Intro for Acting on Principle 5

Acting on Principle, Intro Essay for 2cnd Edition, O O’Neill

1

Acting on Principle, Intro Essay for 2cnd Edition1

1. Empty Formalism and Modern Moral Philosophy

It is now hard to imagine how unpromising the lines of thought in this book

seemed to most people with an interest in philosophical ethics when I first

worked on them in the late 1960s. Many were then still drawn to more-or-

less positivist claims that reasoned approaches to ethical or political claims were

impossible, while those who favoured a reasoned approach usually

proposed some version of ethical naturalism, mostly of a Utilitarian or

Aristotelian variety. There was general agreement that Kant’s claim that

practical reason can guide ethical action was wholly implausible. Although

Kant’s ethical and political philosophy had enjoyed considerable resonance in

the wider world during the post war decades, that can be detected in the

drafting of the Universal Declaration of Human Rights and the European

Convention on Human Rights, and of the West German and other constitutions,

it had few admirers in Anglophone philosophy departments.

This was not because philosophers at that time had no interest in or respect for

Kant’s wider philosophy. Many admired both his metaphysical caution and

the sweep of his arguments about human knowledge and its limits. But the

consensus was that he neither showed how principles could guide action nor

offered adequate reasons for any specific ethical or political principles, so that

both his metaethics and his normative ethics were defective.

These criticisms were not new. They date back to the early days of German

Idealism, and in particular to Hegel’s critique of the ‘empty formalism’ of

1 Acting on Principle grew out of my Ph D dissertation at Harvard, which was supervised

by John Rawls and submitted at the end of 1968 under the title Universalisability. The book

was published by Columbia University Press in 1975 under my then married name, Onora

Nell, and has been unavailable for many years. I am grateful to Columbia University Press

for reverting the copyright to me, and to Cambridge University Press and their readers for

encouraging me to think that it should be made available again. This edition contains the

original bibliography, a selected bibliography of subsequent work on its themes, and a

bibliography of my later work on Kant and Kantian themes. References to my own work in

the footnotes to this introductory essay provide only titles and year of publication; full

details are given in the third bibliography.


2 Kant’s ethics.

2 In the English speaking world less acerbic but substantively

similar criticisms of the core of Kant’s ethics had been made by J.S. Mill in

Utilitarianism, where he wrote

I cannot help referring, for illustration, to a systematic treatise by one of

the most illustrious of them, the Metaphysics of Ethics, by Kant. This

remarkable man, whose system of thought will long remain one of the

landmarks in the history of philosophical speculation, does, in the treatise

in question, lay down a universal first principle as the origin and ground

of moral obligation; it is this: “So act, that the rule on which thou actest

would admit of being adopted as a law by all rational beings.” But when

he begins to deduce from this precept any of the actual duties of morality,

he fails, almost grotesquely, to show that there would be any

contradiction, any logical (not to say physical) impossibility, in the

adoption by all rational beings of the most outrageously immoral rules of

conduct. All he shows is that the consequences of their universal adoption

would be such as no one would choose to incur.3

Curiously, the persistent charge that Kant’s ethics is no more than empty

formalism that prescribes nothing determinate was repeatedly coupled with an

incompatible allegation that it prescribes with rigid insensitivity, so can take

no account of varying circumstances. 4

Some prominent philosophers of the early post war period were even more

dismissive than Hegel. G.E.M. Anscombe, my tutor in Oxford in the

early 60s, published an influential paper titled ‘Modern Moral Philosophy’

in 1958.5 In it she argued that both Kant and the Utilitarians take an

inadequate view of action, fail to understand that acts fall under many

descriptions, consequently do not realise that principles cannot guide

2 G.W. F. Hegel, Philosophy of Right, § 135 "Kant’s … criterion of non-contradiction is

productive of nothing, since where there is nothing, there can be no contradiction either.”

3 J. S. Mill Utilitarianism, ch 1.

4 The charges of formalism and rigourism are incompatible because an ethical position

that is wholly indeterminate prescribes nothing, so will not prescribe with rigid

insensitivity to circumstances. The persistent combination of these incompatible criticisms

of Kant’s ethics is hard to understand. It may be that formalism is seen as a defect in his

metaethics, and rigourism as a defect in his normative ethics—but if his metaethics indeed

has no bite, it can hardly establish normative claims that can be criticised for their

rigourism. 5 G. E. M. Anscombe, ‘Modern Moral Philosophy’, Collected Philosophical Papers of G. E.

M. Anscombe, Vol III, Ethics, Religion and Politics, (Oxford: Blackwell, 1981), 26. Her

influence has been most evident in ‘virtue ethics’, but in my opinion runs far wider.


3 action. Attempts to create an ethics of principles are doomed to fail. She wrote

of Kant that:

... it never occurred to him that a lie could be relevantly described as

anything but just a lie...His rule about universalisable maxims is useless

without stipulations as to what shall count with a view to constructing a

maxim about it

Anscombe levelled the same charges against Utilitarianism:

Mill, like Kant, fails to realise the necessity for stipulation of relevant

descriptions, if his theory is to have content. It did not occur to him that

acts of murder and theft could be otherwise described. He holds that

where a proposed action is of such a kind as to fall under some one

principle established on grounds of utility, one must go by that. 6

She concluded that both Kantian and Utilitarian ethics— the two most

prominent strands of ‘modern moral philosophy’— fail for the same

reasons. At times I have wondered why, given that I was aware of these

powerful lines of thought when I began working on Acting on Principle, I

thought it worth going back to Kant’s ethics.

2. Back to Kant

I suspect that a central reason why I chose to swim against the tide was less that

I was initially drawn to Kant’s practical philosophy, and more that I had

become disillusioned with contemporary accounts of reasoning about action.

