.He does not see this because he holds an instrumental theory of rationality.

I have proposed an alternative theory of rationality that includes notonly forms of correct inference from beliefs to beliefs, and from beliefsand desires to desires and intentions, but also substantive conceptions ofwhat people have reason to do. Such reasons are social constructions inthe following sense: the explanation of why we have the moral reasonswe have is that they play a crucial role in shaping our motivationalsystems for the cooperative life. The kinds of reasons we have are not justout there independently of this function they have in enabling human life.Furthermore, they are constructed in the sense that the specific form ofthe reasons we have is a variable matter that is not dictated by constraintson morality following from its functions and human biological traits.These functions and traits certainly do constrain how we attempt to shapemotivation for the sake of cooperation, but they do not dictate specificreasons. For example, moralities are constrained to require reciprocationfor the help of another, but a rich and diverse set of moral reasons specifythe specific circumstances under which one has a reason to reciprocateand in what manner one should reciprocate.

Moreover, I hold that such functional constraints on morality under-determine what a morality must look like in another sense. In a greatnumber of cases, such constraints do not dictate how a morality guidespeople in cases of conflict between important values such as special dutiesto reciprocate to particular others such as members of one’s family (on theone hand) and impersonal duties to care for and act on behalf of strangers(on the other hand). Nor do the constraints dictate whether a morality isto emphasize the value of belonging to and being responsible to a com-munity that is defined by a set of relationships or the value of being leftto choose for oneself the kind of life one is to live independently of thecommunity to which one belongs and its social interests. Moralities dis-tinguish themselves from one another precisely in how they guide peo-ple’s actions in the case of value conflicts such as conflicts between specialand impersonal duties or between community and autonomy. Guidanceis provided by particular value priorities, and these priorities are con-structed in the sense that they are not there independently of the ways oflife that people construct for themselves.32

32 This theme leads to the way in which my theory is relativistic: moralities that areincompatible in the sense of requiring different patterns of action, at least in some types ofsituations, may correspond to different sets of truth conditions established for moral judg-


Constructed reasons and constructed moralities contribute to con-structed persons. Consider again the idea that human biological naturecoevolved with the capacity to act on cultural norms. Moral reasons arean important subset of cultural norms and play an especially importantrole in fitting persons for the cooperative life. Moral reasons fit personsfor such a life if they get embedded in motivational propensities: forexample, propensities regarding how much persons care for others; whichothers they care for, and in what degree and in what ways; and whetherthe special duties of family take precedence over impersonal duties tostrangers (and, if so, under what circumstances).

Of course, such embedding does not always take place, and there maybe a number of causes for this result. One cause stems from the polymor-phism of human motivational traits. For example, one of the hypoth-esized mechanisms by which altruistic traits could be selected is groupselection, under which altruism is selected because it increases the fitnessof some groups over other groups. Such a hypothesis does not require the“winning” groups to have nothing but altruists, but only a relatively highconcentration of altruists compared to their competitors. Thrasymachusmight be the spokesman for the purely self-interested cohort that groupselection might have left in place. Members of this cohort are, no less thanaltruists, shaped by what reasons are constructed in their societies, but intheir case, the most influential reasons get embedded in self-regardingpropensities. Whether they behave morally might depend crucially on thewillingness of others to invest their resources in punishing noncoopera-tors. Another cause for the failed embedding of moral reasons is thatcultures provide and shape individuals with different kinds of reasons —the prudential kind, for instance, as well as much more specific reasons,such as the beauty of an elegantly formed sentence as a reason for learn-ing the craft of writing. Institutions and practices within a society mayhave their own subcultures and may stress reasons other than the moralkind. The degree of importance they put on such reasons can createtensions and outright conflicts with morality. As an increasing number ofanthropologists have recently come to recognize, cultures are typicallynot harmonious and coherent but internally diverse with many differentcompeting strains, containing corresponding differently motivated indi-viduals and significant motivational ambivalence within individuals.33

ment. These different sets of truth conditions may refer to common types of moral reasons,but embody different value priorities in providing instructions as to how to balance andprioritize conflicts between different moral reasons when they apply to the same situation.See chapter 2 of my Natural Moralities.

33 The anthropologist Bradd Shore gives, in a study of Samoa, a striking example of aculture’s internal diversity and of the resulting ambivalence. Following the violent murderof his father, a young man received public counsel from a village pastor in formal Samoanthat he must resist the temptation to avenge his father’s death, and must keep in mind thevalues of peace and harmony and forgiveness. Yet, later, this same pastor, this time incolloquial Samoan, warned the young man that if he failed to kill the murderer of his father,

258 DAVID B. WONG

Consider that many of the most compelling identities that people pos-sess are social-role identities. To have such identities is to be practicallyoriented so as to perform the responsibilities and exercise the rights of therelevant roles. In this regard, social roles provide their own sets of reasonsfor action —situational features to which people inhabiting these roles areprompted to respond in certain ways. Some social roles provide reasonsthat overlap with or are included within moralities; for example, being aparent provides reasons for responding to children’s needs and nurturingtheir capabilities. However, the reasons provided by such moralized socialroles might not always lead to moral action, as when parents zealouslystrive to meet their children’s needs at the cost of neglecting and under-cutting the welfare of other children. Other social roles, such as thosecreated within economic institutions and practices, create even more poten-tial conflict with moral reasons. Social ideals that are championed byeconomic institutions may compete with moral ideals as well as affect thecontent of these ideals.

Consider persons in whom moral reasons are not well-embedded or areembedded in motivational propensities that are relatively weak com-pared to others. Are such persons irrational for not being motivated bymoral reasons? Here I agree with those who think that irrationality con-notes a more blatant error in reasoning than failure to appreciate some ofthe reasons there are (where failure to appreciate means failure of embed-ding in motivational propensities). Irrationality connotes things such asthe embrace of blatant inconsistencies.34 Failure to appreciate moral rea-sons falls into the realm of the unreasonable, where this means some-thing like not being a suitable cooperative partner, but fellows such asThrasymachus would not be particularly disturbed by that label. Moralreasons, then, have a deep normative grip on the individual in being partof the apparatus of practical reason with which individuals are socializedand made into rational agents. But the depth of the grip does not make itsuch that every individual will be irrational in failing to be motivated bythe normative force of moral reasons.

The conception of practical reason defended in this essay bears animportant resemblance to the kind of expressivism defended most prom-inently by Allan Gibbard.35 He denies, as I do too, that reason is purelyinstrumental. Talk about “what makes sense” to do and feel emerges fromthe efforts of human beings to coordinate with one another. Gibbard

he would not be his father’s son. See Shore, “Human Ambivalence and the Structuring ofMoral Values,” Ethos 18 (1990): 165–79. For an influential critique of the older view of cultureas a static, uniform whole, see Renato Rosaldo, Culture and Truth (Boston: Beacon Press,1989).

34 See John McDowell, “Are Moral Requirements Hypothetical Imperatives?” in McDowell,Mind, Value, and Reality (Cambridge, MA: Harvard University Press, 1998), 77–78; andThomas Scanlon, What We Owe to Each Other (Cambridge, MA: Belknap Press, 1998), 25–26.

35 Allan Gibbard, Wise Choices, Apt Feelings: A Theory of Normative Judgment (Oxford:Clarendon Press, 1992).


recognizes that talk about what makes sense or what we have reason todo and feel has the crucial function of shaping motivation for the sake offurther cooperation (though he does not, to my knowledge, tell a storyabout reasons getting embedded in motivational propensities). He doesnot interpret normative talk as straightforwardly cognitive in content. Ihave articulated a view of how reasons-talk can perform an expressivistand shaping function while being straightforwardly truth-functional. Rea-sons are situational features that require of agents certain kinds of actionsor omissions. Moral reasons are situational features identified in accor-dance with morality’s function of promoting and sustaining social coop-eration. Certain reason types will be universal, required by any moralitythat has a hope of performing its function. Others will be local, requiredby a particular morality but not necessarily by all that can perform morali-ty’s function. I have outlined a way that truth-functional talk of reasonscan perform the function of shaping motivation through the embeddingof moral reasons in the intentional objects of motivational propensities.

VI. Constructed Persons and Reasons of Prudence

The dominance of the instrumental conception of reason makes it harderto see that moral reasons are not simply based on independently existingdesires and inclinations. This section attempts to undermine the instru-mental conception’s dominance by arguing that the concept of a self thatextends over time is a concept constructed to meet the demands of socialcooperation. Prudential reasons are reasons to act on behalf of the self asit extends into the future and are often taken to constitute the paradigmof reasons instrumentally based on desires and inclinations of the indi-vidual. But such reasons, like moral reasons, are constructed to promotehuman cooperation and go into shaping the individual’s motivations, notmerely answering to them.

To begin my case for this conclusion, consider Thomas Nagel’s argu-ment in The Possibility of Altruism that the interests of an agent’s future selfprovide her with reasons to act now, even though no presently existinginterest grounds such reasons.36 Failure to recognize such reasons con-stitutes a failure in prudential rationality to recognize one’s future self asequally real as one’s present self. Nagel makes this argument in order toconstruct an analogy. The failure in prudential rationality is like the fail-ure to recognize others as equally real as oneself. My focus here is not onthe analogy but on Nagel’s argument that it is a failure of prudentialrationality to fail to recognize one’s future self as equally real as one’spresent self. Nagel’s argument is ultimately an appeal to our intuitions:Isn’t it evident upon reflection that one’s future self is as equally real asone’s present self? I do not think this thought is so self-evidently true.

36 Thomas Nagel, The Possibility of Altruism (Oxford: Clarendon Press, 1970), chap. 8.

260 DAVID B. WONG

I agree that reasons of prudence include reasons to act now for the sakeof interests we will have in the future, but such reasons are constructedfor the sake of constructing agents who can integrate and balance amongtheir interests over time. To learn that one has reasons to provide for thesatisfaction of one’s future interests is to learn such integration and bal-ance. The question of what kind of balance is right to achieve is trickierthan simply giving each of one’s future selves equal consideration. Onemight think that all the stages of one’s self are equally real and thereforedeserve equal consideration, but the more distant those stages, the lesscertain one is about their existence, and moreover, the greater the likeli-hood of there being less psychological continuity and connectedness withthose future stages. (Here I use “continuity” in the Parfitian sense to referto memories of past experiences, intentions formed and then acted upon,resemblance over time between sets of beliefs, desires, and goals, andresemblance over time in character traits; and I use “connectedness” torefer to overlapping chains of psychological connections.)37 My far-futureself may be a much more admirable person than I am now, but for thatreason, I may feel less connection with that person and, not unreasonably,may give that person less consideration than my near-future selves. Inputting forward these considerations, I do not mean so much to under-mine this particular conception of prudence as to indicate that the prag-matic purpose for having prudential reasons is consistent with significantvariations in judgment about what we have prudential reasons to do.There is no fact such as the equal reality of various future selves I mightbe, from two minutes hence to twenty years down the line, that requiresa particular conception of prudence.

Derek Parfit, in Reasons and Persons, has argued that, in effect, we mayhave no more reason to care about a future self than we do about otherpeople. This is because there is no deep metaphysical fact that marks ourseparateness from other people. There would be such a fact if we wereseparate Cartesian thinking substances, but if one thinks, as Parfit does(and I do), that there are no such separate substances, the most plausiblealternative view is that being one person over time is nothing beyondhaving a brain and body and a series of interrelated physical and mentalevents (where this relation is characterized in terms of psychologicalcontinuity and connectedness).

Christine Korsgaard has argued in response to Parfit that the require-ments of practical agency in fact confer unity on the self. She agrees withParfit that there is no deep metaphysical fact about being the same personover a stretch of time —no ontological entity or fact beyond having a brainand body and a series of interrelated physical and mental events (inter-related, that is, through connectedness and continuity). Rather, and hereis where she departs from Parfit, the ground for thinking of oneself as

37 Derek Parfit, Reasons and Persons (New York: Oxford University Press, 1984), 206–16.


a unified agent is the raw necessity for eliminating conflict among ourvarious motives and for the unity implicit in the standpoint from whichone deliberates and chooses among these motives. On her view, thisstandpoint involves some principle or way of choosing that one regardsas expressive of oneself and that provides reasons regulating one’s choicesamong desires.38 Furthermore, she observes that most of our projectsextend over periods of time and that we think of our activities and pur-suits as interconnected in various ways, comprising plans of life. Theunity of the self is a consequence of the practical requirements for coher-ence over time and among motives and activities. On her view, then, thepractical requirements of agency ground prudential reasons directly, andnot by way of being founded on a prior notion of unity.

I accept Korsgaard’s argument to this extent: the practical requirementsof agency do contribute to the importance of identity over time, howeverwe conceive of its basis. But the degree of practical coherence we need,and the duration over which we need it, seems a variable matter andsubject to overriding by other practical concerns. A certain degree ofconflict among our motives is an inevitable liability of richness of motiveand interests. Striving to be true to the goals one has set for oneself in thepast makes possible much of what bears great value in human life, butconsistency has a cost. Much of the Daoist text Zhuangzi celebrates a lifeof spontaneity and of concentration on the present moment because aplan-filled, goal-oriented life is one that becomes rigid and creates a filteron one’s perceptions of what there is and what is to be valued in thepresent situation. Those who are highly goal-oriented tend to see theirpresent situation only in terms of what will satisfy their goals, and not interms of those things in their present situation that might satisfy andstimulate but fail to fit with those goals.39 One might agree that the natureof human commitments to projects requires a conception of oneself asunified over certain spans of time, but also point out that a project neednot extend indefinitely into the future. A young man of twenty-two mightcommit to trying to make a go of playing in a rock band until he reachesthirty, but leave the future open after that. He thinks of his life as episodic,as Galen Strawson suggests,40 more along the lines of a series of projects,not necessarily unified by overarching goals or motives. A well-known

38 Christine M. Korsgaard, “Personal Identity and the Unity of Agency: A Kantian Responseto Parfit,” Philosophy and Public Affairs 18 (1989): 109–15.

39 In the first chapter of the Zhuangzi, for example, Zhuangzi’s logician friend Huizi isgiven seeds that give birth to enormous gourds. When Huizi tries to find a use for the shellsof these gourds, he can only think of using them as water dippers or water containers, butthey are too big and heavy for such uses. Failing to find a use for them, he smashes themto pieces. Zhuangzi scolds his friend for having underbrush in his head, pointing out thathe could have lashed the shells together to make a kind of raft to ride upon the lakes andrivers. See Chuang Tzu: The Inner Chapters, trans. A. C. Graham (Indianapolis, IN: Hackett,2001), chap. 1.

40 Galen Strawson, “The ‘Self’,” in Models of the Self, ed. Shaun Gallagher and JonathanShear (Exeter, UK: Imprint Academic, 1999), 3–24.

262 DAVID B. WONG

philosopher, not Strawson, once told me that he did not worry aboutconsistency of theme between his many books. He compared himself toan artist who works on a painting, finishes it, and then goes on to the nextwithout thinking of consistency between them.

I am not denying that the requirements of agency might favor a degreeof diachronic unity, but they might not reproduce the kind of unity we areaccustomed to ascribing to persons over time, nor would they necessarilyproduce the familiar reasons of prudence that extend into the future.More promising, I believe, is John Locke’s suggestion that the concept ofpersonal identity is forensic in nature.41 That is, it enables the holding ofpersons as responsible for their past and future actions, making memoryone of the crucial psychological threads connecting life slices. Further-more, a concept of persons enabling ascriptions of responsibility for pastand future actions (linking, for example, the one who forms intentionswith the one who acts on them) is crucial for the ways in which cooper-ation is encouraged through punishment of norm violators. Deterrence ispossible only with such unity over time. Such enforcement of moral andother cultural norms also seems to require the person as unified to beembodied. It is difficult, at the very least, to imagine a practical way ofidentifying and punishing noncooperators who fail to be embodied.

The concept of the person persisting over time is, I believe, a prototypeconcept. As such a concept, it would include the feature of having ahuman body that persists over time, an embodied consciousness of theenvironment, and memories, thoughts, intentions, and psychological traitsthat are continuous and/or connected in a high degree. The explanationfor why the person prototype has this content and not some other contentis bound up with the cooperative nature of human life and with what ittakes to sustain that nature. It is not just that agents are needed whoconceive of certain intentions as their own and therefore as needing to becarried out, but also that embodied agents need to be identifiable andheld responsible for their cooperativeness or lack of it.42

We run into conceptual puzzles when we consider cases, most of themhypothetical or counterfactual, in which some of the prototype charac-teristics are missing. It then becomes very unclear what the concept ofpersonal identity implies for these cases as to whether the person has

41 John Locke, “Of Identity and Diversity,” in Locke, An Essay Concerning Human Under-standing (1690), ed. P. H. Nidditch (New York: Oxford University Press, 1979).

42 Philip Pettit has raised the question of whether my proposal has a Euthyphro problem:Do we hold a person responsible because he’s the same person, or do we judge him to bethe same person because he is the one in the position to be held responsible? My answer isthat the concept of a person persisting over time evolved because the conditions of humancooperative life require relatively long-term persisting agents to hold responsible (the sec-ond possibility). But once that concept of a persisting person is in place, with accompanyingbodily and psychological criteria for persistence also in place, we hold particular individ-uals responsible because they are the same persons who did the things for which we needto hold someone responsible.


persisted or disappeared or become a different person or two differentpeople, and those who react to the cases appear to put forward a varietyof intuitions or have none at all (for example, so-called “fission” cases inwhich a person splits into two who have the same degree of psychologicalconnectedness and continuity with their ancestor; or teletransportation inwhich one’s body on earth is destroyed but one’s psychological proper-ties and processes are preserved and associated with a new body onMars). I think there is often no single right answer as to whether theperson has persisted in such cases.43 Persons, I believe, do not constitutea natural kind. The concept of persons is constructed at least partly inresponse to practical considerations of the type I have described. To saythe concept is constructed, I again want to make clear, is not to say thatthe very features of the prototype are somehow fabricated. Rather, it is tosay that they were selected as features of the person prototype out of prac-tical considerations. These practical considerations then motivate a con-cept of personal identity over time, which is presupposed by the conceptof prudence and whatever reasons are associated with it. However, thepractical considerations that go into the formation of the relevant proto-type or exemplars may not dictate an answer to whether a person haspersisted in circumstances that go far beyond the context in which theprototype or exemplars are intelligible. That the structure of the personconcept is responsive to social and practical considerations, in other words,might explain why it is very difficult to extend the answer to situationswhere the social practicality vanishes.44

If I am right in my story of how we become selves persisting over time,the reasons we have to provide for our future selves are not based on anecessary and self-evident conception of practical rationality as embody-ing the fundamental aim of satisfying our desires. Such an aim presup-poses a conception of selves persisting over time that is constructed to

43 Or consider the perplexities that Bernard Williams raises in “The Self and the Future,”Philosophical Review 79 (1970): 161–80. On the one hand, if I imagine that A’s memories aretransferred into B’s body, and that A’s body is tortured, and if I further imagine that it wasmy memories that got transferred to B’s body, I would feel fortunate. This seems to favor thepsychological criterion for personal identity. On the other hand, if I am told that I am goingto be tortured tomorrow, but that I will not remember anything leading up to the torture,and that, moreover, my impressions of the past will be quite different from the ones I havenow, I will still be quite frightened and will not take any comfort in the psychologicaldiscontinuity. This seems to favor the bodily criterion of personal identity. Perhaps theseconflicting intuitions are explicable if both the bodily criterion and the psychological crite-rion are present in the prototype without the kind of weighting of these different featuresthat would decide which one is more fundamental to identity.

44 I agree with David Shoemaker’s view, raised in discussion of this paper, that there canbe different concepts of personal identity; e.g., biological criteria of identity might be rele-vant in determining whether I am owed compensation for something that happened to meas a fetus, even if the bodily-psychological concept of the person I have been discussingmight not make me the same person as that fetus. Different concepts of the person mightarise in different social contexts and on the basis of different practical considerations. Andit is possible that different concepts of the person could overlap in their application to thesame context and come into conflict. In that case, conceptual revision might be necessary.

264 DAVID B. WONG

satisfy practical requirements of the social life. These practical require-ments are the source of our criteria for judging which desires are desiresof our persisting selves and not some other selves, and therefore theserequirements are presupposed by the reasons we have for satisfying thosedesires.

VII. Conclusion

Moral reasons are not given by what will satisfy an individual’s desires.Nor are moral reasons simply expressions of social demands on the indi-vidual. Rather, they can have a normative grip that reaches deeply intothe individual. They do so by referring to situational features that call foractions that support the cooperative nature of human life. Through moraleducation and socialization, moral reasons become embedded in the indi-vidual’s motivational propensities and thereby become part of what con-stitutes that individual as a particular person. The requirements of socialcooperation also shape the concept of personal identity over time and thereasons of prudence associated with such a concept. While many havetaken reasons of prudence to be the paradigm of reasons instrumentallybased on desires of the individual, reasons of prudence, like moral rea-sons, are rooted in normative demands (based in the cooperative natureof human life) that go into shaping the motivations of the individual.

The story I have told allows that moral reasons can have a deep nor-mative grip on many (perhaps most) individuals, but not necessarily onall. This is the right result, I believe. The normativity of morality is ines-capable and can reach deeply into the individual’s motivations and notsimply answer to whatever the individual’s motivations happen to be.However, there will be individuals to whom this inescapability does notmatter. These are individuals who are socially constituted, not throughmoral reasons, but through other kinds of reasons. That is why the com-mon conception is that morality is indeed inescapable, that its normativegrip can go deeply into the individual, but also why Thrasymachus is notnecessarily irrational for being indifferent to the demands of morality.

The fact that morality has the function of reconciling self-concern andother-concern also explains the conception of moral reasons as overridingreasons. Moral reasons are grounded in what is purported to be appro-priate consideration of the agent’s self-interest and therefore purport tooverride reasons of self-interest when such reasons conflict with the appro-priate weight given to them in the context of reconciling and balancingself- and other-concern. That is, moral reasons, by virtue of the kind ofreasons they are, purport to give due weight to the agent’s self-regardinginterests and therefore purport to justifiably override those interests whennecessary. At the same time, this explanation of the overridingness ofmoral reasons is consistent with treating the moral perspective as just one


kind of normative perspective45 that, from the agent’s all-things-consideredpoint of view, might not necessarily have overriding force. Moral reasonswill fail to have such overriding force if the agent is one in whom suchreasons fail to motivationally embed at all, or if the agent cares to somedegree about morality but not enough to override other things she caresabout.46 On the one hand, moral reasons purport to override other rea-sons based on their function of helping to organize the individual’smotivational system from the standpoint of promoting and sustainingcooperation. On the other hand, this overriding force is conferred onlyfrom that normative perspective. It is a normative perspective that makeshuman life livable and that goes into the constitution of many or mostpeople, but in the end it can claim superior authority only on the basis ofwhat makes life livable for a community of people. It cannot claim supe-rior authority as a perspective that must motivate each person for thatperson’s life to be livable.

Philosophy, Duke University

45 Other normative perspectives might include the perspective of self-interest, or of othernonmoral ideals to which the agent has subscribed.

46 A question from Michael Huemer prompted me to realize the last possibility: thatmoral reasons might become motivationally embedded but still be overridden by otherembedded reasons.

266 DAVID B. WONG

WHAT DOES MORAL PHENOMENOLOGY TELL USABOUT MORAL OBJECTIVITY?*

By Terry Horgan and Mark Timmons

I. Introduction

Roughly speaking, moral phenomenology concerns the what-it-is-likeness of moral experiences —what it is like, for example, to experienceoneself as being morally obligated to perform an action.1 Some philoso-phers argue that attention to one’s moral phenomenology reveals thatmoral experiences have “objective purport,” in the sense that such expe-riences purport to be about some objective moral reality. In general, how-ever, philosophers have not paid much attention to the rich details ofmoral phenomenology —the details of the what-it-is-likeness of moralexperience.2 Rather, when philosophers appeal to moral phenomenology(as they often do), one typically finds mention of such features asprescriptivity, sense of authority, and the like, but not much in the way ofconcentrated attention on matters of phenomenological detail. Perhapsmoral philosophers think they do not need to bother with such detail,which (one might think) is better left to more qualified experts, includingpsychologists, poets, and playwrights. But if one is serious about havingone’s moral theory (whether in normative ethics or metaethics) fit with, orotherwise accommodate, the facts of moral phenomenology, then gettingthe phenomenological facts straight is crucially important. More specifi-cally, if one wants to make a case for some form of moral objectivity onthe basis of appeal to moral phenomenology, one must take a careful lookat the phenomenology in question.

A fully adequate treatment of moral phenomenology and its bearing onquestions of moral objectivity would require examining a vast range of

* We wish to thank the other contributors to this volume, and its editors, for helpfulcomments. We are especially grateful for comments by Janice Dowell, Michael Huemer,Scott McDonald, Philip Pettit, David Shoemaker, and David Wong. We also benefited fromcomments from Robert Audi, Michael Gill, and Uriah Kriegel.

1 In Section III below, we discuss the subject matter and method of phenomenology ingeneral, and in Section V, we say more about the subject matter and method of moralphenomenology.

2 Nor have they spent much time grappling with questions about the scope, unity, anddistinctiveness of moral phenomenology. One major exception is Maurice Mandelbaum, ThePhenomenology of Moral Experience (Glencoe, IL: The New Press, 1955). More recently, weourselves have broached such questions in Terry Horgan and Mark Timmons, “Moral Phe-nomenology and Moral Theory,” Philosophical Issues 15 (2005): 56–77; and in “Prolegomenato a Future Phenomenology of Morals,” Phenomenology and the Cognitive Sciences 7, specialissue on moral phenomenology edited by Uriah Kriegel (2008).


types of moral experience, including, for example, experiences that haveto do with, or include, judgments of moral obligation and judgments ofmoral value, as well as experiences that involve the moral emotions —guilt, shame, indignation, and the like. In this essay, we restrict ourselvesto questions about the phenomenology of experiences of what we call(following Maurice Mandelbaum) “direct” moral obligation —experiencesin which one is presently confronted with a situation or is facing a set ofcircumstances in which one feels morally obligated to act (or refrain fromacting) in a certain way. About such experiences, we ask two questions.First, what are the introspectively accessible elements of such experi-ences?3 Second, how do these elements bear on the question of whether,and in what way, moral experiences purport to be about some objectivematter; in particular, do such experiences carry objective purport? As weshall explain below, we are here interested in what we will call ontologicalobjective purport —the idea that moral experiences purport to be aboutsome in-the-world moral properties. Given our specific focus, then, thereare three competing answers to consider with regard to the second question:

Affirmative: It is an introspectively accessible fact that direct moralexperiences do carry ontological objective purport.

Negative: It is an introspectively accessible fact that direct moral expe-riences do not carry such purport.

Neutrality: It is not introspectively accessible whether or not directmoral experiences carry ontological objective purport.

The third answer, Neutrality, presupposes that certain questions aboutthe intentional content of direct moral experience do not have introspec-tively accessible answers, despite the fact that the experiences themselvesare introspectively accessible.

A point of clarification is in order concerning Neutrality. This answerdoes not claim that there is no fact of the matter at all about whether ornot direct moral experiences carry ontological objective purport; rather, itonly claims that there is no introspectively accessible answer to the questionabout such purport. Thus, even if Neutrality is right, there might well bea determinate answer to the second question mentioned above; but thisanswer, whatever it is, would not be obtainable just by attending intro-spectively to one’s own phenomenology.

What we will call “the argument from phenomenological introspec-tion” has as its conclusion the Affirmative thesis and thus purports toprovide a pro tanto reason to favor an ontological-objectivist metaethics.

3 By “introspectively accessible elements” we mean elements that are readily introspec-tively accessible by people with ordinary introspective acuity —as distinct, for instance, fromelements that are introspectively accessible only by people with unusually powerful andaccurate introspective skill.

268 TERRY HORGAN AND MARK TIMMONS

Moreover, this argument purports to get that pro tanto reason directlyfrom the introspectively accessible aspects of moral phenomenology. Weourselves find the issue about moral experience and objective purportsomewhat murky, and so much of our work in this essay will be devotedto clarifying the various issues at play in connection with the argumentfrom phenomenological introspection. However, late in the essay (in Sec-tion VIII), we shall offer a brief defense of the Neutrality thesis.4

Here is our plan. We begin in Section II by explaining two conceptionsof moral objectivity that are prevalent in contemporary metaethical dis-cussion, one “ontological” and the other “rationalist.” Because talk ofphenomenology requires clarification, we devote some attention in Sec-tion III to explaining how we propose to use the term. We will then be ina position to characterize more precisely (in Section IV) the kind of phe-nomenological argument we wish to examine —the argument from phe-nomenological introspection. In Section V, we provide a taxonomy oftypes of moral experience so that we can zero in on the phenomenologyof the type we wish to consider —direct moral experiences of obligation —and in Section VI we provide two examples of such moral experience.Next, in Section VII, we turn to the work of Mandelbaum, whose subtlephenomenological characterization of direct moral experiences of obliga-tion we think is correct and provides a basis for exploring the above-mentioned hypotheses regarding ontological objectivity. Having done allof this, in Sections VIII through X we proceed to consider the variouselements of such experiences in order to evaluate the argument fromphenomenological introspection, and in Section XI we respond to someanticipated objections. Our purpose is to explore how a metaethic that isnot ontologically objectivist might provide a basis for challenging theargument from phenomenological introspection. In previous writings, wehave defended what we call “cognitivist expressivism,” a view accordingto which (roughly) moral judgments are genuine cognitive belief states,yet such moral beliefs do not purport to describe some sort of moralreality; rather, they express a certain sort of commitment directed towarda nonmoral state of affairs. We employ our cognitivist expressivism as abasis for exploring a non-ontological-objectivist line of defense againstthe argument from phenomenological introspection.5 As we shall explain,

4 We also believe, although we will not argue for this here, that moral experiences in factdo not carry ontological objective purport. But even if we happen to be wrong about this, westill could be right about the Neutrality thesis. The truth of the Neutrality thesis would beenough to undercut the argument from phenomenological introspection, even if that argu-ment’s conclusion happens to be correct.

5 Terry Horgan and Mark Timmons, “Nondescriptivist Cognitivism: Outline of a NewMetaethic,” Philosophical Papers 29 (2000): 121–53; and “Cognitivist Expressivism,” in TerryHorgan and Mark Timmons, eds., Metaethics after Moore (Oxford: Oxford University Press,2006). The construal of moral phenomenology that we will explore below is not ontologicallyobjectivist; nevertheless, it does recognize and accommodate phenomenological features ofdirect moral judgment that count intuitively as objective. See Section X below.

MORAL PHENOMENOLOGY 269

there is more that would need to be done beyond what we undertake inthis essay in order to mount a full and convincing response.6

II. Two Conceptions of Strong Moral Objectivity

In metaethical discussion, there are two main conceptions of objec-tivity: one ontological, the other rationalist. Both, in their own way, rep-resent what we shall call “strong” conceptions —conceptions that are(respectively) featured in versions of moral realism and Kantian rational-ism. Let us take these in order.

The ontological conception: This conception of moral objectivity is aninstance of the sort of objectivity often associated with ordinary physical-object thought and discourse, according to which (roughly) there is an“objectively existing” world of objects and instantiated properties (includ-ing relations) possessed by (or obtaining among) those objects. Often, thisconception of “objective existence” is expressed in terms of “mind-independence.” 7 But thinking of ontological objectivity in terms of mind-independence is arguably too strong for our present purposes, because,on some views, secondary qualities are response-dependent, and henceare not mind-independent, but nevertheless such properties are objective.

Indeed, in metaethics, the thesis that moral properties (instantiations ofthem) are on an ontological par with response-dependent color properties —that a property like intrinsic value is as much “out there to be experi-enced” as are instantiations of color properties, even if both are response-dependent —would seemingly secure enough ontological objectivity formorality to combat various forms of metaethical noncognitivism, relativ-ism, and subjectivism. So, following contemporary metaethical fashion,we can distinguish two forms or grades of ontological moral objectivism(“realism,” if you like). According to robust versions, moral properties area kind of, or analogous to, prototypical primary qualities in that theirnature and existence (i.e., whether or not they are instantiated in theworld) are not mind- (including human response-) dependent. But accord-ing to more modest versions, moral properties are a kind of, or at leasthave the same sort of ontological status as, secondary qualities —color properties in particular. Their being response-dependent in a wayanalogous to color properties (when colors are construed as “secondary

6 Of course, if cognitivist expressivism cannot accommodate the phenomenological data,and if no other metaethical view that rejects ontological objectivism can do so either, thisdoes not yet guarantee that the argument goes through. For it might be that the phenom-enological data can be accommodated by some form of metaethical rationalism that deniesontological moral objectivism.

7 Some moral realists prefer to characterize realism (including moral realism) in terms of“stance-independence” rather than in terms of “mind-independence.” See Russ Shafer-Landau, Moral Realism: A Defense (Oxford: Oxford University Press, 2004), 15–16. Whetherthis way of characterizing realism allows for a response-dependent account of secondaryqualities to count as realist will depend on how one understands the notion of stance.


qualities”) is still sufficient to say of them that, when instantiated, moralproperties are “there to be experienced.” 8

It is worth emphasizing, too, that the ontological conception of objec-tivity not only claims that moral judgments have descriptive content (i.e.,purport to represent facts consisting of the instantiation of moral prop-erties that are “there to be experienced”) but also claims that such factsare not about certain idiosyncratic psychological states of the agent mak-ing the moral judgment (states such as preferences or attitudes of approval/disapproval). Ontological objectivism is a form of metaethical descriptivism,certainly; but it is to be distinguished from the kind of descriptivism thatearlier in the century was called “subjective naturalism,” which construedmoral facts as facts about the psychology of the morally judging agent.

The rationalist conception: Speaking very generally and somewhat loosely,a realm of thought and discourse is objective, according to this concep-tion, if there is a method of thinking or reasoning whose use would yield(under proper conditions of application) convergence in belief about thesubject matter in question. The primary examples here are mathematicsand logic.9 In metaethics, this view is clearly exemplified by RoderickFirth’s ideal observer theory and Michael Smith’s metaethical rational-ism.10 According to these views, there is a nonmoral conception of ratio-nal choice involving norms of practical reasoning whose proper applicationis sufficient to yield a set of richly determinate moral norms or principles.This form of metaethical objectivity, then, is meant to capture the idea,famously expressed by Immanuel Kant, that “a law, if it is to hold mor-ally, that is as a ground of obligation, must carry with it absolute neces-sity; that, for example, the command ‘thou shalt not lie’ does not holdonly for human beings, as if other rational beings did not have to heedit. . . .” 11 This idea of requirements holding necessarily for all rationalbeings is Kant’s idea of those requirements being categorical.

Even though these two conceptions of moral objectivity differ in impor-tant ways, they do share a uniting theme: they both try to capture a strongconception of objectivity according to which some realm of thought and

8 This is how John McDowell, “Values and Secondary Qualities,” in Ted Honderich, ed.,Morality and Objectivity (London: Routledge, 1985), 170, characterizes a kind of objectivitycharacteristic of secondary qualities (understood as response-dependent) that he contrastswith a stronger kind that does not recognize response-dependent properties as being objec-tive. See also Jonathan Dancy, “Two Conceptions of Moral Realism,” Proceedings of theAristotelian Society, supp. vol. 60 (1985): 167–87.

9 For more discussion of these two conceptions, see Mark Timmons, “Objectivity inMoral Discourse,” in Keith Brown, ed., The Encyclopedia of Language and Linguistics, 2d ed.(Oxford: Elsevier, 2006), 9:5–10.

10 Roderick Firth, “Ethical Absolutism and the Ideal Observer,” Philosophy and Phenom-enological Research 12 (1952): 317–45. Michael Smith, The Moral Problem (Oxford: Blackwell,1994).

11 Immanuel Kant, Grundlegung zur Metaphysik der Sitten (Akademie volume IV, 1985), 389,in Kant, Practical Philosophy, ed. and trans. Mary J. Gregor (Cambridge: Cambridge Univer-sity Press, 1996), 44–45.


discourse is objective only when there is some “reality” —some ontologyor method —that “backs up” or “stands behind” such thought and dis-course and serves to “make true” certain claims in the discourse and“make false” other claims.

Before moving on, we make three observations about the relation-ship between these two conceptions. First, one might hold that theidea of moral objectivity involves a blend or combination of both kindsof objectivity. J. L. Mackie seemed to think so.12 He held that moralthought and discourse are committed to the idea that moral require-ments are categorical (in the sense just indicated) and to the idea thatthere are instantiated moral properties (including moral requiredness)that are strongly objective and thus on a metaphysical par with pri-mary qualities.13 But a secondary quality–like account of moral realismis also (perhaps) likely to attempt a combination: moral properties are“there to be experienced” (ontological claim), and their nature is bestcaptured by a dispositional (response-dependent) account that is basi-cally rationalist.14

Second, it is possible to hold that there are objective moral propertiesin the ontological sense and yet deny, or be neutral with respect to, theidea that morality is also objective in the rationalist sense. Suppose, forinstance, that moral properties are like secondary qualities in that aproper understanding of them requires referring to how human beingsin certain circumstances would respond —without supposing that suchproperties are there to be experienced by all rational beings. Perhaps onecould wax Humean on this particular matter: we humans have a cer-tain constitution —maybe a “moral sense” (in a loose sense of the term) —through which we are able to perceive or respond to values that havean ontological status similar to that of colors. The idea would be tocombine the claim that there are moral properties possessed by itemsof evaluation with a rejection of a strong rationalist commitment toclaims about all rational beings.

Third, at least prima facie it is possible to embrace the idea that moral-ity is objective in a distinctively rationalist way without supposing thatthere are moral properties possessed by items of evaluation in the way

12 J. L. Mackie, Ethics: Inventing Right and Wrong (Harmondsworth: Penguin, 1977),chap. 1.

13 Note that if one claims both (1) that moral properties are analogous to primary qualitiesin being human-response independent and (2) that moral properties are intrinsically reason-providing, then one seems committed to the rationalist claim that recognition of moralproperties provides all rational agents with reasons. This seems to be why Mackie believedthat moral thought and discourse presuppose both ontological and rationalist conceptionsof objectivity.

14 This seems to be the view that Smith is advocating in Michael Smith, “Objectivityand Moral Realism: On the Significance of the Phenomenology of Moral Experience,”reprinted in Michael Smith, ed., Ethics and the A Priori (Cambridge: Cambridge Univer-sity Press, 2004).


that an ontological moral realist might suppose.15 Suppose there are ratio-nal norms of obligation that hold for all rational agents. If so, then one cansay of some action that it is obligatory, but in saying this one does nothave to suppose that in addition to the natural features of the action it hassome further property, obligatoriness, that is either identical to, consti-tuted exhaustively by (but not type-identical to), or otherwise something“over and above” the natural properties on which this special propertysupervenes. One example is R. M. Hare’s metaethical view, according towhich moral judgments are “objective prescriptions.” 16 Hare denied thathis view involving objective prescriptions committed him to a moralontology of the sort that would involve moral properties, though hedefended the claim that certain features of moral language together with(nonmoral) constraints on rational choice yield an essentially utilitarianmoral theory.17 Indeed, contemporary Kantians tout rationalist objectivityas a way of avoiding metaphysical moral realism. Christine Korsgaardwrites: “If ethically good action is simply rational action, we do not needto postulate special ethical properties in the world or faculties in the mindin order to provide ethics with a foundation.” 18

In sum, (1) there are two main conceptions of strong metaethical objec-tivity: ontological and rationalist; (2) the ontological conception coversboth robust and modest conceptions of what can count as an objectiveproperty; (3) it is at least prima facie possible for a metaethical view tocombine either form of ontological objectivity with the rationalist con-ception of objectivity; however, (4) it is also possible (again prima facie)to embrace the rationalist conception without also embracing the onto-logical conception (in either of its versions), and the other way around.

Thus, in light of these conceptions of moral objectivity and their inter-connections, one main question about moral phenomenology is whetherit embodies or “carries” the pretensions of either sort of objectivity. Beforeresponding to this question, there is preliminary work to be done whichwill occupy us in the next three sections.

III. Phenomenology: Subject Matter and Method

In recent philosophy of mind, the term “phenomenology” is used some-times to refer to a type of subject matter, comprising certain aspects of

15 The idea that certain forms of discourse, including mathematics, logic, and ethics, canbe objective —that claims in these areas can be objectively true —without there having to bea realm of objects and properties that make them true is one main theme of Hilary Putnam’sHermes Lectures, published as Part I of Hilary Putnam, Ethics without Ontology (Cambridge,MA: Harvard University Press, 2004). See esp. Lecture 3, “Objectivity without Objects.”

16 R. M. Hare, “Objective Prescriptions,” Philosophy 35 (1993): 1–17.17 Perhaps an objective moral property can be constructed from the components that Hare

posits. But it is enough for present purposes to note, as we do, that embracing rationalistobjectivism but denying ontological objectivism is, prima facie, a metaethical option.

18 Christine M. Korsgaard, The Sources of Normativity (Cambridge: Cambridge UniversityPress, 1996), 311.


one’s mental life that are available to introspection in a distinctive way.But the term is also used to refer to a method of first-person introspectionthat allegedly carries with it a special first-person warrant attaching toone’s first-person judgments about one’s own experiences.19 We think ofphenomenology as involving both the subject matter and the method justmentioned. However, in order to make ourselves clear as we proceed, wewill use the term “phenomenology” mainly to refer to subject matter, andwe will generally use the term “introspection” for the associated method.

In this section, we will make some clarificatory remarks about both thesubject matter and the method. Part of our purpose is to situate moralphenomenology, as we understand it, within the context of various ongo-ing debates in contemporary philosophy of mind. As we will make clear,phenomenology generally, and moral phenomenology in particular, canbe characterized in a way that is noncommittal about competing positionsin several of those debates. Having a characterization of moral phenom-enology that is noncommittal in the ways we will proceed to explain isimportant for our purposes; we want to work with a characterization ofthis area of inquiry that does not beg any questions about potential sub-ject matter and methodology that might taint our investigation of theargument from phenomenological introspection.

A. Phenomenology and phenomenal consciousness

Qua subject matter, phenomenology is closely related to what is oftencalled “phenomenal consciousness” in recent philosophy of mind. Arethey one and the same? That depends. Under some positions concerningthe nature of phenomenal consciousness, they are indeed the same; butunder others, phenomenally conscious states are only a proper part of thefull scope of phenomenology. Let us explain.

One disputed issue concerns the scope of phenomenal consciousness.On some views, its scope is confined to the realm of sensory experience,or perhaps to sensory experience plus sensory imagery. Phenomenal con-sciousness is confined to the what-it-is-likeness of sensations: e.g., what itis like to see a bright red tomato, what it is like to smell rotten eggs, whatit is like to experience a searing pain in the left toe, etc.20 Such viewsmaintain that the range of mental states that are conscious as opposed tounconscious is much wider than the range of mental states that are phe-nomenally conscious. Although one’s occurrent beliefs, for example, typ-

19 Charles Siewert, for instance, in his “Who’s Afraid of Phenomenological Dis-putes?” Southern Journal of Philosophy 45 (2007), understands phenomenology as primarilymethodological.

20 There are various ways that one might widen somewhat the scope of phenomenalconsciousness, while still regarding its scope as fairly restricted. For instance, one mighthold that moods and/or emotions, in addition to sensations and sensory images, havedistinctive phenomenal character.


ically are conscious rather than unconscious, occurrent beliefs have nodistinctive or proprietary phenomenal character; rather, the sense in whichthey are conscious involves their accessibility to the cognitive agent. Suchstates are “access conscious” but not phenomenally conscious, to use NedBlock’s influential distinction. A mental state is access conscious whenthat state is available to the agent for purposes of reasoning.21

On other views, however, the scope of phenomenal consciousness ismuch broader than just sensory. In the strongest form, these views main-tain that all, or virtually all, mental states that are access conscious arealso phenomenally conscious, and, indeed, that their first-person acces-sibility derives from their phenomenal character. (There is a distinctivephenomenal character, for instance, of believing that Hillary Clinton willbe the next president, different from the phenomenal character of hopingthat she will be the next president. Likewise, there is a distinctive phe-nomenal character of believing that she will be the next president, differ-ent from the phenomenal character of believing that she will not be.)22

Another disputed issue concerns whether phenomenal consciousnesspossesses intentionality. On some views (in particular, some views thatclaim that all phenomenal character is sensory or sensory-imagistic), theanswer is no: phenomenal character is inherently nonintentional, andmental intentionality is inherently nonphenomenal. On other views, how-ever, all (or virtually all) phenomenal character is inherently intentional,i.e., inherently represents things as being certain ways.23

Phenomenology, understood as comprising both method and subjectmatter, can be and should be construed in a way that leaves it uncom-mitted about these disputed issues. As method, phenomenology doesassume that much of one’s conscious (as opposed to unconscious) mentallife has a distinctive epistemic status from a first-person perspective: it isdirectly accessible via introspection. But this leaves open whether or not

21 Ned Block, “On a Confusion about a Conception of Consciousness,” Behavioral andBrain Sciences 18 (1995): 227–47.

22 For a defense of this broader conception of the scope of phenomenal consciousness, see,e.g., Terry Horgan and John Tienson, “The Intentionality of Phenomenology and the Phe-nomenology of Intentionality,” in David Chalmers, ed., Philosophy of Mind: Classical andContemporary Readings (Oxford: Oxford University Press, 2002), 520–33; Terry Horgan, JohnTienson, and George Graham, “Phenomenal Intentionality and the Brain in a Vat,” in Rich-ard Schantz, ed., The Externalist Challenge (Berlin: Walter de Gruyter, 2004), 297–317; UriahKriegel, “Consciousness as Sensory Quality and as Implicit Self-Awareness,” Phenomenologyand the Cognitive Sciences 2 (2003), 1–26; David Pitt, “The Phenomenology of Cognition, orWhat Is It Like to Think That P?” Philosophy and Phenomenological Research 69 (2004): 1–36;and Galen Strawson, Mental Reality (Cambridge, MA: The MIT Press, 2004).

