1
Internalism Explained*
Ralph Wedgwood
1. The word ‘rational’ is used in many ways. But when the word is used in the way that is
most common among philosophers, the following intuition seems compelling.
Consider two possible worlds, w1 and w2. In both worlds, you have exactly the same
experiences, apparent memories, and intuitions, and in both worlds you go through exactly the
same processes of reasoning, and form exactly the same beliefs. In this case, it seems, exactly the
same beliefs are rational in both worlds, and exactly the same beliefs are irrational in both worlds.
Now suppose that in w1 you are bedevilled by an evil demon who ensures that many of your
experiences are misleading, with the result that many of the beliefs that you hold in w1 are false. In
w2, on the other hand, almost all your experiences are veridical, with the result that almost all the
beliefs that you hold in w2 are true. Intuitively, this makes no difference at all. Exactly the same
beliefs are rational and irrational in both worlds.1
This intuition seems to support an “internalist” conception of rational belief. According to
this conception, the rationality of a belief supervenes purely on “internal facts” about the thinker’s
mental states — in this example, on facts that hold in both these two possible worlds w1 and w2,
not on facts about the external world that vary between w1 and w2.
Moreover, this seems to be a completely general feature of rationality, since a parallel
claim seems plausible with respect to rational decisions or choices as well. When we assess a
decision as rational or irrational, we are assessing it on the basis of its relation to the agent’s
beliefs, desires, and other such mental states — not on the basis of its relation to facts about the
external world that could vary while those mental states remained unchanged.2 This also seems to
be a special feature of rationality, in contrast to other ways of evaluating beliefs and decisions. All
the other ways of evaluating beliefs and decisions — for example, as “correct” or “incorrect”,
“advisable” or “inadvisable”, and so on — are externalist evaluations. What is distinctive of
“rationality” (at least as the term is most commonly used by philosophers) is that it is an internalist
2
evaluation.
Thus, internalism with respect to rationality seems to have considerable intuitive support.
However, this intuitive support would be undermined if we cannot give an adequate explanation
of internalism. Specifically, we must explain exactly which facts about a thinker count as these
“internal facts” upon which the rationality of a belief or decision supervenes. We must also explain
why rationality should supervene on internal facts in this way.
According to the standard version of internalism, these “internal facts” are defined as
those facts that one is in a position to know “by reflection alone”.3 Many proponents of this
standard version of internalism attempt to explain why internalism should be true by claiming that
to say that a belief or decision is “rational” is just to say that in holding that belief or making that
decision, the thinker is proceeding in a “cognitively blameless” fashion.4
In my view, this standard version of internalism is open to fatal objections. The claim that
rationality is just a matter of “cognitive blamelessness” does not provide an adequate explanation
of internalism;5 and anyway, once “internal facts” are defined in this way, internalism is just not
true.6 But I shall not rehearse these objections to the standard version of internalism here. Instead,
I shall just outline a new version of internalism, which provides a better explanation of what
exactly these “internal facts” are, and of why it is that the rationality of beliefs and decisions
supervenes on them.
As I have remarked, if internalism is true, then it applies just as much to rational decision
as to rational belief. However, the question of whether or not internalism is true has chiefly been
discussed by epistemologists, rather than by theorists of practical reason. For this reason, I shall
focus exclusively on rational belief here, and ignore rational decision altogether. But this is a
purely stylistic choice on my part. All my arguments would work just as well if they focused on
rational decision instead of rational belief.
2. In what follows, I shall use the phrase ‘belief revision’ broadly, so that it includes not only
coming to hold a new belief, but also abandoning or reaffirming an old belief. Even if a belief is
not currently being formed or reaffirmed, one may still hold that belief as a background belief —
3
that is, as a standing belief stored in propositional memory. To simplify the discussion here, I shall
focus purely on what it is for a belief revision to be rational, not on what it is for background
beliefs to be rational.7
It is often claimed that “beliefs aim at the truth.” Clearly, this claim is metaphorical, and it
is obscure exactly how this metaphor is to be interpreted. But I shall assume that there is some
reasonable interpretation on which it is true to say that whenever one revises one’s beliefs, one is
thereby pursuing some aim. It does not matter for my purposes exactly what aim one is thereby
pursuing. But to fix ideas, suppose that whenever one revises one’s beliefs, by forming or
reaffirming or abandoning one’s belief in p, one is pursuing — that is, trying to achieve — the aim
of believing p if and only if p is true.
How is one to try to achieve this aim, of believing this proposition p if and only if p is
true? To try to achieve this aim, one must possess and exercise certain general abilities or
capacities. Specifically, each of the abilities that one must exercise, in order to pursue this aim, is
an ability to revise one’s beliefs in a certain way whenever one is in a certain related condition.
For example, one such ability might be the ability to come to believe p whenever one sees with
one’s own eyes that p is the case.