As a graduate student at Harvard in the late 1960s I joined a small but intense

seminar given by Robert Nozick, which worked through Games and

Decisions by R.D. Luce and H. Raiffa. 7 At first I was beguiled by the

neatness of models of rational choice, and their seemingly manageable

accounts of reasoning about action. But after a few months I concluded that

these approaches to practical reason fail, and that their supposed ethical

implications were illusory. The simplistic assumption that we can

exhaustively list ‘the options’ that agents face seemed open to the very

worries about relevant descriptions that lie behind Anscombe’s criticism of

Utilitarian and Kantian ethics. Even if we could do so, any claim that we can

establish that some option is ‘optimal’ seemed to me to rely on metric,

epistemic and other fictions. To my initial disappointment, I concluded that

6 Page no

7 R. D. Luce and H. Raiffa, Games and Decisions: Introduction and Critical Survey, John

Wiley and Sons, 1957.


4 sophisticated work on consequentialist practical reasoning too was

fractured by metaethical failings and normative deficiencies, which were

cumulatively even more recalcitrant than those Anscombe detected in all

modern moral philosophy.

In rebounding from this brief enthusiasm for models of rational choice and

consequentialist ethics, I optimistically turned initially to some mid-twentieth

century writers who approached ethical reasoning with an explicit focus

on more formal constraints on principles of action, and in particular to

those who argued that some form of universalism was the hallmark of

justifiable ethical principles. I looked at the writings of R.M. Hare, G. M.

Singer and Kurt Baier, but failed to find a convincing account of practical

reasoning in their work. 8

Only then did I begin to think about Kant’s practical philosophy with

more care. For this I was adequately, but not wholly, prepared. Although my

German was fluent, I was neither attracted by the prospect of total immersion in

Kant’s writings, nor inclined to give priority to scholarship over argument. But

at least I had by then read central parts of the Kantian corpus, and after

my rebound from rational choice theory was prepared to take them

seriously.

At Phillipa Foot’s suggestion I had worked through The Groundwork of the

Metaphysic of Morals with some care as an undergraduate (Anscombe

preferred to leave the chore of teaching Kant to her colleague at

Somerville). Later, as a graduate student at Harvard, I had read The Critique of

Pure Reason under Charles Parsons, who sparked my interest in Doctrine of

Method, to which I returned when I began to think more systematically about

Kant’s account of reason. And as Stanley Cavell’s teaching assistant I had

scurried to grasp some of the implications of Religion within the Limits of Mere Reason.

Above all when John Rawls agreed to supervise my thesis I had the good

fortune to start working under a philosopher who thought about and lectured

on Kant’s practical philosophy across his entire teaching career. Rawls later

described his own political philosophy (still mostly unpublished when I began

to work with him) as carrying “to a higher level of abstraction the familiar

theory of the social contract as found in Locke, Rousseau and Kant” 9 and

8 This ground clearing work formed part of my Harvard Ph D, but much of it was

excluded from Acting on Principle.

9 In A Theory of Justice Rawls described his work not as Kantian, but as Contractarian.


5 some of its later versions explicitly as a form of ‘Kantian

Constructivism’.

However, while Rawls’s transformative influence on political philosophy is a

matter of common knowledge, his work on Kant’s practical philosophy and

more broadly on the history of philosophy was not widely appreciated during

his lifetime. His lectures on the history of ethics and political philosophy,

including those on Kant’s ethics, were published only after the millennium. 10

It is now abundantly clear that Rawls’s political philosophy grew out of a

profound engagement not only with its history, but also with the wider history

of ethics, and in particular with Kant’s practical philosophy. He combined a

deep knowledge of the writings of his predecessors, with a commitment to

make as much sense as he could of their work. This approach has had a wide

and in my view beneficial influence on subsequent work and in particular on

explorations of Kant’s practical philosophy. 11

It is no exaggeration to say

that Rawls’s teaching of Kant’s practical philosophy transformed the subject as

much as his work on justice transformed political philosophy.

It was also my good fortune that at the time at which I began to work on

Kant’s practical philosophy better editions and translations of some (but by

no means all) of Kant’s writings in ethics and politics, as well as some

distinguished commentaries, were becoming more readily available, in

particular those by H.J. Paton and L. W. Beck. However, the mammoth

enterprise that became the new Cambridge edition of the works of Immanuel

Kant was planned only in the 1980s, and publication of its successive volumes

began in the 1990s. Consequently the citations of Kant’s writings in Acting

on Principle cite older editions and translations, while those in this introductory

essay refer to the Cambridge translations. 12

This is only one respect in which the work has dated. Others reflect both

wider cultural changes and changing philosophical fashion. I would not now

write as if the masculine pronoun could do general duty for points that are not

10

John Rawls, Lectures in the History of Moral Philosophy, ed. Barbara Herman, Harvard

University Press, 2000, and John Rawls, Lectures in the History of Political Philosophy, ed.

Samuel Freeman, Harvard University Press, 2007.

11

Rawls’s wider influence on the history of ethics is the theme of Reclaiming the History

of Ethics: Essays for John Rawls, ed. Barbara Herman, Christine Korsgaard and Andrews

Reath, Cambridge University Press, 1997, which includes essays on Kant’s ethics by other

former pupils, including Susan Neiman, Adrian Piper, Nancy Sherman, Thomas Pogge as

well as the editors and myself.

12

More detail?


6 gender specific.

13 Nor, I think, would I now take quite so austere and formal an

approach to discussing the structure of maxims as I then did, although I can

still see its advantages.

