. 36
( 38 .)


evaluative judgments of some kind”; he says of this claim that it “seems
clearly false.”65 And immediately after this apparent denial he makes the
weaker claim that “having what is generally called a desire involves having
a tendency to see something as a reason.”66 But one should not inter-
pret this ˜tendency™ as merely reporting a contingent statistical fact. For
soon after these remarks, Scanlon claims that “when a person does have
a desire [in one very common sense of that word] and acts accordingly,
what supplies the motive for this action is the agent™s perception of some
consideration as a reason.”67 And when he later explains what he was
doing in the section from which these last few passages have been taken,
he writes the following:
[t]here is such a thing as a consideration seeming to be a reason for a certain course
of action. . . . I argued in Section 8, in effect, that such “seemings” are the central
element in what is usually called desire.68

It is perhaps because Scanlon gives such pride of place to normative
judgments about reasons, rather than to reasons themselves, or to beliefs
in nonnormative facts that also happen to be reasons, that he de¬nes
rationality as relative to the agent™s own normative judgments. For Scanlon,
it is impossible to act irrationally unless one is acting against one™s own
normative judgment.69 Thus all irrational action conforms to the pattern

64 65 Scanlon (1998), pp. 38“9. 66 Scanlon (1998), p. 39.
Scanlon (1998), p. 24.
67 Scanlon (1998), pp. 40“41. The very common sense here is ˜desire in the directed-
attention sense.™ According to Scanlon™s account of it, this is a sense of ˜desire™ that ¬ts
well with a way the term ˜desire™ is frequently used, and that “captures the familiar idea
that desires are unre¬‚ective elements in our practical thinking” (p. 39). So for Scanlon
even actions based on these unre¬‚ective desires are motivated by the perception of some
consideration as a reason.
68 69 Scanlon (1998), p. 25.
Scanlon (1998), p. 65.

Brute rationality

of weakness of will. But this view has bizarre consequences if Scanlon is
understood as offering an account of rationality in the ˜mental functioning™
sense. Consider, as Scanlon does, a person who is completely indifferent
to the prospect of even the most excruciating pain on future Tuesdays. As
long as this person judges “ based, perhaps, on some extremely strange
theory of well-being “ that such indifference is warranted, Scanlon has
“no hesitation in saying he is not irrational, just seriously mistaken.”70 In
a similar vein, Scanlon writes that:
A person who believes, on general theoretical grounds, that her future interests
should be sharply discounted [or ignored altogether], and who acts accordingly,
may be making a mistake about the reasons she has, but this does not make her
irrational, any more that it does a person who accepts a fallacious argument or
makes some other mistake about the reasons she has.71

I too have no hesitation in saying that such an agent is seriously mistaken.
But she is also irrational, in a very important and intuitive sense. Termi-
nologically, it is very unhelpful to place this irrational agent in the same
basket as someone who overlooks a mistake in a mathematical proof. In
real life, no matter what explanations such a person offered for her attitude,
we would regard her as irrational, and try to get her some psychological
help.72 We would not believe that she was simply in the grip of a false
theory, and try to argue her into a better view.
Moreover, I do not think that the problem here can be dismissed as a mere
difference in terminology. Scanlon has no term that captures a very impor-
tant kind of failure in practical agency. He has his own “irrational,” which
necessarily involves acting against one™s own normative judgments. He has
“what we have most reason to do,” which is analogous to my “objectively
rational,” and which has very little to do with mental functioning at all,
since many of the relevant reasons may be unknown, and unknowable, by
anyone, at the time of action. He has “mistaken,” and “open to rational
criticism,” which indifferently include culpable and nonculpable mistakes
about matters of fact, and mistakes about what counts as a reason. And
he has “unreasonable,” which is always relativized to some goal given by
the context, which goal may not be shared by the unreasonable person,
70 71 Scanlon (1998), p. 31.
Scanlon (1998), p. 29.
72 Being indifferent to such pain is not the same as having suf¬cient reasons to suffer it
willingly. For example, if the agent justi¬ably believed that his suffering pain on Tuesdays
would have some tremendous bene¬t for himself or other people, then it might be rational
to be willing to suffer the pain. But even in such a strange case, the agent is not indifferent
to the pain.

