. 31
( 38 .)


on the reason: that the reason provides a rational requirement.20 Under
this assumption, Owen™s father™s claim certainly looks like bluff, and that
is what Williams regards it as. But once one recognizes the possibility
that some reasons might serve only to justify, one can interpret the exter-
nalist merely as saying that the reason would make it rationally permis-
sible to perform the action, despite the reasons against it. Perhaps even
this is false in the Owen Wingrave example. But there are other exam-
ples the externalist might use, and Williams simply never addresses this
Here is an additional example. John Tilley, in “Motivation and Practical
Reasons,” provides perhaps the clearest illustration of the acceptance of
internalism based on the undefended assumption that all reasons are pro
tanto rational requirements. It is indeed a great virtue of Tilley that he
explicitly articulates the premise that does so much covert work in other
defenses of internalism. He puts the claim in the following form:

(1) Reasons are facts we are rationally required to act upon. That is, if F is a reason
for A to do D, and A is both aware of F and without any reasons that compete
with F, then A is rationally required to act on F and do D, assuming she is not
hindered from so acting.21

20 Williams (1981), pp. 110“11. Williams is clear about what he himself means when he
claims that an agent has a reason to do a particular action: he means that there are rational
processes that could produce a motivation to do the action for which it is claimed that
there is a reason. Interestingly, on a charitable reading this does not mean that the agent, if
rational, will, as a matter of necessity, go through the relevant process, and will therefore,
with equal necessity, be motivated to do the action. So Williams™s own reason claim may
not need to be taken as expressing a rational requirement.
21 Tilley (1997), p. 113.

Brute Rationality

Internalism then follows.22 For if an agent is such that she would not be
moved to action by a given reason when opposing reasons were removed,
then she is disposed to violate a rational requirement, and is not fully
rational. But no defense is offered for (1), even though Tilley acknowledges
that it plays a crucial role in arguments for moral subjectivism, which he
rightly sees as a troublesome conclusion. He simply does not see (1) as
controversial at all, explicitly claiming that even externalists agree with
it. But of course he has in mind the sort of externalists, like Par¬t and
Brink, whom I have placed in the Kantian internalist camp. Had Tilley
appreciated the difference between justifying and requiring, it would have
been far more dif¬cult for him to have regarded (1) as uncontroversial and

conc lu s i on
In current normative theory there are at least two versions of internal-
ism about practical reasons. We may call them ˜Humean internalism™ and
˜Kantian internalism.™ And there is a strong version of externalism that
denies both internalist views. This chapter has made a suggestion about
what the correct view is: that there is no unique correct view that applies
to all types of reasons. Rather, Kantian internalism is true of reasons insofar
as they play a requiring role, while externalism is true of reasons insofar as
they play a justifying role. Regardless of whether each of the arguments
for these particular conclusions goes through, the distinction between
requirement and justi¬cation remains. As was pointed out at pp. 68“69,
this is a logical distinction that does not depend at all upon any controver-
sial substantive normative claims. Despite its logical nature, this distinction
has not been acknowledged or even noted by anyone who advocates any
form of internalism. But because justi¬cation is not the same as require-
ment, it is unlikely that any one form of internalism could possibly be the
whole story. Philosophers who advocate a uni¬ed internalism will have to
show, at least, that requiring and justifying strength necessarily co-vary, or
22 Tilley™s rendering of internalism, on the same page, is the following: “If F is a reason for
A to do D, and A is aware of F, then barring all impediments and practical reasons that
compete with F, agent A will be moved to D by her awareness of F “ assuming she is
23 Christine Korsgaard, in Korsgaard (1996b), makes the same assumption as Tilley, and
infers the same conclusion. She is not as explicit as Tilley, however, since her primary goal
in that paper is not to establish internalism, but rather (correctly) to expose a signi¬cant
undefended assumption behind Humean internalism.

Internalism and different kinds of reasons

that (more surprisingly) their view applies equally well to both. Perhaps in
logic or mathematics justi¬cation and requirement are indeed the same.
That is, perhaps any theorem that is justi¬ed by the rules of logic is also, in a
sense, required by them. But it is a characteristic mistake of philosophers to
take the simplest and most formal models as the purest and most ideal. Any
attention to real practical arguments that are offered about actual matters
of importance will show that practical rationality is completely different
from logic or mathematics. It is something vastly more complex, about
which no simple claims are likely simply to be true.