3. Principles and Acts

My central contention in Acting on Principle was that, despite its spare

formality, the Categorical Imperative could be action guiding: a spare and

formal approach to ethics could be fertile and practical. This was a bold as well

as an unpopular claim, and I bracketed several closely connected topics in

order to focus on essentials.

In particular, I discussed only the Formula of Universal Law (FUL)

formulation of the Categorical Imperative— act only on that maxim through

which you can at the same time will that it should become a universal law—

and virtually ignored other formulations.14

I also set aside questions about

the justification of the Categorical Imperative, and its claims to count as ‘the

supreme principle of practical reason’ (a topic that I began to work on in

earnest in the mid 1980s). My hope, which no doubt seemed rash enough to

others, was simply to show that despite its formality the Categorical Imperative

could guide action in at least some ethically significance respects. However,

I left open many questions about the extent to which it could guide

action, about its justification and about the further reaches of Kant’s

practical philosophy, including his writings on politics, history and religion.

Bracketing these important topics allowed me to concentrate on the relation

between principles and action, and on the ethical implications of the

Categorical Imperative. Looking back at the approach I took, I realise that,

despite my reservations about Anscombe’s conclusions and her view of

principles, I had been deeply influenced by her discussions of act

descriptions and their pivotal role in thinking about action. I too saw

principles and the act descriptions they contain as guiding action by shaping

13

However, from time to time I have been comforted for this failure to anticipate the

Zeitgeist by appreciative comments on my supposed prescience in insisting that the fertility

of ethical theories—their normative potential—matters. 14

Since this was the most formal version of the Categorical Imperative, it seemed the best

test case for a claim that a formal criterion can guide action. Later I argued for a reading

of the several formulations of the Categorical Imperative that supports Kant’s claim that

they are equivalent (G 436). See Consistency in Action 1985; reprinted in Constructions

of Reason .


7 or forming it, so as formal rather than efficient causes of action. In Acting

on Principle I did not address the difficult issues this raises for an account of

freedom of action. Only later did I work on Kant’s efforts to reconcile natural

necessity with human freedom, and propose a way of understanding Kant’s

claims that the standpoints of nature and freedom are both indispensable, but

that neither can be reduced to the other.15

In the meantime I rejected

Anscombe’s claim that the plurality of descriptions true of any given act

creates intractable problems for any ethics of principles.

Anscombe moved from the thought that acts can be described in many ways—

surely incontrovertible—to a claim that principles cannot be action guiding

because many principles will apply to any particular act. This inference

seemed to me mistaken, and I have not revised this view. In practical

reasoning we aim to instantiate or enact practical principles and the act

descriptions they incorporate, rather than to apply principles and the

descriptions they incorporate to already existing instances or cases, as we do

in theoretical (including empirical) reasoning. Practical reasoning,

including ethical reasoning, is not a spectator sport. In practical

reasoning— whether ethical or technical, legal or political—we aim first to

justify principles, and then to enact or instantiate them in actual situations: the

particularity of enactments is necessarily subsequent to action, so neither

can nor needs to be presupposed. If we can justify principles, there will no

doubt be situations in which it is taxing, or even impossible, to enact or

instantiate all of the principles that we have reason to take seriously as well as

we would hope to. But this is not because particular acts can be relevantly

described in multiple ways, but because some situations make it difficult or

even impossible to act on certain principles or combinations of principles.

So I agreed with Anscombe that principles cannot be fully determinate,

but took issue with her conclusion that this indeterminacy means that “their

point as guides to action would be lost”.16

The Categorical Imperative is meant to offer a criterion for rejecting or

accepting ethical principles, and thereby a first move in seeking to act in

ethically acceptable ways. If the criterion is satisfactory, we may be able

to use it to distinguish act types that are ethically required from those that are

not required and those that are ethically permitted from those that are

prohibited. But neither the justification of principles nor the judgements we

make in enacting them require agents to subsume acts that have not yet been

done under principles, or to review all possible descriptions of not yet 15

See Reason and Autonomy in Grundlegung III, in Constructions of Reason.

16

See Acting On Principle p. 11.


8 existing act-tokens. The supposed problem of relevant descriptions can indeed

arise in theoretical, including empirical, judgements of already existing

situations and problems. But it does not arise in practical judgment, where the

relevant act is yet to be performed, so is not there to be judged. In Acting on

Principle I argued in the final chapter that contexts of action were primary and

contexts of assessment secondary. 17

4. Maxims

Kant sees the Categorical Imperative as a second-order principle that agents

can use to work out the ethical status of possible substantive principles of

action. Such principles can be adopted or rejected as maxims by agents at or

through particular times. 18

The notion of a maxim is pivotal to Kant’s

account of the relation between principles and acts. Maxims have

propositional form and structure, so are apt for reasoning, including the

reasoning enjoined by the Categorical Imperative; maxims also articulate

standards that can be used to shape action, so are apt for practical

purposes.

In Acting on Principle I concluded, I now think too baldly, that if maxims

are principles with propositional content that belong to agents at or through

particular times , they are best thought of as agents’ intentions. 19

It is true both

that intentions have propositional structure and content, so are apt for

reasoning, and that they are states of agents at or through certain times,

so may be apt for practical purposes. In construing maxims as intentions

I was saying nothing very new: the long tradition that reads Kant’s ethics as a

‘philosophy of the subject’, and the long running criticism that his ethics is too

individualistic, both commonly construe maxims as intentions.

Yet it is not obvious, either textually or systematically, that maxims can be

equated specifically with intentions, and I noted repeatedly that if maxims are

to be seen as intentions, we must stretch our understanding of intentions.

17

I returned to these issues in The Power of Example, 1986; in Instituting Principles:

Between Duty and Action, 1997 revised version 2002, and most recently in Applied Ethics:

Naturalism, Normativity and Public Policy, 2009.