Brute Rationality

so that there may be no failure in practical mental functioning involved
in her unreasonable behavior.73 What we want is a term that captures the
following: sometimes the agent is aware of facts, or should be aware of
them, and these facts provide reasons that make a certain action irrational “
in a sense different from Scanlon™s. If the agent performs the action any-
way, he is not properly responsive to reasons. It is this notion, and not
Scanlon™s formal one, that is most relevant to real-life questions of moral
responsibility, freedom of the will, disabilities of the will such as phobias,
compulsions, and addictions, competence to give consent, and so on. It
is this notion that includes, as subtypes ordered roughly by degree, ˜silly,™
˜dumb,™ ˜stupid,™ ˜wrong-headed,™ ˜crazy,™ ˜insane,™ and perhaps the collo-
quial understanding of ˜irrational.™74 Scanlon has robbed himself of this
important notion by making action not primarily a response to facts that
may or may not be reasons, but to judgments about reasons.

rat i onal an i mal s
It may seem that the picture of rational action presented in this chapter
implies that the actions of dogs and mice are also, at least generally, rational.
After all, dogs and mice are disposed to avoid pain and death, to seek food
and sex, and so on; there are reasons why they act as they do, and many of
them seem to be adequate reasons for acting in those ways. That is, not only
do they act in a goal-directed way, they respond to the available reasons
in a way that is appropriate, given the normative signi¬cance of those
reasons. At this point it may seem that the implicit normative judgments
of the judgment thesis are part of what is needed to distinguish humans
and other rational beings from arational animals.75 For quite often this
distinction is made by reference to something called ˜the will,™ which is
represented as precisely the power to distance ourselves from our desires
in a way that mice cannot, and to choose which desires to act upon, in
the light of judgments about the values of the ends towards which those
desires are pointed.76 Whether or not this is the correct explanation, surely
some relevant difference must be found between mice and human beings.
When considering the differences between mice and human beings, it
is a theoretical advantage to be able to represent those differences as ones
73 See Scanlon (1998), pp. 32“33 and 191“92.
74 Though again, the question of how untutored people use the word ˜irrational™ is neither
here nor there. See p. 143.
75 76 See Korsgaard (1996a), pp. 91, 113.
See Scanlon (1998), p. 23.

Brute rationality

of degree. And where there is a difference that is not merely a difference
in degree, it is a theoretical advantage to be able to represent such a differ-
ence as arising from a difference in degree. Human beings are animals. At
some point in our evolutionary history our ancestors were more similar
to dogs and mice than they were to us. It does not seem credible that the
characteristics that we currently possess, and in virtue of which our actions
can be assessed as rational and irrational, appeared at one go, as the result
of a single fortuitous mutation. It is true that in possessing hearts and lungs
humans differ from bacteria in more than mere degree. And it may be that
humans possess some entirely new and distinct mental organs completely
absent in dogs or mice. But we are much, much more similar to dogs
and mice than we are to bacteria. Moreover, the weak claim being made
here is only that if a satisfying account of some psychological difference
between mice and human beings can be produced without appeal to an
entirely new and distinct psychological organ, so much the better.
My suggestion is that what is of primary importance in explaining why
human actions can be assessed as rational and irrational, while those of
mice cannot, is that human beings typically see farther into the future, and
represent more possibilities of action and their likely results.77 Since the
selection of one of these possible actions is therefore a far more complicated
process than the process by which a mouse ends up pressing a bar for food,
there are many more ways in which it can go wrong. One way is that
an option that would certainly have been represented in a normal human
being, and that should have been selected, given the relevant reasons,
somehow failed to get represented. Another type of failure is that an
option that should have been selected was not selected, even though it was
represented. It is precisely these sorts of errors that are typically cited when
we explain why we regard someone™s action as irrational: ˜You should have
known that extra helping would make you sick,™ ˜You knew that yelling
at her would just make matters worse.™78 Because we can make these sorts
77 See McDowell (1995), pp. 152“53. McDowell imagines a rational wolf, and makes many
claims about such a being that resonate strongly with the current suggestion. But McDow-
ell turns reason into an independent psychological faculty when he writes of this wolf
that “[h]aving acquired reason, he can contemplate alternatives; he can step back from
the natural impulse and direct critical scrutiny at it.” This is where McDowell goes astray,
assuming that the rational wolf needs something extra to explain its ability to choose
among the various options presented by its increased imagination.
78 The point here is only that these simple forms of irrationality provide a suf¬cient basis
for making one important distinction between humans and other animals. That claim is
consistent with the fact that acting against one™s normative judgments is another way in
which one can act irrationally.