To count as “exercising” this ability, however, it would surely not be enough if it were
simply a fluke that one came to believe p at the same time as one sees with one’s own eyes that p
is the case. It must also be the case that one comes to believe p precisely because one sees with
one’s own eyes that p is the case, and because this belief has the same content — p — as what
one sees with one’s own eyes.8 In general, when one exercises such an ability, one revises one’s
beliefs in the relevant way in response to the fact that one is in the relevant related condition. As I
shall use the phrase, when one exercises an ability of this sort, one is following, or being guided
by, a rule — specifically, by the rule that directs one to revise one’s beliefs in that way in response
to the fact that one is in the relevant condition.
But which rules is one to follow, in order to pursue the aim of believing p (the proposition
in question) if and only if p is true? It would be unhelpful here to say that one should just follow
the rule: believe p if and only if p is true. One could not try to achieve this aim by means of
4
following this rule. Following this rule just is achieving the aim. We need to see what rules one
could follow in order to try to achieve that aim. That is, we need to see what rules there are such
that (if all goes well) one could achieve that aim by means of following those rules.
Among the rules that one could follow will be certain basic rules. “Basic rules” are
analogous to “basic actions”. Basic actions are actions that one performs, but not by means of
performing any other actions.9 Basic rules are rules that one follows directly, not by means of
following any other rules. If one follows any rules at all, then one must follow some basic rules.
Now suppose that there are some such basic rules that one “should” follow, in revising
one’s beliefs, in order to pursue the aim of believing p (the proposition in question) if and only if p
is true. Admittedly, it is not clear what sense of ‘should’ is being used here. I shall return to this
question later. But suppose that there are some such basic rules. Then, I propose, these are the
rules that it is rational to follow. A belief revision is rational just in case it directly results from
one’s following some of these basic rules.10
This proposal may explain another common intuition about rationality. Intuitively, the fact
that one’s belief is true, or counts as knowledge, may be purely a matter of luck; but the fact that
one’s belief is rational cannot be purely a matter of luck — it is something that lies within one’s
control. One way of explaining this distinction, between what “may be a matter of luck” and what
“lies within one’s control”, is based on the notion of “basic actions”. If one is able to perform a
certain basic action, then whether or not one performs that basic action lies within one’s control.
On the other hand, even if one is able to bring about a certain further result by means of
performing that basic action, whether or not one actually brings about that result may be purely a
matter of luck.11 So if rationality is a matter of following basic rules, this could explain why
rationality cannot be a matter of luck.
According to my definition, a “non-basic rule” is a rule that one follows by means of
following some other rule. For example, consider the following rule: “Add salt when the water
starts boiling”. If one follows this rule, then one adds salt to the boiling water, precisely because
the water is boiling. One’s adding the salt is explained by — is a response to — the fact that the
water is boiling. However, it may be that the process whereby one’s action of adding the salt is
5
explained by, or responds to, the fact that the water is boiling can itself be analysed, even at the
folk-psychological level of explanation, into a series of sub-processes. For example, perhaps the
proximate explanation of one’s attempt to add the salt is not the fact that the water is boiling, but
rather one’s belief that the water is boiling. Similarly, perhaps the proximate explanation of one’s
belief that the water is boiling is not the fact that the water is boiling, but one’s having an
experience that represents the water as boiling. In forming this belief in response to having this
experience, one is exercising an ability or following a rule — specifically, a rule that directs one to
form the belief that the water is boiling, in response to having an experience that represents the
water as boiling. Similarly, when one decides to add the salt, in response to one’s forming the
belief that the water is boiling, one is following some other rule (or set of rules). In this way, the
process of one’s following the rule “Add salt when the water starts boiling” is constituted by a
series of sub-processes, including (among other things) the processes of one’s following these
other rules. So, the rule “Add salt when the water starts boiling” is not a basic rule. One follows
this rule by means of following other rules.
On the other hand, for example, consider the rule “Believe p, if one has an experience as
of p’s being the case, and one has no special reason to distrust one’s experiences”. If this is a
“basic” rule, then the process of one’s following this rule cannot be analysed, at the folk-
psychological level of explanation, into a series of sub-processes that include one’s following any
other rule. No doubt, at a “subpersonal” level of explanation, the process of one’s following this
rule can be analysed into numerous sub-processes, perhaps involving various subpersonal
modules’ computing various algorithms. But this is not the sort of explanation that we are
concerned with here. We are concerned with folk-psychological explanations. By this, I mean
explanations that have the following two features. First, these explanations refer to mental states,
like beliefs and desires, of the sort that are referred to in everyday psychological discourse.
Second, these explanations make a certain way of revising one’s beliefs intelligible, as the sort of
thing that we can readily imagine someone in the circumstances finding persuasive or compelling.
At this level of explanation, one’s having an experience that represents the water as boiling is the
proximate explanation of one’s forming the belief that the water is boiling.12 There are no
6
intervening steps that can be captured at this folk-psychological level of explanation.13
When a correct folk-psychological explanation of a belief revision includes all the
intervening steps that can be captured at the folk-psychological level, let us say that it is a “fully-
articulated” explanation. If one follows a basic rule, and that rule directs one to revise one’s
beliefs in a certain way in response to the fact that one is in a certain condition, then a fully-
articulated explanation of that belief revision will identify that fact as the proximate explanation of
that belief revision.