23 These intentionalist views bifurcate into two broad kinds: one kind maintains that allphenomenal character is sensory or sensory-imagistic (while insisting that phenomenalcharacter is also inherently intentional); the other kind maintains that the scope of phenom-enal consciousness is much broader, encompassing all (or virtually all) mental states that areconscious-as-opposed-to-unconscious. The former view is advocated, e.g., in Michael Tye,Ten Problems of Consciousness: A Representational Theory of the Phenomenal Mind (Cambridge,MA: The MIT Press, 1995); and Fred Dretske, Naturalizing the Mind (Cambridge, MA: TheMIT Press, 1995). The latter view is advocated, e.g., in the texts mentioned in note 22 above.


introspective accessibility always rests on first-person apprehension ofphenomenal character. On the one hand, if phenomenal character isrestricted to sensations and/or sensory imagery, then the range of intro-spection extends beyond the bounds of phenomenal consciousness —because certain mental states are introspectively accessible even thoughthey lack any distinctive or proprietary phenomenal character. In thiscase, the subject matter of phenomenology includes aspects of mentalityother than what is phenomenally conscious (while also including phe-nomenally conscious aspects too, of course). On the other hand, if all orvirtually all mental states that are access conscious do have distinctivephenomenal character, then it may well be that introspective access tothese states is always a matter of apprehending their phenomenal char-acter. In that case, the subject matter of phenomenology is just phenom-enal consciousness itself.

B. “What it is like”

In what follows, we plan to make free use of the locution “what it islike.” Its use is very natural in connection with introspective judgmentsabout one’s own mental life, whether or not one takes those judgments tobe grounded in an apprehension of phenomenal character. Consider, forinstance, the question: “How do you know whether you really believethat Hillary Clinton will win the Democratic primary, or merely hope thatshe will?” In response, one should say: “Look, there’s an introspectivelyaccessible difference between just hoping something and actually believ-ing it, a difference in what such states are like. And I know that, in my case,the state is belief —because of what the state is like.” So we are going tomake free use of what-it-is-like talk in what follows.

Is this a broader use of such talk than the common use in philosophyof mind, where the idea of phenomenal character is often glossed as“those aspects of one’s mental life such that there is something it is like toundergo them”? Well, that depends on one’s position concerning thescope of phenomenal character —about which moral phenomenology perse is neutral. Thus, moral phenomenology per se is also neutral aboutwhether its own use of what-it-is-like talk is broader than the use thatapplies such talk only to phenomenal character. (It all depends on thescope of phenomenal character, about which moral phenomenology isneutral.)24

C. Phenomenology and the content of perceptual experience

Another disputed issue in philosophy of mind concerns the intentionalcontent of perceptual experience —in particular, what kinds of content

24 For some discussion of this matter, see Charles Siewert, “Who’s Afraid of Phenomeno-logical Disputes?”


are, or are not, aspects of perceptual experience itself. On some views,perceptual experience per se has only fairly limited and primitive content,even though perceptually grounded beliefs often have much richer con-tent. (Such views typically deny, for instance, that perceptual experienceper se represents causal connections, even though beliefs with causal con-tent often are directly grounded in perceptual experience.) On other views,the content of perceptual experience is very rich —rich enough, for instance,to include the experiential representation of causal connections.25

Phenomenology, as method, is neutral about this dispute too. After all,perceptually grounded beliefs are themselves within the purview of phe-nomenology anyway, insofar as such beliefs are introspectively accessible(as they often are). Of particular relevance here, given our concerns in thisessay, are spontaneous, perceptually grounded moral judgments —as inGilbert Harman’s famous example of rounding the corner and finding one-self “just seeing” (as one says) that those hoodlums are doing somethingwrong in lighting the cat on fire just for fun.26 The moral judgment isspontaneous and, moreover, arises spontaneously from one’s perceptualexperience —that’s what matters. Whether or not the wrongness-content isalready there in perceptual experience itself, as opposed to being a “cog-nitive overlay” that only enters in the cognitive transition from perceptualexperience to belief, is not important for present purposes. And again, moralphenomenology per se is methodologically neutral about this.27

D. Phenomenology and introspective accessibility

As we have characterized phenomenology, both as method and assubject matter, the issue is left open whether or not all aspects of men-tality that fall within the subject matter of phenomenology are reliablyintrospectable. One might think this couldn’t be so, since the subjectmatter is supposed to be that which is available to introspection. How-ever, one needs to distinguish between what is present in experience orexperientially given, and what aspects of the latter are reliably ascertainableintrospectively. Here are two examples that illustrate this point. You mayrecall from a famous essay by Roderick Chisholm the case of glancing ata speckled hen.28 Plausibly, a determinate number of speckles are given in

25 See, e.g., Susanna Siegel, “Which Properties Are Represented in Perception?” in T.Gendler Szabo and John Hawthorne, eds., Perceptual Experience (Oxford: Oxford UniversityPress, 2005), and Siegel, “The Phenomenology of Efficacy,” Philosophical Topics 33 (2005):265–84.

26 Gilbert Harman, The Nature of Morality (Oxford: Oxford University Press, 1977), 4.27 Note, however, that careful application of the introspective method —or of a mixed

methodology that is partly phenomenological and partly abductive —might well be veryrelevant to such disputes. See, e.g., Siegel’s mixed-method argumentation for the claim thatcausation is represented in visual-perceptual experience in Siegel, “Which Properties AreRepresented in Perception?” and “The Phenomenology of Efficacy.”

28 Roderick Chisholm, “The Problem of the Speckled Hen,” Mind 51 (1942): 368–73.


the experience and thus present in experience. But one cannot reliablyascertain that number just via introspection, especially if the experience isfairly short-lived. Here is the second example. Consider the question ofwhether the content of agentive experience is compatible with state-causal determinism. Plausibly, the answer to that question is fixed just bythe nature of the experience itself, as the experience is self-presented tothe agent. Arguably, however, introspection alone will not allow one toreliably ascertain that answer.29

The fact that certain features can be present in experience without beingavailable to introspection is very important for our purposes in this essay,because it underscores certain potential limitations in the extent to whichfacts about experiential character can be ascertained just by means ofintrospectively attending to one’s own phenomenology. Perhaps, forinstance, (1) there is a definite fact of the matter about whether or notdirect moral experiences carry ontological objective purport, and yet (2)this fact of the matter (whatever it is) is not introspectively accessible.Then the argument for phenomenological inspection will be in trouble,even if those who propound that argument happen to be correct (thoughwe ourselves doubt this) in their contention that direct moral experiencesdo carry ontological objective purport.

E. Moral phenomenology characterized

We have just been emphasizing how one might understand phenom-enology in a manner that is noncommital with respect to a number ofcontested issues about its subject matter and methodology. To brieflyreview, the key points are these. The subject matter of phenomenologycomprises occurrent mental states that have a what-it-is-likeness (broadlyconstrued), whether or not all such states possess a phenomenal charac-ter. The methodology of phenomenology is introspection, through whichone has direct first-person access to one’s own conscious mental states.Here the point to stress is that one should not assume that all of one’sconscious mental states (or all conscious elements of them) are directlyintrospectively accessible.

With these caveats in mind, we propose the following as a character-ization of moral phenomenology:

(MP) Moral phenomenology includes as subject matter all of thoseoccurrent mental states with moral content that have a what-it-is-likeness (broadly construed), whether or not all such states(or all aspects of such states) possess phenomenal character.The method characteristic of moral phenomenology involves

29 For more on this theme, see Terry Horgan, “Agentive Phenomenology and the Limitsof Introspection,” Psyche 13 (2007): 1–29.


the use of first-person introspective access to conscious mentalstates with moral content, whether or not all such states (or allconscious elements of such states) are directly introspectivelyaccessible.30

Although our characterization of moral phenomenology is neutral aboutwhether there is more to the subject matter than what is reliably intro-spectively ascertainable, not all the positions mentioned in Section I arethus uncommitted. The thesis called Neutrality, we take it, is committedon this issue (at least on one natural assumption, namely, that directmoral experience either determinately does, or else determinately doesnot, have ontological objective purport). The Neutrality picture is this: onthe one hand, there is a fact of the matter about whether or not directmoral experience carries ontological objective purport, but, on the otherhand, this fact of the matter (whatever it is) is not ascertainable intro-spectively. We will return to this thesis below in Sections IX and XI.31

IV. The Argument from Phenomenological Introspection

As we have said, a main task of this essay is to zero in on a kind ofmetaethical argument that appeals to an alleged fact about moral phe-nomenology that allegedly is introspectively accessible (namely, that moralphenomenology carries ontological objective purport), as a basis for con-cluding that the phenomenology of moral experiences provides a protanto reason to favor an objectivist metaethics. More precisely, we areinterested in a species of phenomenological argument that has the fol-lowing features.

1. The sort of strong objectivity being argued for is ontological. How-ever, it does not matter for our present purposes whether the argument isbeing made in favor of robust ontological moral objectivity or for a moremodest, sensibility form of such objectivity.

30 Our broad characterization of moral phenomenology allows for experiences (or ele-ments of them) that give rise to moral judgments to be included within the purview of moralphenomenology. For example, my experiencing a certain contemplated action as being (inrelation to certain elements of my circumstances) unfitting may result in my morally comingdown on the matter and judging that the action is all-in unfitting and ought not to be done.(For more on the distinction between experiences of prima facie and of all-in unfittingnessand fittingness, see note 56 below.) And, of course, our characterization allows for experi-ences that include (perhaps exclusively) the having or making of a moral judgment. (Thisparticular point about the breadth of moral phenomenology is prompted by some remarksby Philip Pettit in conversation.)

31 The following point bears emphasis. If indeed not all aspects of mentality that belongto the subject matter of phenomenology are introspectively accessible (say, because someaspects of phenomenal character are not thus accessible but are present in experience none-theless), then the distinctive phenomenological method (namely, introspection) will not suf-fice by itself to answer all pertinent questions about the subject matter. Other methods willneed to be brought to bear too, over and above introspection.


2. The argument is supposed to be a distinctive form of metaethicalargument, one that is different from arguments appealing to consider-ations that concern theoretical semantic issues, or to metaphysical andepistemological issues concerning moral thought and discourse.32

3. Part of what makes the argument distinctive is that it relies cruciallyon an appeal to introspection: the argument appeals to putative featuresof one’s own moral experience that one allegedly can detect by turningone’s attention inward and focusing on the relevant kind of experience.Those who appeal to matters of moral phenomenology as a basis fordefending ontological moral objectivism seem to think that whether ornot the phenomenology carries ontological objective purport is some-thing that is introspectively accessible. This entails the denial of the Neu-trality thesis. But suppose that the Neutrality thesis is correct. Presumably,then, mere appeal to introspective awareness will not be sufficient inarguing the pro tanto case33 for ontological moral objectivism; rather, theobjectivist will need to bring to bear other, nonphenomenological consid-erations (e.g., theoretical questions about moral semantics and about moralmetaphysics) to make her case. Once this happens, however, the sup-posed distinctiveness of the argument from phenomenological introspec-tion is lost.

4. Finally, the argument presupposes that people by and large do sharea common moral phenomenology whose elements provide a suitablypretheoretical basis to which the metaethicist can appeal in making a casefor some form of objectivist metaethics. The idea here is that if the argu-ment in question is to have any evidential weight with respect to the protanto plausibility of competing metaethical views, there must be somemetaethical theory-independent facts about the phenomenology to whichan appeal can be made and which clearly are fully compatible with somebut not all existing metaethical theories. Furthermore, for purposes of thephenomenological argument for ontological moral objectivism, there mustbe some pretheoretical elements of moral experience that specifically favoran ontological-objectivist metaethical view, rather than, for instance, sim-ply ruling out (say) crude emotivist views.

To sum up, then, the phenomenological argument we wish to consideris properly understood as an argument from introspection that attemptsto pick out commonly shared pretheoretical elements of moral experi-ences whose ontological objective purport is introspectively accessible.

32 Note that what we are here calling “theoretical” semantic issues are meant to be distinctfrom issues concerning the introspectively accessible content of intentional mental states.

33 The case is pro tanto, rather than complete in itself, because one theoretical option is anerror theory asserting both (1) that it is an introspectively accessible fact that direct moralexperience carries ontological objective purport, and (2) that there are no in-the-world moralproperties or facts —and, hence, (3) that direct moral experience is systematically nonverid-ical in its intentional content. Those who seek to defend moral realism by appeal to theargument from phenomenological introspection need to say something about the theoreticaldisadvantages of such an error theory, in comparison to moral realism.


V. A Taxonomy of Moral Experience

Unfortunately, the term “moral phenomenology,” at least as used byphilosophers, is accordion-like in its usage. Sometimes, it is used verybroadly to refer to any and all of what are considered to be deeply embed-ded features of moral thought and discourse including: (a) its grammarand logic; (b) people’s “critical practices” regarding such thought anddiscourse (including, for example, the assumption that genuine moraldisagreements are possible); and (c) the what-it-is-likeness of variousmoral experiences, including, but not restricted to, occurrent concreteexperiences of morally judging some action, person, institution, or otheritem of moral evaluation. Used very narrowly, the term refers to what wemight call “raw affective feeling states” comprising only a proper subsetof what-it-is-likeness elements in moral experience. As we explained abovein Section III, we construe the scope of phenomenology to range oversensory, cognitive, desiderative, as well as affective experiences. So we doplan to use the term neither very narrowly, nor very broadly, but asreferring to occurrent mental phenomena that seem constitutive of a broadrange of everyday moral experience.

Nonetheless, we need to be more explicit about the kind of moral expe-rience we wish to consider. To do so, we offer a taxonomy of types of moralexperience so that we may triangulate a proper part of moral phenom-enology that will be the focus later in this essay. To anticipate: we are par-ticularly interested in the what-it-is-likeness of (1) judgment-involving(2) first-order (3) experiences of moral obligation that are (4) direct and(5) intuitive. We now proceed to briefly explain these five characteristics.

Judgment-involving moral experiences. Many moral experiences involveas a component coming to have or form a moral judgment. For instance,after thinking it over for a while, one comes to think that the UnitedStates ought not to have invaded Iraq. Part of one’s overall experiencehere is the coming to have or make a moral judgment. Some moralphenomenologists claim that an important variety of moral experiencedoes not involve having or making moral judgments. According toHubert and Stuart Dreyfus, there are cases of what they call “ethicalcomportment” in which a morally skilled individual spontaneously per-forms some morally appropriate action in some circumstance withoutmaking or having a moral judgment.34 Spontaneously reaching out ahelping hand to someone about to slip and fall is perhaps an example.We do not take a stand on whether such cases are properly describedas not involving moral judgment (although frankly we doubt it, becausewe believe that such spontaneous actions are more plausibly construed

34 Hubert L. Dreyfus and Stuart E. Dreyfus, “What Is Morality? A PhenomenologicalAccount of the Development of Ethical Expertise,” in David Rasmussen, ed., Universalism vs.Communitarianism: Contemporary Debates in Ethics (Cambridge, MA: The MIT Press, 1990),237–64.


as the unhesitant effects of moral judgments that are themselves spon-taneous rather than deliberative). But in case such actions do meet thatdescription, our focus is properly characterized as being on judgment-involving moral experiences.

First-order and second-order. Certainly experiences of guilt, shame,indignation, and moral anger are important in the study of moral phe-nomenology. But these moral emotions are typically second-order moralexperiences, since they are directed toward actions and other items ofmoral evaluation that are judged to be morally wrong or bad. One judges,for instance, that Tracy has benefited from her moral wrongdoing, whicharouses in one the moral emotion of indignation. What we are callingfirst-order moral judgments, then, are those more basic judgments ofobligation and value that may prompt some particular moral emotion ofthe sort just mentioned.35

Judgments of obligation and judgments of value. It is common to distin-guish judgments of moral obligation from judgments of moral value.One might suppose that the phenomenologies involving judgments ofone type are much the same as the phenomenologies involving theother type. But Maurice Mandelbaum, in his overlooked 1955 treatiseon moral phenomenology, made a further distinction that challengesthis supposition.

Direct and removed judgments of obligation. According to Mandelbaum, thereis an important phenomenological difference between what he called“direct” and “removed” moral experience. Direct moral experiences arethose in which one is presently confronted with a set of circumstances whichone experiences as “calling for” one to either act or refrain from acting ina particular way on that occasion, and in response to which one comes tohave or make a moral judgment about what one ought or ought not do. Bycontrast, removed moral experiences include those which involve the mak-ing or the having of an ought-judgment about one’s past self or about some-one else, as well as all judgments about the moral goodness or badness ofthe specific character traits and overall character of oneself and others. Man-delbaum claims that one phenomenological difference between such typesof experience is the fact that a direct moral experience

evokes emotion [which], like fear or anger, is experienced as a stateof the self and is directly related to action. [By contrast] the stirredup-ness and pressures which are present in direct moral judgments haveno counterpart in removed moral judgments.36

35 This is not to deny that, in many cases, one’s moral emotions “run ahead” of one’sfirst-order moral judgments: in thinking about some past action of his, John begins to havefeelings of guilt, and it is through his feelings of guilt that he comes to realize the wrongnessof what he did. One way in which this might happen is described by Robert Audi in “TheAxiology of Moral Experience,” The Journal of Ethics 2 (1998): 355–75.

36 Mandelbaum, The Phenomenology of Moral Experience, 127.


For purposes of this essay, we adopt Mandelbaum’s direct/removed dis-tinction. Our focus will be on experiences of the former type, about whichwe will say more in later sections.

Intuitive and deliberative moral judgments. Intuitive judgments are psy-chologically spontaneous in that they occur “without a conscious aware-ness of having gone through steps of searching, weighing evidence, orinferring a conclusion.” 37 Earlier, we mentioned Harman’s example whereyou round a corner and see a group of hoodlums pore gasoline on a catand set it on fire. As Harman says, “you do not need to conclude that whatthey are doing is wrong; you do not need to figure anything out; you cansee that it is wrong.” 38 By contrast, deliberative moral judgments resultfrom such activities as consciously searching, weighing evidence, andthen inferring a moral conclusion.39

Figure 1 presents a “set-aside” sequence that summarizes the variousdistinctions we have been making. As one moves down levels, the arrowspointing to the right point to a category of experience that is being dis-tinguished and set aside from the type of experience under the verticalarrow.

Besides allowing us to specify the type of moral experience that will bethe focus of our investigation, the taxonomy in figure 1 is a reminder thatone should not assume that moral experiences of all types exhibit somecore phenomenological elements that unify these experiences as beingdistinctively moral.40 This is important for questions about moral objec-tivity, because it is possible that some types of moral experience includeelements that support one conception of moral objectivity, while othermoral experiences support a different conception, and perhaps still otherssupport neither conception. Still, it bears emphasis that philosophers whowant to argue that moral experience carries ontological objective purporttend to focus on intuitive moral experiences, while those who are pri-marily interested in defending the rationalist conception tend to focusinstead on the phenomenology of moral deliberation.

Furthermore, and most importantly for our present concerns, thereis the question of whether some type of moral experience —whether

37 Jonathan Haidt, “The Emotional Dog and Its Rational Tail: A Social Intuitionist Approachto Moral Judgment,” Psychological Review 108 (2001): 818.

38 Harman, The Nature of Morality, 4.39 Cases involving intuitive moral judgments are to be contrasted with cases of “ethical

comportment” of the sort that the Dreyfus brothers discuss in “What Is Morality?” in whichone spontaneously responds as a matter of reflex. The latter are cases of experiences that donot seem to involve having or making a moral judgment, not even a spontaneous judgmentthat generates spontaneous, unhesitating behavior.

40 Strictly speaking, moral experiences might exhibit some common features that serve, ina weak sense, to unify them; however, it is possible that such common features do not serveto distinguish moral experiences from certain types of nonmoral normative experiences. Formore on this, see Horgan and Timmons, “Moral Phenomenology and Moral Theory.” Seealso Walter Sinnott-Armstrong, “Is Moral Phenomenology Unified?” Phenomenology and theCognitive Sciences 7 (2008).


intuitive or deliberative —is neutral with regard to such metaphysicalmatters and, in particular, whether by introspection alone one can deter-mine whether the experiences in question carry ontological objectivepurport.

As we have noted, our focus will be on the what-it-is-likeness of thosemoral experiences that combine all of the elements in the left-hand columnof figure 1. Henceforth, we will refer to such experiences simply as directexperiences of moral obligation, even though we mean only to be focusingprimarily on direct and intuitive moral experiences. And, as we have beensaying, our main question about such experiences is whether it is intro-spectively accessible that such experiences carry ontological objectivepurport.

VI. Two Examples of Direct Moral Experience

As we have just explained, “direct” moral experience refers to thoseexperiences in which one presently encounters what one takes to be amorally significant situation, a situation that seems to “call for” one to act

Figure 1. Taxonomy of Moral Experience


or refrain from acting and which involves as a constituent an all-in judg-ment of obligation about what one ought or ought not do in the presentcircumstances. Here are two examples.

A. Sophie’s conversation

Sophie and Audrey have enjoyed a close friendship spanning threedecades. Despite living in different parts of the world for the pasttwelve years, their friendship has not eroded. E-mail has made stayingin touch on a weekly, sometimes daily, basis easy given their familylives and heavy work schedules. Sophie has just received exciting newsthat she is to be promoted from her present position to a vice presidentof the corporation for which she has worked for nearly twenty-fiveyears. This is the sort of news befitting a phone call to Audrey. Shecalls. But instead of being greeted with Audrey’s typical hearty andcheerful “Hello. Audrey speaking,” she hears a weak and decidedlydownbeat “Hello.” Sophie hesitates a moment: “Hello? Audrey, is thatyou? This is Soph.” Short silence. “Are you there? Are you all right?”Audrey: “Yes, it’s me, Sophie, but I’m afraid I have some bad news. . . .I’ve just been let go . . . my job, I mean.” Sophie: “My God! I’m sosorry! This is terrible news!”

This begins a long conversation between Sophie and Audrey duringwhich Sophie is mostly silent, allowing Audrey to express her disappoint-ment, embarrassment, and anger over her job loss. Early on in this con-versation, it is patently clear to Sophie that now is obviously not the timeto share her good news with Audrey —she should wait for a later, moreappropriate occasion. Her judgment about waiting is psychologically imme-diate: she does not consciously rehearse or weigh various considerations;rather, once she hears Audrey’s voice and what Audrey has to say, herjudgment to withhold the news is spontaneous.

B. Rashid’s appointment

It is Friday, the last day of the semester before finals, and Rashid arrivesat the department headed for his office bright and early. In past weeks, hehas been working furiously on a paper due yesterday which he managedto send off in the eleventh hour. As he walks through his building towardhis office, Rashid is experiencing a sense of calm as he reflects with reliefon what he has managed to accomplish during the semester: a paper justcompleted and sent off; a large introductory course with two hundredstudents that was for him a new preparation with many hours spentworking up slide presentations; meeting every other week with teachingassistants; a departmental hire —not to mention the damaged roof at homeand the time spent wrangling with insurance agents, scheduling repairs,training Barkley, his new Lab puppy, and so on. Over the past two weeks,


Rashid has had to ignore some things, including a flood of e-mail whichhe plans to spend the morning sorting through. He thinks he really shouldnot have ignored so much e-mail, but he finds dealing with e-mail a hugedistraction, so on occasion he has to take draconian measures and ignorethe urgent in order to tackle the truly important. Dealing with his inboxought to take about three hours, he guesses, then it’s home again to packfor a short, much-needed vacation. Teaching assistants are giving the finalexamination on Monday; he will be back to submit final grades the fol-lowing Friday. With his e-mail out of the way, Rashid will be able to relax.Ah, sometimes life is good; as he unlocks his office door, he now feelspositively cheery.

Ready to work, Rashid turns on his computer and, as always, hisdaily calendar pops up. He is about to minimize it when he notices anentry for today at 8:30 (in about half an hour) —and he remembers.Many weeks ago (so it seems), he made an appointment with a strug-gling student who had asked Rashid to help him go over comments hehad received from one of his teaching assistants on his most recentpaper. Rashid allows students to rewrite papers, and this one is due nolater than Monday’s final-exam period. Rashid’s cheery mood is replacedby a mild sinking feeling as he begins to realize how much time it willlikely take to provide useful help to this student. For one thing, he willneed to dig out the student’s paper and reread it, and he can predictthat meeting with this student is going to take quite a while. Rashidcould, of course, close his door, turn out the light, and just not answerthe 8:30 knock. But given the present circumstances, which include hisrelationship with and promise to his student, Rashid does not need tofigure out what he ought to do; it is (in light of his recognition of thecircumstances) immediately clear to him. With a deep sigh, Rashid beginsto clear the books from a chair so his student will have a place to sit.He then picks up the paper and begins to read; catching up on e-mailwill just have to wait.

The examples of Sophie and Rashid are cases of direct moral experi-ences. One might think that there is not much to say in addition to thedescriptions of Sophie’s conversation and Rashid’s appointment that wehave already given: in such cases (were one to occupy either of theirroles), one senses what one ought or ought not to do, which is the contentof some occurrent psychological state. Thus (the thought might continue),although there may be a lot to say about the psychology of an individuallike Sophie who spontaneously judges that now is not the time to conveygood news to her friend, what there is to say has to do with psychologicalmatters that are not occurrent and so not part of her on-the-spot phe-nomenology. We deny this. We claim that the phenomenology in suchcases typically includes quite a lot of rich phenomenological detail —detail that one finds in Mandelbaum’s characterization of the phenom-enology of such direct moral experiences.


VII. Mandelbaum’s Phenomenology of DirectMoral Experiences

According to Mandelbaum, the sorts of direct moral experiences had bySophie and Rashid involve two “levels”:41 (1) a “felt demand” which is,as Mandelbaum says, “phenomenologically grounded in” (2) an appre-hension of a contemplated action being unfitting (relaying her good newsin Sophie’s case, not keeping his appointment in Rashid’s case). Let usconsider these levels in order. Mandelbaum writes:

[A] demand is experienced as a force. Like other forces it can only becharacterized through including in its description a reference to itspoint of origin and to its direction. It is my contention that the demandswhich we experience when we make a direct moral judgment arealways experienced as emanating from “outside” us, and as beingdirected against us. They are demands which seem to be indepen-dent of us and to which we feel that we ought to respond.42

Three exegetical and interpretative comments are in order here. First, infurther explaining the “external” source of moral demands, Mandelbaumcontrasts them with nonmoral demands such as those associated withhunger, desire for attention, and sexual arousal, which we experience asbeing “within us.” 43 Second, according to Mandelbaum, this sort of desire/urge–independence grounds the sense of “objectivity” one takes one’sdirect moral experiences to have. In her conversation, Sophie is aware ofvarious features of her present circumstances that she experiences as“calling for” or demanding that she refrain from saying anything to Audreyabout her job advancement. Similarly, Rashid’s present circumstances(including considerations about his promise to a student, the student’scircumstances, and so forth) are experienced by him as placing a demandon his current behavior “emanating from” the desire-independent facts ofthe case. Third, in this phenomenological description, one can distinguishthree elements: (1) the raw affective element, which Mandelbaum describesas a felt “pressure” or “tension”;44 (2) the vector-like force aspect of theexperience, involving, as we have seen, a sense of its origin and direction(which, being directed against the one making the judgment, is “reflex-ive”);45 and (3) the overall motivational pull that the contemplated actionor omission (in the circumstances) seems to exert.

41 “Levels” is our term, but it seems appropriate given Mandelbaum’s talk of felt demandbeing “grounded in” one’s apprehension of fittingness or unfittingness.

42 Mandelbaum, The Phenomenology of Moral Experience, 54.43 Ibid., 55.44 Ibid., 55–56.45 Ibid., 55.


Turning now to the second level of phenomenological description, expe-riences of felt moral demand, according to Mandelbaum, are phenom-enologically grounded in the relational characteristics of fittingness andunfittingness. Mandelbaum writes:

When I experience a demand to keep a promise this demand doesnot issue from me, but is leveled against me: it is not that I want togive X five dollars which motivates me, but the fact that I feel obligedto keep my promise. The promise itself appears as an objective factwhich places a demand upon me whether I want to keep it or not. . . .In this type of case . . . it becomes clear that the element of moraldemand presupposes an apprehended fittingness: the envisionedaction places a demand upon us only because it is seen as connectedwith and fittingly related to the situation which we find ourselvesconfronting.46

Since the notion of unfittingness, unlike fittingness, seems intuitively toinvolve, or be closely related to, the idea of a demand, we take it to be themore basic notion here.47

Let us now put all of this together. Experiences of direct moral obliga-tion (of the sort Mandelbaum is describing) have these three main features:

(1) They are ought-judgment involving: an agent having or under-going such an experience judges of herself that in the presentcircumstances she ought or ought not perform some action.

(2) This ought-judgment is part of an overall moral experience inwhich one experiences a felt demand whose elements include (a) afeeling of pressure, (b) a sense of a vector-like force which has an“external” origin and is directed at oneself, and (c) an associatedmotivational pull toward either performing an action or refrain-ing from performing an action.

(3) This felt demand is experienced as based on an “apprehension” ofunfittingness —that is, of a contemplated action or omission’sbeing unfittingly related to present circumstances (as one takesthem to be).

We submit that Mandelbaum’s phenomenological description accuratelydescribes most, if not all, of the important elements of people’s typicaldirect moral experiences (at least we can speak for ourselves). Thus, in

46 Ibid., 67–68.47 For more on this point, see Horgan and Timmons, “Prolegomena to a Future Phenom-

enology of Morals.” Interestingly, Roderick Chisholm, “Practical Reason and the Logic ofRequirement,” reprinted in Joseph Raz, ed., Practical Reason (Oxford: Oxford UniversityPress, 1978), 118–27, distinguishes two senses of fittingness, a strong and a weak sense, bothof which he defines in terms of requiredness. (Chisholm mentions Mandelbaum.)


examining this phenomenology for any objective pretensions it mayinvolve, one can ask whether any of these elements carry ontologicalobjective purport. Answering this question, we think, is delicate, whichmay explain why one finds disagreement among philosophers about howto properly answer the question. Our strategy in trying to answer it willinvolve two steps.

First, we will describe a metaethical view that denies ontological moralobjectivity, and that we have elsewhere argued is theoretically attractivefor various reasons over and above matters of moral phenomenology; wecall it “cognitivist expressivism.” 48 Second, we will consider the threemain features of the phenomenology of direct moral experience, askingabout each of them whether, alone or in combination, it is an introspec-tively accessible fact that they carry ontological objective purport. In eachcase, we argue that the answer is negative, on the grounds that cognitivistexpressivism, despite repudiating ontological objectivity, can smoothlyaccommodate the introspectively accessible aspects of the phenomeno-logical features in question.

Note that this line of argument assumes only that cognitivist expres-sivism is a credible candidate for being the correct metaethical position —not that it is the correct position. If a credible candidate that eschewsontological objectivity can accommodate the introspectively accessibleaspects of the relevant phenomenology, then that is enough to show thatintrospection alone does not yield a positive answer to the question aboutwhether moral phenomenology carries ontological objective purport.

These two principal steps in our argument will not quite establish theNeutrality thesis set out in Section I, because they do not preclude thepossibility that introspection alone yields a negative answer to the ques-tion about ontological objective purport. Nonetheless, the Neutrality the-sis does emerge as a corollary, given the very plausible claim (which weourselves are happy to concede) that various ontological-objectivist meta-ethical positions can also accommodate the introspectively accessibleaspects of moral phenomenology. By executing our two-step argument,then, we will have provided a defense of the Neutrality thesis —the claimthat introspectively accessible aspects of direct moral experiences do notthemselves determine an answer to the question about whether suchexperiences carry ontological objective purport. And the Neutrality thesisundermines the argument from phenomenological introspection, sincethat argument rests on the contention that it is an introspectively accessiblefact that direct moral experiences carry ontological objective purport. Thus,in order to ascertain whether or not experiences carry such purport, one

48 See Horgan and Timmons, “Cognitivist Expressivism.” The view we describe does treatmoral judgments as objective, in a specific sense to be described in Section X below. But theview is a form of expressivism, and thus it also treats moral judgments as being nondescrip-tive in their overall content. This precludes them from carrying ontological objective purport,because the latter is a species of descriptive content.


must look to other considerations, including perhaps other types of moralexperience and certain nonexperiential metaethical considerations.49

VIII. The Challenge

We think those who defend strong versions of moral objectivism, aswell as those who hold error theories,50 tend to move too quickly fromconsiderations of moral phenomenology to claims about the sort of objec-tive pretensions allegedly possessed by the relevant phenomenology. Wesay this not only because philosophers have tended not to dwell onphenomenological detail, but also because there is one metaethicalposition, “cognitivist expressivism,” that (a) seems very promising in itscapacity to accommodate the facts of moral phenomenology (withoutoverlooking or distorting the facts in question) but (b) denies that moraljudgments are robustly objective. The seeming promise of this view is thebasis for raising a challenge to advocates of the phenomenological argu-ment. We challenge them to point to those putative aspects or elements ofthe phenomenology of direct moral experience whose ontological objec-tive purport is introspectively accessible. More specifically, we challengethem to point to putative phenomenological aspects or elements thatarguably cannot be accommodated by cognitivist expressivism. To fleshout this challenge, we first present a few of the key elements of cognitivistexpressivism, and we then explain why and how this position can smoothlyaccommodate the phenomenology of direct moral experience.

A. A very brief introduction to cognitivist expressivism

Cognitivist expressivism is a metaethical position that is largely over-looked because of a widespread but (we claim) mistaken assumption —what we call the “semantic assumption.” According to this assumption,all beliefs are descriptive beliefs. To be more precise: according to thisassumption, every belief is to be understood as being a kind of commit-ment state —what we call an “is” commitment state —with respect tosome way-the-world-might-be content (i.e., some descriptive content). So,just as, for instance, one’s nonmoral belief (at some time t) that John tookout the trash is to be understood as one’s being is-committed (at t) to aparticular descriptive content —that John took out the trash —so it is withmoral belief (assuming that moral judgments are beliefs). The belief that

49 Here, again, we emphasize that our thesis concerns what is introspectively accessible inorder to distinguish what we are calling the Neutrality thesis from a stronger neutralitythesis according to which moral phenomenology itself is neutral with respect to ontologicalobjective purport.

50 According to an error theory, (1) affirmative moral judgments purport to representobjective moral facts, but (2) since there are no such facts, (3) affirmative moral judgmentsare erroneous (false).


John ought to take out the trash is a matter of being is-committed to thefollowing, putative, descriptive content: that it ought to be that John takesout the trash.

In both nonmoral and moral cases, then, to believe is to be is-committedto a way the world might be, i.e., to a descriptive content. Thus, if moralought-judgments are beliefs, then, according to the semantic assumption,they have moral descriptive content.51 And if they do, then they purportto be about in-the-world (instantiated) moral properties. Furthermore, ifone denies that there are any such properties instantiated by typical itemsof evaluation (actions, persons, institutions), then one is committed to anontological error theory with respect to morals. This was J. L. Mackie’sview.52

Cognitivist expressivism rejects the semantic assumption. This viewrecognizes two importantly different species of belief: in addition tois-commitments (call them “is-beliefs”), there are ought-commitments(“ought-beliefs”). The latter sort of commitment states share much of thephenomenology of is-beliefs, and also share many key functional-rolefeatures possessed by is-beliefs. But what distinguishes these two typesof beliefs is the fact that an ought-belief involves a certain kind ofcommitment —an ought-commitment —directed toward a non-moral-descriptive content. On this picture, to believe that John ought to take outthe trash is to be ought-committed to the non-moral-descriptive way-the-world-might-be content: that John takes out the trash. One might capturethe idea here by saying that, on this conception of moral belief, the oughtis in the attitude, rather than in a descriptive content to which one is (inan ought-ish way) committed.53

This general conception of belief treats moral judgments as genuinebeliefs, but, strictly speaking, it is noncommittal with respect to whethermoral beliefs also have moral descriptive content. Thus, the framework

51 Here we use the term “descriptive,” as applied to content talk, to include content thatpurports to attribute irreducibly normative properties to items of evaluation. On this usage,J. L. Mackie’s claim that moral judgments purport to ascribe to actions the alleged normativeproperty to-be-pursuedness counts as a construal according to which such judgments possessdescriptive content.

52 J. L. Mackie, Ethics: Inventing Right and Wrong (Harmondsworth: Penguin, 1977),chap. 1.

53 On our view, the is in the case of is-belief is also in the attitude. The idea is that thedescriptive content of a belief —of either an is-belief or an ought-belief —is most perspicu-ously expressed not by a declarative sentence but rather by a “that”-clause such as “thatJohn takes out the trash” (or by a nominalized sentence such as “John’s taking out thetrash”). In English, an is-commitment is canonically expressed linguistically by asserting acomplete sentence in the declarative mood —as in “John took out the trash.” An ought-commitment is canonically expressed by asserting a declarative-mood sentence whose pred-icative constituent comprises the modal auxiliary “ought” appended to an infinitival verb —asin “John ought to take out the trash.” On our account, however, the logical structure ofis-beliefs and ought-beliefs is more perspicuously revealed via sentences employing acommitment-operator applied to a descriptive “that”-clause, thus: “It is the case that Johntakes out the trash”; “It ought to be the case that John takes out the trash.”


could be embraced by a moral realist: the claim would be that the beliefthat John ought to take out the trash, for example, is both an ought-commitment to the descriptive content that John takes out the trash and anis-commitment to the putative moral-descriptive content that John ought totake out the trash. However, the framework is also compatible with deny-ing that moral beliefs have moral descriptive content. In some of our pastwritings, we have argued for a non-error version of metaethical irrealism,in which we combine a cognitivist view of moral judgments (i.e., a viewthat construes them as genuine beliefs) with a nondescriptivist/irrealistview about such judgments —hence, what we call “cognitivist expressiv-ism.” On this view, then, there is no such way the world might be morallyas, for example, John’s taking out the trash being an action that ought to occur.But there is a distinctive type of ought-commitment with respect to adescriptive way the world might be, as in John’s taking out the trash. Andthis commitment-state is a belief.

B. The challenge

Of course, there is much to do in defending a cognitivist version ofexpressivism, and elsewhere we have undertaken its defense.54 The pointwe wish to make on this occasion is that this sort of view is a metaethicaloption which promises to accommodate the idea that moral judgmentsare genuine beliefs (a claim, by the way, that is not beyond question), yetdenies that such judgments carry moral descriptive purport (and hencealso denies that they carry ontological objective purport).

Thus, if the phenomenological argument is going to create a presump-tion in favor of an ontological conception of moral objectivity, its advo-cates will need to pinpoint features of concrete moral experience withontological-objectivist pretensions that are introspectively evident. Ourchallenge to the advocates of the phenomenological argument is to pin-point those features of moral experience that supposedly support strongmoral objectivity and to argue that it is introspectively manifest that thosefeatures really do embody ontological-objectivist pretensions.

IX. The Phenomenology of Unfittingness

In addition to the belief-like aspects of ought-judgments that are part ofdirect moral experiences, we have identified two other general features ofsuch experiences: (1) they involve a sense of felt demand whose origin is“external”; and (2) this felt demand is phenomenologically grounded in a

54 Horgan and Timmons, “Nondescriptivist Cognitivism”; Terry Horgan and Mark Tim-mons, “Expressivism, Yes! Relativism, No!” in Russ Shafer-Landau, ed., Oxford Studies inMetaethics, vol. 1 (Oxford: Oxford University Press, 2006); Terry Horgan and Mark Timmons,“Cognitivist Expressivism,” in Horgan and Timmons, eds., Metaethics after Moore (Oxford:Oxford University Press, 2006).


sense (“apprehension”) of fittingness. Do either of these features, perhapsin combination, carry ontological objective purport of a sort that is intro-spectively accessible in direct moral experiences?

Clearly, the affective element involved in feeling a demand, whichMandelbaum describes as a pressure or tension, does not itself carry anysort of objective purport. Feeling a pull to perform an action or to refrainfrom performing one is something one often experiences without theexperience carrying any presumption of objectivity. I feel tempted by thepiece of chocolate cake on the table before me, and it is as if it werescreaming “Eat me!!!” It seems to be tugging me in its direction.55 Buthere one experiences no objective purport.

Of course, the crucial element of felt demand is the “external” originof the vector-like force —the fact that the demand is, as Mandelbaumsays, “emanating from ‘outside’ us,” which he glosses as its being inde-pendent of one’s desires. Moreover, this sort of demand is experiencedas grounded in one’s sense that a particular action is an unfitting responseto those “external” circumstances. The notions of fittingness and unfit-tingness are, of course, relational. As Mandelbaum sometimes puts it,to say that an action is unfitting in relation to a particular set of cir-cumstances is to say that features of those circumstances “call for”one’s refraining from performing the action. So, if direct moral experi-ences carry ontological objective purport in a way that is introspec-tively manifest, then such purport would seem to be located in thefollowing complex experience: a concrete action being called for by desire-independent features of one’s present circumstances. As we explained ear-lier, if a view like cognitivist expressivism, which denies that moralexperiences carry descriptive purport, can accommodate the introspec-tively manifest aspects in question (regardless of whether this type ofview can fully accommodate the phenomenology), and can do so atleast as well as a view that attributes to the phenomenology descrip-tive purport, then we will have secured the Neutrality thesis. To beclear, even if cognitivist expressivism can accommodate the introspec-tively accessible aspects of moral phenomenology as well as any com-peting ontological-objectivist metaethical view, this does not show thatthose aspects don’t carry ontological objectivist purport. Rather, what itshows (we claim) is that whether or not the aspects in question docarry such purport is not something that is introspectively accessible, andthat is enough to establish the Neutrality thesis. We now proceed tosketch two interpretations —a cognitivist-expressivist interpretation andan ontological-objectivist interpretation —to lay bare how they each pur-port to account for Mandelbaumian direct moral experiences.

55 This cake example is from Elizabeth L. Beardsley, “Moral Experience and Ethical Analy-sis,” The Philosophical Review 68 (1959), 519–30 (critical review of Mandelbaum’s The Phe-nomenology of Moral Experience).


We have lately called attention to how cognitivist expressivism under-stands direct moral ought-judgments as kinds of beliefs. What does theview say about direct moral experiences of being called upon by desire-independent considerations such that these experiences would have anobjective feel to them? There are two essential ideas. First, direct moralexperiences qua moral have to do with taking what we will rather vaguelycall a “non-self-privileging” stance toward one’s action and circum-stances. Taking this sort of stance involves being open to being affected bydesire-independent considerations that have largely to do with not harm-ing others. Different normative moral theories spell out this vague notionof non-self-privileging in different ways.

Second, and most crucially, in coming to be and experience oneself asbeing ought-committed to some course of action (or inaction), one expe-riences oneself as (1) becoming ought-committed in a non-self-privileging way,and (2) as becoming so committed because of certain non-normative fac-tual considerations. The idea is that such judgment-involving experiencesof direct moral obligation possess a kind of phenomenological unity: notonly is one aware of some objective, non-normative features of someobject of evaluation (my having promised to take out the trash), and not onlydoes one experience oneself as becoming ought-committed to a certainnon-normative descriptive way-the-world-might-be (my taking out the trash),but one experiences oneself as becoming thus ought-committed becauseof those objective features one is aware of —and, moreover, as ought-committed in a non-self-privileging way. Direct experiences of moralunfittingness, then, are a matter of becoming ought-committed in a non-self-privileging way to a certain action, on the basis of (what one takes tobe) objective in-the-world nonmoral features.56

56 The moral experiences that we are here calling direct ought-judgment-involving moralexperiences are all-in moral experiences of moral unfittingness. We also recognize moralexperiences of unfittingness that are like all-in moral experiences except that one does notcome to feel ought-committed; rather, one only experiences oneself as having a tendency tofeel ought-committed. These are cases in which one would find oneself ought-committed ifthe relevant non-normative feature in question were the only feature one took to be a reason.In contrast to all-in experiences of unfittingness (or fittingness), such cases are those of primafacie unfittingness (or fittingness).

This distinction between two types of experiences of moral unfittingness bears on anexample that Julia Driver described (in conversation) in which one comes across a pigeonlying on the ground (apparently hit by a car) that is in pain and obviously cannot be saved.One can walk away or crush its skull (thereby putting it out of its misery). One judges thatit is morally best (and hence fitting) to crush the pigeon’s skull, but in doing so one takesthe act of crushing to be unfitting. In this case, putting the pigeon out of its misery isexperienced as fitting, but the act of crushing is experienced as unfitting. Driver askedwhether our model of moral experience, featuring as it does experiences of unfittingnessand fittingness, can handle this case. Given what we have just said about the distinctionbetween all-in experiences of moral unfittingness (and fittingness) and experiences of primafacie moral unfittingness, we can say of Driver’s case that, on the one hand, given that theact in question is a crushing of a live animal’s skull, one experiences the contemplated actionas prima facie unfitting. On the other hand, given that the animal is suffering and cannot besaved, one experiences the act as prima facie fitting. On our picture, then, one has a ten-


It is worth emphasizing that this is a two-step construal of direct moraljudgment. (The “steps” are conceptually-discernible aspects, which mightor might not be sequential links in a causal sequence.) The crucial pointto appreciate, in distinguishing this non-ontological picture from an onto-logical picture, is the second step —the becoming committed because step.Here it is important to notice that the “because” is not simply a causal“because.” To construe it this way would be to confuse matters of causalexplanation with matters of normativity. What this picture holds, then, isthat the second step involves what we may refer to as a reasonish-becauseway of coming to have an ought-commitment.

In illuminating this reasonish-because way of coming to be ought-committed, cognitivist expressivism will want to approach the state ofmind corresponding to this way of coming to be ought-committed in amanner similar to the way the view approaches ought-judgments them-selves. As we noted earlier, cognitivist expressivism understands an ought-judgment as a sui generis kind of commitment-state vis-à-vis a coredescriptive content such as John’s taking out the trash. Cognitivist expres-sivism approaches judgments of unfittingness and fittingness in the sameway —as involving a kind of psychological commitment-state vis-à-vis apair of core descriptive contents that is not an is-commitment with respectto a putative in-the-world unfittingness fact. Theoretical understandingof such states comes from what we call “triangulation” —illumination ofsuch psychological states not via some reductive account of them but interms of both their overall phenomenology and their constitutive role(s)in a morally-judging agent’s overall psychological economy. This is aproject that we have partially addressed elsewhere, and it represents acentral task for the expressivist.57

dency to feel ought-committed to crushing the skull, and an opposing tendency to feelought-committed to refraining from crushing the skull. As in typical cases of conflicts ofprima facie duties, one must determine which consideration is all-in or most fitting in thecircumstances.