Brute Rationality

of mistakes, and because we can be trained, by criticism, to minimize
them, we have developed the concepts (˜irrational,™ ˜crazy,™ ˜stupid™) with
which to criticize them. It is because these terms apply to some human
actions, and not to others, that it is correct to say that human actions
are appropriate objects of rational assessment. Because mice do not have
such a sophisticated ability to represent (and, hence, to misrepresent) future
contingencies, their actions will never be correctly assessable as irrational.
Therefore we do not think it appropriate to classify those actions as rational
either. Their actions are outside the sphere of rational assessment, so we
call them ˜arational.™79
It is beyond the scope of this book to provide a complete defense of the
above suggestion. Happily, however, it is not crucial to the current project
that the suggestion turn out to be correct. What is crucial is only that some
relevant difference between humans and arational animals be found that
does not depend on the assumption that rational action requires agents
to make normative judgments. On the above explanation of what makes
human action apt for rational assessment, it remains possible to regard
such action as the product of the interplay of desires, where desires do not
involve any normative assessments.80 In complex choice situations, perhaps
one common desire is the desire to take a moment to re¬‚ect. But even this
desire does not have any normative judgment at its core. Moreover, during
the resulting re¬‚ection, what may often happen is simply that a number
of options are imagined, and one is acted upon, without any evaluation
of it as the best option. Those who are used to thinking of human agency
as involving a conceptually separable will, an entity standing somehow
above our desires and impulses, may protest that this description of human
action is false to the phenomena. They may suggest that we are all aware
of the entity they call ˜the will,™ directing action in the light of normative
judgments based on the relevant reasons. But I must confess, I do not
have even the faintest awareness of any such entity. My own internal
re¬‚ection and, more importantly, my memory of actions done without
re¬‚ection, suggests that the simpler view presented here is much more
true to the phenomena. Typically, I just act. Sometimes I consider the
79 It is true that, for example, brain-damaged mice behave in ways that can, without any
abuse of language, be called ˜crazy.™ But clearly, if one wishes to call the actions of mice
rational and irrational, one is not an adherent of the judgment thesis.
80 Pure cognitivists, and those with views similar to Dancy™s, can substitute ˜capacity to be
motivated by various considerations™ for ˜desire™ here. Nor is anything in the basic picture
opposed to the introduction of other psychological entities, such as intentions, decisions,
and so on.

Brute rationality

options beforehand. Sometimes I consider them carefully. But in the end,
I act, and I am not aware of anything else going on.
Of course there is room in the picture for normative judgments. Not
only can we make them, but they can be part of the explanation for
our actions. There is no limit on the content of our desires that excludes
explicitly normative ends. As a result of upbringings by decent parents
in relatively stable societies, many of us have a desire to do the morally
right thing, or at least to avoid doing the morally wrong one. This desire
explains why moral judgments sometimes play a role in explaining our
actions. And in many nonmoral cases, as in the choice of a career, we may
well ask ourselves ˜What should I do?™ and act on the answer we arrive at.
In many cases this question may merely be a prelude to re¬‚ection on the
options. In such cases the reasons that explain our subsequent action may
lie entirely in the ends of our desires. In other cases, however, the desire
to do what we judge we ought to do may also contribute to the explanation of
our action. When this desire plays the right sort of role, we might wish to
call the resulting action ˜autonomous,™ or to identify it in some other way.
But if we do this, we should keep in mind that autonomy, in this sense, is
not as common as rationality, and is not, for example, a requisite for moral

conc lu s i on
The judgment thesis arises in the course of arguments against the Humean
view that desire is the ground of, or a necessary condition for, an agent
having a normative reason for action. Nevertheless, it is plausible that the
judgment thesis is itself a result of the same theoretical pull that leads
Humeans to make their characteristic claims. This is the pull to establish
a noncontingent connection between desire and justi¬cation. Humeans,
famously, yield to this pull by investing desire itself with normative sig-
ni¬cance. Dancy, Quinn, Raz, and Scanlon may be yielding to it when
they claim that the normal case of desire involves the perception of a con-
sideration as a reason. But in rejecting the Humean picture of normative
reasons, advocates of the objective reasons thesis have the resources to sep-
arate reasons and desires to a more appropriate distance. Such a separation
can help proponents of the objective reasons thesis to avoid the criticism

81 For example, we are morally responsible for many negligent acts that do not involve any
desire to do what we judge we ought to do.

Brute Rationality

that they are committed to ˜queer™ normative properties, the perception
of which is suf¬cient to generate desire, or the more general problem of
explaining how beliefs can motivate independent of antecedent desire.82
These problematic views are avoided if we simply acknowledge that things
such as pain, death, knowledge, pleasure, power, and so on, are extremely
common motives for human action, and that these things are also norma-
tive reasons. These last two claims are of course related. Human language,
and the concepts to which it gives rise, have been formed in the light of
human nature. So it is no surprise that we have a term such as ˜harm,™ that
collects the unvarying core of human aversions. And it is no surprise that
such a term functions in a normative way, to guide and assess action.
Despite the failure of Dancy, Quinn, Raz, or Scanlon to provide any
convincing argument in support of the judgment thesis, those who are


. 36
( 38 .)