Brute rationality

One signi¬cant implication of the view of rationality offered in chapter 7 is
that as long as an action does not stem from the kind of mental malfunction
that would put the agent at increased risk of suffering harms without com-
pensating bene¬ts for anyone else, that person™s action is subjectively ratio-
nal. However, many contemporary philosophers hold that for an action to
be rational in this sense, or even intelligible, it must somehow involve the
judgment, by the agent, that the ends of the action are good. For example,
Jonathan Dancy, Warren Quinn, Joseph Raz, and Thomas Scanlon have
recently and independently presented theories according to which inten-
tional action is action undertaken for a reason, and undertaking an action
for a reason requires that one see something in the action as being of value,
or as being a reason-giving feature. Not surprisingly these philosophers
also hold that we have desires for reasons, at least when these desires are
not simply urges that seize us. Having a desire for a reason involves, for
Scanlon and for Raz, the judgment that the object of the desire is good in
some way, while Quinn holds that the same sort of judgment is required
in order for an action to be rational.1 And Dancy holds that the reasons for
which an agent acts, whether good or bad, must at least be regarded by the
agent as favoring the action.2 Despite many points of disagreement, all of
these philosophers would be able to agree to the following general claim:
rational action involves the making of normative judgments.3 The concern

1 Raz (1999b), pp. 8, 23, 62, 291; Scanlon (1998), pp. 18, 23“24, 33“35, 56“57; Quinn
(1995), pp. 195, 200, 203, 205.
2 Dancy (2000), pp. 129, 136.
3 In what follows I will make the simplifying assumption that the agent is not mistaken or
ignorant about any relevant nonnormative matters of fact, such as whether the liquid in
his glass is gin or gasoline. Even with this assumption, Scanlon would not agree with the
claim to which this note is attached, if ˜rational™ is interpreted in his idiolect. Unfortunately
Scanlon has no term that captures the sense of ˜rational™ as intended here, although it is a
common sense. In Scanlon™s terminology, what is meant by ˜rational™ in the present context
would be expressed by the phrase ˜not open to rational criticism, except with regard to

Brute rationality

in this ¬nal chapter is primarily to deny this view of rational action. In
denying that rational action involves the making of normative judgments, I
will advance the same claim about intentional action more generally. This
is because rational action is a kind of intentional action. And indeed all of
the philosophers discussed here hold their view of rational action because
they hold the same view of intentional action: that it also requires the
making of normative judgments. Rational action, for these philosophers,
is then simply intentional action that actually is justi¬ed by the reasons for
which the agent acted. Thus, the locus of dispute in what follows will often
be the nature of intentional action, even though the ultimate goal is an
account of rational action that does not involve the making of normative
judgments. It will not be possible to specify the notion of rational action
at issue here with great precision, since the four philosophers being dis-
cussed would themselves be unable to agree completely on any proposal,
and would surely disagree with many aspects of the theory of rationality
offered in this book. But we could perhaps all agree that rational action is
intentional action that is free from failures of instrumental rationality, and
is a response to the reasons of which the agent is aware, in a way that is
appropriate, given the normative signi¬cance of those reasons.4
An additional theoretical commitment common to the four philoso-
phers mentioned above is a denial of the basic normative signi¬cance of
desire. Dancy, Quinn, Raz, and Scanlon would all ¬nd something sub-
stantially correct in the following: it is not desires that rationally justify or
require actions; rather, it is the reasons that stand behind those desires. Let
us call this view ˜the objective reasons thesis.™ The objective reasons thesis
is set up in opposition to neo-Humean views of rationality that regard
desires as the source of all of our practical reasons. The view offered in
this book also strongly opposes the neo-Humean view, and I have argued
against it in a number of places in previous chapters. But because the argu-
ments of the present chapter are aimed at philosophers who, like Dancy,
Quinn, Raz, and Scanlon, have also already rejected such views, I will not
repeat these arguments. Rather, the purpose of this chapter is to separate
the objective reasons thesis, which is a view with much merit, from the
nonculpable ignorance or mistake about nonnormative matters of fact.™ Understanding
˜rational™ in this sense, and given the assumption of full and accurate information, Scanlon
could agree with the above claim. See below for more remarks on problems with Scanlon™s
4 Again, for terminological reasons Scanlon would dissent from this as a characterization
of rational action. But he would, I hope, concede that the notion described here is an
important one.