57 In the formal language featured in Horgan and Timmons, “Cognitivist Expressivism,”judgments of the form I ought to do A (now) are rendered as O[A], where “O” expressesought-commitment and “A” expresses a descriptive way-the-world-might-be content (e.g.,that I take out the trash now). Likewise, is-commitments are rendered as I[A]. Thus, “O[A]”is the canonical way of expressing linguistically one’s state of being ought-committed to adescriptive content, which is how cognitivist expressivism understands direct moral ought-judgments. As we have said, the idea is that the “ought” is in the attitude of the psycho-logical state, not in the content toward which the attitude is directed. The formal languagealso generates constructions corresponding to a whole hierarchy of logically complexcommitment-states —e.g., commitment-states of the type {I[ ] or O[ ]}, where such acommitment-state obtains with respect to a pair of descriptive contents expressable by“that”-clauses that would be inserted into the respective bracketed slots. To capture unfit-tingness judgments in such a formal language, a natural idea would be to augment therange of logically complex formal-language constructions to incorporate an adverbial oper-ator that expresses the “reasonish-because” manner of judging. Judgments of the form, Itwould be unfitting for me not to take out the trash now because I’ve promised to do so, or equiv-alently, My promising to take out the trash at this time is a reason for why I ought to take it out now,can be rendered: R(I[B]){O[A]}, which expresses a logically complex commitment-state of


Now that we have sketched this two-step picture, we can compare it tothe three-step picture that is apparently favored by the ontological moralobjectivist. On the three-step construal, the following discernible aspectsare allegedly involved: (1) apprehending some objective, non-normative,fact(s) (my having promised to take out the trash); (2) apprehending anobjective unfittingness fact (my not taking out the trash being unfitting inrelation to my having promised to take out the trash); and (3) becoming ought-committed to some objective non-normative way-the-world-might-be (mytaking out the trash) by virtue of step (2). On the cognitivist-expressivistpicture we are urging, in contrast, there is no such step (2). Although I doindeed apprehend my having promised to take out the trash as a reason totake out the trash, such reason-apprehension is constituted by becomingought-committed in the manner described in the preceding paragraph.And that is a two-step process (albeit a process involving substantialphenomenological unity): it involves (a) apprehending my having promisedto take out the trash, and (b) becoming ought-committed to my taking out thetrash in a way that is both non-self-privileging and based upon step (a) ina reasonish-because way.

Here, then, we have two pictures of the phenomenology of moral unfit-tingness, and the question is whether there is any introspectively acces-sible reason (in direct moral experience) to favor one over the other. Ourmain claim in this essay is that it is not introspectively obvious one wayor the other —that, in order to decide the issue, one has to importnonphenomenological theoretical considerations into the metaethical debatebetween the two camps in question.

X. Moral Objectivity with a Small “O”

Let us return for a moment to Mandelbaum’s claim that the phenom-enology of direct moral experience has an “objective” feel to it. Presum-ably, the objective feel of direct moral experiences emerges from the aspectsof phenomenology we have been describing. Can this sort of feel be

being ought-committed to a certain non-moral-descriptive content (namely, that I take out thetrash now) in a for-the-reason-that I promised to take it out manner. (More precisely, themanner is a for-the-reason-that-it-is-the-case [that I promised to take out the trash] manner.) Thatis to say, this is a commitment-state of the logical type R(I[ ]){O[ ]}, with respect to the pairof descriptive contents that I promised to take out trash and that I take out the trash now. (Weassume here that one takes the reason to be an all-in reason for the ought-judgment.) Ofcourse, this for-the-reason-that-I[ ] manner of becoming ought-committed is also a non-self-privileging manner, and so to capture this in our syntax, one might render the ought-operator thus: Oi[A], for “ought-impartial.” Thus, the full formal rendering of such animpartial, reasonish-because manner of being ought-committed would be this: R(I[B]){Oi[A]}.Again, such a logically complex commitment-state is not an is-commitment with respect toa putative descriptive content (namely, my having promised to take out the trash being a reasonfor my taking out the trash now). On the contrary: according to cognitivist expressivism, thereis no such descriptive content, and there is no such is-commitment to such a putativecontent.


accommodated by cognitivist expressivism? Indeed it can. In capturingthe objective feel in question, there are two elements of such experiencesthat combine to carry a sort of non-ontological-objectivist feel. First, oneexperiences oneself as judging in a non-self-privileging way —taking, as itwere, an impartial stance —and thus as less self-centered and more other-centered. Second, one experiences oneself as becoming committed to whatone ought to do (or refrain from doing) in what we have called a “reasonish-because” way. Thus, one experiences oneself as judging in an impartial,nonarbitrary, reason-based manner. Call this conception of moral objec-tivity small “o” objectivity;58 we think it nicely captures the kind ofobjectivity that is available to introspection when reflecting on directmoral experiences. As we have said, it remains possible that direct moralexperiences also carry commitment to a stronger form of objectivity; but,if so, this is not something that is introspectively accessible.

We claim, then, that cognitivist expressivism —which does not construemoral judgments (or the experiences that embed them) as carrying strongontological objective purport (or rationalist objective purport, for thatmatter) —can capture a notion of moral objectivity that seems to be all thatis needed to fully accommodate the introspectively manifest aspects ofdirect moral experiences.

XI. Objections and Replies

We anticipate various objections to the Neutrality thesis, at least as wehave defended it. Here, in rapid-fire succession, are those objections withour replies.

Mistake: A critic might claim that our phenomenological description ofdirect moral experiences is mistaken, that we leave out or misdescribesome crucial, introspectively manifest element that carries ontologicalobjective purport and that is introspectively accessible.

Reply: We think of our defense of the Neutrality thesis as one that shiftsa burden onto the backs of ontological moral objectivists. Thus, our oppo-nents need to point to those introspectively manifest features of directmoral experiences that cannot be accommodated by cognitivist expres-sivism as well as they can be accommodated by some form of ontologicalmoral objectivism.

Misdescribe: A critic might claim that it is introspectively obvious thatcognitivist expressivism itself just misdescribes the moral phenomenol-ogy in question, and that it is implausible for this reason. To be moreprecise, the claim might be that it is introspectively manifest that (a)moral judgments are beliefs, and (b) beliefs are descriptive in their overallcontent. (To say, for instance, that the belief that John ought to take out the

58 This phrase is inspired by Putnam’s phrase “realism with a small ‘r’.” See HilaryPutnam, The Many Faces of Realism (La Salle, IL: Open Court, 1987), 17.


trash is “descriptive in its overall content” is to say that the “that”-clausethat John ought to take out the trash characterizes a descriptive way theworld might be.)

Reply: Elsewhere, we argue against (b) partly on phenomenologicalgrounds.59 We will not rehearse those grounds here. It is worth noting,however, that even if one insists that genuine beliefs must be descriptive,there is a position that is like cognitivist expressivism that construesought-commitments as, say, quasi-beliefs, which would do the same workfor us in defending the Neutrality thesis as does cognitivist expressivism.It is also worth pointing out here that those who just insist that it isintrospectively evident to them that their phenomenology does carryobjective purport may be guilty of introspective confabulation. Generallyspeaking, it is easy to “read into” one’s phenomenology what is not reallythere. For instance, I see a saguaro cactus that strikes me (based on howit looks) as being about forty-five years old. I say, “Well, it looks aboutforty-five years old.” But here it would be implausible to suppose that theproperty of being roughly forty-five years old is something that is intro-spectively presented to me in my visual experience. Rather, the moreintuitively plausible thing to say here is that, based on my cacti experi-ence, I am able to reliably form beliefs about the age of cacti based onwhat is presented to me in my visual experience. Were I to introspect andconclude that my visual experience presents me with the property ofbeing forty-five years old, I would be guilty of introspective confabula-tion. A similar point holds, we think, for those who would attend to theirdirect moral experiences and claim that it is introspectively obvious thatthey have a divine “stamp” upon them or are of divine origin. And thesame sort of confabulation is going on (we would claim) when some folksclaim that it is introspectively obvious that their own phenomenology ofdirect moral experience (and perhaps their concrete moral experiencegenerally) has ontological objective purport. Making this case is (we admit)beyond the scope of this essay; it is something we plan to explore in ourwork in progress.

Denial: A critic might deny that cognitivist expressivism is an overallplausible metaethical view in ways other than not being able to accom-modate the introspectively manifest features of direct moral experiences.

Reply: This objection is beside the point. It is enough for our purposesthat cognitivist expressivism (assuming it is a coherent metaethical option)can accommodate the introspectively manifest phenomenology of directmoral experiences. In any case, this objection requires going beyond thephenomenology in question and bringing forth various additional theo-retical considerations that are relevant for evaluating a metaethical theory.Elsewhere, we defend cognitivist expressivism as plausible on generalphilosophical grounds.60

59 See Horgan and Timmons, “Cognitivist Expressivism.”60 See note 54.


Limit: A critic might claim that, for all we have said, the phenomenol-ogy in question may carry ontological objective purport, even if suchpurport is not introspectively accessible.

Reply: We completely agree that this is an option. Our only aim was totake on the argument from phenomenological introspection. So even ifmoral phenomenology does carry this kind of purport, we think it isimportant to point out that this putative fact about the phenomenology isnot reliably introspectable. This may explain why moral philosophersdisagree about the metaphysical purport of moral experiences of the sortwe have considered.

Myopia: A critic might claim that there are other types of concrete moralexperience and other deeply embedded features of moral thought anddiscourse (including logical embedding, our critical moral practices, andmoral deliberation), and that some of these features (perhaps in combi-nation) are best accommodated by an ontological-objectivist metaethics.

Reply: Perhaps so. But, as we have said, our aim in this essay has beennarrowly focused in order to examine one specific phenomenologicalargument. We admit that there is more work to be done in defending ametaethical view. It may be, for all we have said, that other types ofconcrete moral experience involve introspectively manifest aspects thatcarry ontological objective purport, though we doubt it. Having made acase that one type of concrete moral experience is neutral on this issue,the burden is squarely on the shoulders of anyone who thinks that someother type of moral experience does carry such purport —all the more sobecause the kinds of moral judgment we have focused on here (namely,direct judgments of moral obligation) are typically the ones most cited byadvocates of the phenomenological argument. As for the thought that onemust look beyond concrete moral experience to other features of moralthought and discourse in order to make a pro tanto case for ontologicalobjective purport, we note that this thought is one way of making ourmain point. As we explained in Section IV, the argument from phenom-enological introspection rests on the assumption that ontological objectivepurport can be read off from the phenomenology of concrete moral expe-riences without appealing to other features of moral thought and discourse.

Variability: A critic might note that, arguably, moral phenomenology istheory-laden: theoretical assumptions, including metaethical ones, satu-rate moral experience. Thus, in all likelihood, the moral phenomenologiesof different individuals will vary with respect to whether their moral expe-riences carry objective purport in general, and whether their direct moralexperiences carry introspectively manifest ontological objective purport.

Reply: This raises large methodological issues that we cannot pursuehere.61 However, this objection is bad news for the argument from phe-

61 Questions about the theory-ladenness of moral experiences were raised in conversationby David Wong. Michael Gill, in his “Variability and Moral Phenomenology,” Phenomenologyand the Cognitive Sciences 7 (2008), pressed this same worry. We reply to Gill in our “Prole-gomena to a Future Phenomenology of Morals.”


nomenological introspection. As we explained in Section IV, the argumentrests on the assumption that the phenomenology of moral experienceinvolves aspects that are pretheoretically there in the phenomenology andare widely shared —aspects about which we can theorize. If this assump-tion is false, then this is a problem for the argument. Furthermore, evenif there is such variability, there may still be a “core” phenomenologicallayer that is common. (We think there is.) If so, then this is where thepresent debate should be focused.

XII. Conclusion

We began with a commonly expressed appeal to moral experience andits phenomenology as a basis for favoring some form of ontologicalmoral objectivism. We also began with the idea that this sort of argument,so far as we can tell, tends to focus on concrete moral experiences ofobligation and value. We decided to focus on a very common sort ofconcrete moral experience —what Mandelbaum calls “direct moral expe-riences of obligation” —as a basis for exploring what we call the argumentfrom phenomenological introspection. According to this argument, thereare introspectively manifest aspects of direct moral experiences that havemoral ontological objective purport. We have challenged this claim. Wemaintain that the introspectively manifest aspects of such moral experi-ences do not determine whether they have this kind of objective purport.This is our Neutrality thesis, which we defended by explaining how ourversion of cognitivist expressivism can fully accommodate the introspec-tively manifest data of direct moral experiences, including the objectivistcharacter of such experiences.

Philosophy, University of Arizona


IMAGINATIVE RESISTANCE ANDPSYCHOLOGICAL NECESSITY*

By Julia Driver

Nature, by an absolute and uncontroulable necessity, has determin’dus to judge as well as to breathe and feel. . . .

—David Hume, A Treatise of Human Nature

I. Introduction

On Hume’s view, there are some things we must believe. In his analysisof and response to skeptical worries in the first book of A Treatise ofHuman Nature, he repeatedly notes that we are so constructed by naturethat we cannot maintain skeptical doubt. Here we have an allusion topsychological necessity. A claim such as “There is a door to my office”(given that I perceive a door to my office, etc.) is psychologically neces-sary for me. It is certainly not logically necessary or metaphysically nec-essary. It is simply something that I cannot avoid believing in practice.

It seems to me that Hume appeals to psychological necessity elsewherein the Treatise (though not so explicitly), as when he discusses our moralcommitments. When Hume notes that belief is more a matter of sentimentthan reason, he is not contradicting earlier claims; rather, he is talkingabout the practical commitment that goes along with belief. A distinctioncan be made between the truth of a proposition and appropriate com-mitment to a belief with that same propositional content. There is a sensein which a good belief is the true one; however, though the point of beliefmay be truth, the point of commitment to belief may not be. Here is wheresentiment properly enters the picture, providing the basis for a sentimen-talist account of our commitment to moral norms.1 On this view, such acommitment does not follow from logic and reason, but from passion.

In this essay, I discuss psychological necessity as it pertains to moralcommitments. I argue that some evidence for a Humean view of thepsychological necessity of certain moral beliefs in practice is provided bythe phenomenon of imaginative resistance. I also argue that this phenom-

* For their very helpful comments on an earlier version of this paper, I would like to thankthe other contributors to this volume. The paper has additionally benefited from conversa-tions with Walter Sinnott-Armstrong and Roy Sorensen. I would also like to thank EllenFrankel Paul for her detailed written comments on the earlier draft.

1 Moral sentimentalism is the view that moral norms derive authority from our passions(what we desire, what we care about). There are many different ways to spell this out inmore detail; for example, sentimentalism can be either cognitivist or noncognitivist.


enon, along with the view that there is a difference between truth con-ditions and commitment conditions, can offer the moral sentimentalist away to respond to a certain family of criticisms intended to portray thesentimentalist as being committed, in principle, to morally outrageousclaims.

II. Imaginative Resistance

When we read fiction, we tend to accept the author’s authority toestablish truth within the text, yet we also tend to reject claims made bythe author that are morally outrageous. The attempt to reconcile thesetwo tendencies presents us with a puzzle. After all, the author seems tohave authority when it comes to descriptively outrageous claims —sowhy not morally outrageous claims as well? We happily go along withtalking mice and time-travel tales, but we balk if recreational torture isendorsed, or presented as truly permissible within the story. This does notdemonstrate that there is an essential difference between the nonmoraland the moral. It simply shows that there are two issues at stake: first,understanding imaginative resistance; and, second, coming up with anexplanation of why, as readers of fiction, we have much more resistanceto things we deem immoral, as opposed to things we deem impossible(but are able to imagine, at least for the purposes of going along with thefiction). In response to the first issue, I would say that what some writershave referred to as imaginative resistance is a specific instance of a moregeneral phenomenon; and, in response to the second, I would say that thedifference in fictional interpretation between descriptive outrageousnessand normative outrageousness is tied to different standards for appro-priate storytelling in each case. We learn from fictions, just as we learnfrom considering some possibilities even if, at first, they seem outrageous.But in the case of moral outrageousness, as Tamar Gendler notes, we shyaway from it because we don’t want to consider the immoral as moraleven within fictional works, where we normally think the author author-itative.2 The reluctance has to do with there being no acceptable moralreasons provided within the story for switching perspectives between thereal world and the work of fiction. Further, moral reasons are taken,generally, to have a special authority over our actions, a practical signif-icance that other reasons lack. As we read a work of fiction, we may wellhave strong views about what a character ought to do —views that wouldbe contradicted by the supposed immoralities of the plot. This is anadditional source of the resistance to accepting those immoralities.

Let us consider the puzzle of imaginative resistance as it has beendiscussed in the literature. The puzzle has to do with the fact that we

2 Tamar Gendler, “The Puzzle of Imaginative Resistance,” The Journal of Philosophy 97,no. 2 (February 2002): 55–81.

302 JULIA DRIVER

refuse to play along when reading fictions in which normative claims thatwe do not share are put forward as true. Brian Weatherson asks us toconsider a story he calls Death, in which two people, Jack and Jill, arekilled by a fellow named Craig because they are blocking traffic whilehaving an argument. The story contains the claim “So Craig did the rightthing, because Jack and Jill should have taken their argument somewhereelse where they wouldn’t get in anyone’s way.” 3 The puzzle here ariseswhen we are asked to imagine that it is fine for Craig to have killed Jackand Jill for blocking traffic. We just can’t do this —which seems oddbecause we are willing to play along with authors of fiction when itcomes to all kinds of other weird things (e.g., magic rings and the like). Ofcourse, we do not think there could be such things as magic rings, but weare willing to imagine them. We are even willing to go along with logicalimpossibilities, such as the adventures of Hal the round square, whosecurvy points get him into all kinds of amusing scrapes. However, we arenot willing to go along with imagining that it is fine to kill people simplybecause they block traffic.

I think that a friend of necessity in ethics should be cheered by thisphenomenon, at least initially. In moral phenomenology, people oftenpoint to the fact that we feel a kind of necessity when faced with moralclaims we take to be true.4 It is not just that claims like “Torturing kittensfor fun is wrong” can be true, or just are true; such claims must be true.Likewise, in the case of the Death story, it seems that we simply cannotimagine that it is okay to kill for such a trivial reason. Might imaginativeresistance signal a kind of necessity for moral commitments? Perhaps, butit would depend on how we understand the workings of the puzzle.Gendler argues that “the primary source of imaginative resistance is notour inability to imagine morally deviant situations, but our unwillingnessto do so.” 5 It is not, presumably, that we can’t go along; we just won’t.Otherwise, we wouldn’t be able to go along with Hal the round square’sadventures either. Hal is an impossible object. But note that we do, Ibelieve, tend to edit out features of the fiction that are descriptively out-rageous and not necessary to advancing the plot. If one knows that onecannot hear explosions in space, one might just ignore that feature in themovie Star Wars. This indicates, I think, a tendency even to resist claimsin stories that are descriptively false (in the sense of violating a law ofnature, for example). It depends on the purposes of the storytelling. Some-

3 Brian Weatherson, “Morality, Fiction, and Possibility,” Philosophers’ Imprint 1, no. 3 (2004): 1.4 For example, in “Moral Phenomenology and Moral Theory,” Philosophical Issues 15 (2005):

56–77, Terry Horgan and Mark Timmons discuss Maurice Mandelbaum’s view that moraljudgments seem to impose demands on us categorically (that is, that they have “objectivepretensions”). Though these authors do not speak in terms of a feeling of necessity, theirdescription of the categoricity of moral experience implies this. The sentimentalist does notdeny that the feeling is there, but only denies that it necessarily indicates something objec-tive “out there” that imposes duties.

5 Gendler, “The Puzzle of Imaginative Resistance,” 56.

IMAGINATIVE RESISTANCE 303

times an author will intentionally flout norms; that is, the author, in a veryobvious way, will ignore or counter some norms or conventions so as topoint out their absurdity. For example, the movie Blazing Saddles floutsconventional Western storytelling norms. However, if (as in the Hal story)it is clear that these norms are being flouted, we go along out of amuse-ment, or out of a desire to see where the story takes us. Perhaps the pointof the Hal story is to illustrate some lessons concerning impossibility, forexample. Even so, one could hold that in the moral cases the resistance ismore extreme.

The imaginative resistance we feel in moral cases reflects our deeplyheld moral commitments, ones that we reflectively endorse. We cannot goalong with moral claims we take to be false, because that would seem toinvolve a rejection of something we care very much about; after all, wegenerally take moral reasons to be much more serious than other sorts ofreasons, even if we do not actually regard moral reasons as overriding.Humean sentimentalism ties this to our feelings of dismay at the thoughtof gratuitous harm. We cannot just excise that from our reaction to thefiction, except under circumstances where the dismay is avoided by theclear flouting of norms in the fiction. Thus, we “cannot” take the morallyoutrageous claims seriously.

Recall that Gendler explains imaginative resistance in terms of ourdesires. On her view, what we are resisting in these cases is being manip-ulated by the author —“manipulated into taking points of view that wewould not reflectively endorse as our own.” 6 I believe Gendler’s desiremodel is substantially correct, though it can be spelled out differently. Amove that Weatherson makes is helpful in this regard. He urges that thefollowing principle, Virtue, is what underlies imaginative resistance:

Virtue: If p is the kind of claim that, if true, must be true in virtue ofsome lower-level facts, and if the story is about those lower-levelfacts, then it must be true in the story that there is some true prop-osition r which is about those lower-level facts such that p is true invirtue of r.7

On Weatherson’s view, Virtue is a “strong default principle of fictionalinterpretation.” 8 I think Weatherson has a good basic approach in that, onhis view, it is understood that for a claim to be true in a story there mustbe something in the story making it true. My gloss on Weatherson’s viewis that there is in the story an implicit commitment to underlying reasonsthat are the justifiers of the claim in question. When this commitment isflouted, as it can be in some fictions, the result is like a Monty Python skit:

6 Ibid.7 Weatherson, “Morality, Fiction, and Possibility,” 18.8 Ibid.

304 JULIA DRIVER

some kind of absurdity is communicated. We do not resist this; indeed,we may go along with it quite happily, precisely because we are pickingup on the absurdity that is the result of flouting the commitment.

When someone judges something to be wrong, they view it to be wrongfor certain reasons. People may differ on whether the reasons need tobe internal or external, but, whatever the details, there are reasons. Andthose reasons can be essentially very different, and yet we still believethat when John says X is wrong (because it displays a lack of respect) andMary says X is wrong (because it causes unnecessary suffering), they areagreeing that X is wrong. In any case, there are reasons that agents areresponding to when they make moral judgments. These reasons may notbe foremost in their minds, and under normal circumstances such reasonswill not be; but acceptance of them still underlies moral judgment.Weatherson alludes to this when he claims: “We are not imagining [thestory] Death if we imagine that Jack and Jill had just stopped arguing witheach other and were about to shoot everyone in sight when Craig shotthem in self-defense.” 9 To imagine this is to imagine a different story, onethat does provide some justification for Craig’s action.10

Weatherson is trying to motivate the assumption that stories typicallyshould be read with a “that’s it” clause in mind. In other words, it isreasonable to assume when reading a piece of fiction that all relevantfacts —particularly any facts that deviate from what the audience wouldbe used to or familiar with —have been provided in the story. I believe thispoints to the fact that at least one factor underlying imaginative resistancein moral cases is that, when we believe something is right or wrong, wetypically believe this is the case for some reason —at least in concretesituations. Further (and I take it this is Weatherson’s point), when we reada work of fiction, we understand, as part of the background, that theworld described by the author is complete for the purposes of understandingthe story. What counts as “understanding” is very loose. When I watch anold episode of Star Trek, I cannot understand Scotty’s explanation of whythe engines have failed. But the author of the script is simply trying toconvey that there is some explanation or other. The same point can bemade with respect to justification. Presumably we do not imagine thestory Death if we imagine that there is some justification for Craig’s action,but we don’t know what the justification is. I cannot speak for everyonehere, but I would have trouble going along with this. It may be that(again) because of the peculiar authority of moral reasons we believe it

9 Ibid., 20.10 Connie Rosati and Scott MacDonald have both suggested to me that what underlies

imaginative resistance in stories like Death might simply be the fact that the story as a wholejust doesn’t seem very convincing. It is not a persuasively told story; thus, we find it diffi-cult to go along with Craig’s claim. This would change the nature of the problem. It wouldstill be puzzling that we go along with poorly told stories that rest on unusual descriptiveclaims.


incumbent upon the author to provide the justification, just as we wouldcall on an agent to provide it. So we don’t go along with outrageousnormative claims even though we let the outrageous descriptive onesslide by (and even enjoy them if they are absurd enough).

Imaginative resistance can be overcome when a work of fiction pro-vides reasons for a change in perspective that are good reasons. Some-times a work of fiction will do this by providing information, or makinginformation salient; but other times it will do this by working on oursympathies, by providing us with alternative points of view. In this way,fiction can be a source of actual moral education —as opposed to beingsimply a reaffirmation of already-held moral beliefs. (Of course, it can bethat, too.) Consider the following review of Harriet Beecher Stowe’s 1852novel Uncle Tom’s Cabin:

We conceive, then, that in writing “Uncle Tom’s Cabin” Mrs. HarrietBeecher Stowe has done more to diffuse real knowledge of the factsand workings of American Slavery, and to arouse the sluggish nationto shake off the curse, and abate the wrong, than has been accom-plished by all the orations, and anniversaries, and arguments, anddocuments, which the last ten years have been the witness of. . . . Aslave-holder might read it without anger, but not easily without asecret abhorrence of the system which he himself upholds. It bringsout, quietly and collaterally, those incidental features of servitudewhich are usually little thought of, but which are the overflow of itscup of abominations. . . . [The novel’s] appeal to our sympathies isgenuine. It artlessly pictures facts, and the facts make us feel. Wehave never read a story of more power. We doubt if anybody has. Thehuman being who can read it through dry eyes, is commended toBarnum.11

This reviewer was merely speculating about the likely effects of UncleTom’s Cabin, but another admirer of the novel wrote a letter to the editorof The National Era (which published the novel in serial form in 1851–52)that included the following claims:

The story of Mrs. Stowe, “Uncle Tom’s Cabin,” is read with interestby persons heretofore violently opposed to everything of an Anti-Slavery nature, and is more or less enlisting their sympathies andremoving their prejudices, more especially among the young. . . . Iknow that there are thousands who would die by inches, who would

11 From The National Era, published in Washington D.C. by Gamaliel Bailey, April 15, 1852,quoting from the Congregationalist. Electronic edition published by Stephen Railton, Institutefor Advanced Technology in the Humanities, Electronic Text Center, Charlottesville, Vir-ginia, 2006; http://www.iath.virginia.edu/utc/reviews/rere01czt.html.

306 JULIA DRIVER

give their lives rather than their support to the infamous law for therecapture of fugitive slaves. They never yet have spoken; their will isnot yet regarded or known; but they will have a will, and a resolutionto back it, only wanting an emergency to make itself known and felt.God grant that that emergency may never come, so that we may beleft in our obscurity, with a consciousness that no duty impels us toemerge from it.12

It seems that this work of fiction galvanized previously hostile or indif-ferent people. Those who started out as indifferent or even hostile to theabolitionist movement were won over once they were presented with astory that made vivid for them all the wrong-making features of slavery.

What sense of necessity is at work here, then? I believe that the desire-based account of imaginative resistance —which, I argue, points to somekind of necessity attached to normative commitments —can give a goodaccount of that necessity by taking a leaf from the “ought” implies “can”literature. There are many different senses of “cannot.” There is the “can-not” one uses to indicate logical impossibility, as in “two plus two cannotequal five.” There is the “cannot” of conceptual impossibility, as in “onecannot imagine a round square.” There is the “cannot” of physical impos-sibility, as in “one cannot fly to the moon.” There is the “cannot” ofmetaphysical necessity, as in “one cannot have had different parents.” Forour purposes, though, the important senses of “cannot” are the two dis-tinct senses associated with psychological possibility. Consider a casefrom the “ought” implies “can” literature: Mary does not know how toplay chess. One might hold, then, that she cannot have an obligation toplay chess, since she cannot play chess given that she doesn’t know therules.13 It is understood that this is psychologically not possible for her.(She can, of course, move the pieces around the chessboard; she is notphysically impaired.) That is, understood in one intuitive way, playingchess certainly seems psychologically impossible for her, since she justdoesn’t know the rules. However, in another sense, it is not psychologi-cally impossible for her, since it is not something incompatible with thelaws of psychology as such. In the second sense, Mary can play chess,since she has all of the cognitive architecture necessary; but, in the firstsense, she cannot, since she doesn’t know how to do it.

Now for necessity. If Mary cannot play chess, then there is a corre-sponding sense of necessity attached to her inability. For the sentimen-talist, it is an interesting question as to which sense is operative with

12 Letter of S. E. M. from Arispe, Bureau County, Illinois, January 29, 1852, to the editorof The National Era, published online by Stephen Railton, Institute for Advanced Technologyin the Humanities, Electronic Text Center, Charlottesville, Virginia, 2006; http://www.iath.virginia.edu/utc/notices/noar01tt.html.

13 Peter Vranas, “I Ought, Therefore I Can,” Philosophical Studies (forthcoming).


respect to our moral commitments (or whether, possibly, both senses areoperative).

We could reasonably hold that both are operative, and this is the viewthat I favor. Thus, there are some things we are committed to given ourpsychological makeup as normal human beings, and then there are somethings we are committed to through accidents of circumstance and upbring-ing. Contrary claims that seem outrageous in the first sense will be muchmore difficult, and perhaps even utterly impossible, to overcome. Butcontrary claims that seem outrageous in the second sense should be mucheasier to overcome. This picture is neutral on the issue of the truth of theclaims in question. Moreover, we will need to tell a separate story aboutthe appropriateness of some moral commitments over others (that is, theissue of commitment is distinct from the issue of truth).

A moral sentimentalist notes that what underlies our commitment tomoral claims is some feeling we have, such as sympathy with othersentient beings. When it comes to the practice of moral judgment, aHumean sentimentalist would hold that we cannot overturn our basic,reflectively endorsed moral commitments at any given time. They arepsychologically necessary for us. One cannot overturn them any morethan one can make oneself into a true practicing skeptic. But here I wouldlike to add a slight twist, borrowing from the work of Frank Jackson. Thetruth conditions of a claim are distinct from its acceptance conditions andfrom its assertability conditions. Some claims may be true, but not assert-able or acceptable. Examples that Jackson discusses are non-robust con-ditionals.14 Appeal to Jackson’s analysis can shed more light on imaginativeresistance and help sentimentalism handle a common sort of criticism.

III. Truth and Robustness

Robustness is a property that some material conditionals have. Gener-ally, when one utters an indicative conditional, one is signaling confi-dence in the antecedent of the conditional (unlike a counterfactual). Butthis is not always the case. It is sometimes the case that if the antecedentwere found to be true, the conditional would be rejected. In those cases,the conditional is not robust, though this is compatible with its truth.

Consider, as an illustration, the following conditional:

(1) If phosphorus does not exist, then neither does thought itself.

Given my belief that phosphorus is a key ingredient to the workings ofthe brain, I may utter (1) in full confidence. However, suppose some-one were to show me that the element phosphorus does not in factexist —that, like phlogiston, it has proved illusory. If one attaches to the

14 See Frank Jackson, Conditionals (New York: Oxford University Press, 1991).

308 JULIA DRIVER

antecedent the probability of 1, would one be willing to stand by theconditional? No. Therefore, it is not a robust conditional. Note, however,that it could still be true. My unwillingness to go along with the condi-tional, my unwillingness to pull the trigger on it, is not the same as itsbeing false.

Frank Jackson’s analysis of what is going on in the case of non-robustconditionals is the following: When an agent asserts the standard materialconditional, she is signaling her confidence in the conditional; she isconventionally implying that it is robust with respect to the antecedent.So the truth conditions of the conditional are different from the conven-tional meaning of the conditional. Thus, a conditional like (1) can be true,but not robust —that is, it may not survive the assumption of the truth ofthe antecedent. The failure of survival will involve imaginative resistancewith respect to the implication.

We can think of a piece of fiction as providing us with a set of claimsthat work as conditionals. Consider:

(2) If killing people merely because they are blocking traffic is per-missible (in the story Death), then what Craig did was perfectlyfine (in the story).

Note that in the story (if we go with Weatherson’s “that’s it” assumption),this claim is presented as part of the story. The antecedent is presented astrue in the story. For the sake of reading the story, we are supposed toassume that it is true. As interpreters, however, we are quite unwilling togo along with the conditional. The conditional is not robust. But it maystill be true (in the story). Gendler and Weatherson’s type of accountleaves it open to regard such conditionals as true in the story, even thoughinterpreters are unwilling to accept such conditionals. I am not sure thisdoes justice to the phenomenology of the puzzle of imaginative resis-tance. This is because it would run counter to the ordinary view that thetruth of moral claims is constant as long as the underlying circumstancesare constant.15

Alternatively, consider the following conditional that appeals to thenorms of storytelling and interpretation (and brings up factors external tothe story itself ):

(2a) If the story has been properly told and read (and the “that’s it”condition is met, among other conditions), then what Craig didwas perfectly fine.

Assign a probability of 1 to the antecedent, and we would not go alongwith this conditional either. Therefore, this conditional is likewise not

15 Someone might hold that conditional (2) is vacuously robust because what is reallydoing the work is the rejection, plain and simple, of the antecedent. Consider, then, (2a).


robust, though it could be true. Standards for “properly told” and “prop-erly read” remain open, but would include the “that’s it” condition under-stood in reading a work of fiction, “coherence” for the story, and so on.

However, robustness only applies to material or indicative condi-tionals, not to subjunctive conditionals. With subjunctive conditionals,one is standardly signaling a belief that the antecedent does not obtain.The issue this presents for fiction is actually rather complicated. Onemight well argue that some fictions are presented more in the spirit ofsubjunctive conditionals, or counterfactuals.

In the fiction case, then, conditional (2) would need to be read in thefollowing spirit: “Of course it isn’t true (in the story), but if it were the case(in the story) that killing people because they are blocking traffic is per-missible, then what Craig did was perfectly fine.”

To the extent that a person experiences imaginative resistance to a story,I believe that he or she is reading it as expressing an indicative condi-tional, where the speaker signals acceptance of the antecedent. This isbecause in stories like Weatherson’s Death the author does not appear tobe rejecting the claim —in the story —that “killing someone simply forblocking traffic is morally permissible.”

Now consider a claim that a sentimentalist (who believes somethingalong the lines that it is the contents of our desire set, plus second-orderattitudes toward features of the desire set, that determine the content ofour moral commitments) would hold:

(3) If it is the case that it is normal for humans to feel approval at theprospect of killing and eating another human, then it is permis-sible to kill and eat another human.

This is rather crude, but what is important to my account is that theantecedent be something the utterer believes unlikely to be the case. Ofcourse, we are horrified and repulsed at the thought of killing and eatingother humans. For any of us to utter (3), we would have to utter it as acounterfactual. Suppose, however, that we somehow discovered the truthof the antecedent, or a good reason to accept it as true; then we could notutter it as a counterfactual. But what would that prove? Here we get backinto imaginative resistance. Suppose the probability of the antecedentwere 1. Suppose, for example, that I was able to show you poll after pollthat demonstrated that human beings, as a whole, do feel it perfectlypermissible to kill and eat other humans —that, in fact, their sympatheticengagement with others is very restricted in scope. We who talk aboutthese issues just happen to be in the small percentage of people who don’tfeel right about it. If we accept that the antecedent of (3) is true, would we,even the sentimentalists, continue to accept the conditional and act on it?No. It is not robust. However, and this is an important methodological

310 JULIA DRIVER

point, that does not mean that (3) is false. Conditional (3) could still betrue. Just because a conditional is not robust does not mean that it is false.

Thus, sentimentalists who bite the bullet have this to fall back on. Thefeeling that there is something very wrong about conditional (3) is explainedby its lack of robustness. But this does not keep it from being true. Andthis is the payoff of robustness in terms of practical deliberation andaction.16 We certainly would not act on (3).

The sentimentalist can argue that, given the way our brains work, wecannot practically question the status of the moral commitments we hold.The commitments impose demands or requirements on us that we, intu-itively, feel to be externally imposed. But for the Humean sentimentalist,at least, there is a sense in which these requirements are externally imposedand a sense in which they are not. When a person experiences a certainsituation as imposing a demand or requirement, what is responsible forthat experience will be his or her desire set. It is because I am sympa-thetically engaged with others that I feel the requirement to help when Iperceive the suffering of others. But there is nothing in their sufferingitself that imposes a requirement. A rational creature with a relevantlydifferent desire set will not experience the suffering of others as imposinga demand.

Of course, this does not mean that the sentimentalist has no standard toappeal to apart from the contents of an individual’s desire set. Humannature (for example, the fact that sympathy seems to be a universal facetof that nature, at least for normal human beings) provides some standardof correctness. Of course, this contingency on human nature strikes manyas deeply unsatisfying, to the point where people such as Richard Joycehold that it is not what we mean by moral demands.17 Moral demands,rather, are deeply categorical, holding not simply due to individual desires,or to desires typical of the normal human. It seems to me, however, thatan error theory can be given for this sense of necessity just as easily as one

16 There are other possible payoffs to be explored. There are other appeals to psycholog-ical necessity made in the literature, and they generally occur in contexts in which practicalmatters are at issue. Do I leave by the window or by the door? Should I give all I own toOxfam, or should I save some resources for my own family? If we are in a certain frame ofmind, we can experience a kind of practical incoherence where some combinations of beliefsand commitments are in conflict. Extreme cases would involve issues of whether to sacrificeone family member to save two strangers. But consider a more normal case: Claudia mustdecide between paying for her daughter’s expensive orthodontia, or giving the money toOxfam. Sacrificing her daughter’s well-being even for the much greater well-being of othersis something she could do, in a sense, since she is quite physically capable of writing thecheck and putting it in the mail, but almost all of her emotional commitments militateagainst it. Thus, there is also a sense in which she cannot bring herself to do it. She knowsthat if she were to do it she would feel like a miserable failure of a mother. She cannot do itwithout feeling she has wronged her daughter. It is against the backdrop of these deepemotional commitments (and the norms they reflect) that it is not a real option for her. And,as long as we believe these commitments are good ones, there will be little pressure forchange. These commitments can be rationally endorsed.

17 Richard Joyce, The Myth of Morality (New York: Cambridge University Press, 2002).


can be given for the view that some moral claims are true. And it is herethat the sentimentalist will appeal to psychological necessity, for whichimaginative resistance provides some evidence. The sense of “cannot”here is not the sense of logical impossibility. The necessity is not nearlythat strong. Certainly Gendler is right here. Instead, the fact that I cannotentertain a certain possibility (e.g., the possibility that killing simply toavoid a traffic jam is permissible) is a matter of its running against deep-seated desires that ground my moral commitments. I cannot do it in thesame sense in which I cannot accept, in reality, someone’s misuse of moralreasons in providing a pseudo-justification of his actions. Again, the issueis not whether the contents of the felt commitments indicate anythingtrue or false about the world. It is a matter of having no choice but toregard them as true or false for practical purposes. But whatever senti-mentalists think about the actual truth or falsity of the moral beliefs thatform a part of our commitments, they will tend to agree that the senti-ments upon which these commitments are based are subject to revisionand correction for the individual; and there is even a possibility that thesesentiments could be subject to revision for the species.18 In the case ofsentimentalists, my view is that they should opt for viewing the feeling ofnecessity that is associated with moral commitments (i.e., the phenom-enology of the moral experience) as being something about us, not some-thing about the external world. This view is supported by the account ofimaginative resistance as an unwillingness, or a weak incapacity, to acceptmorally outrageous claims presented in fiction. As I noted earlier, there isa sense of psychological impossibility that corresponds to weak incapac-ity (this is the sense in which Mary cannot play chess, because she doesn’tknow how). Commitments may be psychologically necessary for us inthat we are not simply unwilling to entertain their negations, but, rather,we do not now have the desires needed to do so. This may be due to a lackof relevant information. In any case, there will be cases where we aresimply unwilling, and cases where a psychological deficit of some sort isresponsible. If this is persuasive, then we have an account of the (purelypsychological) necessity associated with moral commitments that givesus more information, really, about our desire set than about supposeddeontic impositions from our environment.

IV. Conclusion

There are some propositions that we believe, that we are committed tofor practical purposes, independent of any consideration as to their actualtruth or falsity. Such claims may be psychologically necessary for us, even

18 This possibility raises interesting problems for the sentimentalist regarding how todemarcate social groups. There may be concerns here similar to those that crop up in thephilosophy of biology about demarcating species.

312 JULIA DRIVER

though these claims express neither logically nor metaphysically neces-sary truths. I have claimed that imaginative resistance points to the psy-chological necessity of claims we take to be morally true, and thatimaginative resistance itself can be analyzed in terms of the non-robustnessof certain conditional claims (an analysis which, again, is neutral on theissue of the truth of those claims). Moral sentimentalists, who appeal topsychological necessity in order to keep a sense of necessity attached tomoral claims, can also appeal to the non-robustness of certain condi-tionals to counter some arguments that their views imply, in principle,results that are morally outrageous.

Philosophy, Dartmouth College


OBJECTIVISM AND RELATIONAL GOOD*

By Connie S. Rosati

I. Introduction

In his critique of egoism as a doctrine of ends, G. E. Moore famouslychallenges the idea that something can be “good for” someone.1 Mooreasks, “In what sense can a thing be good for me? It is obvious, if we reflect,that the only thing which can belong to me, which can be mine, is some-thing which is good, and not the fact that it is good. When therefore, I talkof anything I get as ‘my own good,’ I must mean either that the thing I getis good, or that my possessing it is good.” 2 As Moore’s remarks indicate,the notion of “my own good” strikes him as strangely proprietary; and, ashe goes on to argue, its proprietary nature precludes it from being nor-mative. In this respect, Moore appears to have found the expression ‘goodfor me’ akin to the peculiar expression ‘true for me.’ 3 Just as the latterexpression suggests a kind of truth that the individual herself could haveexclusive reason to believe, so the former expression suggests a kind ofvalue that the individual herself could have exclusive reason to promote,and that, Moore evidently thought, would be no kind of value at all. As

* I want to thank the other contributors to this volume for constructive discussion of anearlier version of this essay. I also want to thank Richard Kraut and David Sobel for pro-viding me with extremely challenging comments, Ellen Frankel Paul for helpful editorialguidance, and Larry Alexander, Steve Smith, and Michael Jubien for helpful reactions toearlier drafts. The penultimate version of this essay was presented at the 2007 meetings ofthe North Carolina Philosophical Society, as well as to the philosophy department at NorthCarolina State University; many thanks to members of both audiences for instructive com-ments. Finally, I want to thank Don Regan, whose persistence over the years in pressing theMoorean line against “good for” has prompted me to think long and hard about the natureof personal good. Work on this essay was completed during the 2006–2007 academic year,while I was a recipient of the John E. Sawyer Fellowship at the National Humanities Center,Research Triangle Park, North Carolina. I am deeply indebted to the Center for its supportand for the wonderfully collegial environment the Center and its staff have created forscholars working in the humanities.

1 G. E. Moore, Principia Ethica (1903), revised edition, ed. Thomas Baldwin (Cambridge:Cambridge University Press, 1993), sec. 59.

2 Ibid. (Moore’s emphasis).3 See Thomas Hurka, “‘Good’ and ‘Good For’,” Mind 96, no. 381 (1987): 71–73. Hurka,

who agrees with Moore in rejecting talk about the good for, makes the analogy to “true for”explicit. In canvassing the sundry meanings of ‘good for,’ none of which he finds unconfused,Hurka considers uses which treat what is good for a person as what she believes good,either in relation to herself or impersonally. He remarks, “If someone falsely believes that pwe do not say that p is ‘true for her.’ That would go too far toward endorsing her belief, andwe should have the same worry here. If we mean only that someone believes a thing good,that is all we should say, leaving issues of truth or acceptability wide open” (73). I amsuggesting that a related, if not identical, parallel underlies Moore’s thinking.


Moore expresses it, “the only reason I can have for aiming at ‘my owngood’ is that it is good absolutely that what I so call should belong tome. . . . But if it is good absolutely that I should have it, then everyone elsehas as much reason for aiming at my having it, as I have myself.” 4 Theexpression ‘good for me,’ insofar as it denotes something normative, mustsimply be a convoluted way of saying that my possessing a thing is goodtout court.

Moore’s challenge to the notion of “my own good” has implications notjust for egoism, of course, but for the very idea of individual welfare, atleast insofar as the idea of welfare or good for a person signifies somethingmore than that the thing a person gets is good or that the state of herpossessing it is good.5 Insofar as it does have implications for the notionof individual welfare, Moore’s challenge presents a problem not only forwelfare consequentialism, the dominant contemporary rival to his own“ideal” utilitarianism, but for any theory of ethics that assigns us a dutyto promote individual welfare.6 The challenge might fairly be understoodas partly normative, at least in its implications, for ontology circumscribesobligation, but it is, at its core, metaphysical. If Moore is right, the notionof a person’s “own good” or what is “good for her” reflects a fundamentalerror about the sorts of value properties that exist, and, as a consequence,no plausible objectivism about ethics could be welfarist, at least not in thewelfarist’s sense.

4 Moore, Principia Ethica, sec. 59. Jan Narveson has asked why we should bother toconsider seriously Moore’s challenge when the claim he makes —that everyone has as muchreason for aiming at my good as I do —is, as Narveson sees it, so implausible. If X satisfiesone of my desires, Narveson claims, it doesn’t follow that anyone else has a reason to helpme secure X. Because Narveson’s puzzlement might be shared, perhaps it would be helpfulto highlight at least two points on which he and I evidently part company; our differenceswith respect to these points help to explain our differing reactions to Moore’s challenge.First, whereas Narveson is inclined to accept a desire theory of welfare, I believe we havepretty conclusive reasons to reject any such theory. I have argued against certain desiretheories of personal good elsewhere. See Connie S. Rosati, “Persons, Perspectives, andFull-Information Accounts of the Good,” Ethics 105, no. 2 (1995): 296–325; Rosati, “Natural-ism, Normativity, and the Open Question Argument,” Noûs 29, no. 1 (1995): 46–70; andRosati, “Agency and the Open Question Argument,” Ethics 113, no. 3 (2003): 490–527. Forother critical discussion of desire theories, see, e.g., J. David Velleman, “Brandt’s Definitionof ‘Good’,” Philosophical Review 97 (1988): 353–71; David Sobel, “Full Information Accountsof Well-Being,” Ethics 104, no. 4 (1994): 784–810; Don Loeb, “Full-Information Theories ofIndividual Good,” Social Theory and Practice 21, no. 1 (1995): 1–30; and Wayne Sumner,Welfare, Happiness, and Ethics (Oxford: Oxford University Press, 1996), chap. 5. Second,Narveson apparently maintains what I would deny, namely, that good for can (or does?)give rise only to agent-relative reasons. See note 33 below. I am inclined to take Moore’schallenge seriously, because I think we still need an analysis of good for, as well as a betterunderstanding of the sorts of reasons a person’s good might give to her or to anyone else.