18 Kant characterises maxims as ‘subjective principles’, meaning that they are the

principles of subjects or agents (G 4:421n) at or through some time, but does not regard them

as invariably epistemically subjective as opposed to objective—a thought that would have

been incompatible with his view of ethics.

19

AoP 13,35


9 Kant uses the term maxim to cover both agents’ intentions in acting and their

intentions for the future, both their intentions to do specified acts and their

intentions to pursue specified ends or objectives. He also holds that agents

may be unaware of the maxims on which they act, that they may act on

multiple maxims, and that many of the highly specific intentions of daily life do

not count as maxims. I mentioned, but did not fully explore, other respects

in which Kant’s notion of a maxim diverges from and is broader than many

conceptions of intentions, and returned to the topic only at a later point

when others—in particular Rüdiger Bittner and Otfried Höffe— had

investigated it. 20

Only much later did I appreciate the implications of

Kant’s constant emphasis on the limits of self knowledge, so are opaque rather

than transparent to ourselves?

When I returned to consider these matters, I realised that there were other

ways to construe maxims, which do not identify them with agents’

intentions. One approach was to think of maxims as ascribed to agents, in the

way that revealed preference models of rationality ascribe preferences to

agents on the basis of assumptions about their preference orderings and

evidence of what they do. An ascriptive view of maxims readily allows for

the thought that agents need not be aware of their maxims—but at the cost of

uncoupling maxims from practical reasoning, shifting to a third party

perspective on action and counting all the principles that a pattern of action

satisfies as maxims.

However, Kant insists that maxims are ‘determinations of the will’: they not

merely have propositional form and content, but are directed at action. To

adopt a maxim is not just to entertain or propose or imagine doing some type

of act, or pursuing some type of end. Nor is it merely a matter of wishing or

desiring to do some type of act or to pursue some type of end. Agents who

adopt maxims thereby commit themselves to select some adequate means to

enact their maxims when suitable circumstances arise, and to the normal

and predictable consequences of acting on those maxims. Commitments to

action are not a matter of wish, but of will: in adopting maxims agents

commit themselves to taking adequate means and acknowledging the

predictable results of acting on those maxims, should occasion arise. These

links are central to all practical reasoning, and crucial to understanding how the

Categorical Imperative can be deployed.

20

See inter alia Bittner, Rüdiger, 'Maximen', in G. Funke, ed., Akten des 4. Internationalen

Kant-Kongresses, De Gruyter, Berlin, 1974, 485-9; Herman, Barbara, The Practice of Moral

Judgement, Harvard University Press, Cambridge, Mass., 1993; Timmermann, Jens ‘Kant’s

Puzzling Ethics of Maxims’ , Harvard Review , 4, 2000, 38-52.


10

All of this could, I now think, have been said more perspicuously by accepting

that Kant’s maxims are not best construed specifically as intentions, and by

viewing them simply as the practical principles that may be adopted by

agents at or through various times, and are apt both for reasoning and for

practical purposes. This approach would have allowed for a quite general

articulation of the role of maxims in practical reasoning. 21

It would also

have provided a more perspicuous and useful context for understanding

Kant’s repeated insistence on the limits of human self knowledge, and a

vantage point for addressing the temptation to think of Kant’s ethics as

markedly individualistic. In Acting on Principle I repeatedly addressed these

questions but did not, as I now see matters, reach a fully satisfactory view

either of the texts or of the arguments.

5. Duties First

Notoriously Kant’s substantive ethics is not merely an ethics of principles, but

specifically an ethics of principles of duty. Criticism of the very idea of an

ethics of duty was a major and persistent theme of twentieth century

European culture and philosophy. Many writers held that duty offers a

narrow, unappealing, even defective and corrupting way of thinking about

human life and ethical questions, and that it should be discarded in the

dustbin of history. Among philosophers a focus on duty was often thought

to lead a view of ethics that is insensitive to the variety of situations and

circumstances (rigourism again!) , blind to the importance of character and

virtue,22

excessively preoccupied with matters of blame, 23

or too narrowly

focused on action that others can claim as a matter of right. Very often

these lines of criticism were combined. Towards the end of the twentieth

century, as political and philosophical discussions of what ought to be done

21

I compared the merits of these ways of construing maxims in ‘Kant's Virtues’, 1996.

22

Criticisms of Kant’s supposed rigourism and of his supposed neglect of virtue and

character recur in all varieties of virtue ethics, and have been repeatedly rebutted in work

on Kantian ethics in the last 30 years.

23

This more Nietzschean worry was less common. It was put forward perspicuously by

Bernard Williams, who argued that those who make duties central to moral life are

obsessed by reactive attitudes to others’ wrongdoing, typically attitudes of blame, which

he dubs ‘the characteristic reaction of the morality system’. He concludes that if we are

to reject a culture of blame we must reject conceptions of morality that centre on duties.

Bernard Williams, Ethics and the Limits of Philosophy, London: Fontana, 1985, 177.


11 increasingly invoked claims about human rights, a focus on duty was also

sometimes accused (wrongly, in my view) of prioritising agents over those

whom their action affects. 24

In Acting on Principle I approached these questions about the scope and aims

of normative ethics by setting out a reading of Kant’s classification of

duties and arguing that it offered a more perspicuous view of a wider range of

ethical considerations than was often supposed. In all his ethical writings, but

most explicitly in the Metaphysic of Morals, Kant seeks to justify both duties

to perform acts of specific types—duties of right (some of them specifically

duties of justice) and duties to pursue ends of certain types, which he sees as

duties of virtue. He variously characterises the former duties as perfect,

strict or narrow, and the latter as imperfect or wide. Kant’s alleged neglect of

the virtues seemingly reflected a highly selective reading of his work.