3. As I shall now argue, whenever a thinker revises her beliefs through following a rule, a
fully-articulated folk-psychological explanation of that belief revision will always identify the
proximate explanation of that belief revision with some fact about the thinker’s mental states.
Some philosophers may object that it is surely an empirical question whether it is ever
correct to explain a belief revision directly on the basis of an external, non-mental fact. Certainly,
it is an empirical question exactly what sorts of facts can explain a belief revision. But there may
still be certain philosophical limits on which such explanations count as correct fully-articulated
folk-psychological explanations of the relevant sort.
Thus, I am not denying that the formation of a belief can ever be explained in terms of
non-mental facts. There may certainly be correct non-folk-psychological explanations that identify
a non-mental fact as the proximate explanation of why one formed a belief. For example, scientists
might discover that a certain brain state always causes the thinker to believe that he is about to
die. But this would not be a folk-psychological explanation of the relevant sort. It would not make
one’s forming this belief intelligible as the sort of thing that we can readily imagine someone in the
circumstances finding persuasive or compelling. From a folk-psychological perspective, the belief
would still seem opaque and hard to understand.
I am also not denying that it could be the case, for example, that someone’s coming to
believe that I once lived in Malaysia may be explained by the fact that I told her that I once lived
in Malaysia — which is presumably a fact about the external world. This explanation may be quite
correct. It is just not a “fully-articulated” explanation. It is intuitively clear that if this is a correct
7
folk-psychological explanation, there must also be a more detailed correct folk-psychological
explanation, in which the link between my telling her that I once lived in Malaysia and her coming
to believe that I once lived in Malaysia is mediated by intervening facts about her mental states.
Perhaps, for example, in this more detailed explanation, her coming to believe that I once lived in
Malaysia is directly explained by her having the belief that I told her that I once lived in Malaysia
(along with the fact that she had no reason to regard my assertion as insincere, and every reason
to believe that if my assertion is sincere it is probably true); and this belief (that I told her that I
once lived in Malaysia) is itself explained by her having an experience as of my telling her that I
once lived in Malaysia, which is in turn explained by my actually telling her that I once lived in
Malaysia.
Suppose that I claim that someone’s forming a certain belief is explained by a certain
external fact, in a context in which it is unclear how there could be any more detailed correct
explanation in which the link between that external fact and the formation of the belief is mediated
by intervening facts about the believer’s mental states. For example, suppose that I say, “I once
lived in Malaysia, so Janet Reno believes that I once lived in Malaysia”. You would be reluctant
to accept this explanation unless it could be made clear how this link, between the fact that I once
lived in Malaysia and Reno’s believing that I once lived in Malaysia, could be mediated by
intervening facts about Reno’s mental states. You would want to ask, “But how does Reno know
anything about you at all? Did you meet her and talk about your childhood? Did she ask the FBI
to run a background check on you for some reason? Or what?”
This point applies even to perceptual beliefs. Suppose that I claim “Sarah believes that the
flowers in front of her are pink because the flowers are pink”, while simultaneously also claiming
that this link, between the flowers’ being pink and Sarah’s believing that the flowers are pink, is
not mediated by any further facts about Sarah’s mental states. If these claims are correct, then
either Sarah has no perception or experience that represents the flowers in any way, or else, if she
has such a perception or experience, it makes absolutely no difference to whether or not she forms
this belief. But then how can this explanation make this belief intelligible, as the sort of thing that
we can readily imagine someone finding persuasive or compelling in the circumstances?14 How
8
exactly does the mere fact that the flowers are pink make it persuasive or compelling for Sarah to
form precisely this belief, rather than some other belief, or indeed any belief at all? If we want to
know why Sarah found it compelling to form this belief, the most plausible answer is surely to say
something like: “Sarah found it compelling to form this belief because she could see that those
flowers were pink (or: because it looked to her as though those flowers were pink)”.15
In general, then, an explanation of a belief revision that appeals to an external fact can be a
correct folk-psychological explanation only if there is also a “fully-articulated” explanation in
which the link between that external fact and that belief revision is mediated by intervening facts
about the believer’s mental states. So, in any “fully-articulated” explanation, the proximate
explanation of the belief revision is not an external fact, but some fact about the believer’s mental
states.16 Typically, this proximate explanation consists of the experiences, apparent memories,
intuitions or beliefs that are the reasons for which one revised one’s beliefs in that way, along with
the absence from one’s set of mental states of certain sorts of defeating or countervailing reasons.
It is striking how sharply beliefs contrast with experiences on this point. The proximate folk-
psychological explanation of an experience typically does involve an external fact: “The body was
lying there right in front of her, in broad daylight, and her eyes were wide open, so of course she
saw it”. These explanations never appeal to any “reason” for which one has that experience.