5 Michael Smith has argued that Moore’s challenge also has implications for the very ideaof agent-relative value. See Michael Smith, “Neutral and Relative Value after Moore,” Ethics113, no. 3 (2003): 576–98.

6 That would include the broadly Kantian view I favor, a view that makes central theproper valuing of persons and that regards proper valuing of persons as constituted in partby an appropriate regard for their welfare.

OBJECTIVISM AND RELATIONAL GOOD 315

Moore’s argument has recently been revived in a way that makes explicitthe broader challenge it poses to the notion of welfare or good for aperson. Donald Regan, a largely unreconstructed Moorean, elaborates onMoore’s argument, employing it, as he says, to “cast doubt on a propo-sition which many people would regard as the first axiom of moral theory,”namely, that morality is fundamentally concerned with advancing thewelfare or well-being of individuals.7 Regan claims we have no need ofthe notion of well-being or good for a person, which, so far as he can see,captures nothing not already captured by Moore’s idea of good possessedby a person, or as Regan prefers to describe it, good occurring in a person’slife. I will here follow Regan in referring to theorists who treat being goodfor a person, or being good for P, as a distinct normative property, and whomay give normative primacy to individual welfare or personal good, as“good-for theorists,” and I will frequently refer to Mooreans as “goodtheorists,” which is not, I hasten to add, to endorse Regan’s view.8

Regan makes clear that in denying the distinctness and normative pri-macy of good for, he does not mean to deny that we ought to be con-cerned with what goes on in one another’s lives. Moreover, he remarksmore than once that the Moorean concept of intrinsic value may wellencompass everything encompassed by the good-for theorist’s concept ofwelfare: the best account of the good, he assures us, will likely includewhatever is plausibly the good for individuals, even if it includes a goodbit more than this.9 The question that interests him is why we ought to beconcerned with what goes on in one another’s lives. As he expresses it, “I

7 Donald H. Regan, “Why Am I My Brother’s Keeper?” (hereafter “Brother’s Keeper”), inR. Jay Wallace, Philip Pettit, Samuel Scheffler, and Michael Smith, eds., Reason and Value:Themes from the Moral Philosophy of Joseph Raz (Oxford: Clarendon Press, 2004), 202. See alsoHurka, “‘Good’ and ‘Good For’.”

8 I say that these theorists may give normative primacy to good for, because one canbelieve that being good for P is a distinct normative property or concept without believingthat personal good or welfare has primacy from the standpoint of moral theory. My interestin this essay lies just with whether good-for talk concerns a distinct normative property,though what I say on this score no doubt bears on the question of whether good or good for(or neither) is fundamental to morality and moral theory. I explain my preference for theexpression ‘personal good’ in Connie S. Rosati, “Personal Good,” in Terry Horgan and MarkTimmons, eds., Metaethics after Moore (Oxford: Oxford University Press, 2006), 108–9.

9 Regan says, more precisely, that on the “best versions of each theory of value, ‘the good’and ‘the good for’ may have the same extension” (Regan, “Brother’s Keeper,” 209; finalemphasis added). But he sometimes seems to think it likely that they will have the sameextension. In particular, he seems confident that at least his own view of the good, whichincludes pleasure as a part of every state that ought to be promoted, will capture everythingthat good-for theorists deem a good for (210). Indeed, he later considers as a challenge to hisview the suggestion that by treating the good as consisting only in pleasurable experiencesof worthy objects or activities, the Moorean shows himself really to be interested in the goodfor (value for an individual) after all (221). As I have expressed it in the text, what isimportant is that the extension of good, even if broader than that of good for P, is likely toencompass all of the good for. On Regan’s view of the good, the extension of good may infact be no broader than that of good for P. But whether or not this is so will depend on howRegan ultimately spells out the details of his view, as well as on the correct analysis of goodfor P.

316 CONNIE S. ROSATI

do not ask, like Cain, whether I am my brother’s keeper,” but, rather, “why[I am] my brother’s keeper.” 10 The good-for theorist, he maintains, canoffer no coherent answer, not one, at least, that does not rest on a priorMoorean conception of value that renders good for otiose. “So far as I cansee,” Regan tells us, “well-being as a normative concept does not figurein the best account of why we are obligated to care about what happensin others’ lives (or, for that matter, in our own).” 11

One might be tempted to dismiss the broader challenge. Whatever themerits of Moore’s criticism when directed against egoism, one mightargue, it ought not to lead us to abandon common sense.12 After all, whatcould be more obviously true than that things can be good or bad, notmerely in an absolute sense, but for us? What could be clearer than thatour lives can go better or worse, not just, to borrow Henry Sidgwick’smemorable phrase, from the point of view of the universe, but from ourpoint of view? Yet as clear and obvious as this may be, I feel the pull ofRegan’s rhetoric when he virtually dares the good-for theorist to pony upand offer a genuine analysis of welfare, or at least some story that wouldvindicate being good for P as a normative concept or property distinct frombeing good. A striking feature of the rather extensive literature on welfareis, as Regan rightly emphasizes, how little in the way of analysis good-fortheorists provide. As far as I have been able to discover, no real effort hasbeen expended trying to explain, as Regan puts it, what the “for” is doingin the notion of “good for,” or for that matter, what the “good” is doing.13

In this essay, I want to explore what I will from here on call the “Mooreanchallenge,” focusing exclusively on Regan’s recent and more extensivepresentation and development of it.14 I am interested in this challenge for

10 Ibid., 202.11 Ibid., 203.12 In the context of Moore’s critique of egoism as a doctrine of ends, his claims about “my

own good” are preliminary to showing that egoism involves a contradiction. Moore con-strues egoism as the view that each of us “ought rationally to hold: My own greatesthappiness is the only good thing there is.” He claims, as we have seen, that the good of athing cannot be private; the only reason a person can have for aiming at “her own good” isthat it is good absolutely that she possess it, but if it is good absolutely, then everyone hasreason to aim at it. So when the egoist holds that a person’s happiness or interest ought tobe her sole end, this can only mean that it is the sole good and the only thing at whicheveryone ought to aim. “What Egoism holds, therefore, is that each man’s happiness is thesole good —that a number of different things are each of them the only good thing thereis —an absolute contradiction” (Moore, Principia Ethica, sec. 59). For criticism of Moore’scritique of egoism, see C. D. Broad, “Certain Features of Moore’s Ethical Doctrines,” in PaulArthur Schilpp, ed., The Philosophy of G. E. Moore, 3d ed. (La Salle, IL: Open Court PublishingCo., 1968), 43–50.

13 Of those who have offered what they claim to be analyses of personal good, onlyStephen Darwall, as far as I have been able to discover, has said anything that would bearon Regan’s question. See Stephen Darwall, Welfare and Rational Care (Princeton, NJ: Prince-ton University Press, 2002), in which he offers an analysis of a person’s good as consistingin what one ought to want out of concern for her or for her sake.

14 For a discussion of Moore’s challenge in the context of providing a fitting-attitude or“buck passing” analysis of personal value, see Toni Rønnow-Rasmussen, “Analysing Per-


several reasons, two of which I want to mention from the start. First,although much has been written about welfare or personal good, we stilllack a fully adequate theory, and the Moorean’s claims provide a usefulentry into the question of how to understand the basic normative idea ofsomething’s being good for a person.15 Second, if objectivism or realismabout ethics is true, then our most immediate access to ethical facts, Isuspect, is likely to be our access to facts about individual welfare. Askany parent raising a child or any person engaged in caring for an infirmparent or an ailing partner what he or she is doing —anyone who doesthese things without ulterior motive, that is —and however skeptical theymay otherwise be about morality, they will tell you that they are lookingout for their child’s or parent’s or partner’s interests. Ask any personplanning a life for herself why she seeks what she does —any personwhose choices rest not on avowed self-sacrifice, that is —and she willlikely report not merely that she finds those undertakings or pursuitsinteresting, challenging, important, or valuable but that she believes theysuit her and that engagement with them will enable her to lead a happylife.16 If normative facts exist, I conjecture, they get their first grip on usin relation to those we love, including ourselves.17 If these conjectures arecorrect, then making sense of good for will be important to vindicating aplausible objectivism.

sonal Value,” Journal of Ethics 11, no. 1 (2007). As I understand it, the notion of personal valuethat Rønnow-Rasmussen has in mind is not identical to the notion of personal good orwelfare.

15 For preliminary work in this direction, see Rosati, “Personal Good.”16 Regan, more than once, tries to enlist in support of his challenge to the “usefulness” of

good for Joseph Raz’s claim that we do not pursue our goals because doing so will furtheror enhance our welfare. Regan, “Brother’s Keeper,” 203–4, 219. See Joseph Raz, The Moralityof Freedom (Oxford: Oxford University Press, 1986), 316–17. I think the claim is false. I wouldgo this far with Raz: we often undertake the goals we do because we believe them to bevaluable, and when asked why we undertake them, our explanation often appeals to theirvalue rather than to the benefit to us. But I believe these facts are compatible with what I alsothink is clearly true, namely, that we pursue the goals we do because we believe that theywill (or may, where we are uncertain) enhance our welfare. The ordinary person tends notto talk in terms of his or her welfare or well-being, but that is because the terms ‘welfare’ and‘well-being’ are mainly philosophers’ terms. Suppose we ask any ordinary person not “Whydo you pursue X?” but, rather, “Of the many things you believe to be valuable and worthpursuing, why do you pursue X rather than Y?” I conjecture that a typical response wouldbe something along the following lines: “I enjoy X more” or “I believed that, in the long run,X would make me happier.” Of course, a person might also tell us that X would give her alife she found more meaningful, but we often think of the best lives for us as lives that wewill find meaningful.

17 Here I express another disagreement with Regan, who suggests, in acknowledging thatthere is something to the good-for theorist’s position, that “We come to the idea of valuethrough our experiences of wanting and desiring and feeling needs, and so it may seem thatwhen we ascribe value to something, we are necessarily ascribing value ‘for’ ourselves.”Desire, he indicates, seems to him to be “implicated in the aetiology, and to some extent thecontinuing plausibility, of the idea that value is essentially ‘value for’” (Regan, “Brother’sKeeper,” 209). Although our desires and felt needs may have something to do with theaetiology of “good for,” I think a more plausible account would appeal to our experience ofloving people (including ourselves) and seeing them, apparently, suffer or thrive.


Before turning to Regan’s arguments, I want to note that some philos-ophers have argued for a position that is the direct inverse of Regan’s,namely, that nothing is good simply, things can only be good for or goodin a way.18 A few have evidently claimed, even more stringently, thatnothing can be good unless it is good for someone.19 The welfare theoristneed not go so far. On the contrary, she can allow that even if good forturns out to have normative priority from the moral point of view, somestates are simply good and so ought, other things equal, to be promoted.

My aim in this essay is certainly not to argue either that nothing can begood unless it is good for or that good for is less epistemically andmetaphysically problematic than Moorean good.20 Instead, my aim willbe, in the first instance, to explicate the Moorean challenge, the better toclear away certain confusions. Moore, and Regan following him, and nodoubt other neo-Mooreans besides, maintain that good is unanalyzable.

18 See, e.g., Judith Jarvis Thomson, Goodness and Advice (Princeton, NJ: Princeton Univer-sity Press, 2001).

19 Larry Temkin argues persuasively against this view in “Harmful Goods, HarmlessBads,” in R. G. Frey and Christopher W. Morris, eds., Value, Welfare, and Morality (Cam-bridge: Cambridge University Press, 1993). Temkin cites for this view Bernard Williams,“The Point of View of the Universe: Sidgwick and the Ambitions of Ethics,” CambridgeReview (1982), reprinted in Williams, Making Sense of Humanity and Other Philosophical Essays,1982–1993 (Cambridge: Cambridge University Press, 1995); and Philippa Foot, “Utilitarian-ism and the Virtues,” Mind 94, no. 374 (1985): 196–211.The view Temkin criticizes should notbe confused with what Christian Coons has called the “dependence thesis,” which is theview that the value of states of affairs depends on the value of individuals. See ChristianCoons, The Value of States and the Value of Individuals (doctoral dissertation, University ofCalifornia, Davis, 2006). On Coons’s view, states can be good —that is, they can meritpromotion —only for the sake of valuable individuals, and states can be promoted for thesake of an individual either out of respect or concern for her. Coons here exploits Darwall’ssuggestion that concern and respect are the basic normative attitudes we bear towardpersons in order to defend a broader thesis about the structure of value. See Darwall, Welfareand Rational Care.

20 Regan goes to some efforts to explain why he thinks that the concept of welfare facesthe same epistemological and metaphysical difficulties as Moorean good, perhaps becausesome welfare theorists have suggested otherwise. But whether good for faces the samedifficulties as Moorean good surely depends on what we may yet learn about what it is forsomething to be good for someone. So I do not think that the matter can be settled at thisjuncture. In any event, no part of my efforts to understand good for depend on the claim thatgood for is less problematic than Moorean good. Regan’s claims assume that the concept ofwelfare is “irreducibly normative.” He seems to think that means that the concept does notadmit of analysis, and, at least, that it could not be a natural property. He invokes anargument of Stephen Darwall’s which he takes to show that the concept of welfare isirreducibly normative. See Darwall, Welfare and Rational Care, 11. But Darwall thinks that theconcept does admit of analysis, and he thinks he has provided the correct analysis: on hisview, what is good for P is what one ought to want for P out of concern for her or for hersake. Darwall has indicated to me, in an e-mail exchange, that he does not mean to precludethe possibility that the “ought” which figures in his analysis of good for and renders itnormative could be a natural property, in the fashion of the Cornell realists; but it cannot begiven a definitional reduction and is in that sense irreducibly normative. Regan appears tohave a number of related targets in his essay: the idea that good for is better off epistemo-logically and metaphysically than good, the idea that good for has normative primacy, theidea that good for is fully normative, and the idea that good for is independent of (notparasitic on) Moorean good.


One could take the position that good for P, too, is unanalyzable, and if thatis right, then making out that being good for P is a distinct normativeproperty cannot require that one provide an analysis of it.21 I am notinclined to think that good for P is unanalyzable, and, indeed, I have madepreliminary efforts at an analysis elsewhere.22 What I hope to do here isto provide an account of the form and function of good for P, with the aimof characterizing the distinct normative property that apparently figuresin talk about a person’s good. The account will elucidate what ordinarygood-for talk is about, while making sense of what must arguably becommon ground among good-for theorists, regardless of whether theythink being good for P is analyzable and regardless of the form of analysisthey might favor. I cannot make a complete case for good for, of course,any more than the good theorist can make a complete case for good. Ishall not, in any event, be concerned to defend the “axiom” at whichRegan takes aim. My interest lies in clarifying our theoretical alternativesso that we might better decide what to admit into our moral ontology andbetter assess what may be at stake in whether objectivists treat good orgood for as fundamental.

II. The Moorean Challenge

The Moorean challenge is best understood, I believe, on the model ofJ. L. Mackie’s skeptical critique of objective values.23 Recall that Mackiedid not challenge our ordinary understanding of what our normativesentences mean or our understanding of the truth conditions for thosesentences. On the contrary, he allowed that we intend to talk about objec-tively prescriptive properties and that our utterances would express truthsabout the instantiation of just those properties, if only there were any.24

The good theorist likewise does not challenge our understanding ofwhat our sentences mean (of what we intend by our utterances) when wetalk, as we so commonly do, about what is good for people. At least, thegood theorist is not best understood as challenging our understanding ofwhat we mean. After all, ordinary speakers are presumably competentwith both the concepts good and good for. When we talk about what is

21 Regan himself (“Brother’s Keeper,” 207) seems to allow that the good-for theorist couldtake the position that good-for is unanalyzable, though he thinks this would reveal a pointhe means to press, namely, that good for “has no metaphysical and epistemological advan-tage over Moore’s good.” See note 20.

22 See Rosati, “Personal Good.” Good for is, strictly speaking, a relation, and this is, as Iexplain later, important to understanding the nature of personal good; those items thatstand in the requisite relation to P have the property of being good for P. For ease of discus-sion, I sometimes talk simply in terms of good for when I have in mind the property ratherthan the relation, but the context should make clear when I have the property in mind.

23 J. L. Mackie, Ethics: Inventing Right and Wrong (New York: Penguin, 1977).24 Mackie’s skeptical argument is, as has been pointed out in varying ways, premised on

an implausible idea of what objective normative properties would have to be like and ofwhat we are committed to in talking as if there were such properties.


good for someone, we fully intend, as the good theorist recognizes, to talkabout people’s welfare, and the sentences we utter in so doing are bestunderstood as expressing propositions about people’s welfare, not prop-ositions about the good. The good theorist is also not best understood aschallenging our understanding of the truth conditions for our claimsabout what is good for people.25 When we assert the sentence

(a) “X is good for P”

we express the proposition

(b) X is good for P,

which is presumably true if and only if

(c) X is good for P.

It might seem that because the good theorist believes that the only truthsabout value are truths about the good, we must understand his challengedifferently. The good theorist must instead be claiming that proposition(b) is true if and only if

(d) X is good,

or perhaps that (b) is true if and only if, to adopt Moore’s other phrasing,

(e) P’s possessing X is good.

Given the platitude that

(f ) The proposition that p is true if and only if p,

this understanding of the challenge would have the Moorean rejecting thedisquoted truth condition, and so accusing us of being confused about thetruth conditions for our own claims.26 Alternatively, the good theorist

25 See Gilbert Harman, “Moral Relativism,” in Gilbert Harman and J. J. Thomson, MoralRelativism and Moral Objectivity (Oxford: Blackwell, 1996), 3–5. Harman likens his claimabout moral relativity to Einstein’s claim about the relativity of an object’s mass. He con-tends that Einstein’s Theory of Relativity does not involve the claim that people are mis-taken about the meaning of their judgments about an object’s mass; rather, the truth conditionsfor their claims are not as they suppose, “because the only truths there are in the area arerelative truths” (ibid., 4). Harman, in discussing evaluative relativity (ibid., 14–16), treatsgood for as relativistic, but if he means ‘good for’ in the welfarist sense, then the good-fortheorist would contend that he confuses relativism with relationalism.

26 See Paul Boghossian, “What Is Relativism?” in Patrick Greenough and Michael P.Lynch, eds., Truth and Realism (Oxford: Oxford University Press, 2006). I here follow Boghos-sian, who rejects Harman’s ideas about how to characterize relativism in the context of hisown effort to develop a model for how to understand the discovery that truths in a partic-


must be accusing us of repeatedly misspeaking. We mean to say “X isgood,” which expresses the proposition X is good, and is true if and onlyif X is good; or we mean to say “P’s possessing X is good,” which expressesthe proposition P’s possessing X is good, and is true if and only if P’spossessing X is good. But we continually slip and say “X is good for P.”

Putting to one side any difficulties internal to these understandings ofthe Moorean challenge, what is important for present purposes is thateither understanding would fail to capture the Moorean’s claim to bebaffled as to what we could be talking about when we talk about what isgood for an individual. Were we simply confused about the truth condi-tions for our own claims, the correct truth conditions being ones theMoorean well understands, or were we merely misspeaking, our slipsbeing easily identified and corrected, there would be nothing to baffle theMoorean. We intend to talk about what is good for P, and we take thejudgment that X is good for P to be true if and only if X is good for P. Thisis what the Moorean finds unintelligible.

The good theorist surely means to make a claim about the normativefacts: just as Mackie’s more general challenge seeks to call into doubt theexistence of objectively prescriptive properties, so the Moorean challengeseeks to call into doubt the existence of a distinct normative property ofbeing good for P. Of course, we talk as if there were such a distinct property —ordinary normative discourse is littered with claims about what is goodfor us, quite apart from claims about the goodness or value of states ofaffairs. The good theorist argues, in effect, that we are in error. Obviously,the good theorist cannot endorse a global error theory; rather, he advancesa local error theory with respect to good for.

A. The anti-good-for argument

The good theorist’s “anti-good-for argument” can be presented moreformally as follows:

ular domain are relativistic. Boghossian first rejects a model according to which the discov-ery of relativism in a domain is the discovery that the propositions expressed by sentencesin that domain are unexpectedly relational. This model, he claims, would not account eitherfor classic cases of relativistic discoveries in physics or for relativistic claims about morality.He then rejects Harman’s model, which would treat the discovery of relativism not as adiscovery about the propositions expressed by typical sentences in a domain but as adiscovery about their truth conditions. As Boghossian explains, the model would havespeakers misunderstanding the truth conditions for their claims and would amount toabandoning the platitude expressed in (f ). He finally arrives at a model that treats discov-eries of previously undetected relativism as “correcting our view of what the facts are.”Boghossian’s conclusion regarding how to model relativistic claims treats relativism abouta domain, in this regard at least, as akin to Mackie’s skepticism insofar as it corrects for aview about the facts that was “in error.” Of course, the situation I explore in this essay is justthe inverse of the one that interests Boghossian: the “discovery” the Moorean means to pressis not that truths about value are unexpectedly relativistic but that truths about value,contrary to our good-for talk, are never relational in the way we suppose.


(1) The sentence “X is good for P” expresses the proposition X is goodfor P, which is true if and only if X has the property of being goodfor P.

(2) There is no such property as being good for P.(3) So there are no facts about what is good for P.(4) Therefore, all sentences of the form “X is good for P” are, strictly

speaking, not true.

This much gets us to the good theorist’s negative conclusion with respectto good for. Of course, as we have already seen, the good theorist does notmaintain that we point to nothing real when we mistakenly talk as if therewere good-for facts. We can, accordingly, understand the argument tocontinue as follows:

(5) There is, however, an absolute or monadic property good.(6) Moreover, this monadic property can be instantiated by things

that P possesses, or by P’s possessing X, or by things that occurwithin the life of P.

(7) Therefore, among the truths about value are truths of the formX is good and possessed by P,P’s possessing X is good, andX is good and occurs in the life of P.

(8) The only truths about value and persons that approximate towhat we might mean to express by utterances of the form, X isgood for P, are in fact truths of the form

X is good and possessed by P,P’s possessing X is good, andX is good and occurs in the life of P.

(9) Therefore, if we are to speak truly and precisely about value andpersons, we should make only judgments of the form

X is good and possessed by P,P’s possessing X is good, andX is good and occurs in the life of P.27

Having presented his negative case and offered his alternative formula-tions, the good theorist argues, in short, that either the predicate “is goodfor” denotes nothing at all or it denotes “good.” From here on, I will focuson the formulation “X is good and occurs in the life of P” or good occurringin the life of P. I do so not merely because that is the formulation Regan

27 In “What Is Relativism?” Boghossian treats the relativist as urging us to substitute forour claims about monadic properties those relational truths that are the “closest truths in thevicinity.” I am here suggesting that the Moorean should be understood as urging a reformin the opposite direction.


adopts but because I believe that the best construal of it likely providesthe best construal of the other formulations as well.

Now we will want to be careful in how we understand step 9 of theanti-good-for argument. The good theorist need not be understood asadvocating that we purge ordinary discourse of good-for talk (as if wecould do that), or even (more realistically) that we theorists refrain fromsuch talk. After all, the good theorist’s point concerns the facts, and solong as we keep ourselves straight about those, the good theorist canallow that no harm need come from our continuing to use good-for talkas shorthand for his more accurate but awkward locutions.28

I want to focus on two steps in the argument —steps 2 and 8. I willconsider step 8 in Section III below, but I will begin, here, with step 2.Why, according to Regan, should we doubt that there is any such prop-erty as being good for P? Regan argues that good for is a chimera in theclassical sense, a fiction cobbled together from not merely incongruousbut, in this case, contradictory elements.29

Regan relies on certain critical assumptions in setting out his argument.He assumes that when something is good simpliciter, we are required tocare about it: a reason exists to care about it and to promote it.30 The reasonin play is an external, agent-neutral reason, so the “we” means just whatit says: any agent should care about and, ceteris paribus, promote some-thing insofar as it is good. As Regan acknowledges, some will reject hisexternalism about reasons, insisting that an agent has a reason to careabout and promote something only insofar as she takes an interest in it.31

For purposes of the present discussion, however, we can grant Regan hisassumptions. Even if the good-for theorist believes that an individual’sgood gives him agent-relative reasons, she can allow that it also giveshim —and anyone else —agent-neutral reasons. We needn’t worry hereabout whether such reasons are also, strictly speaking, external reasons.The main point is that the good-for theorist can and should concede thatit would rob personal good of the possibility of any important place inmoral theory if facts about what is good for a person only provided

28 Compare Mark Kalderon’s discussion of “revolutionary” fictionalism in Mark EliKalderon, Moral Fictionalism (Oxford: Clarendon Press, 2005), 136–42. See also Boghossian,“What Is Relativism?” On Boghossian’s construal of relativism, the relativist is engaged ina “reforming project” which would urge us to “abandon the absolutist discourse we cur-rently have in favor of a discourse which accommodates [the relativist’s] conviction that theonly facts in the vicinity of that discourse are certain kinds of relational facts.” Of course, wewould expect the impact of such abandonment to vary. It is one thing to give up talk aboutabsolute motion in favor of motion relative to a frame of reference; it is quite another to giveup talk about absolute moral requirements in favor of moral relativism. In the former case,we still have motion, just better understood; in the latter, we (arguably) no longer havemorality. But then, if the relativist is right, we never had it. Still, there may be pragmaticeffects of absolutist moral talk that we would want to retain, even if we became convincedthat the relativist was right about the facts.

29 Regan, “Brother’s Keeper,” 218.30 Ibid., 211 and note 22.31 Ibid., note 22.


agent-relative reasons.32 Ordinary ethical thought, in any case, arguablydoes not treat personal good as purely prudential and normatively agent-relative; and so the good-for theorist would, in my view, make a mistakein treating it as such, at least insofar as she means to capture the notionof good for that figures so prominently in our everyday lives. From thestandpoint of moral theorizing, if we suppose that ordinary ethical thoughtaffords some insight into morality, moral theory must be responsive to it,whether or not, in the end, moral theory properly gives primacy to goodfor.33

32 I am prepared to say, then, that the fact that something is good for a person givesagent-neutral reasons, but I am not prepared to say, as Regan seems to, that we can berequired to care about something. Care is, after all, a conative state, and it is uncertain towhat extent we have control over our carings. Even if it does not make sense to say that weought to care, we can still reasonably say both that it is appropriate to care and that oneought to behave as would someone who did care. On the latter idea, see Darwall, Welfare andRational Care.

33 These last few sentences are intended to provide what I acknowledge to be an all tooquick reply to the surprising resistance I have encountered to the suggestion that facts aboutwhat is good for a person provide agent-neutral reasons. David Sobel has suggested, forinstance, that it is optional whether the good-for theorist treats good for as of great moralsignificance and as giving rise to agent-neutral reasons. I doubt this, at least if the good-fortheorist is interested in whatever property ‘good for P’ expresses and whatever facts theremay be about what is good for persons. I doubt it for the reasons I have gestured toward inthe text. Ordinary moral thought and talk seem to presuppose that the good of other peoplemakes at least some claim on us, and not merely the good of those with whom we havespecial relationships or to whom we have special obligations. Of course, a theorist canstipulate a notion of good for that treats it as purely prudential. But it is unclear what interestsuch a notion could have or how it could be normative. What might it capture that is left outby the notion of good for that figures in our ordinary talk and in the work of most welfaretheorists? And why should the individual herself be interested in her own purely prudentialgood for unless it captures something otherwise missed? Sobel has also suggested that myposition may rule out egoism. I am not convinced that it does. Egoism, in my view, is notproperly classified as a moral theory; rather, it is a theory of what individuals have mostreason to do that competes with morality. The egoist can allow that any person’s good givesone a reason to act, while maintaining that one’s own good alone gives overriding reasonto act. It seems to me, then, that the egoist could take on board what I have to say about goodfor; any difficulties in doing so are internal to egoism itself. Richard Kraut has correctlytaken my remarks to imply that we each have reason to promote the well-being of everyother person. But we each only need the help of certain others, he has reminded me, and sowhy suppose all moral agents should help? I am inclined to think that this sort of resistanceto the idea that personal good gives rise to agent-neutral reasons can be overcome by furtherreflection on what that might involve. P’s good, for example, can give everyone a reason toact, even if each of us still has overriding reason to do something other than act to promoteP’s good. We have reason, for example, to support some division of labor with respect topromoting other people’s good; it makes sense to attend to those we know well and whomwe are best positioned to aid. Not all of us will be equally well-positioned to aid P. More-over, we have reason to leave it to P, ordinarily, to promote her own good without our aidor interference. After all, P is an autonomous agent, and her good will partly consist in herown active pursuit of a way of life. Except in the case of children or adults who are unableto act on their own behalf, acting so as to promote another’s good involves some risk ofsubverting her autonomy. Third parties tend, in any case, to be less well positioned topromote an individual’s good than the individual herself. For these reasons, and not becausegood for gives rise only to agent-relative reasons, the reason third parties have to promoteanother’s good may be overridden or conditional. The fact that good for gives anyone areason for action is consistent with recognizing differences among individuals in their


I want to emphasize that in granting, for purposes of argument, thatwhen something is good we have agent-neutral reason to promote it, I donot take myself to be granting anything special about good. Saying thatwhen something is good we have reason to promote it is, I would argue,just a way of indicating that good is normative. But so are right and ought.And so, the good-for theorist wants to insist, is good for, as Regan wellrecognizes.

B. The dilemma argument

Which brings us to Regan’s argument. The notion of good for, Reganevidently thinks, combines incompatible ideas: the concept of good, ofsomething normative, something that makes a claim on any individual,and an unexplained “for,” which involves a restriction to the individualthat undercuts that universal claim. As a consequence, Regan maintains,the good-for theorist faces a dilemma.34

On one horn of the dilemma, she can concede that ‘good for’ merelydenotes what the Moorean’s ‘good occurring in the life of’ denotes. Thatwould allow the good-for theorist to retain the claim that when X is “good

obligations to promote someone else’s good. A parent does not have exclusive reason topromote his child’s welfare, but he has special obligations that the rest of us may lack, so theduty of seeing to his child’s welfare ordinarily falls, in the first instance, on him. Moreover,he may have, in addition to the agent-neutral reasons anyone has to promote his child’swelfare, agent-relative reasons. Finally, much of what we do, and ought to do, in responseto the claim that others’ good makes on us takes the form of indirect support. We arrangefor public financing or make charitable contributions; we support institutional arrange-ments that provide aid or otherwise support individuals in their self-development andpursuit of their own good. Kraut has suggested as a counterintuitive implication of myposition that since, e.g., a disease is bad for the trees in the forest, we all have reason toensure that trees do not acquire diseases. I do not believe that ‘good for’ in talk about whatis good or bad for trees, artifacts, and so on expresses the same normative notion involvedin talk about what is good for a person. For one thing, I do not believe that trees and artifactsare the sorts of things that have a welfare. An adequate response to Kraut would requireexplaining what things have a welfare and how nonwelfare good-for talk should be under-stood, but I cannot undertake to address these matters fully here, so I merely register myviews. Mark LeBar has questioned my apparent assumption that only agent-neutral reasonscan be fully normative. I do not mean to make any such assumption. For purposes of thearguments in this essay, I assume only that good for would be without much interest fromthe standpoint of moral theory were it merely to give rise to agent-relative reasons. And Iclaim that whether or not there are fully normative agent-relative reasons, and whether ornot good for is also normative in this agent-relative way, good for is normative in the waythe Moorean claims it is not.

34 Regan himself describes the dilemma argument as an argument against the “useful-ness” of good for, and also as an argument for why good for cannot make a claim on people(other than the individual whose good it is), if it is “genuinely independent of Mooreangood.” See Regan, “Brother’s Keeper,” 213. Challenging the usefulness of a normativeconcept seems a peculiar criticism unless whether a concept plays a useful normative rolebears on whether the concept is of a genuinely normative property. I argue later thatjudgments about good for play a certain normative role and that judgments about good (orgood occurring in the life of ) are not well suited to this role. The claim that good for cannotmake a claim on other people is, I believe, best understood as a challenge to the existenceof being good for P as a distinct normative property.


for” P, a reason exists for anyone to care about and promote X for P. Afterall, promoting X for P would just be one part of promoting the goodsimpliciter, which, according to the Moorean, is precisely what is to bepromoted. But it would amount to acknowledging that being good for P isnot a distinct normative property, or as Regan puts it, “That would amountto abandoning ‘good for’ as an independent concept.” 35

On the other horn of the dilemma, the good-for theorist can deny thatgood for is simply good occurring in the life of, aiming to maintain beinggood for P as an independent concept or distinct normative property. ButRegan contends that if the good-for theorist takes this approach, she mustgive up her claim that welfare or good for is normative, because she mustgive up the claim that the fact that something is good for someone pro-vides reasons for anyone, not merely for the individual whose good is inquestion.36 If X is good for P, P may have a reason to promote or obtainX, but no one else has a reason to promote or obtain X for P.

I believe the confusions alluded to earlier (in Section I) lie precisely inRegan’s reasons for drawing the conclusion that being good for P cannot beboth a concept or property distinct from good and a normative concept orproperty. In fairness to Regan, though, this is also precisely where thegood-for theorist needs to do some serious theoretical work. What Reganthinks is that if being good for P is to be a concept or property independentof good, it must involve a relativization not merely of the sort involved intalking about the good that occurs relative to one life or another, but arelativization of goodness —of the normativity —itself. As Regan expresses it,“The good for Abel must be peculiarly Abel’s —the goodness or the valuemust be peculiarly Abel’s —in a way that the mere occurrence of universalgood in Abel’s life does not necessarily satisfy. But if the good for Abel ispeculiarly Abel’s —if its value is somehow essentially a value for Abel —then why indeed should Cain care? We think there is a deep connectionbetween value and reasons. That suggests precisely that if the good forAbel is a matter only of value for Abel, then it creates reasons for Abel, and

35 Ibid., 211.36 After presenting the dilemma argument, Regan goes on to “consider how the dilemma

manifests itself in connection with” certain contemporary theories of well-being. After dis-cussing the views of Wayne Sumner and James Griffin, he remarks, “[T]he joint lesson of thediscussion of Sumner and Griffin is that ‘good for’ is a chimera. If we take seriously thedependence on the agent’s subjective evaluation, which Sumner plausibly claims is the bestrealization of the ‘for’, we get the very implausible conclusion that a life spent in trivialpursuits can be as valuable as Bach’s life or Darwin’s life if the subject is as satisfied, andthat the life of trivial pursuits can make as strong a claim on others to be promoted, so faras intrinsic prudential value is concerned. This is not plausible; the ‘for’, in Sumner’saccount, undermines the ‘good’. In contrast, emphasizing the perfectionist element, as Grif-fin does, seems to attenuate the ‘for’ to the point where it is a mere empirical accident(‘occurring in the life of’) rather than an aspect of the normative concept. In Griffin’saccount, the ‘for’-ness is totally unrelated to the nature of the value” (ibid., 218). See Sumner,Welfare, Happiness, and Ethics, and James Griffin, Well-Being: Its Meaning, Measurement, andMoral Importance (Oxford: Clarendon Press, 1986).


for no one else.” 37 The Moorean idea of good occurring in the life of doesnot, Regan claims, face a similar difficulty: “If the ‘for’ means ‘occurringin the life of’, this gives us an empirical relativization to the agent, but thenormativity involved is still the universal normativity of ‘good’.” 38

It is understandable why Regan thinks that good for involves a relativ-ization of goodness or normativity. As I have already noted, good-for theo-rists say little that would explain what the ‘for’ or the ‘good’ in ‘good for’means, and in the absence of an alternative explanation, Regan’s thoughtseems natural enough. Moreover, theories of welfare —hedonistic theories,desire theories, and even Wayne Sumner’s authentic happiness theory —have tended to be subjectivist.39 (The objective list theory and StephenDarwall’s recent rational care theory stand out partly because they are excep-tions to the dominant subjectivist trend.)40 Regan rightly questions (forexample, in discussing Sumner’s theory) whether an individual’s (subjec-tivist) good could provide any agent other than the individual herself witha reason. After all, according to Sumner, if an individual is satisfied, say,with a life of self-flagellation, then so long as her satisfaction is informedand autonomous, her welfare lies in self-flagellation.41 Perhaps we shouldconcede that such an individual herself has reason to pursue her heart’sdesire, Regan would allow (though he sometimes appears to doubt thatwelfare claims yield even agent-relative reasons); but surely, he would insist,we have no reason to promote her enjoyment of this (worthless) endeavor.

Contrary to what Regan claims, however, the concept of being good forP does not involve a relativization of normativity or goodness, or so I willsuggest. If we can sensibly wonder whether we have reason to promote

37 Regan, “Brother’s Keeper,” 211 (Regan’s emphasis).38 Ibid., 212.39 Although disagreement exists as to how best to understand John Stuart Mill’s moral

philosophy, he claims to advocate a form of hedonism, which he offers as a theory of thegood rather than as a theory of welfare. See John Stuart Mill, Utilitarianism (Indianapolis, IN:Hackett Publishing Co., 1979). Mill claims that the sole good is happiness, and he describesa happy life as one relatively free of pain and rich in enjoyments both in point of quantityand of quality. Although he offers hedonism as a theory of the good, or what he seems tothink is the same thing, the sole end, Mill’s emphasis on the importance of “experiments inliving,” so that individuals may determine wherein their own happiness lies, at least sug-gests that his real concern was with individual good or welfare. See John Stuart Mill, OnLiberty (Indianapolis, IN: Hackett Publishing Co., 1978). For a recent hedonistic theory, alsooffered as a theory of the good rather than as a theory of welfare, see Fred Feldman,Utilitarianism, Hedonism, and Desert (Cambridge: Cambridge University Press, 1997). Fordiffering kinds of desire theories, according to which what is good for a person is, roughly,what she would desire, or desire for herself to desire, under more or less optimal conditions,see, e.g., Griffin, Well-Being; Peter Railton, “Moral Realism,” Philosophical Review 95, no. 2(1986): 163–207; and Peter Railton, “Facts and Values,” Philosophical Topics 14 (1986): 5–31.Sumner offers helpful critical discussion of hedonistic and desire theories of welfare beforepresenting his own theory of welfare as autonomous or authentic happiness; see Sumner,Welfare, Happiness, and Ethics, chaps. 4 and 5.

40 See Darwall, Welfare and Rational Care. See also notes 13 and 20 above. For the idea ofan objective list theory, see Derek Parfit, Reasons and Persons (Oxford: Oxford UniversityPress, 1984).

41 See Sumner, Welfare, Happiness, and Ethics, chap. 6.


a person’s good as some theory of welfare construes it, then we shoulddoubt the adequacy of that theory, rather than the coherence of some-thing’s being good for a person. Regan claims that unless the ‘for’ in‘good for’ signals a relativization of normativity, good for cannot bedistinct from good, though, of course, it is just this relativization thatwould prevent judgments about what is good for a person from makinga claim on agents other than the individual herself. But this is incorrect.After all, we have a number of other terms —for example, ‘right,’ ‘ought,’and ‘rational’ —each of which expresses universal normativity, while alsoexpressing a property distinct from good.

How, then, should we understand good for? What the Moorean seemsto imagine when he criticizes the good-for theorist’s position is that thelatter’s good-for judgments must concern a relational complex, X is goodfor P, with the following structure: X has the monadic property good, butthis goodness, or perhaps X’s being good, stands in a (curious) for relationto P. If this were so, then we could begin to see why the Moorean wouldfind talk about what is good for someone unintelligible. The line of thoughtseems to be that insofar as X has the property good, X is good for P providesagent-neutral reasons, but insofar as the for relation relativizes the nor-mativity or goodness to P, X is good for P provides only agent-relativereasons. Hence, good for is incoherent, combining as it does these incom-patible elements; and being good for P lacks normativity, because being anormative property just is being a property facts about the instantiationof which provide agent-neutral reasons.42

In my view, this line of thought misconstrues good-for judgments. Therelational complex, X is good for P, does not include the monadic propertygood at all. Instead, it includes the relational property is good for P: it hasX and P as relata and is good for as a dyadic relation.43 We should not be

42 More fully, we might say that on the Moorean picture, being a normative concept justis being a concept propositions about the correct application of which provide or entailagent-neutral reasons; and being a normative property just is being a property facts aboutthe instantiation of which provide or entail agent-neutral reasons. One might certainlychallenge this idea. And that suggests a rather different route to addressing the Moorean,though one that will arguably rest on my diagnosis of his argument. That route would takeon board something like the claims I shall make about how to understand the words ‘good’and ‘for’ in ‘good for,’ while holding that personal good provides only agent-relative rea-sons. As I have made clear, however, I believe that facts about what is good for a person alsoprovide agent-neutral reasons.

43 Alternatively, we might say that the proposition expressed by the sentence “X is goodfor P” does not have the logical form

(X is good) for P

but, rather,

x G p

(using ‘G’ to express the relation is good for). I have found Boghossian’s discussion in “WhatIs Relativism?” helpful in clarifying the Moorean’s mistake, as I see it. The Moorean sees the


misled by the fact that the word ‘good’ occurs as part of the name of therelation expressed by ‘is good for.’ We should not be misled here anymorethan we should be misled by occurrences of the word ‘good’ in judgmentsof instrumental value. Consider the sentence, “A hammer is good fordriving nails into wood.” Surely no one supposes that the propositionexpressed by that sentence concerns a relational complex with this struc-ture: A hammer has the property good —the property Moore deemed thecentral subject matter of ethics —and stands in an instrumental for relationto driving nails.44

As an initial matter, then, good for does not involve a relativization ofgoodness or normativity. Good for P is not goodness relative to P but adistinct relational value.45 Of course, simply to assert, as I have thus far,

good-for theorist as making something like the move Boghossian attributes to the relativist,which is to offer “replacing” propositions for our old propositions about morality that arebuilt up out of the old propositions. As Boghossian puts it, “the replacing propositionconsists in the claim that the old proposition stands in some sort of contentual relation to aset of propositions that constitute a [moral] code.” My suggestion is that the good-fortheorist should be understood not as asserting propositions with the sort of logical form therelativist supposes moral propositions to have —a view which is relativism’s undoing —but,rather, propositions with the logical form Boghossian describes as operative in the physicscases. And propositions about what is good for a person take the latter form for a preciselyparallel reason: namely, they express truths about a relational property the concept of whichdoes not contain the concept good as a proper part.

44 Notice that this would yield just the sort of incompatibility that Regan claims for goodfor. The sentence “The hammer is good” asserts a claim about the intrinsic value of thehammer; the added instrumental relation, “for driving nails,” undercuts the claim of intrin-sic value. Lest instrumental uses of ‘good’ seem an aberration, consider occurrences of‘good’ in judgments of something as an instance of a type. Consider the sentence “TheCorolla is a good car.” No one supposes, to put the point differently, that the propositionexpressed by that sentence is built up out of the proposition The Corolla is good (in Moore’ssense) and a categorization as to the type of object, car. That there are distinctions ingoodness and that these are not distinctions across a single monadic property should beevident, e.g., from Christine M. Korsgaard’s classic discussion in “Two Distinctions inGoodness,” reprinted in Korsgaard, Creating the Kingdom of Ends (Cambridge: CambridgeUniversity Press, 1996), 249–74, originally published in Philosophical Review 92, no. 2 (1983):169–95. The expression ‘good for,’ like ‘good,’ appears in a variety of contexts. As I hope isentirely clear, my interest lies solely with uses of the expression in the context of individualwelfare or personal good, what is sometimes described as “intrinsic” “nonmoral” good fora person. See note 33 above.

45 Railton has explicitly described good for as relational goodness. See Railton, “MoralRealism,” 183: “It should perhaps be emphasized that although I speak of the objectivity ofvalue, the value in question is human value, and exists only because humans do. In thesense of old-fashioned theory of value, this is a relational rather than absolute notion ofgoodness. Although relational, the relevant facts about humans and their world are objec-tive in the same sense that such non-relational entities as stones are: they do not depend fortheir existence or nature merely upon our conception of them.” Treating being good for P asa distinct relational property might seem to invite greater problems down the road. In note25, Regan remarks, “The need for good simpliciter seems even clearer when we consider theproblem of how to balance the (reasons created by the) good for Abel and the good for Seth.Balance we must, and that seems to require a ‘good’ (explicit or implicit) which is not a‘good for’” (“Brother’s Keeper,” 213). If good is not a proper part of good for, we wouldseem to have no common value for weighing and balancing. In fact, I think the good-fortheorist has plenty to say to address the worry, but here I merely flag the issue and leavequestions about commensurability and balancing for another time.


that good for is not as the Moorean thought only gets us so far in address-ing the Moorean challenge. My claim responds to a certain argument forthe charge that good for is a chimera —a cobbling together of incompat-ible elements —by offering a diagnosis of the mistake behind that charge.Still, one can reasonably demand a positive alternative to the view I haveclaimed is mistaken. The good-for theorist needs to say more to make itplausible that being good for P is a fully normative property distinct frombeing good.

III. Relational Good

Recall Regan’s conjecture that on the best theory of the good, the notionof good occurring in the life of an individual will likely pick up everyplausible good for. If the extension of good encompasses the good for, thenit might seem not to matter whether being good for P is a distinct normativeproperty, as the good-for theorist maintains. But of course, as Reganstresses, the question is not whether but why I am my brother’s keeper.Even if the good theorist can pick up all the good fors that may be ofpractical, normative concern to the good-for theorist, he will not therebyhave succeeded in answering Regan’s question. The good-for theoristcould still be right in thinking that talk about what is good for someoneconcerns a distinct normative property.

Now I agree with Regan that once we get to this stage in the debate, itis hard to come up with decisive arguments for one side or the other. Heremarks:

[W]e are approaching the point, if we are not there already, where theextensions of the Moorean’s ‘good’ and the ‘good for’ theorist’s ‘goodfor’ are effectively indistinguishable. In such a situation we will notbe able to decide between the concepts by adducing counterexam-ples to one theory or the other —states or events which the propo-nents of one theory can claim the other theory attaches the wrongvalence to. What we need is conceptual arguments, but it is not easyto find arguments that have any purchase on such a fundamentaldisagreement.46

Regan does offer a conceptual argument for the Moorean view —an unsuc-cessful one, I believe —which I will touch on briefly toward the end of thisessay.