On Kant’s account, perfect duties require or prohibit types of act. Many

perfect duties require or prohibit external acts or omissions that can be

enforced by others (for example, by states). However, other perfect duties

require or prohibit types of act that cannot be enforced by others (for example,

duties to refrain from mockery or self abasement, which Kant classifies as

‘ethical duties of omission’)). Here enforcement by others fails because

enforcers cannot target the inward aspects of action that matter in such cases.

Imperfect duties too cannot be enforced by others, but for entirely different

reasons. In this case the impossibility of enforcement arises immediately from

the fact that required or prohibited acts are not sufficiently specified, so that

ways of enacting principles must be selected in context. Duties to seek

specific ends cannot lay down which ways of enacting them may be available or

effective for particular cases. Consequently Kant claims that beneficence is an

overarching principle for imperfect duties to others, but that it is not

possible to lay down in advance which specific helpful acts are required or

whose happiness should be fostered in specific situations. Least of all can

beneficence require agents to perform all the helpful acts that may

severally be available to them. Imperfect duties require agents to adopt

24

See for example Nicholas Wolterstorff, Justice: Rights and Wrongs, Princeton University

Press, Princeton, 2008. I think the focus mistaken because many rights cannot be claimed

unless specific others hold duties. In the case of liberty rights all others must hold the

counterpart duties, in the case of rights to goods and services specifiable others must hold

them. See Children's Rights and Children's Lives, 1988; Women's Rights: Whose

Obligations? 1996; The Dark Side of Human Rights, 2005.


12 certain types of end or policy, but underdetermine the choice of occasions and

means of pursuing them.

Equally, Kant’s overarching principle for imperfect duties to self, the duty

to develop one’s potential (hence duties of self development or self perfection),

also cannot be parsed as a determinate duty to develop particular capacities

to some prescribed level, but only as an indeterminate duty to cultivate powers

of mind and body that fit one to pursue the variety of aims one may have

reason to adopt.25

However, as Kant sees it, neither the pursuit of one’s

own happiness nor the development human potential in others is a matter

of duty. Our own happiness is not a matter of duty because we are

naturally inclined to pursue it; the development of others’ potential is not a

matter of duty because (less plausibly) this is not a task for others. By

the standards of today, my discussion of these topics was less than

complete, but fortunately others have since been more patient and done far

more to articulate the details of Kant’s account of the virtues. 26

In subsequent writing I repeatedly returned to these central Kantian themes.

Sometimes I took an explicitly Kantian approach; at other times I spared

readers the trek through textual details and tried to reach wider audiences. I

argued in many contexts that rights cannot be prior to the duties, and that to

assume otherwise is usually a matter of mere assertion or declaration rather

than argument. I also argued that adequate ethical theories must provide

an account of indeterminate ethical requirements—imperfect duties— and

that theories that offer an account only of action to which others can lay

rightful claim are not merely incomplete, but morally inadequate. I continue

to think that good character is not optional, but a matter of (imperfect) duty. 27

This remains an unpopular position. It seemingly invites criticisms both from

virtue ethicists, who think that it does not take virtue or good character

seriously enough, and from theorists of justice, who often deny that

25 MS 6 392 wrongly footnoted as 391 at A on P 89 26

See Mark Timmons, ed., Kant’s Metaphysics of Morals: Interpretative Essays, Oxford

University Press, 2002 ; Denis Lara, ed, Kant’s Metaphysics of Morals, CUP, 2010; and

Andreas Trampota, Oliver Sensen and Jens Timmermann, eds., Kant's Tugendlehre: A

Comprehensive Commentary, de Gruyter, ….

27 Over the years I have probably spent more time writing on these issues than on any

others. The closest to a unitary treatment is Towards Justice and Virtue: A constructive

account of practical reasoning, Cambridge University Press, 1996.


13 perfect duties are more fundamental than rights, and doubt whether anything

can be said about imperfect duties or about virtue. 28

6. Instrumental Rationality and Universalism

If the Categorical Imperative is to offer a test for the ethical status of maxims,

it must do more than articulate generic standards for practical reasoning; but

equally it cannot ignore those generic standards. Practical reasoning of all

types includes instrumental and consequential reasoning, but even taken

together they are insufficient for reasoned ethical claims. Kant notes that

instrumental reasoning yields definite conclusions only if combined with

other claims about what matters morally, for which no adequate reasons can

be given. Ethical systems they appeal to, the satisfaction of desire, the

demands of Church or State, the metaphysics of perfectionism, or the tyranny

of the majority or other extraneous standards rely on arbitrary assumptions and,

as he sees it, can reach only heteronomous, partially unreasoned accounts of

ethics. Reasoned ethical conclusions need more than instrumental

rationality.

This conclusion does not mean that ethical reasoning can ignore instrumental

rationality. Like other forms of practical reasoning, ethical reasoning is

not merely about wishing for certain things, but about commitment,

determination, willing, and in this case specifically about commitments,

determinations and willings to change the world (in some small part) to fit the

demands of ethical reasoning. Kant’s good will is not a matter of merely

wishing things to go well, but of seeking to take effective means to make them

go well when the situation allows. Two standards will be generically

required in all practical reasoning. The first is a commitment to take some

adequate means to enact one’s maxims if and when circumstances allow; the

second is a commitment to see any normal, predictable results of acting on

that maxim as one’s own doing. Although I did not then label these

commitments respectively instrumental and consequential, these labels are

apt, and both are part and parcel not only of ethical but of other forms of

practical reasoning.