There is one final move available to the externalist here. The externalist may concede that
correct fully-articulated folk-psychological explanations always identify the proximate explanation
of the formation of a belief with some fact concerning one’s mental states. But he may insist that
in some cases, these mental states include so-called “factive mental states”, like the state of
knowing that p is the case, or seeing that p is the case. The defining mark of a “factive” mental
state is that it must consist in standing in some relation to a true proposition. If one knows or sees
that p is the case, then p must actually be true. Internalists have always excluded such “factive
mental states” from the inventory of “internal” states that determine the rationality of beliefs.17
For many purposes, it is harmless to say that these “factive states” are mental states.
However, such factive states cannot be the proximate explanations of the formation of beliefs in
correct fully-articulated folk-psychological explanations. Suppose that we want to explain why a
9
thinker comes to believe p. One candidate explanation identifies the proximate explanation of the
thinker’s coming to believe p as the fact that the thinker knows q; another candidate explanation
identifies this proximate explanation as the fact that she believes q. In this case, if either
explanation is correct, it is the second explanation not the first. It is highly plausible that the
thinker’s knowing q is partially constituted by the thinker’s believing q; and if the thinker had
merely believed q, and not known q, she would still have come to believe p, in exactly the same
way that we are trying to explain. So, the thinker’s knowing q has the effect of producing the
belief in p only because her knowing q is partially constituted by her believing q. This seems to
show that it is the thinker’s believing q, not her knowing q, that really explains her having the
belief in p that we are trying to explain.18
This is an application of a plausible general principle about explanation. If one fact is
partially constituted by a second,19 and a certain effect would still have been produced even if the
second fact had obtained and the first fact had not, then if either fact explains that effect, it is the
second fact rather than the first. The first fact contains elements that are irrelevant to explaining
the effect; it is the second fact that really does the work in explaining that effect.20
A parallel argument applies to other “factive states” as well.21 It seems then that correct
fully-articulated folk-psychological explanations always identify the proximate explanation of a
belief revision with some fact about the believer’s mental states, not including “factive mental
states” of the kind that I have just discussed. Let us call such facts “internal facts”.
As I explained earlier, if one follows a basic rule, and that rule directs one to revise one’s
beliefs in a certain way in response to a certain fact, then a correct fully-articulated explanation
will identify that fact as the proximate explanation of that belief revision. So, these basic rules will
only ever direct one to revise one’s beliefs in response to such “internal facts”. Let us also say that
the fact that a certain belief revision is explained by certain “internal facts” of this kind is itself an
“internal fact” of this kind. Then, for each of these rules, whether or not one is following the rule,
on a given occasion, depends purely on these “internal facts” about one’s mental states.
4. I have not yet explained why internalism is true. Even if, as I have just argued, “basic
rules” only ever direct one to revise one’s beliefs in response to “internal facts” of this kind, it
10
could still be that what makes these rules the rules that one “rationally should” follow is some
external fact about these rules — such as the fact that these rules are highly reliable at yielding
true beliefs and avoiding false beliefs. As I shall put it, this position would combine “belief
internalism” with “rule externalism”.
The closest parallel in the literature to my argument for belief internalism is the “refutation
of belief externalism” that is given by John Pollock and Joseph Cruz.22 After giving their
refutation of belief externalism, Pollock and Cruz turn to “rule externalism”. Specifically, they
consider the claim that the “basic rules” that we rationally should follow are those rules that are as
a matter of fact most reliable at reaching the truth, “regardless of whether we know these facts
about reliability”. They reply to this claim as follows:23
What could the point of this claim be? It cannot be taken as a recommendation about howto reason, because it is not a recommendation anyone could follow. We can only alter ourreasoning in response to facts about reliability if we are apprised of those facts.
Here, Pollock and Cruz seem to infer from the premiss ‘We can only alter our reasoning in
response to facts about reliability if we are apprised of those facts’ to the conclusion ‘No one can
follow the recommendation to reason in the most reliable way’. If this inference were valid, then
we could also infer from the premiss ‘We can only add salt to the water in response to the fact
that the water has started boiling if we are apprised of the fact that the water has started boiling’
to the conclusion ‘No one can follow the recommendation to add salt when the water starts
boiling’. But that inference cannot be valid. Even if the premiss is true, it is obviously possible to
follow the recommendation to add salt when the water starts boiling.
Pollock and Cruz seem to assume that a “recommendation that someone could follow”
must be a recommendation that we can always follow whenever it applies to us. But how many
recommendations are there of which that is true? Take the simplest of logical precepts: “From
‘p & q’ infer p”. We are not always able to follow this recommendation: some conjunctions are
too complex to be recognized as such; or, more simply, we might suddenly die, or go insane, or
fall asleep, before we have completed the inference; and so on. The only general recommendations
that we can always follow, whenever they apply to us, are recommendations that are specifically
11
restricted to those cases in which we are able to follow them — for example, “From ‘p & q’ infer
p, whenever you are able to follow this recommendation”. But recommendations of this sort
might be externalist recommendations, such as “Form your beliefs by reliable methods, whenever
you are able to follow this recommendation”. So the idea of a recommendation that we are always
able to follow does not support internalism. To support internalism, we need a different idea —
the idea of a recommendation that we follow directly, not by means of following any other
recommendation.