One consideration might seem, at the outset, to favor the good-fortheorists’ position. As a matter of ordinary observation, different items,activities, undertakings, and relationships seem to suit or fit differentindividuals, and they do so in such a way that some (and not others) at

46 Regan, “Brother’s Keeper,” 210.


least appear, intuitively, to benefit those individuals. But the Mooreanpresumably recognizes something close to this thought, for he will surelyacknowledge that certain goods can “occur in” some individuals’ livesbut not in others. The good that consists in the pleasurable experience ofplaying a piano sonata will occur in the life of the person capable ofplaying the piano and of taking pleasure in playing it, but not in the lifeof the person who lacks either the ability to play or to appreciate classicalmusic. Thus, the debate seems even more intractable than Regan sug-gests: not only might good and good for share an extension, but, at a moretheoretical level, the good theorist and good-for theorist can each appar-ently account for key features of our experience.

These considerations suggest a different obstacle to recognizing beinggood for P as a distinct normative property. Normative concepts figure, indiffering and characteristic ways, in the regulation of attitude and action.One reason to doubt the existence of a supposed normative propertywould be if that property (or the concept of it) seemed to have no dis-tinctive normative role to play. But this, the good theorist might suggest,is precisely the case with good for. Even if the concept of being good for Pis not incoherent, why believe there is such a relational property as beinggood for P when all the normative work is apparently already done by good(occurring in the life of )?47 I want to approach this question from twodirections: the first returns us to step 8 of the “anti-good-for argument”(discussed above in Section II.A) and involves thinking about what ispicked up by good for that might be missed by good occurring in the lifeof; the second involves, much more briefly, thinking directly about thenormative role apparently played by good for. As I will try to show,contrary to what Regan claims, the Moorean’s good and good for are not“effectively indistinguishable,” and we can shed light on the distinctnessof good for by attending to its special —and pervasive —normative role.

A. Good for versus good occurring in a life

Regan claims to find deeply puzzling the ‘for’ in ‘good for P,’ and Ihave undertaken to address his perplexity. But I want to register my ownpuzzlement as to what could be meant by such expressions as ‘goodpossessed by P’ and ‘good occurring in the life of P.’ It would be amistake, in my view, to think that good for is problematic in a way thatgood occurring in the life of is not. So the question with which I want tobegin is this: What might be missing from good occurring in the life of P thatis captured by good for P?

47 I think Regan hints at this worry when he raises doubts about the “usefulness” of goodfor, but he never articulates the point of the usefulness challenge as fully as would bedesirable, and I believe expressing the problem simply in terms of “usefulness” obscures thefact that the underlying issue is ontological. See note 34.


Suppose that a new parent approaches you, a seasoned parent, andasks for advice about child-rearing. You begin to advise him about how todiscern and promote what is good for his child. He stops you short andsays, “Now wait a minute, I’m a Moorean. Believe me, I love my childevery bit as much as the next parent, but I’m not interested in what is‘good for her.’ In fact, I don’t understand this ‘good for’ talk. I’m inter-ested in promoting good occurring in her life.” You would no doubt bepuzzled, and calling attention to this fact is meant to stress a point madeearlier in passing, that common sense is on the good-for theorist’s side. Ofcourse, common sense might be confused, but it might instead be trackinga species of value omitted from the Moorean picture, and it is this speciesof value that the good-for theorist presumably seeks to investigate.

So what advice might you offer if you wanted to answer the Mooreanparent in terms he would find intelligible? To figure this out, we first needto understand the Moorean notion of good occurring in the life of, and forthis, we need a working account of the good. Moore himself famouslythought that many things are good, although the greatest goods are cer-tain “organic unities.” As Moore explains,

there is . . . a vast number of different things, each of which hasintrinsic value; there are also very many which are positively bad;and there is a still larger class of things which appear to be indiffer-ent. But a thing belonging to any of these three classes may occur aspart of a whole, which includes among its other parts other thingsbelonging both to the same and to the other two classes; and thesewholes, as such, may also have intrinsic value. The paradox, to whichit is necessary to call attention, is that the value of such a whole bearsno regular proportion to the sum of the values of its parts.48

Moore contrasts the relation among parts of an organic unity with therelation between means and ends. A change of means would leave theintrinsic value of the end unchanged. In contrast, a change of any partwould alter the intrinsic value of the whole or organic unity to which thepart contributes, even if the part itself has no intrinsic value. Moore tellsus that the relation between parts and whole is not causal.49 Nor is itadditive; the intrinsic value of the whole bears no regular relation to thesum of the values of its parts.50 Otherwise, though he provides numerousexamples, Moore says little to clarify the nature of organic unities, and, infact, this is in keeping with his view of good as a simple, nonnatural, suigeneris property.51 No more can be said to explain why some items com-

48 Moore, Principia Ethica, sec. 18.49 Ibid., sec. 22.50 Ibid., sec. 18.51 For excellent discussion of Moore’s theory of organic unities, see Thomas Hurka, “Two

Kinds of Organic Unity,” Journal of Ethics 2, no. 4 (1998): 299–320.


bine so as to produce value in excess of their value as parts than can besaid to explain why good supervenes on some things rather than others.

I will return to the matter of organic unities later. For now, I want tofocus on the fact that Moore allows that a variety of things individuallyhave at least some intrinsic value. In particular, I want to make use brieflyof one of Moore’s more controversial claims about the value of an indi-vidual thing. Moore famously maintained that a beautiful object has atleast some intrinsic value apart from consciousness of it, even if the valueof the whole comprised of being conscious of a beautiful object far exceedsthe value of the beautiful object alone.52 In a thought experiment, intendedas one part of his refutation of hedonism, Moore asks us to imagine twoworlds, a beautiful world filled with mountains, sunsets, forests, water-falls, and everything we most admire, and a filthy, garbage-strewn worldfilled with all manner of things we find utterly revolting. He asks us toimagine further that no one will be around to experience either world.Still, Moore asks, wouldn’t it be rational to produce the former worldrather than the latter? Moore thinks it would: the existence of the beau-tiful world is better in itself “quite apart from its effects on any humanfeelings.” 53 Although Moore maintained that ‘good’ expresses a simple,unanalyzable nonnatural property, he sometimes characterized being goodas being such as “ought to be” or “ought to exist for its own sake.” 54 Inconcluding that the beautiful world is better in itself, Moore effectivelyconcludes that it is such as ought to exist.

But now consider that you (and the rest of us) are about to be annihi-lated. You stand in the “Create-A-World” booth with two buttons on thepanel in front of you. You push the button marked “beautiful,” and poof!There it is (and there we aren’t). Only, as it turns out, this beautiful newworld is not empty of human life —one person is in it, just one, and,indeed, as befits the world around her, she is a beautiful person, only sheis endlessly sleeping. Now the question is whether beauty is occurring inher life. In one sense, the answer must be yes; after all, she’s alive, beautyexists, and in the very time and place that she exists —it’s all around her.If we are not prepared to say that the beauty of the world around her isoccurring in her life, then at least we might be prepared to say that herown beauty is occurring in her life. And if Moore is right that beauty hasat least some intrinsic value, that would seem to be good occurring in herlife. Good occurs in the life of P when good occurs in the time and placein which P lives. Call this the time-and-place, or “TP,” characterization ofgood occurring in the life of.

If good occurring in a life required only spatiotemporal coincidence,then individuals could sleep through lives in which a good deal of good

52 Moore, Principia Ethica, sec. 50.53 Ibid.54 Ibid., sec. 69.


occurs, and this would have peculiar results for what we ought to do.55

We ought, other things equal, to redecorate meticulously the rooms of thepermanently comatose, pipe beautiful music into their rooms, send in theclowns. To be sure, we will promote more good by expending our ener-gies elsewhere.56 But the suggestion that we could have any reason topromote good occurring in the lives of the permanently comatose, at leastin this sense, is dubious at best. If this were all the Moorean had in mind,then I think we could reasonably doubt that good occurring in a lifemakes a claim on us just as much as we can doubt that the self-flagellator’sSumnerian good for makes a claim on us. Good occurring in the life of P,in the bare sense sketched by TP, does not capture what the good-fortheorist has in mind.

If coexistence in time and place, even to the extent of physical embodi-ment, is not enough for occurring in a life, what more is required? Sup-pose our sleeping beauty awakens, and she becomes conscious of theworld she inhabits. The question is not whether her consciousness of herworld introduces additional value.57 The question is whether her con-sciousness of it is enough to make what she is conscious of “occur in herlife,” so that if what she is conscious of is good, then there is “goodoccurring in her life.” 58 Consciousness might seem to be necessary. We

55 However problematic in other regards, this way of understanding good occurring inthe life of would at least have the virtue of ruling out alleged cases in which a personbenefits from events that occur after her death.

56 Moore remarks, “In any actual choice we should have to consider the possible effectsof our action upon conscious beings, and among these possible effects there are alwayssome, I think, which ought to be preferred to the existence of mere beauty. But this onlymeans that in our present state, in which but a very small portion of the good is attainable,the pursuit of beauty for its own sake must always be postponed to the pursuit of somegreater good, which is equally attainable” (ibid., sec. 50).

57 Moore himself at least seems to have thought consciousness to be of rather insignificantvalue. In his discussion of “the ideal,” Moore remarks, “By far the most valuable things,which we know or can imagine, are certain states of consciousness, which may be roughlydescribed as the pleasures of human intercourse and the enjoyment of beautiful objects”(ibid., sec. 113). But in introducing the notion of organic unities, he tells us that becauseconsciousness does not always greatly increase the value of a whole, “we cannot attributethe great superiority of the consciousness of a beautiful thing over the beautiful thing itselfto the mere addition of the value of consciousness to that of the beautiful thing. Whateverthe intrinsic value of consciousness may be, it does not give to the whole of which it formsa part a value proportioned to the sum of its value and that of its objects” (ibid., sec. 18). Andin the case of the organic unity involving contemplation of beauty, Moore makes it clear thatthe most valuable cases of aesthetic appreciation involve not “bare cognition of what isbeautiful in an object,” but also an emotional response appropriate to the kind of beauty inquestion (ibid., sec. 114). Moore later concluded that it is common to all intrinsic goods thatthey contain “both some feeling and some other form of consciousness.” See G. E. Moore,Ethics (New York: Oxford University Press, 1978), 107.

58 In a further challenge to the hedonist’s claim that pleasure is the sole good, Moore asks,“Can it really be said that we value pleasure, except in so far as we are conscious of it?Should we think that the attainment of pleasure, of which we never were and never couldbe conscious, was something to be aimed at for its own sake?” Pleasure, Moore concludes,is “comparatively valueless without the consciousness,” and thus is not the sole good.“[S]ome consciousness must be included with it as a veritable part of the end.” Moore,


think of the life of a person as something she lives, and not merely bybeing present, but by functioning as an agent. And so it would seem thatfor good, or anything else for that matter, to occur in her life, it has to besomething that at least enters into her consciousness and in that waymakes contact with her. Perhaps we should say, then, that good occurs inthe life of P when good occurs in the time and place in which P lives andP is conscious of the occurrence. Call this characterization of good occur-ring in the life of the time-place-consciousness characterization, or “TPC.”We can imagine more stringent characterizations than TPC, depending onwhat P must be conscious of for good to occur in her life. But because itmakes no difference to the arguments to come, let’s leave as an openquestion whether P must be conscious merely of the occurrence of some-thing that happens to be good or whether P must be conscious not merelyof something’s occurrence but of its good-making qualities or even ofboth those and its goodness.59

At this point, I want to suggest, we can begin to see where the goodtheorist and the good-for theorist would part company. Of course, somegood-for theorists might well say that TPC in fact captures all they evermeant to say when they talked about what is good for a person. Somewho accept an objective list theory of welfare, for example, might takesuch a view. To my mind, however, if they do take that view, then theyought to be good theorists rather than good-for theorists, because theywill have accepted a view that falls short of what I think good-for theo-rists must be tracking —or better, as we will see, attending to —in treatinggood for as distinct from good. So what does good occurring in the life ofleave out that good for might capture? We should notice two points ofdivergence between good occurring in the life of, in the TPC sense, andgood for.

First, not all good-for theorists believe that in order for something, X, tobe good for P, its occurrence or existence must enter into P’s conscious-ness. To take some stock examples, some have maintained that P could bebenefited by posthumous fame or that P could be harmed by the betrayalof her faithless partner, even if she never learns of his or her faithlessness.Yet both those good-for theorists who allow for such good (and bad) forsand those who deny them, take themselves to be arguing about what isgood for an individual, and so about where good for P is instantiated. If allthat good-for talk were about was good occurring in the life of as pre-scribed by TPC, good-for theorists who do not accept a consciousnessrequirement would not be talking about good for, despite what they maythink. These good-for theorists could be mistaken, but I do not think their

Principia Ethica, sec. 52. But this is not, as Moore’s later discussion of organic unities makesclear, because consciousness itself has much intrinsic value. See note 57 above.

59 Moore himself, in talking about the organic unity that comprises beauty and the con-templation of it, suggests something like this latter view. See Moore, Principia Ethica,secs. 115–21.


position should be ruled out prior to our efforts to advance and assessanalyses of good for P.

This first divergence might seem to suggest that the Moorean ought tohave stuck with TP as a specification of what it is for good to occur in thelife of someone. But even good-for theorists who reject a mental-stateview of the good for would and should contend that TP does not captureall that is involved in being good for someone.60 In any case, good andgood for diverge in a second way, even if we suppose that goods mustenter into consciousness to occur in a life.

Return to our beauty, who is now awake and conscious of the beautifulworld she inhabits; add, if you like, that she is conscious of being con-scious of beauty. Even if that were enough for good to occur in her life, itseems not enough, intuitively, for good for her to occur. We might cash outthis intuition in various ways, and which way makes most sense bears onwhat a correct analysis of good for P would look like. Suppose she isconscious just as described, yet takes no pleasure in the beauty aroundher; it is a matter of indifference to her. Or suppose her apprehension ofbeauty, even if accompanied by immediate pleasure, is promptly fol-lowed by crippling prickly sensations or by an overwhelming sense of herown inadequacy that engenders suicidal ruminations. Examples like thesesuggest that TPC, with its consciousness condition, not only requires toomuch but too little. As I will try to explain, the problem is not simply thatTPC is incomplete but that good occurring in the life of has entirely thewrong structure to get at what fundamentally concerns the good-fortheorist.

In response to the first claimed divergence between good and good for,the good theorist might simply deny that good can occur in a life whereP is not conscious of a good, even if good can also occur outside of a life.As for the cases in which our beauty apprehends beauty but takes nopleasure in it, or takes immediate pleasure promptly followed by physicalor emotional suffering, good has not, in fact, occurred in her life; or, moreprecisely, in the latter cases, the good occurred fleetingly only to be over-whelmed by the bad. Such possibilities do not show, however, that theidea of good occurring in the life of misses something captured by goodfor; they merely indicate that we do not yet have the right account of thegood.

As mentioned earlier, Regan himself thinks that the good is pleasurableexperience of objectively appropriate activities. For Regan, then, the goodconsists in something like Moorean organic unities.61 Of course, the mys-

60 For one thing, TP does not allow us to capture common-sense ideas about when goodseither do or do not make a difference to how an individual’s life goes. It seems all the sameto the permanently comatose patient’s life, whether the walls of her hospital room remainbare or display a Picasso.

61 Moore remarks that “Pleasure does seem to be a necessary constituent of most valuablewholes” (Principia Ethica, sec. 55), which, of course, leaves open the possibility of valuable


tery, on the Moorean picture, is just the one mentioned in passing in ourearlier consideration of organic unities, namely, what explains the com-plexity: why do some things but not others hang together in such a waythat good supervenes on them?

Regan considers a more specific version of the question, as it pertainsto his own claims about the good. One might ask, he conjectures, whetherby limiting the good to pleasurable experiences, the Moorean is not reallyconcerned with good or value for the individual after all. “Why elseshould an experience of a beautiful sunset, or a great mathematical theo-rem, or a Bach cantata, need to be enjoyable in order to be valuable?”Regan responds:

I cannot answer this point fully here. The short answer is that whatis really valuable (non-relatively) is the appreciative engagement ofthe subject with a worthy object. The subject’s pleasure is relevantbecause pleasure is an inevitable concomitant, and therefore a sign,of the right sort of engagement. But it is the engagement of thesubject and appropriate object that is valuable. To my mind, whenthere is the right sort of engagement, we could as well say that thevalue created is value ‘for’ the sunset, or the theorem, or the cantata,as insist that the value is ‘for’ the subject. The object of appreciationis brought to life by being attended to and appreciated. But in fact,the real value is neither ‘for’ the subject nor ‘for’ the object. The valueis just there, in a whole to which both subject and object make anindispensable contribution.62

As far as I can see, this response merely repeats in varying ways theMoorean position, without answering the underlying question. The ideaof an object of appreciation being “brought to life” by being pleasinglyappreciated seems to me a mere metaphor. As for the claim that we couldjust as much say that the right sort of engagement creates value for the sun-set, this strikes me as exposing a dilemma for Regan. If the expression ‘valuefor’ has to be understood in the same way in connection with the sunsetas it does in connection with a person, and it can only plausibly be under-stood as an abbreviated expression for good or value occurring in the lifeof, then value for the sunset would have to be value occurring in the lifeof the sunset. But what could this mean? Either good occurring in thelife of amounts to something along the lines of TPC, in which case good

wholes that do not include pleasure. He later came to the view that pleasure is a part ofevery intrinsically good whole, though he rejected the idea that intrinsic value is alwaysproportionate to the quantity of pleasure. See Moore, Ethics, chap. 7, pp. 103–7. The passagefrom Ethics quoted in note 57 continues: “and, as we said before, it seems possible thatamongst the feelings contained must always be some amount of pleasure.”

62 Regan, “Brother’s Keeper,” 221. I am not aware of any other place in his publishedwork that Regan addresses the issue in full.


cannot occur in the life of a sunset and thus be “for” the sunset, or goodoccurring in the life of is nothing but good, full stop, in which case Regancannot plausibly claim that good occurring in the life of captures all thatthe good-for theorist could sensibly mean by good for, even if the goodincludes all good fors.

If Regan’s response provides less illumination than we might havehoped, it nevertheless affords a very useful starting point for getting aclearer grip on the key differences between the pictures offered respec-tively by the good theorist and the good-for theorist. One differenceconcerns whether complexity or structure is to be found in substantivevalues or goods, or whether it is to be found in the value property itself.A view like that of Moore and Regan builds structure into its theory of thegood leaving good itself a simple, unanalyzable nonnatural property. And,indeed, that goods have a complex structure seems the only thing thatcould give the good theorist a prayer of making good on the claim thatthe good will include any plausible good for. The good-for theorist, incontrast, finds structure in the value property. ‘Good for P’ expresses akind of relational value, and to give an analysis of the relational propertyof being good for P —of welfare or well-being —is to give an analysis of acertain normative relation that holds between persons and other things.63

Whether (and this is a point to which I will return briefly) those things arethemselves good in something like the Moorean sense, whether, moreprecisely, they are intrinsically valuable, is a question about the good forrather than a question settled by the nature of the relational property.

In the case of both complex goods or organic unities and the good for,a relation of some sort holds between items. On Regan’s view, the relationis between a worthy object or activity and a subjective response, one thatfits or “brings to life” that object or activity. The subjective response is a“sign of” appropriate engagement, but we might say, more strongly, thatit fits and completes the object or activity so as to produce a whole that isgood. Although Regan says that what is “really valuable” is appreciativeengagement with a worthy object, in calling objects worthy and activitiesobjectively appropriate he must be supposing that such things have atleast some intrinsic value. It’s just that the values that are to be promotedare the wholes consisting of the worthy item or activity and the comple-mentary response. It is important to be clear about the relata in the com-plex goods that constitute the to-be-promoted. Regan makes clear that hedoes not subscribe to the view that individuals have value.64 The relata,

63 By a “normative relation,” I simply mean that the relation between the person and thething gives rise to reasons and obligations or, more generally, that it entails a variety ofnormative judgments. That X is good for P entails, ceteris paribus, that we have reason topromote X for P, that we are obligated not to deprive P of X, that P’s pursuit of X is rational,and so on.

64 Regan, “Brother’s Keeper,” 225–30. See also Donald H. Regan, “The Value of RationalNature,” Ethics 112, no. 2 (2002): 267–91. See also notes 67 and 69 below. Could a Moorean


then, are not the worthy object and the person herself but, rather, theworthy object and the appropriate response or engagement (pleasurableexperience, in Regan’s view, being an indicator thereof ) that happens tobe had by an individual. She figures in the mix, it seems, just because itis only agents who can respond in a fitting way to worthy objects orengage in objectively appropriate activities. When the relata are in place,the simple nonrelational property good supervenes, or, if you prefer, thepredicate “good” applies.65

What are the relata in the relation that is the focus of the good-fortheorist? As I have already noted, good-for theorists have said extraor-dinarily little about this.66 I believe, however, that the entire project ofattempting to discern the elements of personal good or individual welfarerests on something like the following picture. For the good-for theorist,the relation of interest holds between a valuable being —a person or agent —and an object or activity or another being.67 It might turn out that the onlyobjects or activities that can stand in the good-for relation to a person areones that are themselves worthy, but as far as I can tell, whether that is sois not settled by the very concept of personal good. The relata, then, arepersons, valuable beings, and objects or activities or other beings thatmay or may not themselves be valuable. The question still to be answeredby the good-for theorist concerns the nature of the relation that holdsbetween these relata.

subscribe to the idea that persons have value? Perhaps he could, so long as he was notcommitted to the view that states of affairs are the sole bearers of intrinsic value, in whichcase the idea that persons have value would involve a kind of category error. If the Mooreancould, then he could also allow that states of persons getting what is good for them haveintrinsic value and even that such states are among the most valuable.

65 See Donald H. Regan, “How to Be a Moorean,” Ethics 113, no. 3 (2003): 651–77, esp.651–57, where he seems to express some reservations about treating good as a property.

66 Some have, nevertheless, viewed themselves as engaged in providing an analysis ofwelfare. See, e.g., Sumner, Welfare, Happiness, and Ethics, ch. 6; and Darwall, Welfare andRational Care.

67 In what does the value of persons consist? This is, in my view, one of the million-dollar questions in ethics, and I confess to having no especially interesting answer to it.Still, we can get a feel for the idea of persons as valuable beings by reflecting on someof the ways in which philosophers have been or might be tempted to answer it. Onemight say that the value of persons just consists in their being appropriate objects ofcertain attitudes, such as love, concern, or respect. (Coons, following Darwall’s sugges-tions in Welfare and Rational Care about care and respect as attitudes we take up towardpersons, adopts a view something like this. See Coons, The Value of States and the Valueof Individuals, chap. 2.) Or for those who, like me, do not favor “fitting-attitude” analysesof value, one might point to features of persons in virtue of which certain attitudes arefitting. Perhaps the most common feature philosophers have adverted to is, roughlyfollowing Kant, their “rational nature.” (But see note 68.) I am tempted to say, as a firstapproximation, and will simply assert in the text of this essay, that the value of per-sons consists in their being intrinsic sources of legitimate normative demands. Cf. JohnRawls, “Kantian Constructivism in Moral Theory: The Dewey Lectures 1980,” Journalof Philosophy 77, no. 9 (1980): 515–72, p. 543: “People are self-originating sources ofclaims. . . .” But I have no illusions that this suggestion takes us very far in answeringthe question.


Another difference between the pictures offered by good and good-fortheorists should, at this juncture, already be apparent: the relations respec-tively involved in the good and in good for involve differing directions offit. For the good theorist, fit is object- or activity-directed —attitudes orresponses are to fit worthy objects or appropriate activities. Certainresponses are, we might say, “called for” by the nature of the valuableitem and complete a valuable whole as to-be-promoted. For the good-fortheorist, in contrast, fit is person- or agent-directed —items (valuable or not)are to fit or suit a valuable being —are “called for” by her nature —and soare to be promoted in relation to her.68 The views differ, we might say, inwhat they deem to be the center of normative gravity.69

68 When I say that, in the case of good for, items are to fit the person, I mean to allow thatthis fit to the person can be achieved in various ways, some of which will involve changesin the individual. Either way, however, the direction of fit normatively speaking is towardthe person —she is the center of normative gravity. I have talked about a person’s nature as“calling for” certain items. But her nature includes her capacities as an autonomous agent,and this fact, I have argued elsewhere, has important implications for the nature of personalgood. See Rosati, “Personal Good.”

69 As I have explained in a commentary on Darwall’s Welfare and Rational Care, I believethat the deep insight behind his rational-care theory of welfare is that a person’s welfare isin some sense what one ought to want for her out of an appreciation of her value. SeeConnie S. Rosati, “Darwall on Welfare and Rational Care,” Philosophical Studies 130, no. 3(2006): 619–35. Of course, this is not precisely how Darwall formulates his own view, but Ibelieve that the appeal of his view lies in how it reflects this idea. In advancing his rational-care analysis, Darwall is quite explicit that the direct object of care is the person herself, andhe claims that concern for the person is “primitive” (Darwall, Welfare and Rational Care, 71).A few philosophers, including Darwall, have suggested that the normativity of welfare restson the value of persons —our welfare matters because we matter. In addition to Darwall, seeElizabeth Anderson, Value in Ethics and Economics (Cambridge, MA: Harvard UniversityPress, 1993), 26; J. David Velleman, “A Right of Self-Termination?” Ethics 109 (1999): 606–20;and Rosati, “Personal Good.” Of course, the historical roots of this idea lie in Kant’s dis-tinction between conditioned and unconditioned value. See Immanuel Kant, Foundations ofthe Metaphysics of Morals, trans. Lewis White Beck (Upper Saddle River, NJ: Prentice-Hall,1995), 393–95 (Prussian Academy pagination). I believe that the underlying insight, as wellas the insight about the normativity of welfare, are best understood as indicating a featureof the relational property, what I have been calling its “direction of fit.” I have explainedelsewhere why I do not think that Darwall’s insightful account succeeds as an analysis ofwelfare. See Rosati, “Darwall on Welfare and Rational Care”; see also Rosati, “PersonalGood.” For a reply by Darwall, as well as other commentary on his book, see StephenDarwall, “Précis of Welfare and Rational Care,” Philosophical Studies 130, no. 3 (2006): 579–84;Fred Feldman, “What Is the Rational Care Theory of Welfare?” Philosophical Studies 130, no. 3(2006): 585–601; and Thomas Hurka, “A Kantian Theory of Welfare?” Philosophical Studies130, no. 3 (2006): 603–17. Regan examines Darwall’s theory in the course of making his caseagainst good for. I do not address his discussion of Darwall, in part, because I do not fullyagree with his characterization of Darwall’s position; but, more important, I do not believethat most of Regan’s criticisms of Darwall’s view apply to what I say. The one claim hemakes that would be relevant to assessing the position sketched in this essay is that he seesno special dignity or worth in persons. Regan has critically examined arguments for the ideathat rational nature has value; see Regan, “The Value of Rational Nature.” Although I thinkhe is right to find fault in extant arguments, I do not believe the shortcomings of thosearguments by any means end the matter. In any case, my aim in this essay is not to arguefor the value of persons but to explain how ‘good for P’ might express a distinct, fullynormative property. For a reply to Regan, see David Sussman, “The Authority of Human-ity,” Ethics 113, no. 2 (2003): 350–66.


Notice that this difference in the center of normative gravity tends toupend Regan’s inquiry, or at least his framing of it. As he expresses it, hisinquiry concerns not whether but why we are our brother’s keeper. Thegood theorist, he claims, thinks that we ought to be concerned with whatgoes on in one another’s lives, that we are our brother’s keeper; he merelyoffers a different and, Regan thinks, better account of why we are ourbrother’s keeper than the one offered by the good-for theorist. But nowwe can see that this is not, strictly speaking, true. Being concerned withwhat happens in our brother’s life is not being concerned for our brother,and, indeed, if our brother is not himself valuable, such direct concernwould appear to be misguided. On the good theorist’s view, we neverought to act, to borrow Stephen Darwall’s edifying formulation, for ourbrother’s sake, but only for the sake of what goods may occur in his life.Our brother and his responses will be an “inevitable concomitant,” to besure, and we cannot promote the good without supporting him as a partof that process, but if the question is whether I am my brother’s keeper, theanswer must be no. And so, contrary to Regan’s framing of his inquiry,the question “Why am I my brother’s keeper?” can never really arise forthe good theorist.70

The difference in what I have been calling the center of normativegravity reflects, in turn, a difference between two types of normativereasons, each of which has the universal normativity that Regan claimsgood has but that good for, if distinct from good, lacks. Some normativereasons derive from the value of worthy objects when appropriately appre-ciated or from the value of objectively worthwhile activities appropriatelyengaged in by an agent. Other normative reasons derive from the valueof agents or persons themselves. The normative reasons provided by factsabout what is good for a person are of the latter sort, and because thevalue of persons is an objective value, not a merely relative one, reasonsderiving from what is good for a person are reasons for anyone.71

These considerations point us to an important fact about the nature ofgood for that helps to make clear why the monadic property good is noproper part of good for P and that helps us to see what the ‘for’ is for.‘Good for’ expresses a relation; the good for is a form of objective value.But whereas ‘good’ in the good theorist’s scheme signals the intrinsic

70 For related reasons, Regan’s view has a problem that is the mirror opposite of awell-known problem for desire theories. Whereas desire theories, as Mark Overvold hasargued, seem to leave no room for self-sacrifice, Regan’s view makes no room for self-interested action, because it recognizes no meaningful notion of self-interest. One could acton one’s desires or what one values, contrary to promoting the good, but that would not beto act self-interestedly in the usual sense. See Mark Overvold, “Self-Interest and the Conceptof Self-Sacrifice,” Canadian Journal of Philosophy 10 (1980): 105–18. But see Regan’s discussionof the self-sacrifice problem, which no doubt reveals how he would respond to my remarksin this footnote: Regan, “Brother’s Keeper,” 224, note 52.

71 This assumes, of course, that the good-for theorist is correct in presupposing the valueof persons.


value of an item or whole, ‘good’ in the expression ‘good for’ does notsignal the intrinsic value of the item deemed to stand in the good-forrelation or to instantiate the relational property of being good for P. We canmake the point clearer by reminding ourselves of an important distinc-tion, to which Christine Korsgaard has called our attention, between twoways in which something can have value.72 It can have intrinsic value,which is the value a thing has in itself, a value which has traditionallybeen taken to depend on a thing’s intrinsic properties.73 It can also haveextrinsic value, or value that derives from some other source. Whereas thegood theorist’s good is intrinsic value, the good-for theorist’s good for,although just as normative, is a kind of extrinsic value; the good for givesanyone a reason to act, but the source of the “good-for value” of any itemthat is good for a person is its relation to a being with value. Of course,an item with good-for value might also have intrinsic value, but that isnot the sort of value our good-for talk concerns.74

Although good for is a form of extrinsic value, we can still meaning-fully distinguish between those extrinsically valuable things that are goodfor a person directly, or just in themselves, and those that are merelyindirectly or instrumentally good for her.75 Those things that are instru-mentally good for her are those that are means to the things that aredirectly good for her. Those that are directly good for her are those thatstand in a relation to her of the sort that the good-for theorist still needsto specify —those, we might say, that “fit” her. Insofar as these claimsabout the structure of the value tracked by our good-for talk are correct,the ‘good’ in ‘good for’ signals not intrinsic value but the existence of areason for promoting or supporting X in relation to P. Again, the item inquestion might itself have intrinsic value, but the occurrence of ‘good’ in‘good for’ does not signal the intrinsic value of that item.76 So how shouldwe understand the ‘for’? If ‘good for,’ in talk about what is good for aperson, expresses a relation of “fit” between a person and some object,

72 See Korsgaard, “Two Distinctions in Goodness.” See also Anderson, Value in Ethics andEconomics, 26. Rae Langton provides helpful refinement of Korsgaard’s distinctions in herexcellent essay “Objective and Unconditioned Value,” Philosophical Review 116 (2007).

73 For a challenge to this traditional view, see Shelly Kagan, “Rethinking Intrinsic Value,”Journal of Ethics 2, no. 4 (1998): 277–97.

74 Here might be as good a place as any to acknowledge that it might also be good thatpersons have what is good for them. But the fact, if it is one, that some or even all valuablestates of affairs are states of persons getting what is good for them would not show thatbeing good for P is not a distinct normative concept or property.

75 I talk in terms of things that are directly good for a person rather than in terms of ends,because not everything that is good for a person plausibly counts as an end, at least in thesense of something she could undertake to pursue or promote.

76 I should add here an important qualification to the line of argument I have beenpressing. The qualification is that the concept good for does not contain the concept good asa proper part if we understand good in the Moorean’s sense. That leaves open the possibilitythat, on other understandings of the concept good, we might plausibly understand propo-sitions about what is good for a person as, in some sense, built out of propositions aboutwhat is good.


activity, or being, the ‘for’ indicates the direction of fit.77 As a first approx-imation, I would suggest that when we say that X is good for P, weindicate that there is a reason to promote X with P as the beneficiary of theaction in light of or out of regard for the value of P.

An analogy may help to shed light on the structure of the relationalproperty of being good for P. Consider a valuable work of art —valuableaesthetically speaking, that is. When we appreciate the value of a work ofart, or what we take to be a valuable work of art, we undertake to pre-serve it in its valuable condition. That will involve preserving not simplythe features of it in virtue of which it is aesthetically valuable, but otherfeatures and surrounding conditions which may be necessary to preserveits overall condition. What this will require will vary, of course, with thesort of art work it is; preserving a marble sculpture will involve actionsdifferent from those involved in preserving a medieval fresco. Neverthe-less, these undertakings have in common an orientation toward the valu-able work of art.

Good for a person operates in a structurally similar way, presupposingthe ethical value of persons. Of course, persons are not themselves “to bepromoted”; we face no general duty to produce persons, as many believewe do to produce or bring about the good. Persons are, however, valuablein the sense that their nature renders them sources of a legitimate demandfor our considerate and respectful treatment.78 We perhaps most readilyappreciate the value of persons in the context of personal affections, andwe may be moved to act toward our intimates out of love or concern inprecisely the ways we ought to act just out of a regard or respect for themas valuable beings. When we appreciate, and act out of regard for, thevalue of a person, we seek to preserve her condition as the valuable beingshe is. That will involve not simply attending to those features she shareswith all persons and in virtue of which persons are valuable, but attend-ing to her sundry other features and her circumstances, all of which maybear on what preserves, sustains, or nurtures her. What regard for aperson will require will vary, at least in important ways, with the varyingfeatures and circumstances of different individuals. Nonetheless, as in thecase of a valuable work of art, our orientation in considering what is goodfor a person is the person herself.

Let me quickly sum up the discussion thus far. Good-for value is rela-tional, not relativist value.79 It is objective in that it gives anyone a reasonfor action, though the good for might, for all that has been said herein, be

77 See Rosati, “Personal Good,” for more discussion of the idea of ‘good for’ as expressinga relation of “fit.”

78 See note 67.79 A third reason for my interest in the Moorean challenge is a suspicion that relational

properties may figure more centrally in ethics than is commonly appreciated, and so a betterunderstanding of such properties would be helpful. Perhaps, for example, actions are neverright simply but always right in relation to a context and an agent at a time.


otherwise subjective; perhaps the normative relation good for itself involvesa mental-state requirement, or perhaps the relational property being goodfor P is instantiated only by items that produce particular mental states.Finally, good for is a form of extrinsic rather than intrinsic value.

B. The normative role of personal good

Peter Railton has observed, in developing his informed-desire accountof good for a person, that our discourse includes alongside the languageof the desired, wanted, or preferred, the language of the desirable, valu-able, or good.80 These discrete kinds of talk at least purport to do differentwork; in particular, talk of what is desirable or good purports to providea critical stance on the merely desired or preferred. I think a like point canbe made here. Our discourse includes alongside the language of what isgood or valuable or worth wanting, the language of what is good forsomeone, what is in her interest, what is worth wanting or doing for hersake. These discrete kinds of talk at least purport to do different, if notunrelated, work. Our discourse thus itself provides prima facie evidencefor the existence of distinct notions, doing distinct normative work.

What normative work might that be, in the case of good for? Assumingthat ordinary ethical talk is a reliable guide, good for functions norma-tively in at least three ways. First, judgments about what is good for aperson guide child-rearing. The vast bulk of what effective parents do inraising their children, including what they do in fostering their capacitiesas autonomous agents and bringing them into the moral community,aims to benefit their children, if not in the present moment, then over thecourse of a lifetime.81 Second, good-for judgments guide individualdecision-making with regard to how to build a life. Of course, in decidingwhat to pursue and how to live, we attend not merely to our desires butto what we believe to be of value and to “what we owe to each other,” toborrow Thomas Scanlon’s apt expression.82 But except where moralitymakes demands on us that require that we forgo a benefit, and exceptwhere an individual undertakes or is forced into a life of self-sacrifice, anindividual rationally forms these judgments with an eye to her long-range happiness or life satisfaction. Finally, good-for judgments guide thesetting of social policies. Indeed, their bearing on social policies is evident

80 See Railton, “Facts and Values,” 11.81 For extended discussion of effective parenting as a model for framing a theory of

preference-formation, see Connie S. Rosati, “Preference-Formation and Personal Good,”Royal Institute of Philosophy Supplements 81, Supp. 59 (2006): 33–64. See also Tamar Shapiro,“What Is a Child?” Ethics 109, no. 4 (1999): 715–38.

82 T. M. Scanlon, What We Owe to Each Other (Cambridge, MA: Harvard University Press,1998). Of course, my view of personal good and, in particular, my claims about the functionof personal good, indicate some disagreement with the central claims of chapter 3 of Scanlon’sbook, which presents a revised version of Scanlon’s 1996 Tanner Lecture, “The Status ofWell-Being,” Michigan Quarterly Review 36, no. 2 (1997): 290–310.


not only when it comes to questions of public health, education, andwelfare, but even in the basic operation of the law, at least in places wherethe legal system takes seriously the value of persons.

The property of being good for a person, as I have characterized it, wellmatches the normative role of judgments about what is good for a person.Whether from the first-person or third-person standpoint, our focus inmaking these judgments is on the person herself and what preserves andadvances her, presupposing her value and attending to the particulars ofher nature and circumstances.

The good theorist’s position, in denying the coherence of good for, isconspicuously at odds with the way we find it natural to talk and thinkabout ourselves and others.83 No parent, except perhaps someone like ourbewildered Moorean parent, would say that he sees to his child’s educa-tion because he seeks to bring about the valuable state of affairs thatconsists in his child’s pleasurable appreciation of knowledge. A parentaims, in his child-rearing activities, not at the good but at what is good forhis child —it is good for the child that she be educated; and he sees to hereducation, at least in the first instance, out of love and concern for her, notout of a concern for realizing value in the world.84 That a parent aims atthe good for his child rather than at the good simpliciter helps to explainwhy he focuses on his child and what happens to her in a way theMoorean might well regard as quite out of proportion to the overall goodto be realized.85 Vast swathes of our social lives, including family law,pediatric medicine, and primary and secondary education, reflect theparental orientation, with its focus on children and their welfare. Of course,a parent may, in aiming at what is good for his child, direct her towardwhat he thinks is good simply; but it is for the child’s own sake that hewants her not to miss out on valuable pursuits; it is for her sake that hewants the valuable state of affairs that consists in her taking pleasure inlearning to obtain. No person, except perhaps a Moorean, thinks of herquest for happiness as a quest to produce valuable states of affairs. Aperson aims, in seeking to lead a satisfying life, not to increase the quan-tity of value in the world but to enhance for her own sake the characterand quality of her own passage through time. And this is so, even if sheenhances her own life to a very large extent by engagement in activitiesthat increase the amount of value in the world as well. Our tendency totalk as we do and to orient much of our thought and action around whatbenefits ourselves and others seems utterly natural and, for reasons I willsuggest shortly, nearly inescapable.

83 True, we may be radically mistaken; we may not be valuable; there may be no norma-tive property of being good for P. But, then, there may be no property of being good either.

84 Darwall’s discussion of welfare, in Welfare and Rational Care, well reflects these sorts ofideas. See also Rosati, “Preference-Formation and Personal Good” and “Autonomy andPersonal Good: Lessons from Frankenstein’s Monster” (manuscript).

85 Thanks to David Sobel for urging me to stress this point.


One might be tempted to think that the good theorist will still routinelyissue the prescriptions of a good-for theorist. After all, if, say, the plea-surable responses of individuals are constituents of any valuable whole,then producing the good will be pretty much all about promoting goodoccurring in the lives of individuals. And as a necessary means thereto,there must be functioning persons; so we will feed, educate, and heal inall the familiar ways. But good occurring in the life of is in fact not wellsuited to the normative role that good for plays. Even if, in order to pro-mote the good for, we must preserve and promote persons as a necessarymeans thereto, there is no reason to think that would involve us in all thatwe ordinarily undertake on people’s behalf. The difference in direction offit between good and good for carries with it an orientation toward dif-ferent ends, and thus a difference in what is treated as a means and whatis treated as an end. What’s more, because the preservation and promo-tion of persons must be fundamentally instrumental for the good theorist,he confronts, in principle, the trade-offs that give rise to familiar, vexingproblems for utilitarianism. Why, for example, waste any resources onpersons whose capacities are such that developing them promises to yieldlittle in our efforts to promote the good?

IV. Agency and Value

Trade-offs of some sort there will inevitably be, of course, and so inmaking the foregoing claims, I do not mean to prejudge the question ofthe normative priority of good versus good for, though we can perhapsbegin to see what may be at stake in how we settle the question. I havebeen aiming simply to make it plausible that ‘good for P’ expresses adistinct, fully normative property and so one with a legitimate place inour moral ontology, whether or not it gets pride of place.86

Regan does, however, offer an argument to settle the question of nor-mative primacy against good for and in favor of good. His argumentappeals to a developmental story about agency. Regan’s tale goes roughlyas follows. We acquire our idea of value as a result of experiencing ourown felt wants, desires, and needs. The origins of our idea of value may,in fact, explain why many are inclined, mistakenly, to think that “valuemust be relative to a valuing agent.” 87 But even if our ideas of value havetheir beginnings in our desires and needs, we can develop into completeand competent agents only insofar as we come to see that we must lookbeyond these inner drives in order to arrive at the whole truth aboutvalue, including the truth about value occurring in our lives. The devel-opment from our initial, inner promptings to a full understanding ofvalue occurs incrementally. At first, we may experience problems of con-

86 This is assuming, of course, that there are any objective values at all.87 Regan, “Brother’s Keeper,” 222; see also 209.


flicting desires, and so we learn to restrain some desires in the interest ofothers. As a next or associated step, we may come to assess our lower-order desires from the standpoint of our higher-order desires. Ultimately,however, we will come to ask about the justifiability of our most funda-mental desires and even our entire set of desires: are those desires “jus-tified and directed at worthy ends”? Regan explains: “This is the step,raising what I have elsewhere called the ‘desire-transcending question,’that makes us agents. When we ask this question, we are asking for astandard for our desires, a standard that cannot be grounded in ourdesires themselves. . . . [A]cknowledging this question and confronting it. . . is the crucial passage to agency.” 88

At the point at which we realize that we must “transcend our desires,”Regan asks, does an agent who is deliberating about what is to be pro-moted think in terms of good or good for?

So far as I can see, there is no reason why he should assume at thisstage, when he first begins his survey of the world and his search forthe to-be-promoted, that the to-be-promoted is necessarily limited tooccurrences within people’s lives —and even less reason why he shouldassume that the to-be-promoted is necessarily limited to occurrenceswithin people’s lives that might seem in some intuitive sense toinvolve benefit ‘for’ them. But these limitations are what would cor-respond to the claim that the question the agent asks himself shouldbe understood as a question about the ‘good for’. If the agent doesnot begin with these a priori limitations, then we should understandhim as asking about the ‘good’. Even if it turns out that ‘the good’ isentirely constituted by things in people’s lives, and even things thatmight be thought of as ‘benefiting’ them in some sense, still, theagent’s question is about ‘the good.’ 89

Indeed, Regan claims, nothing would compel an agent “to attend to thefact of her own particularity as an agent.” 90

Although I am not convinced by all of Regan’s developmental story, Isee nothing in that story itself that a good-for theorist need deny.91 Thegood-for theorist might well allow that we come to the idea of valuethrough our own experiences of needing and wanting things, that wegradually come to restrain, assess, and, ultimately, seek full justificationfor our desires, as we develop into full-fledged agents. But if good for isas I have described it here, then full justification of an agent’s desires will,

88 Ibid., 223. See also Regan, “How to Be a Moorean,” sec. 1, for more discussion along thesame lines.

89 Regan, “Brother’s Keeper,” 223–24.90 Ibid., 224–25. Regan says this in response to an objection which stresses that it is the

agent, after all, who must decide what to do.91 But see note 17.


in some instances, be provided by facts about her own good rather thanfacts about the good. What’s more, the good-for theorist would remind usthat Regan’s developmental story, while acceptable as far as it goes, omitsa critical part: namely, the process or activity of raising full-fledged agents,a process that requires that we attend not merely to the good but to thefledgling agents themselves.

Indeed, given the characterization of good for on offer here, we shouldexpect that a fully developed agent’s question regarding the to-be-promoted will largely concern the good for. To be sure, it need not, andshould not, in the course of an ordinary human life, concern only her owngood for. If she is to be a faithful friend, a loving partner, an effectiveparent, and a full participant in the moral community, she will inevitablyask how to promote not simply her own good and not even simply thegood but also —and often —another’s good. Nonetheless, as an agent, as abeing with an interest in the “full truth about value,” the question of herown good cannot fail to press, and legitimately so. I would even beprepared to go further and say inescapably so; if she is to be able tofunction effectively as an agent in the first place, she must operate froma standpoint that presupposes and responds to her own value. I will nottry to defend these stronger claims here, however.92 Perhaps it is possiblefor a fully developed agent rationally and willingly to regard her life asa mere vessel of valuable states, and herself as merely a vassal of thegood. But if that is possible, it is only, I would suggest, because in thecourse of becoming an agent, other people did not so regard her.