Equally clearly, instrumental and consequential rationality are not sufficient for

ethical reasoning. Kant’s further and distinctive claim is that ethical 28

The latter is the central tenet of so-called ‘political liberalism’: a position that John Rawls

took up in his later work when he queried more comprehensive forms of liberalism. Political

Liberalism denies that we can say anything reasoned about the wider good for man, about

virtue, or about ethical matters other than justice.


14 principles must be principles that ‘can be willed as universal laws’. He

suggests that we can make the idea of willing a principle ‘as a universal law’

vivid by thinking of ourselves as acting in a world in which everyone acts

on the same principle, which can then be considered as if it were a natural law.29

This simple thought experiment reveals that, given the generic commitments to

instrumental and consequential rationality, certain rather tempting practical

principles cannot be willed as universal laws. Anybody who seeks to act on

principles that can serve for all must reject these principles of action. Anyone

who adopts such principles rejects universalism.

In Acting on Principle I tried to provide a systematic and lightly formalised

reading of Kant’s approach, but examples can make the central points more

intuitive. If I try to conceive of a possible world in which all adopt

maxims of deception or betrayal, the thought experiment undermines

the very trust needed for my own proposed deception or betrayal.

In such a world, acts and practices that support trust will wither or

collapse, undermining the possibility of taking advantage of that trust by

deception or betrayal. Analogously, if I try to conceive of a possible world in

which all adopt maxims of destruction or coercion, the thought experiment

undermines the forms of interaction and trust that my own proposed

deception or betrayal requires. The general point illustrated here is that

these principles cannot be universalised in any possible world, because in

such cases individuals’ instrumental reasoning conflicts with the

consequential reasoning to which universal action on that very maxim

would point. Some entirely coherent maxims cannot be willed as universal

laws. This conclusion was sufficient to show that the Categorical Imperative is

no empty formalism, and to establish a link between Kant’s metaethics and his

normative ethics.

It is easier to see how these types of argument work than it is to understand why

Kant presents them as the key to ethical reasoning. One way of making the

underlying thought more intuitive is to note that action on maxims that cannot

be universalised amounts to treating oneself as a special case:

If we now attend to ourselves in the transgression of any duty, we find

that we do not really will that our maxim should become a universal law,

since that is impossible for us, but that the opposite of our maxim should

instead remain a universal law, only we take the liberty of making an

29

This is the point of the Formula of the Law of Nature version of the Categorical

Imperative that invites agents to ask whether if the action they propose were to take place ‘by

a law of the nature’ they would regard it as possible to will it. Critique of Practical Reason,

5:69; Groundwork 4:421; AoP 62


15 exception to it for ourselves (or just for this once) to the advantage of our

inclination. G 4:424

By contrast, if we are to offer others reasons for acting on one or another

principle, we must propose principles of action that are intelligible to and can

be adopted by those others as maxims. The point is not that others must

actually be able to act on a proposed principle, since in many cases they may

find no occasion or means do so: Kant does not suggest that ethics is a matter

of action on principles that it is feasible for everybody to enact, let alone on

principles that it is feasible for everybody enact simultaneously or at the same

place.

Kant’s focus on failures of universalisability brings to the surface a

fundamental feature of his ethical reasoning. The thought that some maxims

are fit to be universal laws but others are not, provides a criterion in the first

place for ethically unacceptable action, that is for maxims that should be

rejected, whose adoption would violate duty. The ethical conclusions that

emerge from failed attempts to will maxims as universal laws distinguish

universalisable from non-universalisable maxims. However showing that a

maxim is universalisable does not show that it is a matter of duty. As Kant

sees it, establishing that a maxim could be universally adopted shows only that

acting on it is permissible. To show that a maxim is one of duty, it is necessary

to show both that the maxim can be universalised and that its rejection cannot.

Where a maxim and its rejection can both be universalised, neither rejecting

nor adopting that maxim is a matter of duty, and action on such maxims is

merely permissible. The formalisms are simple, if a bit tedious.

7. Contradictions in Conception, Contradictions in the Will

When I wrote Acting on Principle I was unsure how far these lines of argument

could be taken. How full an account of ethical duties could they support? It

was at least a start to identify arguments to show that if we are not to make

exceptions of our own case—if we are to treat others as persons like

ourselves—we must reject principles of aiming to destroy or threaten

agency, such as principles of deception and coercion. But could the

categorical imperative establish more? One way of exploring the limits of

these arguments was to consider the extent to which they support Kant’s

normative ethics: how good was the fit between the demands of the Categorical

Imperative and Kant’s classification of duties? Once again I stayed mainly

with the Formula of Universal Law.

Kant claims that the Formula of Universal Law can be used in two parallel

ways that are relevant to different ranges of duties. When applied to maxims of


16 action it may show up contradictions in conception: it may show that certain

principles cannot coherently be adopted by all. Contradictions in conception,

as the phrase suggests, arise when a maxim cannot be willed as a universal

law because the instrumental and consequential commitments that such willing

require contradict one another. When applied to maxims of ends the

Formula of Universal Law may show up what Kant terms contradictions in

the will. A world in which such maxims are universally adopted is not

inconceivable: Kant accepts that we can coherently think of worlds without

beneficence, and of worlds in which nobody develops their potential. But

it is not instrumentally rational for finite beings, who know that they will

sometimes be unable to pursue their aims unaided, to neglect either

beneficence or the development of their own potential. If they do so, they

fail to will conditions that will sometimes—or often!—be indispensable

for their own pursuit of ends. I noted that simpler prudential arguments might

also reach these conclusions: a line of thought since explored more

thoroughly by others.30

A world of agents who are indifferent to one

another’s needs, or to the development of human potential, is a world in which

the instrumental conditions of action will be systematically lacking or at risk.