As I explained earlier, when one follows a rule, one is exercising an ability that one has, to
revise one’s beliefs in a certain way in response to being in a certain corresponding condition;
one’s disposition to revise one’s beliefs in that way, in response to being in that condition, has
been activated. If the rule is a basic rule, then the activation of that disposition is not mediated by
one’s following any other rule. But we can still ask what explains why one has that disposition at
that time.
I proposed earlier that the rules that it is rational for one to follow, in revising one’s beliefs
with respect to a proposition p, are precisely those rules that one “should” follow in order to
pursue the aim of believing p if and only if p is true. Intuitively, however, it seems that for
rationality, it is not enough if it is simply a fluke that one is disposed to follow the rules that one
should follow. It must be the case that one is disposed to follow these rules precisely because they
are the rules that one should follow: this disposition must be sensitive to, or be explained by, the
fact that these are the rules that one should follow.24
Under favourable conditions, one’s disposition to follow certain rules, in revising one’s
beliefs about p, may indeed be sensitive to the fact that these rules are reliable ways of reaching
the truth. Intuitively, however, one’s disposition to follow these rules could not be directly
sensitive to that fact. It could only be indirectly sensitive to that fact, by means of being sensitive
to some other fact. This is because a similar point applies to the explanation of one’s disposition
to follow a certain rule as to the explanation of one’s forming a certain belief. A correct fully-
articulated folk-psychological explanation of this disposition will always identify, as the proximate
explanation of that disposition, some internal fact concerning one’s mental states.25 Perhaps, for
12
* An earlier draft of this paper was presented to the philosophy department at Stanford University.I thank the members of that audience, and also Steve Yablo, Tim Williamson, James Pryor, JohnGibbons, and Alexander Bird, for helpful comments.1 Several epistemologists have proposed theories that are incompatible with this intuition (at leastassuming that a belief is “justified” if and only if it is rational). See e.g. Timothy Williamson,“Knowledge as Evidence”, Mind 106 (1997), 717–41, and Alvin Goldman, “What is JustifiedBelief?” in George Pappas, ed., Justification and Knowledge: New Studies in Epistemology(Dordrecht: Reidel, 1979), and “Strong and Weak Justification”, Philosophical Perspectives 2(1988), 51–69. Even Goldman, however, felt the pull of this intuition, when he suggested that therules that we ought to follow, in forming and revising our beliefs, are the rules that are mostreliable in “normal worlds” — that is, those worlds in which our general beliefs are true; see his
example, one is disposed to follow a certain rule because one believes that the rule is a reliable
way to reach the truth. Or perhaps one is disposed to follow the rule simply because one
possesses certain concepts, and a disposition to follow the rule is a necessary consequence of
possessing those concepts. At all events, the proximate explanation of this disposition is an
internal fact, not a fact about the external world.
This suggests the following interpretation of the term ‘should’, as it appeared in my
proposal that the rules that it is rational for you to follow, in revising your beliefs with respect to
p, are the rules that you “should” follow, given the aim of believing p if and only if p is true. You
are following the rules that you “should” follow, in this sense, only if the fact that you “should”
follow these rules is the proximate explanation of your disposition to follow them.26 If the
proximate explanation of your disposition to follow a rule must itself be an “internal” fact
concerning your mental states, then the fact that you “should” follow a certain rule must also be
such an “internal” fact. Let us also say that the fact that a disposition to follow a rule is explained
by such an “internal fact” is itself an “internal fact”. Then, we should conclude, if my proposal is
correct, the rationality of a belief revision supervenes purely on “internal facts” about the believer.
5. Thus, my proposal about the rules of rational belief revision, along with these theses about
folk-psychological explanation, explains which facts about a thinker count as the relevant “internal
facts”, and also explains why it is that the rationality of belief revisions supervenes on them. The
intuitive support that internalism enjoys is not undermined: internalism can be adequately
explained.