V. Conclusion

If the claims I have been making about how to understand good for arecorrect, then ‘good for P’ indeed expresses the concept of a fully norma-tive property distinct from good —and even from good occurring in the lifeof an individual. Given the difference in direction of fit, we have reason todoubt that in fact the extension of good will include all the good for,although significant overlap will no doubt exist.93 One insufficientlyexplored question concerns why that would be so. Of course, it stillremains for good-for theorists to offer a compelling analysis of the fun-damental relation that has as its focus the value of the person. I hope tohave at least made it plausible that their efforts would not be wasted.

Philosophy, University of Arizona

92 For initial steps in that direction, see Rosati, “Autonomy and Personal Good: Lessonsfrom Frankenstein’s Monster.”

93 See Darwall, Welfare and Rational Care, chap. 4, for an intriguing discussion of theimportance to individual welfare of what he calls “valuing activities.”


FOUNDATIONS IN AQUINAS’S ETHICS*

By Scott MacDonald

I. Introduction

Aquinas likes foundations. He famously endorses a form of causalproof for God’s existence because he believes that one cannot go on toinfinity in series of what he calls “per se causes.” 1 We might say that hethinks there have to be causal foundations, causes that possess their causalpower in and of themselves and do not acquire it from other causes. Inepistemology, he argues that one cannot go on to infinity in chains ofjustification; there must be epistemic foundations that are known just byvirtue of themselves and require no further epistemic support.2 In hismoral theory, too, Aquinas likes foundations. Here is an application offoundationalist principles in his moral psychology:

In all cases of things that have a per se order with respect to oneanother, it must be that, when the primary thing has been takenaway, the things that are ordered toward the primary thing (ad pri-mum) are taken away. . . . Now in the case of ends one finds two sortsof order: the order of intention and the order of execution, and theremust be something primary in each sort of order. For a kind ofprinciple that moves the appetite is primary in the order of intention.Thus, when that principle has been taken away, the appetite wouldbe moved by nothing. . . . But the principle in intention is the ulti-mate end. . . . [T]herefore in this case it is not possible to proceed toinfinity, because nothing would be desired (appeteretur) if there wereno ultimate end.3

* I read earlier versions of this paper to the philosophy departments at Purdue Uni-versity, UCLA, and Texas A&M University, and at a conference at the University ofToronto. I am grateful to the audiences on those occasions for stimulating discussion. Iam grateful, too, for helpful comments from the other contributors to this volume, andits editors.

1 See the first two of Aquinas’s so-called five ways: Summa theologiae (hereafter, ST ), partI, question 2, article 3. (Hereafter references to ST will cite part, question, and article asfollows: Ia.2.3. “Ia” indicates the first part [ prima pars] of ST, “IaIIae” the first part of thesecond part [ prima secundae], and “IIaIIae” the second part of the second part [secundasecundae].)

2 See Aquinas’s Commentary on Aristotle’s Posterior Analytics book 1, chapter 4.3 ST IaIIae.1.4. Translations are my own except where noted.


So just as God is Aquinas’s first mover, ultimate ends are his first motivators.4

Foundationalist principles and reasoning appear prominently in othercontexts in Aquinas’s moral theory, and in this essay I consider one ofthese contexts. Aquinas sees a close analogy between practical reasoning —the sort of reasoning by which moral thought proceeds —and theoreticalreasoning —reasoning about matters of fact. He frequently exploits thatanalogy to make a point about the nature and structure of practical rea-soning, as in this passage:

Human reasoning, since it is a kind of movement, proceeds from anunderstanding of certain things —namely, things that are known nat-urally without any inquiry on reason’s part —as though from a kindof immovable principle. Moreover, it comes to a stop at an under-standing insofar as, by means of principles known naturally per se,we form a judgment about the things we arrive at by reasoning. Now,it is clear that just as theoretical reason reasons about theoreticalmatters, practical reason reasons about things that are to be done.Therefore, there must be things that are impressed in us naturally:principles regarding things to be done, just as much as principles oftheoretical matters. . . . [T]he principles regarding things to be donethat are impressed in us . . . pertain to a specific natural dispositionwhich we call synderesis. (ST Ia.79.12)

I will have more to say about the details of this passage shortly. But whatis immediately clear is that Aquinas thinks that practical reasoning pro-ceeds from foundations: practical principles that are impressed in usnaturally, known naturally per se, and possessed by synderesis.5

What I want to suggest here is that Aquinas’s foundationalism aboutpractical reasoning —the sort highlighted by Aquinas himself in passagessuch as the one I have just quoted —is easily misunderstood. Commen-tators often suppose that Aquinas takes this foundationalist account ofpractical reasoning to license a moral theory, and in particular a theory ofnatural law, with thick foundations: universal, objective, self-evident, sub-stantive moral principles from which all morality can be derived.6 But

4 I have discussed these various applications of foundationalist reasoning in variouspapers. See Scott MacDonald, “Aquinas’s Parasitic Cosmological Argument,” Medieval Phi-losophy and Theology 1 (1991): 134–73; MacDonald, “Theory of Knowledge,” in NormanKretzmann and Eleonore Stump, eds., Cambridge Companion to Aquinas (Cambridge: Cam-bridge University Press, 1993), 160–95; and MacDonald, “Ultimate Ends in Practical Rea-soning: Aquinas’s Aristotelian Moral Psychology and Anscombe’s Fallacy,” The PhilosophicalReview 100 (1991): 31–66.

5 Aquinas takes the technical term “synderesis” from the tradition. It appears to be acorruption of the Greek “suneisis” (“understanding”). What exactly Aquinas means by it willemerge in Section III below.

6 See, for example, Martha Nussbaum’s description of Aquinas’s moral theory in Nussbaum,Aristotle’s De motu animalium (Princeton, NJ: Princeton University Press, 1978).

FOUNDATIONS IN AQUINAS’S ETHICS 351

some foundationalisms are thinner than others, and I will argue thatAquinas’s foundationalism about practical reasoning is among the thinones and is thinner than these interpreters suppose. Since I think thatif one is going to be a foundationalist, thin is better, I will take myself tobe going at least some way toward defending Aquinas against thickinterpreters.7

In order to get the relevant texts and issues in front of us, I am going tofocus on a paper of Terence Irwin’s.8 Irwin is not a thick interpreter ofAquinas; he thinks Aquinas is conflicted as between thick and thin foun-dationalism about practical reasoning, concluding that Aquinas’s positionis inconsistent. I will reject Irwin’s conclusion, but I think his analysisusefully brings the issues into focus.9

Irwin characterizes the conflict he finds in Aquinas not in terms of thickversus thin foundations but in terms of wide-scope versus narrow- orrestricted-scope deliberation. But Irwin’s focus on the scope of what Aqui-nas calls “deliberation” is just another way of coming at the issue offoundations, as will become clear. Irwin argues that certain features ofAquinas’s moral theory commit him to thinking that deliberation haswide scope in our practical thought whereas other features of his accountcommit him to restricting deliberation’s scope more narrowly. In order tosee what this means, we need to have before us a basic picture of Aquinas’sunderstanding of practical reasoning.

II. Practical Reasoning: The Basics

Practical reasoning is reasoning about things that can be done(operabilia, agibilia) or are to be done (agenda). (For present purposes,we can leave aside practical reasoning’s application to things that canbe made [ factibilia].) Thus, one condition on practical reasoning —partof what distinguishes it from other sorts of reasoning —is its object:things that can be done, actions or courses of action that can be under-taken. But some of our thinking or reasoning about action is not entirelyor fully practical —it is theoretical or merely hypothetical thinking aboutaction. To think through the process of changing a flat tire in prepara-tion for writing an automobile owner’s manual, or to muse about whatone will do with the winnings, should one win the lottery, is to reasonabstractly and theoretically about things that can be done. If reasoning

7 My project here requires supplementing with an account of what Aquinas calls the“precepts of natural law.” I intend to undertake that account in another paper.

8 T. H. Irwin, “The Scope of Deliberation: A Conflict in Aquinas,” Review of Metaphysics 44(1990): 21–42.

9 Irwin himself no longer endorses the conclusion he draws in “The Scope of Delibera-tion.” See T. H. Irwin, “Practical Reason Divided: Aquinas and His Critics,” in GarrettCullity and Berys Gaut, eds., Ethics and Practical Reason (Oxford: Clarendon Press, 1997),189–214; and the chapters on Aquinas in Irwin’s forthcoming History of Ethics (Oxford:Oxford University Press).

352 SCOTT MACDONALD

about action is to be full-fledged practical reasoning, it must also satisfy oneother condition: it must begin from or be undertaken in the service of apresent desire or intention that one believes would be fulfilled or advancedto some extent by the action about which one reasons. If one desires some-thing (or intends to bring about some state of affairs or obtain or promotesome end), then when one reasons about how to satisfy that desire (bringabout that state of affairs, promote that end), one is reasoning about howone will proceed in that regard, what one should do, what course of actionto undertake. The desire or intention that precipitates the reasoning makesit fully practical. One’s reasoning in that case is not only about what can ormight be done but is also actually directed toward acting, toward pro-ceeding in the direction of fulfilling one’s intention. When reasoning of thissort, in these conditions, has run its course, the typical outcome, Aquinastells us, is a judgment about what course of action to undertake and a cor-responding volition or act of will —what Aquinas calls “choice.” On thispicture, then, the cognitive enterprise of reasoning about action is fully prac-tical when it is integrated with one’s appetitive states and acts, arising fromdesire or intention and giving rise ultimately to volition or choice.10

The point of fully practical reasoning is to enable us to act in aid of ourdesires and intentions by discovering or constructing ways of advancingor fulfilling them. The need for reasoning of this sort is especially clear incases where the objects of our desires or intentions are relatively generalor indeterminate or remote —too general or indeterminate or remote for itto be entirely obvious how to realize them. Reasoning of this fully prac-tical sort, then, connects our relatively general or indeterminate desires orintentions to more specific or determinate ones, and ultimately to con-crete courses of action we can undertake to satisfy or realize them. Aqui-nas develops this picture at some length and in considerable detail earlyon in the second part of Summa theologiae. In schematic form, the centralpart of his complex analysis looks like this:

Appetitive act: intention of some end ECognitive process: deliberation about how to attain or achieve E

(giving rise to a determinate judgment that somecourse of action Ø-ing is to be done [or somemore determinate end E* is to be pursued])

Appetitive act: choice to Ø [or to pursue E*]11

10 For fuller discussion, see my “Practical Reasoning and Reasons-Explanations: Aquinas’sAccount of Reason’s Role in Action,” in Scott MacDonald and Eleonore Stump, eds., Aquinas’sMoral Theory (Ithaca, NY, and London: Cornell University Press, 1999), 133–60.

11 Here are three simple examples:

(A) intention: to incorporate regular exercise in my lifedeliberation: walking to work every day would be a good way of incorporatingregular exercise in my life; judgment: so I should start walking to workchoice: to walk to work every day (or today) (continued)


(Terms in caps are Aquinas’s preferred technical terms. There is consider-ably more complexity in Aquinas’s account than this basic schema presents,but the additional complexity does not affect the present discussion.)12

As I have been describing it, practical reasoning is the cognitive pro-cess, the process of reasoning, that spans the distance between the precip-itating intention and the concluding choice. This way of characterizingfully practical reasoning focuses attention on the reasoning process towhich Aquinas gives the name “deliberation” and its immediate result, ajudgment about what is to be done. If we think of deliberation (plusjudgment) as what makes practical reasoning reasoning, then we mightthink of the flanking acts or states of appetite or will as what makespractical reasoning practical or fully practical. And we might think of thisthree-part schema as displaying the fundamental components (and therelations among them) required for a basic unit of fully practical reason-ing. (Henceforth, I will be using the expression “practical reasoning” tomean “fully practical reasoning.”)

III. The Scope of Deliberation

Once we see the basic schema, we might wonder whether the picture itpresents is complete. We might wonder, that is, whether there is more tothe reasoning part of practical reasoning than what is captured in theschema by “deliberation” (plus “judgment”). One way of pressing thatfamiliar question in this context would be to ask whether some rationalprocess or activity can play any role in our acquiring our ends, in ourcoming to have the intentions we have, the intentions from which delib-eration begins. If the answer to that question is “no,” then we mightsuppose that the basic schema I have presented is essentially the wholestory of practical reasoning. And in that case we might say that thereasoning in these practical contexts is subordinate to and wholly in theservice of antecedent desire. (“Wholly in the service of antecedent desire”sounds a bit more dignified than “slave of the passions.”)

Aquinas, however, clearly believes that the answer to our question is“yes”: reason or intellect can —in fact, must —play a role in our discov-ering or constructing the ends that are presupposed by deliberation. In

(B) intention: to walk to work every daydeliberation: if I’m going to walk to work every day, I’ll need to get up earlier;judgment: so I should get up earlierchoice: to get up earlier

(C) intention: to be happy (possess the complete good, live the best life)deliberation: happiness consists in a life containing an appropriate balance ofpleasure and virtue; judgment: so I should live that lifechoice: to pursue a life containing an appropriate balance of pleasure and virtue

12 Aquinas introduces these technical terms at ST IaIIae.11–13. The full discussion in STIaIIae extends from q. 6 to q. 17. For more detail, see my “Practical Reasoning andReasons-Explanations.”

354 SCOTT MACDONALD

fact, he seems tempted to offer two different accounts of the role reasoncan play, one of which ascribes a wide scope to deliberation and the otherof which restricts deliberation’s scope more narrowly. On the one account —the wide-scope account —deliberation can be about things that are for thesake of our ends (as the schema makes clear), but also, in significant waysto be explained shortly, about the ends themselves. We can deliberateabout what will be our ends. This is what it means for deliberation tohave wide scope.

On the other, narrow-scope account —the account suggested by Aquinas’sfoundationalism about practical reasoning —deliberation is restricted toreasoning about things that are for the sake of intended ends and cannotalso be about the ends themselves. This is what it means for deliberationto be restricted to a narrow scope. For Aquinas, however, this sort ofrestriction on the scope of deliberation does not amount to a restriction onreason’s scope in our practical lives, because deliberation is not the onlyrational process or activity that bears on practical matters. As we haveseen, he identifies a nondeliberative cognitive capacity which he some-times calls “synderesis,” a disposition by which we grasp fundamentalpractical principles —ends —from or on the basis of which deliberationcan proceed. On this account, synderesis and deliberation together —notdeliberation alone —constitute the rational components of practical reason.

Now, Irwin thinks it is clear that Aquinas sometimes ascribes widescope to deliberation and sometimes (more often, perhaps, when he isemphasizing the foundationalist structure of practical reasoning and therole of synderesis in it) restricts deliberation’s scope. That is the conflictIrwin finds in Aquinas’s account. Let’s look at the grounds for ascribingthese two different positions to Aquinas, beginning with the case for arestricted-scope reading.

A. The case for nondeliberative practical understanding (withrestricted-scope deliberation)

As we have seen, Aquinas has clear foundationalist commitments, andthey inform not only his account of theoretical cognitive activity but alsohis explanation of the nature of practical reasoning. In a passage welooked at earlier (ST Ia.79.12), he develops an extended analogy betweenthe structures of theoretical and practical reasoning, which I have set outin table 1.

In theoretical thought, we know some things because we have reasonedto them —they are conclusions of theoretical inferences. But theoreticalknowledge of this sort must rest ultimately on foundations: propositionsthat we know just in virtue of themselves and not by having reasoned tothem from other propositions. Aquinas calls these foundational proposi-tions “first principles”; he claims that we know them just in virtue ofthemselves (they are known per se); and he argues that, because of the


sorts of creatures we are, we have a natural capacity (which he calls“understanding”) for grasping first principles. Theoretical reasoning, there-fore, begins from the understanding of first principles and proceeds toconclusions which we can know by virtue of their inferential connection —inthis case, typically their deductive or demonstrative connection —withfirst principles.13

Aquinas insists that practical reason manifests the same sort of foun-dationalist structure.14 In some cases when we judge that certain thingsare to be done, our judgment is the conclusion of a process of reason-ing, deliberation. But judgments of this sort must rest ultimately onfoundations —first practical principles —that we grasp just by virtue ofthemselves. Aquinas says that, like theoretical first principles, these prac-tical foundations must be known per se, and it must be that we have anatural capacity —practical understanding or synderesis —for apprehend-ing them as such. Like theoretical reason, then, practical reason mustbegin from first principles grasped by a kind of understanding andproceed to judgments which we can know by virtue of their inferentialconnection —in this case, typically a nondeductive connection —with firstprinciples.15

With the foundationalist structure of practical reason in place, Aquinasasserts that nondeliberative practical reason —synderesis —must providethe ultimate starting points for practical reasoning. There must thereforebe a cognitive act, the natural understanding of first practical principles,antecedent to the first component in our original basic schema (set out inSection II above). Deliberation (a cognitive process) presupposes inten-tion of an end (an appetitive state), as the schema indicates. But synderesis(a nondeliberative cognitive act or state) must ultimately lie behind theintention of an end. In this structure, deliberation does not extend (ulti-mately) to our ends; synderesis accounts for our grasp of our (ultimate)

13 Aquinas takes the demonstrative syllogism, a particular form of deductive inference, asthe paradigm of theoretical reasoning. See my “Theory of Knowledge.”

14 See also ST IIaIIae.47.6.15 On the nondeductive nature of practical inference, see my “Practical Reasoning and

Reasons-Explanations.”

Table 1. The Analogy between Theoretical and Practical Reasoning

Theoretical Reasoning Practical Reasoning

Natural (unmediated) apprehen-sion of principles known per se

Understanding(intellectus)

Synderesis

Reasoning from principles toconclusions

Theoretical reasoning (typi-cally, demonstration)

Deliberation

Mediated apprehension of con-clusions

Theoretical knowledge(scientia)

Practical judgment

356 SCOTT MACDONALD

antecedent ends. Deliberation’s scope is therefore restricted to things thatare for the sake of any ultimate antecedent ends that fall within synderesis’sscope. For these reasons, attributing a restricted-scope account of delib-eration to Aquinas seems above reproach.

B. The case for wide-scope deliberation

Aquinas claims that, strictly speaking, deliberation is only about thingsthat are for the sake of an end and cannot be about ends themselves. Butthat claim can sound more restrictive than Aquinas intends it to be. Heallows, for instance, that basic units of practical reasoning of the sort wehave identified admit of being linked together in longer chains of prac-tical reasoning.16 In chains of this sort, deliberation will be about thingsthat are for the sake of our antecedent ends, things which, in turn, becomeour ends, once deliberation has led us to choose them, and can be thestarting points for further deliberation. Chains of reasoning of this sortprovide a way for deliberation to be about ends, but only about subsequentends, and then only indirectly, by being directly about things that are forthe sake of distinct antecedent ends.

Aquinas also allows another, extended sense in which deliberation canbe about antecedent ends, that is, about the ends presupposed by theprocess of deliberation. Some of our ends are abstract, general, or inde-terminate in various ways. Practical reasoning that begins from an end ofthat sort must first specify or make determinate or identify the constitu-ent parts of the end before it can proceed to consider concrete actions thatmight realize the end.17 Aquinas’s central discussion of the nature ofhappiness makes this point, too. All human beings desire happiness, thecomplete good, Aquinas believes. But the fundamental desire for happi-ness is wholly indeterminate; no one can pursue happiness without firstforming a conception of what happiness, the complete good, is, what itconsists in, what its essential components are (if it has component parts).Aquinas supposes that forming a determinate conception of happinesscan be a process of practical reasoning in the form of deliberation, movingus from an antecedent desire for happiness to a more determinate desirefor a life of civic virtue, say, or beatific vision, or a life with an appropriatebalance of pleasure and virtue.18 Moreover, it is reasonable to think thatwhen one deliberates about happiness in this way, one is deliberatingabout happiness, about what happiness is. But since happiness is the ante-cedent end presupposed by deliberation in this sort of case, it will followthat Aquinas is committed to thinking that we can deliberate about ourantecedent ends in these kinds of cases. Practical reasoning that begins

16 Examples A and B in note 11 link together in this way.17 Example A in note 11 is a case of this sort.18 See example C in note 11.


from a desire for happiness must, on Aquinas’s view, begin with deliber-ation about happiness. And so he allows important cases in which delib-eration’s consideration of things that are for the sake of an end amountsto or involves consideration of the end itself in a certain respect. In impor-tant cases, then, deliberation not only can but must have wide scope.19

The general point here is that deliberation needn’t be solely about exter-nal means —causal, productive, or preliminary means —to some end. (Iwill have a bit more to say about these special cases in Section V below.)

Thus, Aquinas ascribes to deliberation sometimes a wide scope andsometimes a narrow scope. Now, these two accounts pretty clearly pull indifferent directions, but it is worth pointing out that nothing we have seenabout them so far shows them to be strictly incompatible. Perhaps, forexample, Aquinas intends to maintain that practical reasoning is ulti-mately foundational in its structure with nondeliberative synderesisexplaining our access to first practical principles. But where other endsare concerned —ends that are not ultimate or first principles —deliberationcan have a wide scope of the sort we have seen. In that case, deliberationcould be about our antecedent ends just so long as it is not the ultimatestory about our grasp of ends.

IV. Irwin’s Conflict

Irwin argues, however, that when we attend to Aquinas’s broader moraltheory and the ways in which he employs his understanding of the natureand structure of practical reasoning in developing the theory, conflictemerges.

Aquinas’s account of the cardinal virtues —prudence and the three moralvirtues, justice, temperance, and courage —is built on the general accountof practical reasoning we have been developing. The moral virtues aresettled dispositions in the appetitive part of the soul that incline a personto the human good, where the human good consists in action or activityin accordance with right practical reason:20

[I]t must be said that the end of the moral virtues is the human good,and the good of the human soul is to be in accordance with reason (asis clear from Dionysius, De divinis nominibus IV). (ST IIaIIae.47.6)

19 “We can derive things from the natural law in two ways: in one way as conclusionsfrom its first principles; in a second way as specifications of certain general principles. . . .The second way is like the way that craftsmen in the course of exercising their skill adaptgeneral forms to specific things. For example, a builder needs to adapt the general from ofa house to this or that shape of a house.” ST IaIIae.95.2, trans. Richard J. Regan, Aquinas:Treatise on Law (Indianapolis, IN: Hackett, 2000).

20 “Right practical reason” refers to practical reason that correctly identifies ways of actingthat lead to or constitute genuine human good.

358 SCOTT MACDONALD

Aquinas goes on in this passage to explain what he means by “in accor-dance with reason”:

Thus it is necessary that the ends of the moral virtues pre-exist inreason. For just as certain things exist in theoretical reason as natu-rally known (which understanding has to do with) as well as certainthings —conclusions —that come to be known through them (whichtheoretical knowledge [scientia] has to do with), so certain things pre-exist in practical reason as naturally known principles. The ends ofthe moral virtues are of this sort because an end holds the place in thecase of things that can be done that a principle holds in theoreticalmatters, as was held above. Moreover, there are certain things inpractical reason that are like conclusions. Things that we get to start-ing from the ends themselves —things that are for the sake of theend —are of this sort. Prudence, which applies universal principles toparticular conclusions where things that can be done are concerned,has to do with these sorts of things. For that reason it does not pertainto prudence to establish the end for the moral virtues. It pertains toprudence only to deal with (disponere) things that are for the sake ofthe end. (ST IIaIIae.47.6)

Prudence, then, is the virtue associated with deliberation, and since delib-eration is only about things that are for the sake of the end, so is pru-dence. The right practical reason required for moral virtue, therefore,requires right deliberation about things that are for the sake of the end,and prudence is the virtue grounding right deliberation. But right prac-tical reason requires more than prudence; it requires in addition thatnaturally known (correct) first principles be in the soul antecedently, onthe basis of which prudence can proceed to a correct judgment aboutwhat is to be done. The natural understanding that grasps these practicalprinciples is synderesis, as Aquinas goes on to explain:

[I]t must be said that natural reason, which is called “synderesis,” hasestablished the end for the moral virtues . . . , not prudence, for thereason already given. (ST IIaIIae.47.6.ad1)

Thus, moral virtue requires synderesis and prudence. Moreover, pru-dence requires moral virtue, for correct apprehension of first practicalprinciples is not sufficient for right judgment about what is to be done. Ifthe appetites are not oriented by stable dispositions toward their ends,they can distract or deflect deliberation in its movement from right endsto judgments about things that are for the sake of those ends. Moral virtueand prudence, then, are codependent.

Consider now the following theses drawn from this brief account of therole of practical reasoning in virtue:


(P) Prudence (the virtue associated with deliberation) is right prac-tical reason about things that are for the sake of an end, not aboutends.

(S1) Synderesis (natural understanding of first practical principles),rather than prudence, apprehends the right ends (the ends of themoral virtues).

(V) Moral virtue requires being directed toward the right end (whichdistinguishes virtue from vice).

As Irwin points out, (P) and (S1) together appear to commit Aquinas to arestricted-scope conception of prudence as far as the ends of moral virtueare concerned. Moreover, (S1) and (V) entail

(1) That a virtuous person is directed toward the right ends (and isthereby distinguished from the vicious person) is explained bythe activity of synderesis.

As Irwin goes on to argue, however, the details of Aquinas’s understand-ing of synderesis undercut this result. That is to say, given Aquinas’sconception of it, synderesis cannot play the role that (1) claims for it. Thisis because Aquinas accepts

(S2) Synderesis (a) is possessed by everyone and (b) functions bynatural necessity.21

If Aquinas accepts (S2), then, given his commitment to (S1), he is com-mitted also to

(2) Everyone necessarily apprehends the right ends distinctive ofvirtue.

But if Aquinas’s understanding of synderesis —expressed in (S1) and (S2) —commits him to (2), then it is clear that he cannot also endorse (1).

Moreover, (2) and Aquinas’s view that the apprehension of the right endsdistinguishes the virtuous person from the vicious one together entail

(3) No one is vicious,

which is a happy conclusion, but not one Aquinas would be happy toendorse.

21 “The natural law directs human beings by certain general principles . . . and so thenatural law is one and the same for everyone” (ST IaIIae.91.5.ad3, trans. Regan).

“Everybody knows truth to some extent, at least regarding the general principles of thenatural law” (ST IaIIae.93.2, trans. Regan).

“We should say that the natural law regarding general first principles is the same for allpersons both as to their rectitude and as to knowledge of them” (ST IaIIae.94.4, trans.Regan).

360 SCOTT MACDONALD

There is one more important difficulty, as Irwin notices, if we introduceinto this mix Aquinas’s views about free choice.22 Aquinas holds that agiven volition or action is up to us, in our power, only if it is or is subjectto what he calls free choice. Moreover, he argues that we have free choiceonly about things that are subject to deliberation. But these two claims,namely,

(FC1) Something can be up to us, in our power, only if it is a matterof free choice, and

(FC2) Something can be a matter of free choice only if it is subject todeliberation,

entail that a crucial component of moral virtue —whether or not we aredisposed toward the ends of moral virtue —is not something that is up tous or in our power. That is,

(4) Our relation to the ends of moral virtue is not something that isup to us, in our power.

This is because (by [1]) whether or not we apprehend the right ends(which apprehension is necessary for moral virtue) is determined bysynderesis. But because synderesis is nondeliberative natural understand-ing (by [S2]), its products do not result from deliberation. Thus, our appre-hension of the right ends is not something that is in our power. Moreover,because synderesis is natural understanding, its functioning and its resultsare necessitated (by nature). So, again, they will not be in our power. Butthese results —expressed in (4) —are not acceptable to Aquinas, who holdsthat it is up to us whether we value the ends of virtue.

So there are several conflicts that emerge from Aquinas’s particular useof his account of practical reasoning in his theories of prudence and moralvirtue. Irwin finds the root of the difficulties in Aquinas’s two differentaccounts of deliberation, one of which gives a wide scope to deliberationand the other of which restricts deliberation’s scope. Given those twoaccounts, it is open to Aquinas to take a wide- or restricted-scope view ofprudence, the deliberative virtue. (P) itself is open to either interpretation.But Aquinas opts for a restricted-scope view of prudence (at [S1]), assign-ing to synderesis alone the role of setting the ends of moral virtue, a rolethat synderesis (as Aquinas understands it at [S2]) is unsuited to fill. In fact,Aquinas’s double-mindedness about the scope of deliberation is mirroredin a double-mindedness about synderesis. The conception of synderesisexpressed at (S2) must understand synderesis’s deliverances to be “thin,”highly general, nonsubstantive. That must be the case if synderesis is pos-sessed by and functions by nature in everyone, virtuous, vicious, and moral

22 These views are developed by Aquinas in ST IaIIae.6–13.


middlings alike. But the conception of synderesis expressed by (S1) requiresthat its deliverances be “thick” —substantive enough to distinguish the rightends constitutive of moral virtue. So (S1) and (S2) express, respectively, athick and a thin conception of synderesis. And here we are, back to the issueof thick versus thin foundations for practical reasoning.

In the discussion of prudence and moral virtue we have been trackingwith Irwin, Aquinas appeals to thick synderesis to do the work left undoneby restricted-scope prudence. Aquinas would have done better, Irwinthinks, to give prudence wide scope and give up the thick conception ofsynderesis. Abandoning (S1) and giving (P) a wide-scope reading wouldhave left Aquinas with a consistent set of fundamental principles and amore attractive theory of practical reasoning. According to Irwin,

Aquinas relies so heavily on a wide conception of deliberation in hisclaims about free choice, happiness, and the specification of the gen-eral principles of natural law that he ought also to rely on it toanswer the questions about virtue and prudence.23

V. Aquinas Unconflicted

What Irwin thinks Aquinas ought to have done, I think Aquinas actu-ally does. That is, I think Aquinas does assign prudence wide scope insetting the ends for moral virtue, and so has no need to —and does not, infact —appeal to a thick conception of synderesis. Aquinas does not makeit as easy to see this as he might have done, and thus he leaves himselfopen to misunderstanding. Nonetheless, I think that a close look at acrucial text or two will show that we are not forced to read Aquinas asIrwin does. In order to make my case, I will need to say something about(P), (S1), and (V), starting with (P).

A. Regarding (P)

Aquinas clearly asserts (P). But he also clearly holds that (P) is com-patible with prudence’s having wide scope. Prudence (and deliberationgenerally) must have an antecedent end from which to begin. But ourreasoning can be about that very antecedent end in certain sorts of cases,namely, when the end is general or indeterminate and we reason abouthow to specify it, or when the end is formal or abstract and we reasonabout what it consists in or how it might be concretely instantiated.24

23 Irwin, “The Scope of Deliberation,” 41.24 As in examples A and C in note 11. Example A is a case in which the end intended is

intended under a general description; example C is one in which the end intended isintended under an abstract or formal description.

362 SCOTT MACDONALD

The idea in these sorts of cases is that there is a sense in which adescription of an end in purely formal terms and a description of it interms of what satisfies or instantiates the formal description are differentdescriptions or conceptions of one and the same end. Similarly, there is asense in which the description of an end in general terms and a descrip-tion of it that specifies it are descriptions of one and the same end.

To illustrate, consider a case involving prudence where issues of tem-perance are in play. Let the intended end be acting in accordance with rightreason, and let prudence deliberate about what acting in accordance withright reason consists in, reaching the judgment that right reason requires,in general, eating healthily and, in this particular case, declining thedouble chocolate cheesecake.25 This case, like the others we have lookedat, has the right sort of features to count as a case in which prudence haswide scope. I will put this case to work in just a moment. But the firstpoint to see is that Aquinas’s asserting (P) is neutral with respect towhether he assigns prudence wide or only restricted scope, or commits toa thick or only a thin conception of synderesis.

B. Regarding (S1)

Now consider (S1). Aquinas clearly asserts that synderesis apprehendsthe ends of the moral virtues. But now suppose that prudence has widescope with respect to the ends of moral virtue, as in the case I have justdescribed of the double chocolate cheesecake.26 Given these assumptions,it seems that synderesis’s apprehension of the ends of moral virtue wouldbe compatible with prudence’s having wide scope with respect to theends of virtue provided that the ends that synderesis apprehends are for-mal or general or indeterminate and that prudence reasons about whatthese ends consist in or how they are to be specified.

If, as in the case before us, the end apprehended by synderesis were act-ing in accordance with right reason, then there would be room for prudenceto reason about what course of action in general or in some particular cir-cumstances might be one that is in accordance with reason. Not only mightprudence reason in that way, it seems that it must. For without some spec-ification or determination of the end acting in accordance with right reason,one cannot get started pursuing that end. In order to pursue it, one has todo something specific. Supposing that the prudent person in this case rea-sons as I have suggested and chooses ultimately to decline the cheesecake,then prudence will have found a way of pursuing the end presupposed by

25 intention: to act in accordance with right reasondeliberation r judgment: right reason requires eating healthilychoice: to eat healthilyintention: to eat healthilydeliberation r judgment: eating healthily requires declining the cheesecakechoice: to decline the cheesecake

26 See previous note.


the person’s reasoning, the end of acting in accordance with right reason.Prudence will have identified something for the sake of that end but alsowill have successfully reasoned about the end (in the sense of having set-tled what the end itself consists in, in general or here and now).

This way of characterizing the ends that synderesis apprehends —assufficiently formal or general —is just to say that wide-scope prudence iscompatible with synderesis, provided synderesis is conceived of as appro-priately thin.

So why think that Aquinas commits to a thick conception of synderesis?Irwin claims to find Aquinas committing himself to both a thick and a thinconception of synderesis in this passage:

[I]t must be said that [i] being conformed to right reason is theproper end of each moral virtue. [ii] For temperance aims at this,so that a person does not turn away from reason on account of hisconcupiscences, and similarly in the case of courage, so that aperson does not turn away from the right judgment of reason onaccount of fear or bravado. [iii] This end has been established forhuman beings in accordance with natural reason. For natural rea-son dictates to each person that he should act in accordance withreason. [iv] But how and by means of what things a person attainsto the mean of reason in his acting pertains to the arrangement ofprudence. For although attaining to the mean is the end of moralvirtue, the mean is found through the right arrangement of thingsthat are for the sake of the end. (ST IIaIIae.47.7)27

Irwin claims to see the thin conception in Aquinas’s claim, at [iii], that“natural reason dictates to each person that he should act in accordancewith reason” but the thick conception at [ii]: “temperance aims at this,so that a person does not turn away from reason on account of hisconcupiscences, and similarly in the case of courage, so that a person doesnot turn away from the right judgment of reason on account of fear orbravado.” This bit at [ii] points to thick dictates of synderesis because itidentifies ends that distinguish the virtues from one another.

But I think the passage need not be read in the way Irwin suggests.Natural reason (that is, synderesis) dictates acting in accordance withreason (at [iii]); this is a thin dictate. And Aquinas tells us at the beginningof the passage (at [i]) that this thin dictate articulates the proper end of

27 I have inserted the lowercase roman numerals in order to be able to refer to parts of thispassage in the discussion that follows.

Irwin translates the passage a bit differently at [ii]: “For temperance aims at this, namely,that a human being should not deviate from reason because of appetites; and similarly, <theaim of> bravery is that a human being should not deviate from the correct judgment ofreason because of fear or rashness” (Irwin, “The Scope of Deliberation,” 27–28). Irwin’s“namely” in the first half of [ii] renders Aquinas’s ut. I take it differently, as introducing aresult clause, with “so that.” See the discussion that follows in the text.

364 SCOTT MACDONALD

each moral virtue. (The demonstrative “this” at the beginning of part [iii]refers us back to part [i], not to what has just been said in part [ii]. Theclarifying second sentence in part [iii] makes that clear.)

Now, what are we to make of the bit of the passage (part [ii]) thatmentions the specific virtues of temperance and courage? I think we neednot read it as Irwin does, as identifying two thicker dictates of naturalreason, one for temperance (that a person does not turn away from reason onaccount of his concupiscences) and one for courage (that a person does notturn away from the right judgment of reason on account of fear or bravado). Wecan read the two parallel “so that” (ut) clauses as result clauses ratherthan as parts of the intentional objects at which temperance and courageaim. So part [ii] might be read as telling us that temperance aims at actingin accordance with reason where the result of temperance’s aiming in thatway will be that the temperate person does not turn away from reason onaccount of his concupiscences. Read that way, part [ii] of the passage,consistently with the rest of the passage, identifies only a thin dictate ofnatural reason. Aquinas need not be switching here back and forth andback again, from thin to thick and back to thin conceptions of synderesisand its results. Moreover, Aquinas refers elsewhere and often to versionsof this thin dictate.28 That provides more evidence that the consistentthin-dictate reading is correct here.

It should also be said that part [iv] of the passage can be taken asconfirming the wide-scope prudence hypothesis I have been suggesting.Temperance (and each of the other virtues) aims at the mean of reason.That (namely, the mean of reason), like the expression with which it ismeant to be synonymous, acting in accordance with right reason, is purelyformal. And what Aquinas says here in part [iv] is that specifying themean of reason is the work of prudence. So when prudence determineswhat the mean actually consists in in circumstances where concupiscencesare in play (in general terms or in a particular situation), prudence will bedetermining what the end of the virtue of temperance consists in. Hence,prudence will be exercising its wide-scope powers on a thin end presup-posed by prudence and apprehended by synderesis. That is, in fact, thepicture I have sketched in the double chocolate cheesecake case we havebeen considering.29

Thus, (S1) requires clarification:

28 “It is the law for human beings, which is allotted by God’s ordination according to theircondition, that they act according to reason” (ST IaIIae.91.6, trans. Regan).

“The virtue of the irascible and concupiscible powers consists of being duly obedient toreason” (ST IaIIae.92.1, trans. Regan).

“Things to which nature inclines human beings belong to the natural law, as I have saidbefore, and one of the things proper to human beings is that their nature inclines them to actin accord with reason” (ST IaIIae.94.4, trans. Regan).

“It is correct and true for all persons that they should act in accord with reason” (STIaIIae.94.4, trans. Regan).

29 See note 25.


(S1*) Synderesis (natural understanding of first practical principles),

rather than prudence, apprehends the (thin) right end of themoral virtues, namely, acting in accordance with right reason.

C. Regarding (V)

If we take Aquinas’s conception of synderesis to be consistently thin,then we need to say a bit by way of explanation with regard to (V). Moralvirtue requires being directed toward the right ends, and the right endsdistinguish the virtuous person from the vicious person. If we follow outthe line I have been pushing, then we might say this: There are two sortsof right ends, call them thin and thick (or thin and thicker) ends. Themoral virtues aim at a thin end —the mean of reason, or acting in accor-dance with reason. But this thin end does not distinguish virtue from vicebecause, Aquinas supposes, everyone aims at this thin end. Moreover,given (as I have argued) that this thin end is apprehended by synderesis,which all of us have in common, this is the result we would expect.

But the moral virtues also aim at thicker ends —specifications and deter-minations of the mean of reason. These thicker ends will distinguish virtuefrom vice. The vicious person will, like everyone else, aim at acting inaccordance with reason, but will be misguided about what acting inaccordance with reason actually consists in, and hence will care about andbe directed toward the wrong thick ends. But these thicker ends, insofaras they are specifications or determinations of the thin end of acting inaccordance with reason, will be determined or specified by deliberation —prudence in the case of the virtuous person, bad deliberation in the caseof the vicious person. It seems to me that part [iv] of the passage we havejust looked at supports this way of looking at it.30

Thus, (V) requires clarification:

(V*) Moral virtue requires being directed toward the (thick) rightend (which distinguishes virtue from vice).

Now, when (P), (S1), and (V) are understood in the ways I have suggested,they do not conflict; that is, they do not lead to the contradictions Irwinidentifies. We should conclude, therefore, that Aquinas himself consis-tently endorses wide-scope prudence and a correspondingly thin concep-tion of synderesis.

30 See also ST IaIIae.94.3, trans. Regan: “Since the rational soul is the specific form ofhuman beings, everyone has an inclination from one’s nature to act in accordance withreason. And this is to act virtuously. And so in this regard, all virtuous acts belong to thenatural law, since one’s own reason by nature dictates that one act virtuously. But if weshould be speaking about virtuous acts as such and such, namely, as we consider them intheir own species, then not all virtuous acts belong to the natural law. For we do manythings virtuously to which nature does not at first incline us, but which human beings bythe inquiry of reason have discovered to be useful for living righteously.”

366 SCOTT MACDONALD

VI. Conclusion

I want to conclude by calling attention to a moral that might be drawnfrom my reading of Aquinas’s foundationalism about practical reasoning.It is this: If Aquinas consistently maintains that synderesis’s dictates arethin —and, as I have told the story in my brief account, they are exceed-ingly thin —then the grounds for finding in his foundationalism the basisfor a robust kind of intuitionistic moral epistemology are also exceedinglythin. On my reading, the deliverances of the natural understanding thatbelongs to practical reason are perhaps purely formal but certainly entirelygeneral and, for that reason, practically useless by themselves. The realwork of practical reason begins with deliberation, and deliberation, Aqui-nas tells us, is complex, messy, contingent, and not intuitionistic. In thisrespect, the foundationalist model and the analogy with theoretical rea-soning which he deploys tirelessly can mislead. The practical foundationsAquinas’s account provides are, in fact, quite meager. That is not to saythat there are none or that he does not really believe the model or theanalogy. It is to say that we should not bring to the model or the analogyrobust expectations (perhaps acquired elsewhere) of what moral founda-tions will be like and what substantive work they will do for us in moraltheory.

This particular moral, in fact, precisely parallels a moral I have drawnin another, related context.31 Aquinas argues in the treatise on happiness(ST IaIIae.1–5) that some one desire, a desire for a single ultimate end, ispresupposed by and provides the foundation for all our end-directedthought and action. That looks like an untenably strong claim about ourfundamental desires. Properly understood, however, this claim turns outto be quite thin: he means only that all our end-directed thought andaction is directed toward a purely formal ultimate end, living the best life.The real work in pursuing the best life begins with reasoning about whatthe best life consists in. And that is a question for practical reasoningitself —wide-scope deliberation —not something given to it as part of itsultimate, unquestioned and unquestionable foundations.

It seems to me, then, that Aquinas’s understanding of practical reason-ing and its scope and limits is consistent throughout his theory of endsand his theory of the virtues. Deliberation has wide scope. Nevertheless,there are some ends that are presupposed even by wide-scope delibera-tion. Those ends, however, are exceedingly thin.

Philosophy, Cornell University

31 See my “Ultimate Ends in Practical Reasoning.”


REVISIONARY INTUITIONISM*

By Michael Huemer

I. A Conservative Metaethics?

Ethical intuitionism is often associated with a conservative approach tonormative ethics, an approach that embraces common-sense morality withat most minor revisions. This association seems to be borne out by W. D.Ross’s blandly conventional characterization of our “prima facie duties,”which, in his view, include such duties as keeping promises, showinggratitude to benefactors, improving oneself, avoiding injury to others,and so on. When these prima facie duties come into conflict, one mustsimply exercise one’s best judgment, intuitively, to decide which obliga-tion is more pressing in the circumstances. This system of duties, Rossexplains, derives from the prephilosophical convictions of “the plain man”:“The main moral convictions of the plain man seem to me to be, notopinions which it is for philosophy to prove or disprove, but knowledgefrom the start.” 1 H. A. Prichard, the father of twentieth-century intuition-ism, took a stance no less conservative. In his view, the characteristicmistake of moral philosophy is that of seeking further justification for ourprereflective moral convictions; in reality, the only knowledge moral phi-losophy can bestow on us is the knowledge of the self-evidence of theprinciples of obligation comprising common-sense morality.2

This alliance between intuitionism and common-sense morality at firstseems natural, perhaps even inevitable. The core epistemological thesisfrom which intuitionism takes its name is that all moral knowledge derivesits justification from certain “intuitive” moral truths. One natural way ofunderstanding the notion of an intuitive moral claim is simply as a claimthat seems correct, prior to reasoning. If morality is to be based on suchseemingly-correct moral claims, it is only reasonable to take into accountwhat seems true to most people: if I take my own moral intuitions asevidence of moral reality, then presumably I should recognize the moralintuitions of others as equally evidence of moral reality, absent specialevidence of cognitive defects on their part or exceptional cognitive abil-ities on my part. And once one takes this epistemological stance, it seems

* I would like to thank the other contributors to this volume, and its editors, for theircomments on an earlier draft of this essay.

1 W. D. Ross, The Right and the Good (Indianapolis, IN: Hackett, 1988), 21–22.2 H. A. Prichard, “Does Moral Philosophy Rest on a Mistake?” in Prichard, Moral Obli-

gation (Oxford: Clarendon Press, 1957), 16.

DOI: 10.1017/S026505250808014X368 © 2008 Social Philosophy & Policy Foundation. Printed in the USA.

natural, perhaps inevitable, that one will more or less embrace the moralbeliefs that seem right to most people —in short, common-sense morality.

Pace Prichard, some revisions to common-sense morality may be calledfor. For it may emerge that some of our prephilosophical moral beliefs aremutually inconsistent or paradoxical. In such a case, one would expectthe ethical intuitionist to endorse whatever resolution of the moral par-adox is best supported by intuition —which is to say, to recommend thesmallest revision to common-sense morality that restores coherence toour moral belief system, taking into account both the number of beliefsthat must be revised and the strength of the relevant intuitions. Thisseems to be the essence of the widely accepted method of “reflectiveequilibrium.” 3 We would expect that such a method, if it generates anymoral system, would lead to a moral system reasonably close to common-sense morality. In short, it seems that the epistemology of intuitionismsupports the method of reflective equilibrium, which in turn supportscommon-sense morality, with at most minor revisions.

This, I intend to argue, is a mistake. Intuitionists need not embraceanything close to common-sense morality. Instead, intuitionists may, andprobably should, adopt revisionary ethical views, rejecting a wide rangeof commonly accepted, prephilosophical moral beliefs. Indeed, I believeintuitionists are better positioned than partisans of most alternative con-ceptions of ethics to motivate a rejection of common-sense morality. Thisfact derives partly from the staunch realism of ethical intuitionism, andpartly from the evidence, often cited by critics of intuitionism, of non-rational biases contaminating common moral judgments.

In the following discussion, after briefly describing the doctrine ofethical intuitionism, I shall review some important challenges to the valid-ity of intuition as a source of moral knowledge. According to these chal-lenges, many of our ethical intuitions can be explained as products ofemotional and other nonrational biases, and for that reason cannot betaken as pointing us toward any objective moral truths. Traditionally, thissort of challenge is thought to support the conclusion, either that ethics issubjective, or that, if there are objective moral truths, we are not in aposition to know those truths. I shall argue that the intuitionist mayrecognize the seriousness of the challenges to the reliability of intuition,without giving in to subjectivism or skepticism. The proper response for

3 The method of reflective equilibrium, as initially described by John Rawls, calls for oneto consider intuitively plausible general theories that come close to systematizing one’smoral judgments about particular situations, and then to adjust both one’s moral judgmentsabout particular situations and one’s moral theories to bring them into harmony with eachother. See John Rawls, A Theory of Justice (Cambridge, MA: Harvard University Press, 1971),48–49. The method is now often understood in a broader sense, to include any process ofweighing conflicting beliefs against each other and renouncing the less plausible beliefs inorder to restore coherence to one’s belief system. For a defense of the method, see MichaelDePaul, “The Problem of the Criterion and Coherence Methods in Ethics,” Canadian Journalof Philosophy 18 (1988): 67–86.