Consequently, maxims of indifference to others’ needs or to the development of

human potential cannot be coherently adopted by all in a world of finite agents

whose natural endowments need the support of others’ help and acquired

competences.

8. Beyond Principles

Like anybody finishing an absorbing but focussed piece of writing, I knew that

Acting on Principle had left many relevant issues undiscussed. I had said

little about other formulations of the Categorical Imperative. I had not

discussed Kant’s claim that it is the supreme principle of practical reason. I

had said almost nothing about the idea of persons as Ends in themselves, which

is central to the Formula of the End in Itself version of the Categorical

Imperative, and did not consider the relationship of the Formula of

Universal Law to other versions of the Categorical Imperative. 31

Nor had I

discussed the basis of moral standing or reasons for taking one or another

view of the scope of principles of universal form. I had mentioned

30

Alison Hills, book or articles? 31

Others have dealt with this extensively, including several who worked under John Rawls.

Thomas E. Hill, Jnr., Christine Korsgaard, Andrews Reath, Kerstein add. It is difficult to set

a boundary here because there is so much on treating others as persons, on not treating them

as mere means. Some of it remotely Kantian.


17 autonomy only in passing, and then (like many others) misconstrued Kant’s

conception of autonomy as a version of individual autonomy, which has

been so central to late twentieth century liberal thinking. I said nothing

about dignity, which the human rights movement then and since sees as one of

Kant’s fundamental insights, and which is receiving a lot of attention from Kant

scholars today. 32

Although I said a bit about judgement, I did little more than

protect a space for saying something more substantive. And I hardly referred to

Kant’s views on justice and institutions, on politics, history or religion.

Of the themes explored in Acting on Principle, one turned out to be central

not only to later work on Kant but also to much of my subsequent work in

political philosophy and ethics. This was the contention that practical

reasoning should focus in the first place on agents and their duties, and treat

recipients or claimants and their rights as a derivative topic. I have explored

this line of thought in many contexts, often with little reference to Kant’s

specific treatment of these themes. These explorations proved uphill

work in a culture that increasingly proclaims rights, while ignoring the

duties without which those rights remain gestural and unrealisable. However,

for present purposes I shall comment briefly on other, more Kantian

continuations of the lines of thought explored in Acting on Principle .

9. Upstream and Downstream

It would be absurd to think that a work on connections between reason and

action could be completed in any tidy or final way. But while I never aimed

to cover the topic exhaustively, I have followed the starting points explored

in Acting on Principle in a number of directions.

Much of the more Kantian parts of my subsequent work addressed two ranges

of questions. In the first place I tried to make sense of Kant’s claim that the

Categorical Imperative, in its various formulations, is the ‘supreme principle of

practical reason’. Secondly, I tried to show how the principles of duty

identified by the Categorical Imperative can be deployed in action. I have often

thought of these two lines of inquiry as exploring the arguments that lie

respectively upstream and downstream from the topics discussed in Acting on

Principle. Upstream issues include Kant’s spectacular claim that a vindication

32

But see Oliver Sensen, Kant on Human Dignity, De Gruyter, 2011. Samuel J Kerstein is the

book out?


18 of reason is possible; downstream issues include consideration of ways in

which normative standards, among them principles of duty, can inform practical

judgement and be used to shape action.

Upstream lie the headwaters of Kant’s critical philosophy, and in particular his

challenging thought that critique and vindication of reason are possible: an

enterprise that many have condemned as pointless, because evidently circular.

As I explored Kant’s discussions on reason, which lie scattered throughout his

writings, it became increasingly clear to me that his underlying thought is that

reasons must be the sorts of things that agents can give and receive, accept or

reject. Reasoning must be fit to connect a plurality of (potential) reasoners,

whether in communication or in collaboration. This is why Kant repeatedly

deploys metaphors of construction and of public reasoning to suggest that

anything that is to count as reasoning must respect the necessary conditions for

the possibility of communication or collaboration, and also explains the

seeming oddity that many of his comments on the vindication of reason are

located in his writings on culture, politics and history.

During the last half century, many other philosophers, most prominently Jürgen

Habermas and John Rawls, have explored lines of thought that are at least

superficially similar to Kant’s, and have often spoken of their positions as

Kantian. However twentieth century accounts of public reason and of the

construction of reason all, so far as I know, depend in part on considerations

that Kant would have seen as arbitrary, such as preferences, or the search for

agreement, or procedures of discourse. They see public reasoning as a matter

of proceeding in ways that support freely entered discourse and agreement,

and are typically closely allied with conceptions of democratic reasoning. In

Kant’s terms, however, they propose heteronomous conceptions of public

reason.

As I worked through Kant’s scattered discussions of the vindication of

reason I came to see that his account of public reason was more austere, but

also more interesting. 33

What counted as reasoning for Kant was not freely 33

Constructivisms in Ethics 1988-89 ; Enlightenment as Autonomy: Kant's Vindication of

Reason 1990; Vindicating Reason 1992.; Kommunikative Rationalität und praktische

Vernunft, 1993; Vier Modelle praktischer Vernunft, 1994; Practical Reason and

Possible Community: A Reply to Jean-Marc Ferry, 1994; Political Liberalism and Public

Reason: A Critical Notice of John Rawls, Political Liberalism, 1998; Within the Limits of

Reason 1997; Kantian Constructivisms, 1999 ; Kant’s Conception of Public Reason, 2001;

Autonomy and the Fact of Reason in the Kritik der praktischen Vernunft, 30-41, 2002;

Constructivism in Rawls and Kant 2003; Constructivism vs. Contractualism, 2003; Kant:

Rationality as Practical Reason 2004; Autonomy, Plurality and Public Reason, 2004 ; Self-

Legislation, Autonomy and the Form of Law, 2004, 13-26; Orientation in Thinking:

Geographical Problems, Political Solutions, 2011.