13
Epistemology and Cognition (Cambridge, Mass.: Harvard University Press, 1986), pp. 107–09.2 This formulation is designed to be compatible with externalism with respect to mental content —that is, with the view that the content of a thinker’s mental states does not supervene on narrow,intrinsic properties of the thinker alone, but may also depend on the thinker’s relations to herenvironment. For the classic argument in favour of externalism with respect to mental content, seeHilary Putnam, “The Meaning of ‘Meaning’”, reprinted in his Mind, Language, and Reality(Cambridge: Cambridge University Press, 1975).3 James Pryor, “Highlights of Recent Epistemology”, British Journal for the Philosophy ofScience 52 (March 2001), § 3.1. This definition of “internal facts” is not universally accepted. In“Internalism Defended” (forthcoming in an anthology on internalism and externalism edited byHilary Kornblith), Earl Conee and Richard Feldman define internalism as the view that therationality of a belief is determined by the believer’s “internal states”. They distinguish twoaccounts of these “internal states”: “accessibilism” defines the believer’s “internal states” as statesto which the believer has a special sort of epistemic access, while “mentalism” defines these“internal states” simply as the believer’s mental states. Conee and Feldman tentatively prefermentalism over accessibilism. Ernest Sosa draws a similar contrast between “Chisholmianinternalism” and “Cartesian internalism” in “Skepticism and the Internal / External Divide”, TheBlackwell Guide to Epistemology, ed. J. Greco and E. Sosa (Oxford: Blackwell, 1999), p. 147.4 For an illuminating discussion of the possibility of explaining internalism in this way, see WilliamAlston, Epistemic Justification (Ithaca, New York: Cornell University Press, 1989), Essay 8.5 It is doubtful whether this claim — that epistemic rationality or justification is just a matter of“epistemic blamelessness” — entails that internalism is true; on this point, see especially Goldman,“Internalism Exposed”, Journal of Philosophy 96 (1999), 271–93. Moreover, this claim actuallyseems to be false. This is because the claim overlooks the distinction between justification andexcuse. Even if an action is not justified at all, the action can be “blameless” because it isexcusable. Similarly, a belief can be blameless because it is excusable, even if it is not in any wayjustified or rational. For this point, see James Pryor, “Highlights of Recent Epistemology”, §4.3.Alvin Plantinga also makes much of the fact that some very strange epistemic practices may be“blameless”, if they result from brain damage or the like; see his Warrant: The Current Debate(New York: Oxford University Press, 1993), p. 39.6 The main reason for this is that, for every member of the set of facts that determine whether ornot a belief is rational, it is not a necessary feature of that fact that one is in a position to knowthat fact by reflection alone. (For some powerful arguments for this point, see Williamson,“Cognitive Homelessness”, Journal of Philosophy 93 (1996), 554–73, and “Evidence andScepticism”, Philosophy and Phenomenological Research 60 (2000), 613–28.) Thus, thefollowing case seems possible. In both w1 and w2, you believe p on the basis of certain reasons, butin w1 you are in a position to know by reflection alone that you believe p on the basis of thosereasons, while in w2 you are not in a position to know this; otherwise you are in just the samemental states in both w1 and w2. So, according to the standard version of internalism, the fact thatyou believe p on the basis of these reasons may be part of what makes the belief rational in w1, butit cannot be part of what makes the belief rational in w2. Hence, this version of internalism mustsay that it could be the case that this belief is rational in w1 but not rational in w2. But then the factthat you are in a position to know about the basis for your belief in w1 is itself one of the facts onwhich the rationality of the belief supervenes. The possibility of higher-order access to these factsis not merely a consequence of these facts’ determining the belief’s rationality. It is itself one ofthe facts that determines the belief’s rationality. So, if there is a set of facts that determinewhether or not a belief is rational, then for every fact F that belongs to that set of facts, that setmust include, not just F, but also the fact that one is in a position to know F by reflection alone,the fact that one is in a position to know by reflection alone that one is in a position to know F byreflection alone, and so on, ad infinitum. This makes it doubtful whether there is any set of factsthat determines whether or not any belief is rational at all.7 The notions of a rational background belief and of a rational belief revision are connected in thefollowing way. So long as it is rational for a thinker to hold a certain background belief, then nobelief revision on the part of the thinker will be irrational merely because it is based on thatbackground belief. I focus on belief revisions here, not because I doubt that internalism is true of
14
background beliefs as well, but purely because background beliefs introduce further complicationsthat I cannot pause to analyse here. For an illuminating discussion of when such backgroundbeliefs are rational, see Alan Millar, Reasons and Experience (Oxford: Clarendon Press, 1991),chapter 6.8 This is still only a necessary condition, not a sufficient condition, for exercising the ability inquestion. To get a sufficient condition, we would somehow have to rule out cases of “deviantcausal chains”; see Donald Davidson, “Freedom to Act”, reprinted in his Essays on Action andEvents (Oxford: Clarendon Press, 1980), pp. 78–79. I am optimistic that this problem can besolved, but I cannot attempt to solve it here.9 On the concept of “basic actions”, see especially Arthur Danto, “Basic Actions”, reprinted inAlan White, ed., The Philosophy of Action (Oxford: Oxford University Press, 1968), pp. 43–58.10 Much more needs to be said about this conception of rationality. But I shall not go into thesequestions here. For a more detailed account of rationality, see my “The A Priori Rules ofRationality”, Philosophy and Phenomenological Research 59 (1999), 113–31.11 This is, of course, only one way of distinguishing between what “may be a matter of luck” andwhat “lies in one’s control” — or, more precisely, perhaps, what “lies in one’s direct control”.There are many other ways of understanding this distinction on which non-basic actions maysometimes be “within one’s control” and one’s performance of a basic action may sometimes be“a matter of luck”.12 I am assuming here that if the process of one’s revising one’s beliefs through following a rulecan be analysed into a series of sub-processes (at the folk-psychological level of explanation), thenthe final member of this series of sub-processes must be the process of one’s exercising someability for revising one’s beliefs — that is, the process of one’s following some other rule. Itfollows that the process of one’s following a “basic rule” cannot be analysed (at the folk-psychological level of explanation) into any sub-processes at all. This is why, if the rule inquestion really is a “basic” rule, one’s having an experience that represents the water as boiling isthe proximate explanation of one’s forming the belief that the water is boiling.13 I should point out that I am using the terms ‘explain’ and ‘explanation’ in a systematicallyambiguous way. When I speak of giving an “explanation” of why internalism is true, I have inmind a philosophical explanation of a necessary truth. When I speak of a “psychologicalexplanation” of a belief revision, I have in mind an empirical explanation of a contingent truth.Finally, I also use the term ‘explanation’ to refer both to the explanatory accounts that are givenby theorists, and to a fact which those theorists could correctly cite as the explanans of whateverthey are trying to explain. This ambiguity should cause no confusion in context.