REVISIONARY INTUITIONISM 369

an intuitionist, rather than renouncing the possibility of objective moralknowledge, is to adopt a critical methodology in which the kinds ofintuitions that are most subject to bias are discounted, while intuitionsthat are less prone to bias are given more weight. The end result will mostlikely be a revisionary ethical theory.

II. The Commitments of Ethical Intuitionism

A number of doctrines have gone under the name of “ethical intuition-ism,” from G. E. Moore’s conception of primitive, “non-natural” ethicalproperties, to Ross’s theory of multiple basic “prima facie duties,” toPrichard’s doctrine of the self-evidence of obligations.4 Here I shall focuson two claims central to the version of ethical intuitionism I defend.5 Thefirst of these claims is moral realism: the sort of intuitionism in which I aminterested takes at least some evaluative claims to be objectively true, thatis, true in virtue of facts existing independently of our attitudes towardthe objects of evaluation. For instance, “Torturing puppies is wrong” ismade true by the fact that torturing puppies is (in normal conditions)wrong, which holds regardless of how we feel or what we believe aboutthe torture of puppies. The thesis is both ontological and semantic: objec-tive moral facts exist, and moral language is about them.

Second, my intuitionism embraces an epistemological doctrine to theeffect that all justification for evaluative beliefs derives ultimately fromintuition. There are at least two ways in which the notion of intuition maybe understood. On one account, intuitions are a species of beliefs, distin-guished from other beliefs by the special way in which they are justified.Roughly, the idea is that for some propositions, one’s understanding ofthem can itself be an adequate source of justification for believing them.6

For example, once I understand the proposition “2 is less than 3,” I amjustified in believing it. This intuitive belief, then, requires an exercise ofintelligence —I must intellectually grasp the proposition “2 is less than3” —but it does not require inference or reasoning. That is, I do not derivethe proposition “2 is less than 3” from any other propositions.

On another account, intuitions are a type of cognitive state distinctfrom belief, a state that one sometimes avows by statements of the form,“It seems to me that p,” or “It appears that p.” 7 More specifically, intu-itions are initial intellectual appearances, that is, states of its seeming to onethat something is the case upon intellectual consideration (as opposed to

4 G. E. Moore, Principia Ethica (Cambridge: Cambridge University Press, 1903); Ross, TheRight and the Good; Prichard, “Does Moral Philosophy Rest on a Mistake?”

5 See my Ethical Intuitionism (New York: Palgrave Macmillan, 2005).6 See Robert Audi, The Good in the Right (Princeton, NJ: Princeton University Press, 2004),

33–36, 48–49.7 See Huemer, Ethical Intuitionism, 99–104; George Bealer, “A Theory of the A Priori,” in

James Tomberlin, ed., Philosophical Perspectives 13: Epistemology (Cambridge, MA: Blackwell,1999), 30–31.

370 MICHAEL HUEMER

sensory observation or introspection), but prior to reasoning. When onethinks about a proposition p, one often has the experience of seemingly“seeing” it to be true. This “seeing” is intellectual, rather than perceptual,but, unlike most intellectual cognition, it does not require reasoning fromany further premises. Advocates of this account argue that intuitions aredistinct from beliefs, since it is possible to have the intuition that p with-out believing that p, perhaps because one takes one’s intuition to beunreliable or because one takes oneself to have overriding evidence againstp. For example, most people who consider the comprehension axiom ofnaive set theory find it intuitive (it seems right), even those who know theaxiom to be false because of the paradoxes it engenders.8

Although I favor the second account of intuition, either account willsuffice for my purposes here. The key point for what follows is thatintuitions are taken to be cognitive, intellectual states with propositionalcontents. Intuitions thus contrast with emotions, which are noncognitive;with sensory observations, which are nonintellectual; and with statesof liking or dislike, which are noncognitive, nonintellectual, and perhapsnonpropositional.

How are these two principles —moral realism and the epistemologicaldoctrine of intuitionism —related to each other? Given moral realism, thecorrect evaluative judgments are independent of what our intuitions are.Our having the intuition that x is wrong, no matter how many share theintuition, is compatible with x’s actually being permissible. The relation-ship between our intuitions and the moral facts is, in some importantrespects, analogous to the relationship between our observations and thephysical facts about our environment: physical facts exist independentlyof our observations, but observations are our way of knowing about thephysical facts; the function of observation, which it usually fulfills, is tocorrespond to the physical facts. Similarly, moral facts exist indepen-dently of our intuitions, but intuitions are our way of knowing about themoral facts; the function of ethical intuitions is to correspond to the moralfacts. Sometimes our intuitions may deceive us, just as our senses maydeceive us. Intuition may, in fact, be less reliable than sensory observa-tion; nevertheless, enough of our intuitions are accurate that we can con-struct a substantial body of ethical knowledge.

The analogy between observation and intuition does not hold in allrespects. The experience of seeing an object differs from that of having anintellectual intuition, and, more importantly, the explanation for the reli-ability of sensory observation presumably differs from the explanation forthe reliability (such as it is) of intuition. Observations are reliable indica-tors of physical facts because of the way our sense organs are causally

8 The comprehension axiom states that for any well-formed predicate, there exists a setcontaining all and only the things to which that predicate applies. This leads to a contra-diction when the well-formed predicate “is not a member of itself” is introduced.


connected to phenomena in the external world. The explanation for thereliability of intuition is more controversial. Perhaps the best explanationadverts to the notion of our grasping abstract objects; in any case, itpresumably is not the same as the explanation for sensory observation.9

This does not, however, defeat the point of the analogy I have drawn. Thepoint I have made is simply that moral realism fits together with anintuitionist moral epistemology just as realism about external objects fitstogether with a broadly empiricist epistemology of the external world.Importantly, just as we must be on guard against sensory illusions, theintuitionist must be on guard against moral illusions.

III. Four Skeptical Challenges

In this section, I review four main kinds of challenges that have beenposed to the validity of intuition as a source of moral knowledge. My aimhere will not be to rebut these challenges, but rather to set out the moralskeptic’s case as clearly as possible, and thus to indicate the seriousnessof the problems that intuitionists face.

A. The incoherence of our intuitions

Some philosophers argue that our moral intuitions about many specificscenarios are influenced by factors that seem morally irrelevant. Thus,Peter Singer suggests that our intuitions about obligations to assist othersin need are improperly influenced by the physical proximity of thoseothers, so that we think we have much stronger obligations to people whoare in front of us than to people who are thousands of miles away.10 Thismay explain why most people consider it seriously wrong to refuse tosave a child who is drowning in a shallow pond, just to avoid gettingone’s clothes wet and missing a lecture, yet few consider it seriouslywrong to decline to save starving Third World children just to avoidgiving up a few luxuries that one enjoys. On a similar note, Peter Ungerargues at length that our intuitions about obligations to assist other peo-ple are improperly influenced by the conspicuousness of others’ suffering.He goes on to argue that our intuitions about various cases also reflectsuch implausible rules as the following:

First, when serious loss will result, it’s harder to justify moving aperson to, or into, an object than it is to move the object to, or into,

9 Some philosophers argue that one’s understanding of the natures of abstract objects —for example, the nature of knowledge, or of value —must lead one to have generally reliableintuitions about the properties of and relationships among these abstract objects. See Bealer,“A Theory of the A Priori,” and my Ethical Intuitionism, 122–27, for discussion.

10 Peter Singer, Practical Ethics, 2d ed. (Cambridge: Cambridge University Press, 1993), 232.

372 MICHAEL HUEMER

the person. Second, when serious loss will result, it’s harder to justifychanging the speed of a moving object, or changing its rate of motion,than changing the object’s direction of motion. Third, when there’ll bebig loss, it’s harder to justify speeding up an object than slowingdown an object. Fourth, it’s a lot harder to justify taking an object atrest and setting it in motion than to justify taking an object in motionand increasing its speed. . . . [Fifth,] it’s harder to justify imposing asubstantial force on an object than it is to justify allowing a forcealready present ( just about) everywhere, like gravitation, to work onthe object.11

Unger uses a complex series of hypothetical cases to argue that our intu-itive reactions are best explained by the above sorts of rules. He findsthese rules silly, and accordingly finds many of our unreflective moralreactions seriously flawed.

An intuitionist might observe that both Singer and Unger themselvesrely on intuitions, even in their criticisms of other intuitions. Singer countson the intuition that physical proximity is morally irrelevant, just as Ungercounts on the intuition that the factors invoked in the five rules he describesabove are morally irrelevant. Thus, their arguments do not impugn thereliability of intuition in general; Singer and Unger simply want to arguefor preferring certain intuitions over others.12

However, a moral skeptic might well pose an argument less friendly tonormative ethics than those of either Singer or Unger. Many of our moralintuitions clash with one another. For instance, most people share thefollowing intuitions:

(a) Refusing to save the child in Singer’s Shallow Pond example isseriously wrong.13

(b) Refusing to donate to famine relief is not seriously wrong.(c) The conspicuousness of someone’s suffering is not morally rele-

vant to the obligation to assist them.

These intuitive claims stand in tension once we accept Unger’s argumentthat conspicuousness explains the difference in our intuitions about the

11 Peter Unger, Living High and Letting Die: Our Illusion of Innocence (New York: OxfordUniversity Press, 1996), 101–2. Unger does not, however, argue for a general rejection ofintuition. Rather, he believes that through philosophical reasoning, we can correct the dis-torting influences on our intuitions.

12 Although Singer does not describe the moral premises he favors as “ethical intuitions,”I think that is in fact what they are. See the discussion in my essay “Singer’s UnstableMetaethics,” in Jeffrey Schaler, ed., Singer Under Fire (Chicago, IL: Open Court, forthcoming).

13 In this example, you see a child in danger of drowning in a shallow pond. Even thoughit will mean getting your clothes muddy and missing the lecture you were on your way togive, you clearly have a strong obligation to pull the child out of the pond. Singer thinks thisis comparable to your obligation to donate money to save Third World children frommalnutrition and disease (Singer, Practical Ethics, 229–46).


Shallow Pond and about famine relief. Singer and Unger would have usembrace (a) and (c) while rejecting (b). A moral skeptic, however, wouldhave us withhold judgment. The skeptic might argue, to begin with, thatSinger and Unger’s position is not obviously the most intuitive resolutionof the puzzle. On the Singer-Unger analysis of the cases, it is not just thatwe ought to donate some money to famine relief; rather, the failure todonate almost all of one’s wealth and income to charity organizations ismorally comparable to murder.14 This view is not obviously much moreplausible than either the denial of (a) or the denial of (c). Therefore, weshould withhold judgment on what the proper resolution of the paradoxis, if there is any proper resolution. Furthermore, the skeptic might arguethat conflicts and tensions of this sort among our ethical intuitions are socommon that we should infer that ethical intuition is an unreliable sourceof information, or at least that we are not justified in treating it as reliable.If this is true, then we should withhold judgment on ethical matters, evenin cases in which our intuitions do not conflict with one another.

B. Cultural indoctrination

A second argument for the unreliability of ethical intuition (on theassumption that objective moral truths exist) points to the influence of cul-ture on individuals’ ethical intuitions, together with the wide variability ofcultures.15 In contemporary Western society, almost everyone considers suchpractices as polygamy, infanticide, and slavery to be clearly wrong. Yet eachof these has been accepted in many cultures throughout world history.16

Similarly, some other cultures have, without any sense of wrongness, prac-ticed human sacrifice or cannibalism. While these facts about accepted prac-tices do not directly prove anything about what intuitions people have, itseems very likely that members of those cultures practicing polygamy,infanticide, or slavery would have very different intuitions about those prac-tices from ours. Most members of contemporary Western societies have an

14 The reasoning is roughly as follows: Imagine that you have already donated most ofyour income to charity organizations working to relieve world poverty. Now, while walkingpast a shallow pond on your way to a philosophy lecture, you see a small child drowning.You could pull the child out, but doing so would get your clothes muddy and make youmiss your lecture. Even so, it would be seriously wrong not to pull the child out. But, asSinger and Unger would argue, failure to save the drowning child in this circumstance ismorally comparable to failure to save another starving child by sending money to UNICEF.Thus, even when you have already given away most of your money, you are still obligatedto give more (Unger, Living High and Letting Die, 60–61, 135–39). Furthermore, since Singerrejects the moral significance of the distinction between killing and letting die, he holds thatallowing people in the Third World to die is comparable to murder (Singer, Practical Ethics,222–29).

15 See John L. Mackie, Ethics: Inventing Right and Wrong (New York: Penguin, 1977), 36–38.16 J. Patrick Gray reports a total of 1,045 societies practicing at least occasional polygamy,

compared with 186 exclusively monogamous societies; see Gray, “Ethnographic Atlas Code-book,” World Cultures 10 (1998): 90.

374 MICHAEL HUEMER

intuitive, negative evaluation of those practices, and it is hard to believethat the members of other societies would persist in those practices if theytoo had this intuitive reaction. It seems clear, then, that one’s culture hasa strong influence on one’s ethical intuitions.

For a cultural relativist or subjectivist, this poses no problem; it justshows that the moral truth varies from one society to another. For a moralrealist, however, a serious epistemological problem arises. The variationin moral intuitions across cultures strongly suggests that many or mostintuitions —including, of course, one’s own —are explained more by his-torical accident than by objective ethical truths, even if such truths exist.It seems that one should distrust many of one’s own intuitions, unless onecan somehow argue that one’s own culture is special, having somehowdeveloped a rapport with the moral truth that other cultures did not. Oneneed not be a card-carrying multiculturalist to have doubts about theprospects for such an argument.

One might argue that modern Western culture has one very importantadvantage over most other cultures: a long and sophisticated tradition ofrational moral philosophy. Modern Western values, it might be said, arelargely a product of rational reflection by philosophers over a period ofcenturies, and therefore represent progress relative to value systems thatheld sway earlier in our history, as well as the value systems of societieswithout such a tradition of carefully reasoned moral philosophy.

The moral skeptic might respond that the only reason for taking theWestern tradition of moral philosophy to give us access to moral truthrests on faith in the efficacy of moral reasoning, but that moral reasoningcan do little in the way of securing moral truth if the starting premises ofour reasoning —namely, our ethical intuitions —are compromised. SharonStreet articulates the skeptical argument as follows:

[W]hat rational reflection about evaluative matters involves, inescap-ably, is assessing some evaluative judgements in terms of others. . . .Thus, if the fund of evaluative judgements with which human reflec-tion began was thoroughly contaminated with illegitimate influence. . . then the tools of rational reflection were equally contaminated,for the latter are always just a subset of the former. It follows that allour reflection over the ages has really just been a process of assessingevaluative judgements that are mostly off the mark in terms of othersthat are mostly off the mark. And reflection of this kind isn’t going toget one any closer to evaluative truth, any more than sorting throughcontaminated materials with contaminated tools is going to get onecloser to purity.17

17 Sharon Street makes these remarks regarding the biases supposedly generated bynatural selection; see Street, “A Darwinian Dilemma for Realist Theories of Value,” Philo-sophical Studies 127 (2006): 124. Her ultimate conclusion favors some form of subjectivism.


C. Biological programming

A third problem afflicting our moral intuitions is that of biases pro-duced by evolutionary forces. Given the importance of ethical intuitionsto human behavior, and presumably to our reproductive fitness, it isreasonable to speculate that natural selection might have favored predis-positions toward certain kinds of intuitions over others. When we look atthe content of most people’s moral intuitions, many of them indeed looksuspiciously like the sorts of intuitions natural selection might produce,and rather unlike products of impartial, intellectual reflection. Most peo-ple would probably find each of the following evaluative claims plausible:

(a) Individuals have much stronger obligations toward their ownkin, especially their children, than toward others.

(b) Loyalty and devotion to one’s own society, as above other soci-eties, is a virtue. (In modern times, this virtue takes the form ofpatriotism.)

(c) Human beings are far more important than any other species,and human interests count, morally, for vastly more, if not infi-nitely more, than animal interests.

(d) Incest is inherently wrong, even in cases of mutual, informedconsent.

(e) Sexual promiscuity is desirable for a male, but highly undesirablefor a female.

It is difficult to explain why the above evaluations should be true, but itis easy to see why natural selection would have rewarded those whobelieved them.18

The fact that evolutionary pressures explain our having some cognitiveor perceptual faculty does not in general undermine trust in that faculty.Evolution explains why we have eyes and ears, why we have the capacityfor reasoning about the physical world, and why we have the capacity tolearn languages. In none of these cases does an evolutionary account ofthe origin of our faculty undermine trust in that faculty: no one arguesthat our vision, reasoning, or language apprehension is unreliable becauseit is a product of natural selection. So why should ethics be any different?The answer is that in the case of nonmoral faculties, there is a reason whyaccuracy should be selected for. Accurate nonevaluative beliefs are usu-ally useful for attaining one’s goals; therefore, if a conscious organism is

18 On (d), see Michael Ruse, Taking Darwin Seriously (Amherst, NY: Prometheus, 1998),145–47. On (e), see Richard Dawkins, The Selfish Gene (Oxford: Oxford University Press,1999), chap. 9. Jonathan Haidt has studied attitudes toward harmless, consensual incest, aswell as other victimless alleged wrongs; see Haidt, “The Emotional Dog and Its RationalTail: A Social Intuitionist Approach to Moral Judgment,” Psychological Review 108 (2001):814–34.

376 MICHAEL HUEMER

generally well-adapted, so that its desires are generally in line with whatwould promote its own reproductive fitness, then, as a rule, more accu-rate factual beliefs will increase its fitness. For instance, if one wants toavoid predators, then correct beliefs about where predators are locatedwill benefit one.

However, this reasoning cannot be extended to values, even if a realmof objective values exists. If an organism has generally accurate noneval-uative beliefs, it cannot be assumed that, in general, its also having objec-tively correct values will increase its reproductive fitness. Rather, anorganism’s reproductive fitness would seem to be best promoted by itshaving values skewed in a certain direction: by the organism’s taking itsown reproductive success, or things normally correlated with one’s ownreproductive success, to be good, whether or not those things are objec-tively good. The crucial asymmetry is that nonevaluative beliefs typicallyfunction to help us select the correct means of achieving our goals, whereasevaluative beliefs typically influence what goals we seek. Natural selec-tion could be expected to favor individuals whose goals are in line withthe “goals” of evolution, and who take the correct means of achievingtheir goals.

For this reason, it would seem that if the values toward which naturalselection biased us coincided with the objectively correct values, thiswould be sheer coincidence. Such a coincidence cannot reasonably beexpected. We should assume, therefore, that to the extent that biologyinfluences our ethical intuitions, it leads us away from the objective truth,if objective truth exists in this area.19

D. Personal biases

Finally, as Walter Sinnott-Armstrong argues, our moral intuitions maybe biased by emotions and personal interests.20 It is easy to see how ourinterests are affected by what evaluative propositions we and others accept.On the one hand, if I reject the moral significance of the distinction betweenkilling and letting die, I may feel obliged to donate large amounts ofmoney to the poor; if I accept that the interests of animals matter morally,I may feel obliged to give up some of my favorite foods. On the otherhand, I stand to benefit from others’ acceptance of the obligation to assistothers (particularly if I am poor) and of the obligation to refrain fromharming others. These facts may bias me toward certain evaluative con-clusions and away from others.

Even when our personal interests are not at stake, moral issues canarouse strong emotions. Even those who never expect to be eligible for

19 Ruse, Taking Darwin Seriously, 252–54; Street, “A Darwinian Dilemma.”20 Walter Sinnott-Armstrong, Moral Skepticisms (Oxford: Oxford University Press, 2006),

195–204.


capital punishment, and never expect to know anyone who is, may evincepassionate opinions about the justice of capital punishment. Likewise,both pro- and anti-abortion forces evince strong emotions when it comesto the morality of abortion. Though these are extreme cases, emotionalreactions to moral issues are common. Emotions are known to impairjudgment with respect to (other) factual questions, so, assuming the truthof moral realism, it is prima facie reasonable to assume that emotionsimpair our moral judgment as well.

In a criminal trial, friends or family members of the accused wouldnever be allowed to serve on the jury, nor would the defendant’s enemies.We cannot be sure that their judgments would be compromised by theirfeelings about the accused, but the danger of this is great enough that itwould be foolish to trust their judgments. Similarly, people who, forwhatever reason, have strong feelings about the defendant’s alleged crimewould not be allowed to serve. Thus, a rape victim would be excusedfrom jury service for a rape trial. Even though, in this case, the would-bejuror’s interests would not literally be at stake, her emotions could beexpected to compromise her objectivity. Again, we do not know that thiswould occur, but the danger of it is strong enough to render us unjustifiedin relying on the judgment that such a person would render. In the sameway, the moral skeptic argues, when we are emotional about moral issuesor when our own interests are at stake, the danger that our intuitions willbe biased is too great for us to be justified in relying on those intuitions,in the absence of independent corroboration.

Taken together, the arguments of Sections III.A through III.D suggestthat a wide range of ethical intuitions are subject to distorting factors andshould not be relied upon. Unless we can somehow separate the contam-inated intuitions from the trustworthy ones, we will have no way ofconstructing a rational moral system, and moral skepticism will prevail.

IV. How to Build Castles on Sand: A Reply to the Skeptics

One possible reaction to the arguments of Section III is to give up onnormative ethics, despairing of ever identifying sufficiently reliable intu-itions on which to base an ethical system. Another reaction is to resolvesimply to apply the traditional reflective methods of ethics very carefully,in the hope that they will weed out the worst biases and distortions in ourethical judgments. The latter reaction strikes me as overly complacent,while the former is overly defeatist. The arguments of Section III mayshow that we have a good deal less moral knowledge than is commonlysupposed, but I see little plausibility in the suggestion that they show thatno one knows whether Ted Bundy’s murders were wrong.21 An inter-

21 Ted Bundy killed dozens of young women across the United States during the 1970s,becoming one of history’s most notorious serial murderers.

378 MICHAEL HUEMER

mediate reaction is called for: the arguments of Section III motivate a shiftin the methods of ethics, a shift that leaves us with some moral beliefs, butperhaps with a very different set from those that more traditional meth-ods lead to.

Consider again Sharon Street’s objection to the possibility of overcom-ing ethical biases through rational reflection:

[I]f the fund of evaluative judgements with which human reflectionbegan was thoroughly contaminated with illegitimate influence . . .all our reflection over the ages has really just been a process ofassessing evaluative judgements that are mostly off the mark in termsof others that are mostly off the mark. And reflection of this kind isn’tgoing to get one any closer to evaluative truth. . . .

This objection has force on a simple model of ethical reflection: Supposewe consider a single ethical intuition whose veracity is in question, andsuppose (i) that our only method of evaluating it is to compare it with asingle other ethical intuition, and (ii) that each other ethical intuition isequally suspect as the original intuition. In that case, it seems clear that noethical progress will be made.

That simple model, however, is inadequate in at least two importantways. First, it fails to take account of the probative value of coherence,understood, as in the coherence theory of justification, in terms of mutualsupport relations and explanatory relations among a large set of believedpropositions.22 If suspect intuitions can be assessed in terms of a body ofother ethical judgments, then —even if most of the other ethical judgments arefalse —it may be possible to make a reliable assessment. To see how this ispossible, consider a simple analogy. Suppose a detective interviews sixeyewitnesses to a robbery. All of them claim to have seen the robbersdrive away in their getaway car. The detective interviews the witnessesseparately, giving them no opportunity to confer with each other. Nowsuppose that two of the six witnesses agree that the getaway car hadlicense plate number X78 41A, while the other four witnesses report fourdifferent license plate numbers. In this case, even though most of thewitnesses are wrong, the detective could still conclude that the correctlicense plate number was probably X78 41A. The reason is, in essence, thatit is extremely unlikely that even two witnesses would agree on a specificlicense plate number, in the absence of collusion, unless that number were

22 According to the coherence theory of justification, a belief is justified, if at all, by virtueof the way it fits together with the rest of one’s belief system. This “fitting together” isusually understood as a matter of supporting and being supported by other beliefs, beingexplainable in terms of one’s other beliefs, and so on. See Laurence BonJour, The Structureof Empirical Knowledge (Cambridge, MA: Harvard University Press, 1985), 93–101.


correct.23 In this example, the witnesses are less than 50 percent reliable —two-thirds of the witnesses report incorrect license plate numbers. Nev-ertheless, by relying on coherence —trusting the two witnesses whoseanswers agree with each other —one can attain a conclusion that is muchmore than 50 percent likely to be correct.

Analogously, suppose that only a third of our ethical intuitions wereaccurate, the others being skewed in various directions by various factors.We might nevertheless be able to identify the correct intuitions, since thecorrect intuitions would tend to cohere with each other, while the othertwo-thirds of our intuitions would generally fail to cohere either with thecorrect intuitions or with each other. If we found that the largest coherentsubset of our intuitions comprised one-third of our intuitions, while therewas no other coherent subset anywhere near as large, then we would beprima facie justified in regarding that largest coherent subset as roughlyaccurate. The point here is not that such a coherent set of intuitions wouldbe guaranteed to be true or close to the truth. Rather, the point is that, paceSharon Street, even if our moral intuitions are unreliable taken singly, itdoes not follow that ethical reflection cannot produce conclusions that arehighly likely to be true.

Second, and more importantly, it is a mistake to suppose that an ethicalintuition can be criticized only by appeal to other ethical intuitions. Intu-itionists who accept the is-ought gap (myself included) will grant that arebutting defeater for an ethical intuition must derive from other ethicalintuitions; however, an undercutting defeater for an ethical intuition mayderive from nonevaluative premises.24 A rebutting defeater for an ethicalintuition is something that provides prima facie justification for denyingthe content of that intuition; that is, it justifies a contrary evaluative claim.Since intuitionists hold that all justification for evaluative claims derivesultimately from ethical intuition, a rebutting defeater for an ethical intu-ition must derive from other ethical intuitions. An undercutting defeaterfor an ethical intuition, however, need not provide justification for anyevaluative claim; rather, an undercutting defeater simply constitutes

23 The probability of as many as two of the six witnesses picking the same incorrectlicense plate number by chance (assuming random selection from six-digit alphanumericsequences) is about one in 145 million.

This kind of argument is advanced by BonJour, The Structure of Empirical Knowledge,147–48, and C. I. Lewis, An Analysis of Knowledge and Valuation (La Salle, IL: Open Court,1946), 346. See Erik Olsson, Against Coherence (Oxford: Clarendon, 2005), for an extendeddiscussion of the conditions under which coherence produces confirmation. Note that theargument requires that the witnesses be more reliable than chance, but not that they be morethan 50 percent reliable. Note also that I do not hereby embrace a coherence theory ofjustification, since I do not claim that coherence is either necessary or sufficient for justifi-cation; I claim only that coherence can ratchet up the level of justification that intuitions startwith.

24 See John Pollock and Joseph Cruz, Contemporary Theories of Knowledge, 2d ed. (Lanham,MD: Rowman and Littlefield, 1999), 195–96, for the distinction between rebutting and under-cutting defeaters.

380 MICHAEL HUEMER

grounds for doubting the reliability of the intuition in question. For exam-ple, suppose I have the intuition that incest is wrong. Suppose I thenacquire good evidence that my intuition is a product of cultural or bio-logical programming. This may function as an undercutting defeater,since it gives me a reason not to rely on my intuition that incest is wrong;however, it does not provide a rebutting defeater, since it gives me noevidence that incest is permissible.

The arguments of Section III provide prime examples of undercuttingdefeaters for some ethical beliefs. If the biases discussed there affected allintuitions equally, and produced distortions in random directions, thenthe biases would be difficult to correct for. But neither of those things isthe case. In the case of each of the factors said to distort our ethicalintuitions, intuitions would be expected to be skewed in a specific direc-tion. As a corollary, not all intuitions are equally open to the accusation ofbias. For instance, biological evolution would be expected to produce abias toward favorable evaluations of things that promote one’s own inclu-sive fitness; intuitions that do not imply favorable evaluations of thingsthat promote one’s own inclusive fitness are not candidates for beingproducts of this particular bias. Similarly, cultural conditioning would beexpected to produce a bias toward favorable evaluation of the practices ofone’s own culture; self-interest would produce a bias in favor of positiveevaluations of oneself, one’s own practices, or things that benefit oneself;and emotions would produce biases in favor of positive evaluations ofthings that give one positive emotions, and negative evaluations of thingsthat give one negative emotions. Ethical intuitions that do not conform tothe relevant expectations are not open to the charge of being produced bythese biases.

When we put together the foregoing considerations, a cautious andcritical intuitionist methodology emerges. While our ethical theory mustbe based on intuitions, we should not accept or reject an ethical principlesolely on the basis of an uncritical appeal to a single intuition, nor shouldwe assess an intuition’s probative value solely in terms of its subjectivestrength. Rather, we should screen intuitions according to the followingcriteria:

(1) Seek a substantial body of ethical intuitions that fit together well,rather than placing great weight on any single intuition.

(2) Eschew intuitions that are not widely shared, that are specific toone’s own culture, or that entail positive evaluations of the prac-tices of one’s own society and negative evaluations of the prac-tices of other societies.

(3) Distrust intuitions that favor specific forms of behavior that wouldtend to promote reproductive fitness, particularly if these intu-itions fail to cohere with other intuitions.


(4) Distrust intuitions that differentially favor oneself, that is, thatspecially benefit or positively evaluate oneself, as opposed toothers.

(5) Distrust intuitions that line up with strong emotions one hasabout the things evaluated.

If these precepts are followed in the construction of an ethical system, theresulting system will be exempt from the skeptical challenges of Section III.

To some degree, these methodological precepts are already being imple-mented. A prime example is Peter Singer’s qualified defense of infanti-cide. After arguing that abortion is permissible because fetuses lack aright to life, Singer confronts the charge that his arguments lead to thepermissibility of infanticide, since newborn infants would also lack aright to life on Singer’s criteria. What is of interest here is not whetherSinger’s initial arguments regarding fetuses’ alleged right to life succeed,but how he responds to the common intuition that infanticide is obvi-ously unacceptable. Singer mentions three sources of bias with regard tothe question of infanticide: First, “there are no doubt very good evolu-tionary reasons why we should instinctively feel protective” toward humanbabies. Second, we have emotions “based on the small, helpless, and —sometimes —cute appearance of human infants.” Third, Singer observesthat other societies and other moral philosophers —including such refinedthinkers as Plato, Aristotle, and Seneca —endorsed the killing of deformedinfants. The suggestion here seems to be that our current horror at thethought of infanticide is a product of our particular culture.25 For thesethree reasons, the intuition that infanticide is wrong is an especially strongcandidate for being a product of bias. If an otherwise acceptable moraltheory conflicts with that intuition, this should not be taken as a strongreason for rejecting the theory. I take it that this is essentially Singer’spoint.26

Nevertheless, Singer’s way of reasoning in this passage is not typical ofthe tradition of Western philosophical ethics. Much more common arerelatively uncritical appeals to intuition. When ethical intuitions conflict,alternative resolutions are typically assessed purely in terms of the num-ber and strength of the intuitions each resolution can accommodate. Thelessons of the skeptics regarding the unreliability of certain kinds of intu-itions are rarely heeded, and were almost never heeded prior to thetwentieth century. To this extent, revisionary intuitionism represents anew approach to normative ethics.

25 Singer, Practical Ethics, 170–73.26 Presumably, Singer would not back this point up, as I would, with a realist conception

of ethics, since his sympathies lie more in line with noncognitivism. I discuss the resultingtension between his metaethical views and his ethical methodology in my “Singer’s Unsta-ble Metaethics.”

382 MICHAEL HUEMER

V. Abstract, Concrete, and Formal Intuitions

We have just seen that the intuitionist’s most natural response to theskeptical challenges of Section III is to attempt to distinguish those intu-itions that are most likely to be reliable from those that are less likely tobe reliable, and to base her ethical theory on intuitions of the former kind.This casts new light on a long-standing dispute among intuitionists con-cerning which kind of intuitions should be given most weight or shouldplay the greatest role in the forming of our ethical beliefs. Each of thefollowing types of intuitions has had its defenders:27

(1) Concrete intuitions: These are intuitions about specific situations,such as the intuition that in Singer’s Shallow Pond example,one is obligated to rescue the drowning child, or that in theTrolley Car problem, one should turn the trolley away from thefive bystanders toward the one.28

(2) Abstract theoretical intuitions: These are intuitions about very gen-eral principles, such as the intuition that the right action is alwaysthe action that has the best overall consequences, or that it iswrong to treat individuals as mere means.

(3) Mid-level intuitions: These are intuitions about principles of anintermediate degree of generality, such as the principle that, otherthings being equal, one ought to keep one’s promises; that oneought to show gratitude for favors done to one; or that it is moreimportant to avoid harming others than it is to positively helpothers.

Initially, it is unclear why preference should be given to any of these typesof intuitions over the others. What does level of generality have to dowith how likely an intuition is to be correct?29 As I shall argue in thissection, level of generality matters because intuitions of different levels ofgenerality differ in their susceptibility to various kinds of error. When wetake this into account, the effort to accommodate the challenges of theskeptics naturally leads one, for the most part, to prefer certain kinds ofabstract theoretical intuitions over concrete and mid-level intuitions.

Concrete and mid-level intuitions are particularly susceptible to thekinds of biases discussed in Section III. One reason for this is that wetypically have stronger emotions about concrete cases and mid-level

27 See, respectively, Jonathan Dancy, Moral Reasons (Oxford: Blackwell, 1993); Henry Sidg-wick, The Methods of Ethics, 7th ed. (Indianapolis, IN: Hackett, 1981); and Ross, The Right andthe Good.

28 In this example, one must choose between allowing a runaway trolley to run over andkill five people, and flipping a switch to send the trolley onto another track, where it will runover and kill one person. See Philippa Foot, “The Problem of Abortion and the Doctrine ofthe Double Effect,” Oxford Review 5 (1967): 5–15.

29 DePaul expresses doubt on this score in “The Problem of the Criterion and CoherenceMethods in Ethics.”


generalizations than about very abstract principles. Compare the emo-tional impact of the statement “Killing deformed human infants is accept-able” to that of the statement “A being has a right to x only if that beingis capable of desiring x.” 30 The latter, abstract principle is much lesssusceptible to emotionally-based bias. In addition, concrete intuitions aremore likely to be influenced by biological programming, because thebiases with which evolution is most likely to have endowed us are biasesfavoring relatively specific forms of behavior that would have promotedour ancestors’ inclusive fitness. Biological evolution is unlikely to haveendowed us with biases toward embracing very abstract principles, sinceour biological ancestors probably engaged in little abstract reasoning. Forinstance, attitudes toward incest, human offspring, and social hierarchiesare more likely to be influenced by biology than are intuitions aboutprinciples of additivity in axiology.31 Finally, culturally generated biasesare more likely to affect concrete and mid-level judgments than highlygeneral ethical judgments, because our culture has a complex set of rel-atively specific rules —rules governing who is allowed to marry whom,how one should greet a stranger, how one should interact with one’s boss,and so on. What rules, if any, our society accepts on the most abstractlevel is extremely unclear. Does our culture endorse the Categorical Imper-ative? What general criterion of rights does it endorse? The obscurity ofthe answers to these questions prevents cultural conditioning from directlydetermining our intuitions about the Categorical Imperative or the gen-eral criterion of rights. In contrast, it is perfectly clear how our culturemight bias judgments about, for example, the acceptability of polygamy.

Abstract theoretical intuitions, in contrast, are prone to the simple butwidespread problem of overgeneralization. This is the tendency to judgethe truth of a generalization in terms of typical cases, or the sorts of casesthat are easy to think of. Confronted with the generalization that all A’sare B, one will have a tendency to judge the generalization true if alltypical A’s are B, even if some A’s of a sort that do not readily come tomind are not B.32 For example, the following generalization seems ini-tially plausible:

(C) For any events x and y, if x was the cause of y, then if x had notoccurred, y would not have occurred.

30 Michael Tooley employs the latter principle in his defense of abortion and infanticide;see Tooley, “Abortion and Infanticide,” Philosophy and Public Affairs 2 (1972): 37–65.

31 Axiological additivity principles claim that value can be added along some dimension —for example, that the value of a pair of people’s lives is equal to the value of the first person’slife plus the value of the second person’s life; or that the value of some event is equal to thevalue of the first half of the event plus the value of the second half of the event.

32 This is a special case of the “availability heuristic,” discussed in Amos Tversky andDaniel Kahneman, “Availability: A Heuristic for Judging Frequency and Probability,” inDaniel Kahneman, Paul Slovic, and Amos Tversky, eds., Judgment under Uncertainty: Heu-ristics and Biases (Cambridge: Cambridge University Press, 1982), 163–78.

384 MICHAEL HUEMER

But now consider the following case:

The Preemption Case: Two mob assassins, Lefty and Righty, have beenhired to assassinate FBI informant Stoolie. As it happens, both ofthem get Stoolie in their sights at about the same time, and both firetheir rifles. Either shot would be sufficient to kill Stoolie. Lefty’sbullet, however, reaches Stoolie first; consequently, Lefty’s shot is theone that actually causes Stoolie’s death. However, if Lefty had notfired, Stoolie would still have died, because Righty’s bullet wouldhave killed him.

This shows that there can be a case in which x is the cause of y, but if xhad not occurred, y would still have occurred.33 Notice that, before weconsider the Preemption Case, claim (C) seems plausible; but after weconsider the Preemption Case, (C) no longer seems plausible. In fact, thehistory of analytic philosophy is littered with examples of generalizationsthat initially seem true, until one is confronted with recherché counter-examples. In these cases, the generalization loses its intuitive appeal assoon as the counterexample is discovered; its appeal depended upon ourhaving only certain typical kinds of cases in mind. Because of this, evenwhen no counterexample has yet been devised, most generalizations rightlyoccasion a lingering suspicion that a clear counterexample may be justwaiting to be discovered.

Mid-level generalizations appear to give us the worst of both worlds:they are sufficiently concrete to be susceptible to biases with an emo-tional, cultural, or biological source, while at the same time they aresufficiently general to be susceptible to overgeneralization. For instance,the belief that adultery is wrong is open to the suspicion of being partlya product of emotional, cultural, and/or biological bias. At the same time,it is a sufficiently general claim that one may evaluate it by thinking oftypical cases, perhaps overlooking some atypical cases of adultery. Thelatter problem, that of possible overgeneralization, may be remedied byadding a qualifier to the principle, resulting in a claim such as “Adulteryis prima facie wrong” or “Adultery is wrong in typical cases.” This does,however, have the disadvantage of rendering the principle less useful,since the principle does not tell us in which atypical cases, if any, adulteryis not wrong all things considered.

All three types of intuitions, then, have their own problems. This doesnot mean that no intuitions can be relied upon. What it means is that wemust consider more than an intuition’s level of generality. As I indicatedin Section IV, we must consider an intuition’s content to determine whetherit is a plausible candidate for being a product of one of the common typesof biases discussed in Section III. In addition to this, however, there is a

33 David Lewis discusses this kind of case in “Causation,” in Lewis, Philosophical Papers,vol. 2 (Oxford: Oxford University Press, 1986), 159–214.


particular species of abstract ethical intuitions that seems to me to beunusually trustworthy. These are what I call “formal intuitions” —intuitionsthat impose formal constraints on ethical theories, though they do notthemselves positively or negatively evaluate anything. The following areexamples of such formal ethical intuitions:

(1) If x is better than y and y is better than z, then x is better than z.(2) If x and y are qualitatively identical in nonevaluative respects,

then x and y are also morally indistinguishable.(3) If it is permissible to do x, and it is permissible to do y given that

one does x, then it is permissible to do both x and y.(4) If it is wrong to do x, and it is wrong to do y, then it is wrong to

do both x and y.(5) If two states of affairs, x and y, are so related that y can be

produced by adding something valuable to x, without creatinganything bad, lowering the value of anything in x, or removinganything of value from x, then y is better than x.

(6) The ethical status (whether permissible, wrong, obligatory, etc.)of choosing (x and y) over (x and z) is the same as that of choos-ing y over z, given the knowledge that x exists/occurs.

These kinds of intuitions are particularly plausible candidates for beingproducts of rational reflection. They are not plausibly regarded as prod-ucts of emotional bias, cultural or biological programming, or self-interested bias.

What of the threat of overgeneralization? It seems to me that these prin-ciples are not the result merely of considering some typical kinds of casesand then evaluating just those cases. Rather, we seem to be able to see whyeach of these things must be true in general; these principles seem to berequired by the nature of the “better than” relation, the nature of permis-sibility, the nature of ethical evaluation, etc. Accordingly, if someone wereto describe a proposed counterexample to one of these principles, our reac-tion would not be (as with the Preemption Case discussed above) to sim-ply give up the principle in question without protest; rather, our reactionwould probably be to call the case a “paradox.” For example, consider thefollowing series of cases, derived from an essay by Stuart Rachels:34

Case 1: You have a year of the most intense ecstasy imaginable.Case 2: You have two years of slightly less intense pleasure.Case 3: You have four years of pleasure slightly less intense than in

Case 2. . . .

34 Stuart Rachels, “Counterexamples to the Transitivity of Better Than,” Australasian Jour-nal of Philosophy 76 (1998): 71–83. Rachels takes his series of cases to provide a counter-example to the transitivity principle. I take it, instead, to illustrate certain biases in ourevaluations of the cases.

386 MICHAEL HUEMER

If we continue in this way for many more stages, eventually we arrive at:

Case N: You have millions and millions of years of barely noticeablepleasure.

In each succeeding case, you get a slightly less intense pleasure than theprevious case, but it lasts for twice as long. Most people have the intuitionthat each case in the series is better than the previous one; for example, itis better to have two years of great pleasure, than it is to have only oneyear of slightly greater pleasure. However, most also have the intuitionthat the final case, in which one gets millions of years of barely noticeablepleasure, is not better than the first case. This seems paradoxical. In con-trast, however, the Preemption Case discussed earlier does not strikeanyone as “paradoxical”; rather, we simply accept it straightaway as acounterexample to claim (C), that if x caused y, then if x hadn’t occurred,y wouldn’t have occurred. I think this difference is explained by the factthat, whereas the intuition supporting claim (C) is a result of merelyconsidering some typical cases of causation, the intuition supporting theprinciple of the transitivity of “better than” is a result of an insight intothe nature of the “better than” relation as such.

It seems to me, then, that formal ethical intuitions should be givenspecial weight in moral reasoning. These formal intuitions are not suffi-cient to generate any substantive ethical system. Nevertheless, they ruleout some otherwise attractive (combinations of ) ethical views and are forthat reason useful in resolving some ethical disputes.35

VI. How Ethical Revision Is Possible

The dominant antirealist approaches in contemporary metaethics can-not support a revisionary ethical theory. Moral nihilism, of course, sup-ports a sort of ethical revisionism, but its revisionism is too total: it rejectsall first-order ethical theories.36 Cultural relativism, in contrast, requiresone to accept the norms of one’s own society. At most a small amount ofethical revision is possible, in the event that some of the social norms areinconsistent, but the goal of ethical theorizing must nevertheless be tofind the consistent ethical system that is closest to the prevailing social

35 See my “Non-Egalitarianism,” Philosophical Studies 114 (2003): 147–71, for an argumentagainst welfare egalitarianism based mainly on formal intuitions. See Derek Parfit, Reasonsand Persons (Oxford: Clarendon, 1987), 419–30, for an argument (though Parfit does notendorse the argument) based mainly on formal intuitions for “the repugnant conclusion”that, for any world of very happy people, some world with a much larger population ofpeople with lives barely worth living would be better. See also Unger, Living High and LettingDie, 88–94, for an argument that our intuitions about sacrificing individuals to producegreater benefit violate the principle of independence of irrelevant alternatives.

36 Mackie, Ethics, 30–35.


norms.37 Subjectivism and noncognitivism would both have one acceptthe consistent ethical system that is closest to what one’s own attitudes,desires, and/or feelings would support. While this might lead to an uncon-ventional morality in the case of atypical individuals, it cannot lead to asubstantial revision of one’s own moral attitudes, and for most people itwill result in an endorsement of something very close to conventionalmorality.

What these antirealist approaches lack is the notion that some of our (orour society’s) ethical attitudes may be biases —illegitimate influences onour ethical thinking that lead us away from the moral truth —since, forantirealists, there is no moral truth independent of our attitudes. It is thisnotion of bias that allows intuitionists to select out different classes ofintuitions for different treatment: the intuitions that are most subject tobias are to be discounted, while those that are most clearly products ofintellectual reflection are to be preserved as a basis for ethical theory.Depending on how prevalent ethical biases are, this can lead to a signif-icant revision of common-sense morality.

What areas are most ripe for ethical revision? The area of sexual moral-ity is probably the clearest case, since it is an area in which common moralattitudes exhibit multiple signs of unreliability. First, people tend to havestrong feelings about such things as homosexuality, bestiality, or polyg-amy. Second, on many issues of sexual conduct, we tend to have attitudesthat conform to parochial cultural mores. For instance, attitudes towardprostitution, homosexuality, and polygamy in contemporary America aremuch harsher than those prevalent, respectively, in modern Japan, ancientGreece, or Imperial China —suggesting that most Americans’ attitudestoward these practices are determined more by the time and place inwhich they happened to be born than by any objective ethical truths.Third, many of our attitudes in this area that are not culturally specific aresubject to sociobiological explanations. Our feelings about incest and aboutfemale sexual promiscuity are cases in point. For these reasons, conven-tional sexual morality should probably be rejected more or less wholesale,excepting those aspects that are mere applications to sexual behavior ofgeneral principles of benevolence and respect for others. Traditional moralproscriptions against activities between consenting adults —includinghomosexuality, various forms of “unnatural” sexual activity, polygamy,and incest —are probably unjustifiable prejudices. In each case, of course,specific arguments put forward by the advocates of such proscriptionsmust be examined. The important point here is that one’s initial, intuitiveopposition to those arrangements or forms of sexual activity ought not tobe treated as serious evidence of their wrongness.