19 conducted discourse that aims at or secures agreement—although reasoned

discourse, like reasoned action, has indeed to be freely chosen, and may if

things go well secure agreement—but freely conducted discourse that is

guided by principles that could be principles for all, so can be followed by any

potential audience or collaborator. The claim that the Categorical Imperative is

the supreme principle of practical reasoning is ultimately no more—but also

no less—than the modal claim that we do not reason with others unless the

considerations we offer them are ones that (we take it) they can follow.

That is why the basic principle of practical reasoning demands action on

principles that can be followed by all: anything else may fail to offer others

anything that they can view as a reason. Despite the sometimes bombastic

rhetoric Kant sometimes uses, his modal conception of the demands of reason

turns out to be quite modest.

During a long period of intermittent work on Kant’s conception of practical

reason I came to realise—with many thanks to discussion of the topic by

Thomas E. Hill, Jnr—how closely Kant’s account of autonomy was connected

to his account of practical reason. Kantian autonomy cannot be identified

with any of conceptions of autonomy as individual independence that became

widely influential in the later decades of the twentieth century. 34

In retrospect

I think that I should have noticed this far sooner, for the textual evidence is

substantial. In Groundwork Kant articulated the so-called Formula of

Autonomy version of the Categorical Imperative as ‘the idea of the will of

every rational being as a will giving universal law’: it is clearly very close to

the Formula of Universal Law formulation of the Categorical Imperative. In

the Conflict of the Faculties he remarks that “the power to judge

autonomously—that is freely (according to principles of thought in general) –is

called reason”35

, while in the Critique of Practical Reason, he asserts that

“the moral law expresses nothing other than the autonomy of pure practical

reason.” 36

In short, what interests Kant is the autonomy of reasoning, rather

34

See Allen Wood’s useful list of versions of the Formula of Autonomy in Kant's

Ethical Thought, 1999, 163-4, also .the references in note 33, Autonomy and Trust in

Bioethics, 2002 and Autonomy: The Emperor’s New Clothes, 2003.

35

Immanuel Kant, The Conflict of the Faculties, 7:27.

36

Immanuel Kant, Critique of Practical Reason, Theorem IV, 5:33.


20 than that of individual reasoners. He does not construe autonomy as

some version of individual independence. He does not speak of autonomous

persons or selves. The central position he accords to a conception of

principled autonomy depends on seeing it matter of acting on principles that

combine law like form and universal scope, so on reasoned principles. 37

Downstream considerations have also proved absorbing. The indeterminacy

of principles of action that led Anscombe (and many other proponents of virtue

ethics) to conclude that principles lack practical implications provides the

context for Kant’s discussion of practical judgement and the enactment of

principles. Kant insisted that there are and can be no complete rules for the

exercise of judgement— the very point that Anscombe concludes will derail

any ethics of principles. Judgement, including practical judgement does not

follow any algorithm; rather it is a “special talent that cannot be taught but

only practiced.” 38

Practical judgement is exercised by bringing reasoning , and

with it principles of duty, to bear on action, but it does so not by subsuming

particulars under those principles, but rather by enacting or instantiating them

in particular situations. Kant discussed this task and its difficulties in several of

his later works, but never assumes that acting on principles of duty will be a

matter of applying principles to already existing particulars. Rather, the

exercise of practical judgment is a matter of finding at least some way to respect

and enact a variety of principles of duty—and other practical demands— in

particular situations. 39

The fact that we often need to enact a plurality of different practical principles,

including a plurality of principles of duty, can be problematic. For the most part

it is not impossible to combine truth-telling with refraining from injury, keeping

promises with helping others. But there are cases when one or more

practical, including ethical, demand is hard or impossible to reconcile with

others. In some cases, practical judgement may fail to discover any adequate

way of meeting all ethical requirements, or of meeting them well, and some of

37

See in particular Self-Legislation, Autonomy and the Form of Law, 2004. 38

CPR A133 B172

39

See The Power of Example, 1986; Principles, Judgment and Institutions, 1997; Instituting

Principles: between Duty and Action, 1997, revised 2002; Practical Principles and

Practical Judgement, 2001; Modern Moral Philosophy and the Problem of Relevant

Descriptions, 2004; Experts, Practitioners and Practical Judgement, 2007; Normativity and

Practical Judgement 2007; Applied Ethics: Naturalism, Normativity and Public Policy, 2009.


21 them may be short-changed. This seems to me no more—and no less—than a

reflection of the way things are, and not a demonstration of the inadequacy of

Kant’s ethics, or of any other ethical system. The possibility of searing

dilemmas or conflicting ethical claims is real, and remedies are incomplete. But

there are partial remedies. They include the construction and maintenance of

effective social and political practices that reduce and head off the number of

conflicting demands. They also include the practices of self discipline that Kant

discusses under the heading of ethical duties of omission. And they include

practices for acknowledging the claims of unmet—sometimes unmeetable—

claims of duties. The claims of unmet, sometimes unmeetable duties has been

illuminated by seeing them as remainders that represent undischarged ethical

demands. However, remainders are often seen rather narrowly as calling for

certain attitudes to unmet duties—classically regret remorse and guilt.

However, we need a broader account of ways of dealing with the claims of

unmet duties, that covers action as well as attitudes. Unmet obligations may

call for a gamut of action ranging from apology, punishment and

forgiveness, to ways of making amends and wider forms of reconsideration and

reconciliation.

Documents

Intro for Acting on Principle 5