14 It is hard to imagine what this belief could be like, especially if we suppose that Sarah does nothave any experience that represents the flowers in any way. But perhaps “blind-sight” cases wouldbe an example. It is certainly not at all clear that it is rational to form beliefs on the basis of “blind-sight” in the same way as it is rational to form ordinary perceptual beliefs.15 Some philosophers will resist the idea that the experience is an “intervening mental state”,mediating between the fact that flowers are pink and Sarah’s belief that the flowers are pink.Instead, these philosophers might suggest, the experience and the belief are merely independenteffects of a (non-mental) common cause. If this suggestion were correct, however, then we couldgive a correct explanation of the belief merely by appealing to this non-mental cause, even if nointervening mental state (such as an experience) played any role in explaining why Sarah formedthat belief. As I have argued here, this explanation could not be a correct folk-psychologicalexplanation of the belief: it would fail to make the belief intelligible, as something that we canreadily imagine someone in the circumstances finding persuasive or compelling. But surely Sarah’sbelief can be given a correct folk-psychological explanation of this kind.
Some other philosophers might suggest that, even if a correct folk-psychologicalexplanation of the belief will imply that an experience of the relevant kind occurred, nonetheless,the belief is not explained by the experience. For example, some philosophers might suggest thatthe experience is a mere enabling condition, rather than a full-fledged explanation of the belief inquestion. This suggestion is hard to understand, but it can probably be ruled out. An “enablingcondition”, in a sense that contrasts with a genuine explanans, must just consist in the absence of
15
factors that would prevent the explanans from having its effect. But it is clear that the occurrenceof an experience of the relevant type is not an enabling condition of this sort.
Other philosophers might suggest that the experience just is the perceptual belief, so thatthe belief cannot be explained by the occurrence of the experience. This suggestion can also beruled out, because I am understanding both the explanandum and the explanans here as involvingmental state types, not mental state tokens. The mental state type believing that the flowers arepink is undoubtedly distinct from the type having an experience that represents the flowers aspink. One might believe that the flowers are pink without having any experience that representsthe flowers as pink; and vice versa. (For example, a blindfolded person might feel some flowers,and irrationally believe for no reason that they are pink. Or someone might have an experiencethat represents the flowers as pink, but distrust her own experience for some reason and so refuseto believe that the flowers are pink.)16 A similar argument, for a similar sort of “internalism” about the explanation of action, iscriticized by Rowland Stout; see his Things That Happen Because They Should (Oxford:Clarendon Press, 1996), pp. 33–36. (Stout calls the argument the “Argument from the Impotenceof Unrepresented Facts”.) Stout thinks that the argument depends on the assumption that a set offacts can supply an adequate explanation of a phenomenon only if those facts constitute a“logically sufficient condition” for that phenomenon (p. 35). But my argument does not dependon this assumption. It depends on the point that any attempt to explain the formation of a beliefon the basis of an external non-mental fact will count as a correct folk-psychological explanationonly if there is also a more detailed correct explanation in which the link between the external factand the formation of the belief is mediated by intervening facts about the believer’s mental states.17 The most notable externalist to insist on including these “factive” mental states among themental states that can determine the rationality of a belief is Williamson; see especially his“Knowledge as evidence”; “Is knowing a state of mind?” Mind 104 (1995), 533–65; and“Conditionalizing on Knowledge”, British Journal for the Philosophy of Science 49 (1998),89–121. Compare also John McDowell, “Knowledge and the Internal”, Philosophy andPhenomenological Research 55 (1995), 877–93.18 I am not denying that knowledge ever plays a significant role in folk-psychological explanations.Williamson gives a number of examples where it seems clear that knowledge does play such a role(see “The Broadness of the Mental: Some Logical Considerations”, Philosophical Perspectives12 (1998), 389–410). For example, perhaps one keeps on digging because one knows that thismine contains gold. Believing, even truly believing, that it contains gold would not have beenenough; for then one might have inferred this belief from a lemma whose falsity one might easilyhave discovered while digging, in which case one would have abandoned the belief and stoppeddigging. In this case, the explanandum — one’s keeping on digging — consists of one’sinteracting with one’s environment in a certain way. So it is only to be expected that theexplanans will also involve the agent’s relations to her environment. (For this point, see especiallyChristopher Peacocke, “Externalist Explanation”, Proceedings of the Aristotelian Society 93(1993), 