37 See Gilbert Harman, The Nature of Morality (New York: Oxford University Press, 1977),94–95. However, I argue in Ethical Intuitionism (176–79) that even demands for consistencyand coherence are problematic for ethical antirealists.

388 MICHAEL HUEMER

A more philosophically challenging area is that of the dispute betweenconsequentialist and deontological approaches to ethics. There is littledoubt that common-sense morality is deontological. More specifically,there are many cases in which we regard it as wrong to sacrifice anindividual for the greater benefit of others. Yet these ethical beliefs mayrequire revision in the light of ethical argument. Consider the followingrelatively weak deontological principle:

(WD) There is a way of harming people, which we may call “deonti-cally constrained harming,” such that it is wrong to harm aperson in that way, even to produce a slightly greater benefitfor others.

For example, suppose that I am driving two people to the hospital. Inorder to get them there in time to save their lives, I will have to run overand kill a single child who is playing in the street. (There is no time tomove the child.) Most people have the intuition that, even though doingso would result in a lesser loss overall, it is impermissible to run over thechild. It is this sort of intuition that motivates Weak Deontology.

WD is a modest deontological principle, compatible with a wide rangeof nonconsequentialist views. WD leaves open, for example, that theremay be more extreme cases in which running over the child is justified —for instance, perhaps I might run over the child if doing so were necessaryto get to a nuclear bomb set to blow up New York City in time to disarmit. In other words, though harming the child in this way to produce aslightly greater benefit (two lives saved) is impermissible, WD allows thatharming the child in this way to produce a vastly greater benefit may bepermissible. WD also leaves room for different ways of specifying what isthe relevant way of harming others. This “way of harming others” may bedefined in part by the kind of harm produced, by the circumstances inwhich it occurs, or even by the intentions of the agent.

To see why even this very weak form of deontology may require revi-sion, imagine a situation in which I have two actions available, A1 and A2,where I can perform either action singly, perform both actions, or performneither action. Each action will affect just two people, P1 and P2. ActionA1 will have the effect of harming P1 while producing a greater benefit forP2. Action A2 will harm P2 while producing a greater benefit for P1. If,however, I perform both actions, the net effect will be a benefit for bothP1 and P2:

Action Effect on P1’s Utility Effect on P2’s UtilityA1 −1 +2A2 +2 −1

(A1+A2) +1 +1


It seems that, in addition to what we have stipulated so far, it is possiblethat the harms produced by each of A1 and A2 might be of the deonticallyconstrained kind. If so, then WD implies that A1 is impermissible and A2is impermissible. Yet it does not seem that (A1+A2) —the “conjunctiveaction” of performing both A1 and A2 —is impermissible, for (A1+A2)benefits both P1 and P2 while harming no one.

There are a number of ways a Weak Deontologist might respond to thissituation. One might respond by denying the plausible formal constraintthat, if it is wrong to do x and it is wrong to do y, then it is wrong to doboth x and y. Or one might argue that it is wrong to do both A1 and A2 inthe situation described. Or one might try to delineate the class of deonti-cally constrained harms in such a way that the sort of scenario envisionedbecomes impossible (for instance, perhaps because A2 ceases to be a deonti-cally constrained harm in a situation in which A1 is also performed). Eachof these approaches merits further study, which must await another occa-sion.38 My purpose here is simply to illustrate in principle how a revisionof a major tenet of common-sense morality could be justified within anintuitionist framework. Suppose it turns out that the intuitions motivat-ing Weak Deontology support an account of deontically constrained harmsthat makes the sort of scenario I have described possible, so that each ofA1 and A2 would be judged wrong, independently of whether the otheris performed. Suppose also that this intuitively supported account impliesthat an action such as (A1+A2), if we recognize such conjunctive actions,would not impose any deontically constrained harm. Then a consequen-tialist intuitionist would be in a position to argue that our Weak Deon-tological intuitions should be rejected, since they lead to a conflict withplausible formal ethical constraints. One plausible formal constraint isthat if it is wrong to do x and it is wrong to do y given that one does x,then it must be wrong to do both x and y. Another plausible formalconstraint is that the moral acceptability of a form of behavior should notbe made to depend on one’s method of individuating actions (for instance,on whether some bit of behavior counts as one action or two).

Our concrete and mid-level deontological intuitions conflict with theabstract, theoretical intuitions of many philosophers, which are conse-quentialist in nature. In addition, it seems likely that consequentialistethical theories will have desirable formal properties that the main formsof deontology worth considering lack. If so, the best resolution may wellbe in favor of consequentialism, despite the fact that this entails a rejec-tion of many widespread and salient ethical intuitions. Whether this is, infact, the best resolution will depend on the results of future ethicalinvestigation —so far, I have merely supposed hypothetically that thingsturn out in the way I would consider favorable to consequentialism.

38 I discuss the issue in greater detail in my “A Paradox for Moderate Deontology”(unpublished manuscript).

390 MICHAEL HUEMER

VII. Conclusion

The question of methodology in ethics has hitherto received too littleattention. Nonintuitionists have made little effort to draw out the meth-odological implications of their metaethical theories, while intuitionistshave done little more than enjoin us to rely upon our ethical intuition.Though intuitionists have often expressed preferences for one sort ofintuition —abstract intuitions, concrete intuitions, or mid-level intuitions —over others, these have appeared as little more than personal preferences.Yet the field of ethics is surely in need of methodological scrutiny. Thereare few conclusions that can be held up as established results of ethics,and there are few nontrivial arguments in the field that are generallyaccepted as sound. Skeptics suggest that perhaps there is no ethical knowl-edge, that our moral beliefs are largely prejudices that cannot be reliedupon. Furthermore, even among nonskeptical philosophers, radically dif-ferent conceptions of the nature of the ethical enterprise abound. It wouldbe surprising if these fundamental differences entailed no significant dif-ferences in how one should approach first-order ethical questions. For allthese reasons, it seems especially important to attend to the question ofhow in general one should proceed in ethics.

Perhaps the most common approach to normative ethics is a kind ofnarrow reflective equilibrium: a variety of ethical intuitions or judgmentsare canvassed; in cases of conflict, they are weighed against one anotheron the basis of the strength of our initial inclination to accept them; andadjustments are made in an effort to produce a coherent ethical systemwith the least revision possible. Often, for example, an ethical theory iscriticized by appeal to cases in which the theory can be shown to conflictwith relatively strong moral intuitions; since our intuitions about thecases are stronger than our initial inclination to accept the moral theory,it is urged that we should reject the theory. This method may be expectedto lead to something like W. D. Ross’s pluralistic system of prima facieduties.

This method, as I have urged, is oversimplified. The key point in aproperly critical intuitionist methodology is that not all intuitions arecreated equal. Intuitions that are controversial or that may easily beexplained as products of bias have relatively little evidential value. Intu-itions that are widely shared and are not plausible candidates for beingmere prejudices have much greater evidential value. Among this latterclass, there are a number of formal ethical intuitions —which do not en-tail any specific evaluations but which place constraints on systems ofevaluations —that are particularly strong candidates for being products ofintellectual reflection rather than bias.

The preferential treatment I have advocated for certain kinds of intu-itions paves the way for a degree of ethical revisionism not otherwisepossible. How far such revision should go remains an open question.


Almost certainly, for example, the correct view of sexual morality is highlyrevisionary. It is less clear whether the correct views on such matters asjustice and individual rights are similarly revisionary. Nevertheless, thecritical intuitionism I have advocated probably holds out the best hopefor advocates of revisionary ethical theories, such as utilitarianism orethical egoism.

Philosophy, University of Colorado at Boulder

392 MICHAEL HUEMER

MORAL OBJECTIVITY

By Nicholas Rescher

I. Introduction

The aim of this essay is to set out an argument for moral objectivity. Abrief sketch of the considerations at issue should help make it possible tokeep sight of the forest amid the profusion of trees. Overall, then, the lineof thought that is being set out here runs as follows:

• To validate moral objectivity, it must be shown that an impersonalmatter of fact (rather than a personal opinion or feeling) is at issue.

• A key step in this direction emerges from the consideration thatmorality is a functional enterprise whose aim is to channel peo-ple’s actions toward realizing the best interests of everyone.

• This makes morality into something quite different from meremores geared toward communal uniformity and predictability.(After all, morality is not a matter of anthropology; it addresseswhat people should do rather than what they actually do.)

• The inherent generality of moral principles means that they oper-ate at a level of universality that transcends the limits of societalvariation.

• This circumstance militates decisively against moral relativism.• Nevertheless, general moral principles can (and should) lend some

degree of support to the characteristic (and potentially idiosyn-cratic) claims of our own community.

• In consequence, morality is rooted in the very nature of rationalityand thereby provides the moral enterprise with an objectively cogentrationale.

So much for the general line of argument. I will first take up the issue ofmoral relativism, since it is such a pervasively held position.

II. Moral Relativism

There are two distinct modes of moral relativism. Both agree that allmoral codes are of equal validity —that each code is as good as any other.However, one mode, “indifferentist relativism,” sees all codes as equallyvalid, and the other, “nihilism,” sees all as equally invalid. The former,DOI: 10.1017/S0265052508080151© 2008 Social Philosophy & Policy Foundation. Printed in the USA. 393

syncretist approach takes the line of an indiscriminate openness and accep-tance; the latter, nihilistic approach takes the line of a total rejection.Either way, the prospect of a reasoned endorsement of one moral positionover others is excluded.

Both doctrines are deeply problematic, however. On the one hand,indifferentist relativism is caught up in the evident implausibility of hold-ing that any moral code, any set of moral rules whatsoever, is as good asany other for us here and now, in the circumstances in which we actuallyfind ourselves in our interactions with others. Moral nihilism, on theother hand, is caught up in the no less striking implausibility of thecontention that no moral code whatsoever is valid, that no code can makea warranted claim to effectiveness in safeguarding the interests of people.Both forms of relativism are deeply enmeshed in difficulties, but theformer version represents the more common threat of a subjectivist rel-ativism that maintains: “To each his own; such differences as there arebetween moral codes merely lie in the mind of the exponent.” On such asubjectivist view, morality is a matter of attitudes or tastes: we have ours,and they have theirs, period. In effect, there simply is no fact of the matterabout a moral thesis such as “Stealing is wrong.” The subjectivist insists:“That’s just what people happen to think in our set or our society.” Onthis view, the sole reason why moral judgments ultimately matter for usis one of “keeping in step” with the rest of our group so as to conform ourbehavior to socially accepted norms. Any code of conduct is as good asany other: it is simply a matter of “When in Rome, do as the Romans do.”All moralities are created equal.

From such a standpoint, there is no place for moral objectivity. Thus,there nowadays prevails a widespread but nevertheless unfortunate ten-dency to deny the possibility of rational controversy about moral matters,to relegate morality to the never-never land of matters of taste, feeling, orotherwise discursively insupportable opinion.

While such a line of thought has been envisioned by philosophers sincethe ancient skeptics, it is nowadays commonplace.1 To take a single exam-ple, Charles L. Stevenson’s “attitudinal expression” theory is an influen-tial representative of the general approach.2 Stevenson holds that a moralevaluation merely characterizes the subject and not the object that is atissue. My contention that “You acted wrongly in stealing that money” is,Stevenson holds, a mere compound of two assertions, one a factual con-tention (“You stole that money”) and the other a personal avowal (“I

1 The classic exponent of the position is David Hume. See Hume, A Treatise of HumanNature (London, 1738), and Hume, An Inquiry Concerning the Principles of Morals (London,1752), Appendix 1, sec. 1. Antecedents can be found in the compendia of skeptical viewsassembled by Sextus Empiricus.

2 See Charles L. Stevenson, “The Emotive Meaning of Ethical Terms,” in his Facts andValues (New Haven, CT: Yale University Press, 1963), and also his earlier Ethics and Language(New Haven, CT: Yale University Press, 1944). Two works particularly useful on the “emo-tive theory” are G. J. Warnock, Contemporary Moral Philosophy (London: Macmillan, 1967),and J. O. Urmson, The Emotive Theory of Ethics (London: Hutchinson, 1968).

394 NICHOLAS RESCHER

disapprove of your doing this and urge you not to do similar things in thefuture”). The former component makes an objective, descriptive claim;the latter expresses the asserter’s attitude toward it. The point of makingmoral contentions, Stevenson insists, “is not to indicate facts but to createan influence. Instead of merely describing people’s interests [moral con-tentions endeavor to] change or intensify them. They recommend an inter-est in an object, rather than state that the interest already exists.” 3 Moralclaims are primarily designed to exhort people to channel their actionsinto generally beneficial modes.

Such an approach denies that actions are ever actually wrong in them-selves, but insists that people merely think they are. Moral languageis only “used to express feeling about certain objects, but not to makeany assertion about them.” 4 Alternatively, moral judgments are onlyprescriptions —oblique injunctions designed to incite others to action.5 Inany event, they do not really express evaluations which, as such, are rightor wrong, appropriate or inappropriate. Rather, they reflect the circum-stance that people only attach or attribute value to human actions, ascrip-tions which are always made on an entirely subjective basis reflecting themakers’ feelings, wishes, or attitudes —always without any real founda-tion or warrant in the nature of the object. There is nothing more to moralpraise or condemnation than personal feeling. Rightness or wrongnesssimply lies in the view of the individual (or the group) as a mere expres-sion of personal (or social) disapproval.

Notwithstanding its widespread endorsement, such a position is deeplyproblematic. Let us consider why.

III. Morality in Functional Perspective

The key to moral objectivity lies in the very nature of the moral enter-prise. For the reality of it is that morality is a functional enterprise thatexists for the sake of an end and purpose: to foster modes of action andinteraction that facilitate the realization of human interests and, in par-ticular, that channel people’s actions in ways that make people’s liveswithin their communities more beneficial and pleasant.

Accordingly, a moral system is instituted within the human communityto guide behavior in ways that are beneficial to the best interests of thecommunity collectively and its members distributively, that is, to makepeople’s lives more satisfactory and fulfilling. The pivotal question, then,is this: Will a purportedly ethical mode of behavior, if systematicallyadopted, enhance the satisfaction of life and lessen its burdens and neg-ativities? Does it, or would it, make people healthier, happier, more suc-cessful in the realization of those things that make life fuller and more

3 Stevenson, Facts and Values, 16.4 A. J. Ayer, Language, Truth, and Logic, reprint ed. (New York: Dover, 1952), 108.5 Compare R. M. Hare, The Language of Morals (Oxford: Oxford University Press, 1952).

MORAL OBJECTIVITY 395

satisfying and rewarding? Does it, or would it, facilitate the realization ofpeople’s best interests?

The functional nature of morality —its being geared to serving the inter-ests of people —makes for a “moral realism” which maintains that thereare indeed moral facts.6 It does not, however, underwrite that mode of“realism” which maintains that moral rightness is a certain sort of prop-erty of acts (be it a “nonnatural” property, or a “supervenient” property,or the sort of preternatural property discernible only by some specialpower of moral intuition).7 Rather, the functional approach indicates thatwhat is at issue with “rightness” is not a “property” in the ordinary senseat all, but a contextual feature of a relational sort which turns on the placeof the act in question within a wider framework of relevant circum-stances. Being “(morally) right,” like being “average” in size or “inexpen-sive” in price, is a contextual feature relating to the setting of one itemwithin its embracing environment. While it is not a property of an iso-lated item, such a feature is nevertheless one whose possession is objec-tively discernible, albeit only within the setting of that larger context.Thus, the moral status of an act is not the sort of thing that is a propertyor quality at all, but a relational feature whose determination involves awide variety of contextual issues: agents, circumstances, motives, alter-natives, and the like.

Subjectivist relativism stands committed to the idea that morality issimply a matter of local convention. It loses sight of what is really at issuewith morality —the proper consideration of people’s real interests. Andthereby it makes a travesty of morality by restricting the idea of whatpeople ought to do to what the particular customs of their society require.It confuses morality with mores. Mores are, indeed, simply matters ofcustom and convention, like table manners and dress codes; but moralityinvolves the adaptation to local conditions of universal principles regard-ing the safeguarding of people’s interests. A crucial divide thus separatesmorality from mores; a difference in kind is at issue. After all, moralchoice is a matter of opting not for what people may happen to prefer butfor what is (morally) preferable —or what can reasonably be claimed to beso with a view to the determinative aim of the whole enterprise, namely,the safeguarding of the real interests of people in general. This normativedimension means that the variability involved in the variation of moralrules from one society to another is a mere surface phenomenon that doesnot reach down to the level of fundamentals.

6 A comprehensive bibliography of the subject is given in Geoffrey Sayre-McCord, “MoralRealism Bibliography,” The Southern Journal of Philosophy 24, Supplement (1986), 143–59.

7 Critics of moral realism often suppose, quite wrongly, that ethical appraisals can reflectmatters of fact only if ethical characteristics represent supernatural properties that aresomehow discernible by a special faculty of moral intuition —a peculiar “moral sense,” as itwere. Only recently has this far-fetched view attracted the criticism it deserves. See RichardBoyd, “How to Be a Moral Realist,” in Essays on Moral Realism, ed. Geoffrey Sayre-McCord(Ithaca, NY: Cornell University Press, 1988), as well as some of the other essays of thatanthology.


IV. Uniformity Despite Diversity: Morality Versus Mores

A critic might object at this point that moral objectivity is surelycounterindicated by the circumstance that one cannot validly criticize themoral code of a society by any “external” criteria. But this is by no meanstrue. Whether a certain operational code is intended within the ambit of itssocial context to operate as a moral code may well be a proper subject ofdiscussion and controversy. Nonetheless, once it is settled that it is indeeda moral code that is at issue, then in view of that very fact one can certainlybring principles of critical evaluation to bear. For at this point the ques-tion becomes paramount whether, and how effectively, this code accom-plishes for its society that function for which moral codes are institutedamong men —to constrain their interactions in ways that safeguard theirbest interests.

When we are assessing a moral code in this way, we are not simply exer-cising a cultural imperialism by judging it against our own code; we arenot asking how concordant or discordant it is with the prevailing mores-correlative standards of our society. We are judging it, rather, against thoseuniversal and absolute standards in terms of which the adequacy of anycode, our own included, must be appraised. The evaluation of appropri-ateness is not one of our code against theirs, but one of judging both oursand theirs by a common, generic standard. What makes an action right orwrong (as the case may be) is just whether doing the sort of thing at issueprotects or injures the interests of all the agents concerned. (To reempha-size: morality is by definition geared to the benefit of rational agents, evenas refrigeration is by definition geared to cooling.)8 This is part and parcelof the very meaning of “morally right” and “morally wrong,” and it ren-ders judgment in these matters factual, objective, and rationally disputable.

Morality, again, is a particular, well-defined sort of purposive projectwhose cohesive unity as such resides in its inherent function of moldingthe behavior of people in line with a care for one another’s interests.Despite the diversity of the substantive moral codes of different societies,the basic overarching principles of morality are uniform and invariant;they are inherent in the very idea of what morality is all about.

The long and short of it is that the anthropological reduction of moral-ity to mores just does not work. Some things are wrong in an absolute anduniversal way: murder (i.e., unjustifiably killing another person); takingimproper advantage of people; inflicting pointless harm; lying and decep-tion for selfish advantage; betraying a trust for personal gain; breakingpromises out of sheer perversity; and misusing the institutions of one’ssociety for one’s own purposes. Local custom to the contrary notwith-

8 Does this way of viewing the matter put subrational animals outside the pale of moralconcern? By no means. For it matters deeply to rational agents how other rational agentstreat animals, or for that matter, how they treat any other sorts of beings that have interestscapable of being injured. We have a substantial interest in how others comport themselveswho also belong to the type to which we see ourselves as belonging.


standing, such things are morally wrong anytime, anywhere, and foranyone. Their prohibitions are moral universals —parts of morality assuch. (Thus, these prohibitions hold good not just for us humans but forall rational beings.)

To be sure, different societies operate with different moral ground-rules at the procedural level. Some societies deem it outrageous forwomen to expose their faces, their breasts, their knees; others view thisas altogether acceptable and perhaps even mandatory. Nonetheless,behind this variation stands a universal principle: “Respect people’ssensibilities about the appropriate and acceptable appearance of fellowhumans by conforming to established rules of proper modesty.” Thisoverarching principle is universal and absolute. Its implementation withrespect to, say, elbows or belly buttons is, of course, something thatvaries with custom and the practices of the community. The rule itselfis abstract and schematic; it stands in need of implementing criteria asto what proper modesty and due decorum demand. The matter is oneof a universal principle with variable implementations subject to locallyestablished standards and criteria that are grounded in the particularcustoms of the community.

Thus, while the concrete strictures of morality —its specific ordi-nances and procedural rules of thumb —will, of course, differ from ageto age and culture to culture, nevertheless the ultimate principles thatserve to define the project of “morality” as such are universal. Theuniform governing conception of “what morality is” suffices to estab-lish and standardize those ultimate and fixed principles that governthe moral enterprise as such. At the level of fundamentals, the variabil-ity of moral codes is underpinned by an absolute uniformity of moralprinciples and values. At this highest level alone is there absoluteness:the rejection of appropriate moral contentions at this level involves alapse of rational cogency. But at the lower levels there is almost alwayssome room for variation, and dispute as well. (How concern for thewell-being of one’s fellows can be brought to effective expression, forexample, will very much depend on the institutions of one’s society,and also, to some extent, on one’s place within it.)

To be sure, anthropologists, and even, alas, philosophers, often saythings like, “Members of the Wazonga tribe deem it morally proper (oreven mandatory) to sacrifice first-born female children to the tribal gods.”But there are big problems here. This way of talking betokens lamentablyloose thinking. Compare:

(i) The Wazonga habitually (customarily) sacrifice first-born femalechildren.

(ii) The Wazonga think it acceptable (or perhaps even meritorious)to sacrifice first-born female children.

(iii) The Wazonga think it morally acceptable (or mandatory) to sac-rifice first-born female children.


It must be noted that, however true and incontestable the first two con-tentions may be, the third is just untenable. For compare (iii) with:

(iv) The Wazonga think it mathematically true that dogs have tails.

No matter how firmly convinced the Wazonga may be that dogs havetails, thesis (iv) —taken as it stands, with its overt reference to mathematicaltruth —is firmly and squarely a thesis of ours, and not one of theirs! That“mathematically” is clearly something that we ourselves have interjectedinto the picture. For this very reason, it is in deep difficulty unless thecondition is realized —as in the circumstances seems highly implausible —that the Wazonga have an essentially correct conception of what math-ematics is, and, moreover, are convinced that the claim that dogs havetails belongs among the appropriate contentions of this particular realm.Analogously, one cannot appropriately maintain thesis (iii) unless one isprepared to claim both that the Wazonga have an essentially correct con-ception of what morality is (correct, that is, by the defining standards ofmorality), and furthermore that they are convinced that the practice inquestion is acceptable within the framework of this (moral) project as soconceived. The concatenation of these two conditions is not only implau-sible in the circumstances, but even paradoxical, seeing that the firstcondition powerfully counterindicates the second. The salient point isthat the mere fact of seeing this practice as mandated by custom (as partof “what is expected of us and what we have always done”) does notmake it part of the Wazonga’s morality.

The anthropological route to moral relativism is, to say the least, emi-nently problematic. For only a rationale that is conjointly articulated interms of the definitive telos of the moral enterprise can render a custom-ary practice —however compulsory —into a moral one. There is no diffi-culty whatever about the idea of different social customs, but the idea ofdifferent moralities faces insuperable difficulties. The case is much likethat of saying that the tribe whose counting practices are based on thesequence “one, two, many” has a different arithmetic from ourselves. Todo anything like justice to the facts, one would have to say that they donot have arithmetic at all, but just a peculiar, and very rudimentary, wayof counting.9 The situation of those exotic tribesmen of ours is similar. Onthe given evidence, they do not have a different morality; rather, theirculture just has not developed to a point where they have a morality at all.If they think it is acceptable to engage in practices like the sacrifice offirst-born girl children, then their grasp on the conception of morality is,

9 The earliest Mesopotamian counting notation was a matter of context-variable numer-ical indicators, with one symbol indicating “10” when sheep were at issue, “6” whenreferring to containers of grain, and “18” when referring to fields or plots of arable land.Whatever these symbols were, they were not numbers, and such combinational manipula-tions as they admitted of do not constitute arithmetic. But this fact pivots on what wethink —namely, on how we understand what is at issue with “arithmetic.”


on the face of it, somewhere between inadequate and nonexistent. It is( just barely) conceivable that sense could be made of a locution like “TheWazonga believe that morality requires them to act X-wise [i.e., to engagein child-sacrifice, or in some other practice that we ourselves deem mor-ally outrageous].” But the story needed to make sense of this would haveto be complex and (to our ears) extremely implausible. (For example, wemight imagine that the Wazonga believe that this practice affords the onlyway to prevent some evil demon from wreaking havoc on the commu-nity.) Of course, the basis of such a story must be a principle which (likesaving the community from anguish, agony, and annihilation) envisionsa patently moral purpose from our point of view and would thus (in thecircumstances) serve as a moral motive for us as well.

This example underscores my earlier point that there is a fundamentaldifference between morality and mores, a difference which relativismsimply ignores in its tendency to identify the two. Relativism proclaims:“Other societies have their moral convictions (rules, standards, values)and we have ours. One is every bit as good as the other. To each his own.Nobody is in a position to criticize or condemn the moral views of oth-ers.” But to take this line in moral matters would, in fact, be to abandonthe very idea of morality. To be sure, mere custom is something else again.Variability does indeed obtain with respect to mere mores: we eat withcutlery, they eat with chopsticks; we sleep on beds, they sleep in ham-mocks; we speak one language, they speak another —each with equalpropriety. However, this indifference does not hold for morals, for mattersof moral principle, where universality does and must obtain. It is non-sense to say: “We treat the handicapped humanely; they drown them atsea to spare themselves the inconvenience of special accommodation. It’sall just a matter of local custom.” This sort of thing is just not viable fromthe moral standpoint! If crass selfishness, pointless maltreatment, wantondeceit, or the infliction of needless pain is wrong for us —as indeed itis —then it is wrong for them too; and conversely. This is the fundamentalinsight that endows morality with its emphasis on an absolutist univer-sality which has it that, at the level of fundamentals, matters of moralprinciple are the same universally and for everyone.

So here stands the characteristic, and ultimately right-minded, insightof Kantian ethics: that when it is indeed morals rather than mores that arein view, then the issue has to be seen as one that is universal and absolute —that in this domain of morality proper, what is appropriate for one mustbe so for all. At the level of deeper principles, there is indeed a strictequality before the moral law.

V. The Hierarchy of Principles

Philosophers and social scientists who advocate moral relativism oftenfall prey to the oversimplification of seeing moral norms at all levels as


being of a piece. They fail to see the distinction between established mores,which indeed are variable and context-dependent, and higher-level valuesand principles, which, I contend, are fixed, universal, and unchanging. Rec-ognition of the hierarchical stratification of moral norms is essential to aproper understanding of morality. The fact that there are uniform andunchanging principles at the top of the hierarchy —principles that inherein the very conception of morality itself —is quite compatible with plural-ity, variation, and even some measure of conventionality in the moral normsof the lower levels. The multilevel structure of moral norms provides thekey to reconciling the inherent absolutism of morality with the “culturalrelativity” of moral codes by showing how the relativistic variation of suchcodes is perfectly compatible with the absolutism of moral fundamentals.Plurality and variability in regard to lower-level norms is in no way at oddswith an absolutist uniformity of higher-level principles.

It is, in the end, the controlling role of higher-level principles inherentin the very idea of what “morality” is all about that saves morality froma destructive fission that extends an open invitation to indifferentism.These higher-level moral principles ground such fixed and absolute moralvalues as: the value of the human person, grounding (inter alia) his or herrights to safety and security; the dignity and respect of the individual;liberty, individual freedom, and the right to self-determination; and peo-ple’s rights to consultation in matters affecting their well-being and inter-ests. We cannot at one and the same time remain within the moral purviewand avoid acknowledging these fundamental desiderata as pivotal to themoral enterprise, seeing that they are constitutive of this enterprise assuch. To abandon them is not to contemplate “a different morality,” but toabandon the moral domain itself, to change the subject of discussion (e.g.,to mores and customs). And when these rights are infringed or abrogatedin a certain culture or society, that culture or society proclaims its moralfailings in virtue of this very fact. There is simply no room for negotiationhere. Different cultures may have different mores, but they do not —cannot —have different morals at the level of basic principles.

The uniformity of the higher-level norms determinative of moralitymeans that different families of (appropriate) moral rules —different moralcodes —simply represent diverse routes to the same ultimate destination.Moreover, given the functional integrity of morality —as an endeavorgeared to safeguarding and promoting the best or real interests of rationalagents as such —this is exactly how it should be. The mere fact that asingle enterprise —morality —is at issue means that, despite the pluralityof moral codes, we have to deal with a single uniform family of funda-mentals; it means that the variability of moral rules is underpinned by anabsolute uniformity of moral principles, the plurality of valid moral codesnotwithstanding.

All moral codes have important elements in common simply in view ofthe fact that morality is at issue. As such, all moral codes are bound toencompass such morally fundamental considerations as the following:


• What people do matters. Some actions are right, others wrong;some are acceptable and some are not. There is an important dif-ference here.

• This is not just a matter of convention, custom, and the way thingsare done. Violations of moral principles are not just offenses againstsensibility but against people’s just claims in matters where peo-ple’s actual well-being is at stake.

• In violating moral rules, we inflict injury on the lives, welfare, orotherwise legitimate interests of others —either actually or by wayof putting them at unjustified risk.

Attunement to considerations of this sort is, by definition, essential to anysystem of morality, and serves to provide the basis for such imperativesas “Do not simply ignore other people’s rights and claims in your owndeliberations!”; “Do not inflict needless pain on people!”; “Honor thelegitimate interests of others!”; “Do not take what rightfully belongs toothers without their appropriately secured consent!”; “Do not wantonlybreak promises!”; and “Do not cause someone anguish simply for yourown amusement!” In the context of morality, principles and rules of thissort are universal and absolute. They are of the very essence of morality;in abandoning them, we would withdraw from a discussion of moralityand would, in effect, be changing the subject. What we would say mightbe interesting, and even true, but it would deal with another topic.

From the moral point of view, the empirical search for “cultural invari-ants,” as pursued by some ethnologists, is thus entirely beside the point.10

When such investigators embark on a cross-cultural quest for “moraluniversals” or “universal values” amid the variation of social customs,they are engaged in a search which, however interesting and instructivein its own way, has nothing whatever to do with the sort of normativeuniversality at issue with morality as such. Moral universality is not amatter of cross-cultural commonality but of a conceptually constraineduniformity. (It would be just as pointless to investigate whether anotherculture’s forks have tines.)

How can this fixity of the conception of morality and of the fundamen-tal principles that are at issue within it —inherent in the monolithic uni-formity of “what morality is” —be reconciled with the plain fact of a diversityof (presumably cogent) answers to the question: “What is it moral to do?”How can such an absolutism of morality’s fundamentals coexist with thepatent relativity of moral evaluations across different times and cultures?

10 See, e.g., Clyde Kluckhohn, Culture and Behavior (Glencoe, IL: The Free Press, 1962);Kluckhohn, “Ethical Relativity; Sic et Non,” The Journal of Philosophy 52 (1955): 663–77; R.Redfield, “The Universally Human and the Culturally Variable,” The Journal of GeneralEducation 10 (1967): 150–60; Ralph Linton, “Universal Ethical Principles: An AnthropologicalView,” in R. N. Anshen, ed., Moral Principles in Action (New York: Harper, 1952); and Linton,“The Problem of Universal Values,” in R. F. Spencer, ed., Method and Perspective in Anthro-pology (Minneapolis: University of Minnesota Press, 1954).


The answer lies in the fact that several intermediate levels or stratainevitably separate the overarching aims and principles of morality fromany concrete judgments about what it is moral to do. We have, in fact, todeal with a descending hierarchy of the form indicated in table 1.

At the topmost level (level 1), we have the defining aims of morality,the objectives that identify the moral enterprise as such. At this level, weare instructed to act with a view to safeguarding the valid interests ofothers and to avoid harming them. These defining aims of morality expli-cate what is at issue when it is with morality (rather than, say, basket-

Table 1. Illustrations from the Implementation Hierarchy of Morality

Level 1: Defining Aims of Morality• To support the best interests of people.• To avoid injuring other people.

Level 2: Fundamental Principles (Controlling Values)• Do not cause people needless pain (gentleness).• Do not endanger people’s lives or their well-being unnecessarily (care

for safety).• Honor your genuine commitments to people; in dealing with people,

give them their just due ( probity).• Help others when you reasonably can (generosity).• Do not take improper advantage of others ( fairness).• Treat others with respect.• Do not violate the duly established rights and claims of others.• Do not unjustly deprive others of life, liberty, or the opportunity for

self-development.• Do not deliberately aid and abet others in wrongdoing.

Level 3: Governing Rules• Do not kill.• Do not lie.• Do not cheat.• Do not steal.

Level 4: Operating Directives• Killing in self-defense, in war, or when executing a legal judgment is

justifiable.• Be candid when replying to appropriate questions, but lying to save

another’s life is permissible.• Do not play with unfair dice or defraud others in trade.• Stealing when one’s family is starving is allowable.

Level 5: Concrete Rulings• Jones, return the money you borrowed from Smith.• Mr. CEO, do not pollute this river; dispose of your sewage elsewhere.• Mrs. Smith, do not let these children play with those matches.


weaving) that we propose to concern ourselves. In spelling out the purposeof morality, what it is all about, these top-level norms provide the ulti-mate reference points of moral deliberation. Moreover, they are unalter-ably fixed; they are inherent in the very nature of the subject.

These aims also fix the fundamental principles and controlling values(level 2) that delineate the moral virtues (fairness, generosity, probity, andthe rest). Such values define the salient norms that link the abstract defin-ing aims of morality to an operational morality of specific governingrules. The norms embodied in level 2 are universal and absolute, servingas parts of what makes morality the thing it is. Accordingly, these level-2principles also lie fixedly in the very nature of the subject.

At the two topmost levels, then, there is simply no room for any dis-agreement about morality. Here disagreement betokens misunderstand-ing: if one does not recognize the fundamental aims, principles, andvalues that characterize the moral enterprise as such, then one is simplytalking about something else altogether. In any discussion of morality,these things are simply givens provided by the very terms of reference atissue. However, this situation changes as one moves further down thelevels and takes additional steps toward concreteness.

At level 3, we encounter the governing rules and regulations that directthe specifically moral transaction of affairs. Here we have generalities ofthe usual sort: “Do not lie”; “Do not cheat”; “Do not steal”; etc. At thislevel, we come to the imperatives that guide our deliberations and deci-sions. Like the Ten Commandments, they set out the controlling rules ofthe moral practices of a community, providing us with general guidancein moral conduct. Here variability begins to set in, for these rules imple-ment morality’s ruling principles at the concrete level of recommendedpractices in a way that admits of adjustment to the changeable circum-stances of local conditions.

A generalized governing rule proceeds on the order of the injunction“Do not steal” (i.e., “Do not take something that properly belongs toanother”). This is, in itself, still something abstract and schematic. It stillrequires the concrete fleshing out of substantive implementing specifica-tions to tell us what sorts of things make for proper ownership. Thus, thenext level, level 4, presents us with the ground rules of procedure oroperating directives that furnish our working guidelines and criteria forthe moral resolution of various types of cases. (For example: “Killing inself-defense, in war, or when executing a legal judgment is justifiable.”)At this level of operating directives, the variability of local practice comesto the fore, so that there is further room for diversification here; weourselves implement the level-3 rule “Do not lie” by way of the level-4directive “Be candid when replying to appropriate questions,” but a soci-ety of convinced skeptics could not do so.

The operating directives of level 4 incorporate the situation-relativestandards and criteria though which the more abstract, level-3 rules gettheir grip on concrete situations. Those general rules themselves are


too abstract —too loose or vague —to be applicable without further direc-tions to give them a purchase on concrete situations. They must be givenconcrete implementation with reference to local, and thus variable,arrangements.11

Finally, at the lowest level (level 5), we come to concrete rulings, indi-vidual resolutions with respect to the specific issues arising in particularcases. (For example: “It was wicked of Lady Macbeth to incite her hus-band to kill the king.”)

In the implementation hierarchy laid out in table 1, we thus descendfrom what is abstractly and fixedly universal to what is concrete andvariable. The principles of level 2 can be viewed as direct implications ofthe aims of level 1, but as we move downward past level 3 to the oper-ating directives of level 4, there is, increasingly, a looseness or “slack” thatmakes room for the specific and variable ways in which different groupsimplement the particular higher-level objectives at issue.

As we have seen, the entire hierarchy begins with the ruling impera-tive: “Support the best interests of people.” This overarching purposedoes not itself stand subordinate to anything else. It is the defining aim ofthe enterprise that gives unity and determinativeness to the justificatoryventure. The process of validating lower-level considerations in terms ofhigher-level ones must come to a stop somewhere.

Overall, then, we have to deal with a chain of subordination-linkagesthat connect a concrete moral judgment —a particular moral act-recommendation or command (at level 5) —with the ultimate definingaim of the moral enterprise (level 1). We confront a hierarchy of imper-atives that place different injunctions at different levels of fundamentalityin the moral enterprise, with some (the defining aims and fundamentalprinciples) absolute and others (the concrete rulings) variable and relativeto context and circumstance. Concrete moral rulings must derive theirvalidity through being appropriate instantiations of an overarching prin-ciple of universal (unrestricted) validity.

To be sure, there is a “slack” that leaves room for increasing variabilityand dissensus, especially at the hierarchy’s two lowest levels. Operatingdirectives and concrete rulings will vary with situations and circum-stances. We cannot expect to encounter any universal consensus acrosscultural and temporal divides: moralists of different eras (like physicians)are bound to differ, and the same is true of different cultures. There is,inevitably, substantial variability among particular groups, each with itsown ideas influenced by locally prevailing conditions and circumstances.

11 The analogy of natural law is helpful: “Theft, murder, adultery and all injuries areforbidden by the laws of nature; but which is to be called theft, what murder, what adultery,what injury in a citizen, this is not to be determined by the natural but by the civil law . . .”(Thomas Hobbes, De Cive, chap. IV, sec. 16). St. Thomas Aquinas holds that appropriatehuman law must be subordinate to the natural law by way of “particular determination” —with different human laws, varying from place to place, nevertheless representing appro-priate concretizations of the same underlying principle of natural law. See Summa Theologica,IaIIae, questions 95–96.


Nonetheless, the impact of lower-level variation is mitigated by the factthat justification at lower levels proceeds throughout with reference tohigher levels where uniformity prevails.

At the level of aims and fundamental principles, then, morality is abso-lute; its strictures at this level hold good for everyone, for all rationalagents. Lower-level rules, directives, and rulings must —if valid —preservea “linkage of subordination” to those highest-level abstractions. The valid-ity of concrete rulings is always a matter of their attuning global (andabstract) prescriptions to local (and concrete) conditions. Without thatlinkage to the fixed, highest-level absolutes (levels 1 and 2), the linkage tomorality itself is severed. For a concrete ruling to be a proper moral rulingat all, there must be a suitable moral rationale for the action —a pathwayof subordination that connects it in a continuous manner all the way upto the defining aims of the moral enterprise. Varying practices and codesof procedure only possess moral validity insofar as they are implemen-tations of a fixed and determinate set of moral principles. Moral validitymust always be rooted in a moral universality that is constrained by aconceptual fixity.

VI. The Claims of Our Own Community

How does the moral code of any society secure its obligating hold onme? Why is it that I should see myself as duty-bound to follow theestablished rules of my community rather than some other rules?

The answer is straightforward. A certain (potentially variable) ruleobtains its binding grip (its deontic hold, as it were) upon me preciselythrough representing the contextually appropriate way of implementinga fixed moral principle or value in the particular social context in whichI find myself. Local morality is a matter of substantially limited imple-mentation of universal principles, and its bearing upon ourselves is rootedin the fact that it is our particular circumstance and situation that is atissue.

The salient point is that the moral code of our own community is,through this very fact, the one that is paramount for us. This moral codeis the one that defines and specifies the rights and expectations of thoseothers with whom we are (ex hypothesi ) co-situated in a context of mutualinteraction. After all, in morality, as elsewhere, the universal is availableto us only through mediation of the specific: one can only pursue ageneralized desideratum via its particularized (and variable) concretiza-tions. Communication is universal, language is specific; eating is univer-sal, cuisine is local; morality is universal, particular concrete moral codesare variable and diversified. Nonetheless, such variability does not under-mine or abrogate validity. The universals of morality not only permit butrequire adjustment to local conditions, and at the local level we are con-cerned not with the validation of morality as such, but with the justifi-cation of a particular moral code for a particular group in particular


circumstances. The concrete moral code of our community is the onlyway in which implementation of the higher-level demands of morality isavailable to us, given the realities of the particular context in which welive and labor.

To be sure, the behavioral practices of a society (ours included) are notabove criticism. Anything that people do can be done badly —the shapingof a moral code included. Confronted by any set of purported “moralrules of behavior,” we can and should ask: How well do these rulesimplement the fundamental principles that articulate the aims of themoral enterprise as we do (and must) understand it? Any system, forexample, that authorizes the infliction of pain on people for no betterreason than affording amusement to others deserves flat-out condemna-tion and rejection. It would be a decisive objection to any system of“morality” that it deems acceptable (let alone worthy of approbation) amode of behavior which is immoral on the conscientious application ofour standards —for example, a system that approves pointless lying, orwanton cruelty, or any other practice that undermines the legitimate inter-ests of people.

Accordingly, the moral code of one’s own community can be subjectedto moral criticism and reevaluation. The question can always be pressedwhether the concrete moral practices and rules of one’s own communitydo indeed implement effectively the definitive values of the moral enter-prise and, above all, whether these practices and rules satisfactorily servethe best interests of people in general. The concrete moral code of ourcommunity can be found defective or deficient in point of morality. (Thinkof Nazi Germany, for example.) Socially accepted principles of action areclearly not beyond criticism. However, this criticism can, if appropriate,only be developed from the vantage point of those aims and fundamentalprinciples that characterize the moral project as such.

Of course, no belief system, no framework of thought, no system ofnorms —moral or otherwise —is an absolute, detached from the worldand delivered to mankind in unchangeable perfection by the world-spiritfrom on high. We can do no more than utilize the local, particularized,diversified instruments we humans can manage to develop within thelimitations of our place and time. Thus far, what moral relativists saymakes sense. But this emphatically does not engender an indifferentistsubjectivism, for pluralism does not mean that there are no applicablestandards, that with morality one can throw things together any whichway. As I have argued, an overarching purpose is operative in the veryconception of what morality is all about, and this prevents morality frombecoming unraveled as a rational enterprise through the pluralism andvariability that relativists mistakenly see as somehow destructive of moraluniversality.

Moreover, the existence of differences over the recognition of the fun-damental facts at issue is not relativism but merely a perfectly rationalsituational contextualism. People must feed themselves and house them-


selves. Nature dictates no single, unique process for accomplishing suchends: we must make the best use we can of the possibilities that place andtime put at our disposal. The same holds true with respect to morality’srequirements of due care for the interests of people. Here, too, one mustsimply do the best one can, striving within the conditions and circum-stances of one’s setting to establish practices and rules that align humaninteractions in a productive harmony from which everyone can benefit.

VII. Conclusion

What, then, of moral objectivity? Let us go back to basics. What is it thatmakes something objective? The objectivity of an issue lies in its being amatter of fact that, in principle, can be determined to be so by anyone,because what is at issue is not a matter of opinion or of custom but ratherobtains impersonally, independently of what individual people may thinkor prefer. Objective matters do no lie in the eyes of the beholder but pivoton the actual facts. This being so, consider the salient question that arisesin regard to morality:

Would the prevalence of such-and-such a way of behaving amongthe members of the community at large effectively conduce to peo-ple’s best interests in making their lives more secure, more pleasant,and/or more rewarding and satisfying?

The matter at issue here is not a matter of what I like or what wouldplease me; it is not my attitude or my reaction or my own personalinterests that are at issue —or indeed yours or anybody’s. The question isinherently general, relating to the reaction of people at large, and it relatesnot to what they want but rather to what makes them better off by wayof being conducive to their well-being. The question concerns the condi-tion of people in general, not on the basis of what you or I or some groupor other do think about this, but on the basis of what people should, andsensible people would, think. Specifically, it is a question of what makessomeone better off in terms of their real or true interest —what conducesto their health and well-being, their security and safety, their opportuni-ties for self-development and self-expression. All these issues, and otherslike them, are substantially matters of objective fact.

Accordingly, what renders morality objective is the fact that moralevaluations can —and should —be validated as cogent through consider-ation of how the practices being evaluated advance the aims of the enter-prise for whose sake morality is instantiated in human affairs. Morality assuch consists in the pursuit, through variable and context-relative means,of invariant and objectively implementable ends that are rooted in acommitment to the best interests of people in general. To claim that some-one ought (or ought not) to act in a certain way is thereby to commitoneself to the availability of a good reason why one should or should not


do so —and a reason that is not only good but good in a certain mode, themoral mode, in showing that this sort of action is bound up with due carefor the interests of others. Whether an action exhibits due care for theinterests of others is something open to general view, something that canbe investigated by other people as readily as by the agent himself. Sincepeople’s (real or true) interests are rooted in their needs, the morallycrucial circumstance that certain modes of action are conducive (andothers harmful) to the best interests of people is something that can beinvestigated and sensibly assessed by the standards generally prevalentin rational discussion. These matters are not questions of feeling or taste,but represent something objective about which one can deliberate andargue in a sensible way on the basis of reasons whose cogency is, orshould be, accessible to anyone. The modes of behavior of people thatrender life in their communities “nasty, brutish, and short” (or even merelymore difficult and less pleasant than need be) generally admit of straight-forward and unproblematic discernment.

The fact that thievery, vandalism, boorishness, arrogance, and rudenessare ethically inappropriate is not rooted in some individual’s or group’sdislike of such things, but rather in the (perfectly objective) fact that suchmodes of behavior will, as they become more prevalent, increasinglydegrade the quality of life of the community by creating circumstances inwhich the pursuit by individuals of their life-plans and objectives becomesincreasingly difficult.

Philosophy, University of Pittsburgh


Documents

Objectivism, Subjectivism, and Relativism in Ethics: Volume 25, Part 1 (Social Philosophy and Policy) (v. 25)