Brute Rationality

view that rational action requires agents to make normative judgments,
however implicit, unconscious, or inchoate. This latter view, which we
might call ˜the judgment thesis,™ misrepresents and overintellectualizes the
vast majority of our everyday choices, desires, and actions.

what are normat ive j udg m e nt s ?
The plausibility of the judgment thesis, and the plausibility of attributing
it to Dancy, Quinn, Raz, and Scanlon, will depend a great deal on what
we take normative judgments to be. For purposes of present discussion,
normative judgments are judgments that something is good or bad, or
of value, or that something provides a reason.5 But is important to make
clear at the outset that these judgments need not be explicit or conscious.
None of the philosophers mentioned above hold the extreme view that
explicit normative judgments are required for desire or for action. And
yet the judgments at issue cannot be mere dispositions to agree to certain
normative propositions. They actually play a role in generating motivation,
and so they must be regarded as psychologically present in some fairly
strong sense.6
Perhaps the weakest interpretation of the claim that an agent is making
a normative judgment is that something ˜appears of value™ to the agent, or
˜appears to be a reason.™ By way of comparison, it is certainly possible for
something to appear blue to a perceiver without that perceiver having the
concept ˜blue™ in any sense other than that in which babies or apes could
be said to have that concept. If the normative judgments that Dancy,
Quinn, Raz, and Scanlon have in mind are like this, then they could
certainly avoid the charge of overintellectualizing our desires and choices.7
The question of importance for this weak understanding of normative

5 There are important questions about the relation between value and reasons. But since the
arguments of this chapter will deny that any sort of normative or evaluative judgment is
typically involved in desire or intention, the distinction between value and reasons will not
affect any of the arguments that follow.
6 The word ˜generating™ is here meant to be neutral as between causal and noncausal accounts
of the relation between reasons (or beliefs about reasons) and motivation. For Dancy, for
example, the above claim is meant to be compatible with the idea that an appropriate
normative judgment is one of the ˜enabling conditions™ for a reason to be an agent™s
7 On the other hand, and as we will see, such a characterization runs the risk of blurring the
line between cognitivist and noncognitivist accounts of motivation. For example, Philippa
Foot and Simon Blackburn both place ˜recognition of reason for action™ among the attitudes
and feelings that are characteristic of noncognitivist views. See Blackburn (1995), p. 36.

Brute rationality

judgments is whether it makes sense to think of something ˜appearing
good™ or ˜appearing to be a reason™ in the same way that it makes sense to
think of something appearing blue.
One argument in favor of this view comes from Scanlon.8 We all under-
stand what it would be like for someone to seem, for example, trust-
worthy, when we actually judge that she is not. This suggests that there
can be normative appearances that are distinct from explicit normative
judgments, and that can even con¬‚ict with them. Scanlon takes this case
of seeming trustworthy to involve a normative appearance that parallels
the “vague appeal to a normative category” that he asserts is involved
in typical desires. But what is going on in such cases? When something
appears blue although we explicitly judge that it is not really blue, we
can explain what is going on in the following way: the agent™s visual
experience is as if the object perceived were blue. But we cannot explain
Scanlon™s example in the same way unless we posit a special faculty that
perceives trustworthiness, or, at the very least, a distinctive phenomenol-
ogy of trustworthiness. Then we could explain normative appearances by
saying that the appearances produced by this faculty were as if it were
mediating the perception of someone who was really trustworthy, or that
the phenomenological experience of the perceiver was of the distinctive
sort typically produced by trustworthy people. But such a special faculty
represents quite a hard bullet to bite: the arguments of this chapter will
have accomplished quite a bit if they compel anyone to bite it. And the
idea that there is a distinctive phenomenology of trustworthiness is hardly
more appealing.9 Moreover, there is a straightforward alternative. We can
say that because of the unconscious in¬‚uence of features of the seemingly
trustworthy woman™s behavior, facial expression, and so on, we are strongly
disposed to believe what she says, or to follow her advice, or otherwise
to respond to her in ways that are appropriate for genuinely trustworthy
people. If we also have the explicit belief that she is not trustworthy, then
if we notice that we often (for example) form beliefs based on what she
8 It is unclear how this solution would be applied to account for the “standing normative
judgments” to which Scanlon (1998), p. 24 appeals in explaining our unre¬‚ective desires
and choices.
9 This point is very similar to the Wittgensteinian point that there is no distinctive or essential
phenomenology of ˜grasping™ or ˜understanding™ a rule. Rather, what is important is that
one can go on in the right way. And since one may think that one can go on in the right
way, or one may con¬dently try to go on in the right way, when one cannot, there is a
place in our language for phrases like ˜it seemed to me that I understood, although I did
not.™ The fact that there may sometimes be a certain felt experience at the moment of
comprehension is no argument against Wittgenstein.

Brute Rationality

says, this is an important piece of information, worth being able to report


. 31
( 38 .)