203–30.) In the cases that I am principally concerned with, however, the explanandum isone’s coming to believe p. Even if one cannot come to believe p unless one has interacted in theright way with a suitable environment, one’s coming to believe p does not itself consist of one’sinteracting with the environment in any particular way. So there is no reason to think thatproximate folk-psychological explanation of one’s coming to believe p must involve one’srelations to one’s external environment.19 This clause is important, to get round the objection that this counterfactual test will always leadone to prefer the most disjunctive explanations possible. The truth of a proposition is not“partially constituted” by the truth of a disjunction of which that proposition is a disjunct —whereas knowing p is partially constituted by believing p. (The principle about explanationappealed to in the text is analogous to a principle about causation that is defended by StephenYablo. See “Cause and Essence” Synthese 93 (1992), 403–49, esp. 413–23, and “WideCausation”, Philosophical Perspectives 11 (1997), 251–81. Some closely related ideas aboutexplanation are defended in Williamson, “The Broadness of the Mental”, pp. 395–406.)20 Rowland Stout criticizes a broadly similar argument about the explanation of action, in ThingsThat Happen Because They Should, pp. 23–32. He rejects this argument on the grounds that it
16
relies on the dubious principle that if there are two distinct explanations of the same phenomenon,one of these explanations must be a proper part of the other. But my argument does not rely onthat principle; it is entirely compatible with the claim that the same phenomenon may have twodistinct explanations neither of which can be slotted into the other. My argument relies on aspecial relation that holds between an explanation of a belief that appeals to what the believerknows and the corresponding explanation that appeals to what the believer believes — viz., thefact that the thinker’s knowing q is partially constituted by the thinker’s believing q.21 To apply a parallel argument to the case of perceptual beliefs, we need the assumption thatseeing that p is the case is “partially constituted” by having a visual experience as of p’s beingthe case. This is denied by those who hold a “disjunctive” view of perception. For a classicstatement of the disjunctive view, see John McDowell, “Criteria, Defeasibility and Knowledge”,Proceedings of the British Academy 68 (1982), pp. 455–79. For criticism of some of thearguments that are used to support this disjunctive view, see Alan Millar, “The Idea ofExperience”, Proceedings of the Aristotelian Society 96 (1996), 75–90. The main argumentagainst the disjunctive view is the “Argument from Hallucination” — that is, the argument that itis only if there is a common factor in veridical perception and hallucination that we can explaincertain subjectively seamless transitions between perception and hallucination, as well as certainstriking similarities in the causal powers of the perception and the hallucination. For a powerfulrestatement of the Argument from Hallucination, see Mark Johnston, “The Obscure Object ofHallucination” (unpublished manuscript).22 See Cruz and Pollock, Contemporary Theories of Knowledge, 2nd edition (Lanham, Maryland:Rowman & Littlefield, 1999), pp. 130–37. Their argument goes roughly as follows. Epistemicnorms are “procedural norms” that can be “internalized”. When norms are internalized, thisenables “our cognitive system to follow them in an automatic way without our having to thinkabout them”. So, the circumstance-types in response to which these norms tell us to do somethingmust be “directly accessible to our system of cognitive processing”; that is, “our cognitive systemmust be able to access them without our first having to make a judgment about whether we are incircumstances of that type”. States that are in this way “directly accessible to our cognitivesystem” are what Cruz and Pollock call “internal states”. The problem with this argument, as I seeit, is that not enough is said about what it means for “our cognitive system” to “follow a norm inan automatic way”, or to “access” a circumstance-type. This makes the argument hard toevaluate.23 Contemporary Theories of Knowledge, p 140.24 Compare Christopher Peacocke’s claim that a method “selected by spinning a roulette wheel”will hardly make the belief that results from that method count as knowledge; see his Thoughts:An Essay on Content (Oxford: Blackwell, 1986), p. 141.25 To recall a point that I made in note 18 above, the fact that one is disposed to follow this rule atthis time does not itself consist of one’s interacting with one’s environment in any particular way.So there is no reason to think that the proximate folk-psychological explanation of this fact mustinvolve one’s relations to the wider environment.
26 Why accept this interpretation of my proposal? We should accept it precisely because it givesthe best explanation of certain powerful intuitions — viz., the intuition that exactly the samebeliefs are rational in the two worlds w1 and w2 that I mentioned at the beginning, and the intuitionthat rationality cannot be purely a matter of luck. There is nothing objectionable in my relying onthese intuitions in this way. My goal here is not to show that these intuitions are true, but toexplain why they are true.