. 4
( 38 .)


before they act on them. These theories then go on to argue that any
such being is somehow implicitly committed to acting morally. As in
the case of contractualist theories, it is important for Kantian theories
that the initial characterization of rationality be given in nonmoral terms.
Otherwise the view becomes trivial or circular. Moreover, it is important
for these views that the notion of rationality they make use of be one
that is somehow inescapable. That is, when a Kantian offers her account
of rationality, it would be bad for her if many people could sincerely say
˜Oh, well, if that™s what you mean by “rational,” I really don™t care much
whether I generally act rationally.™17 The notion of rationality should be
one such that virtually no one would ever want to act in an irrational
way: or at least, it should be one such that no one could think that she
herself had an adequate reason for acting irrationally.18 The problem with
this avenue of escape is that identifying rational action with moral action
also identi¬es immoral action with irrational action. But typically if a
person habitually acts in signi¬cantly irrational ways, we think that it is
appropriate to call the person himself irrational, in a sense that is supposed
to be relevant to questions of moral responsibility. Thus, on Kantian views,
the more egregiously immoral one is, the less morally responsible one
would seem to be. In fact, it is also a contingent fact about many Kantian
moral theories that they identify irrational action with action that is not
completely free or autonomous. This also suggests that grossly immoral
action, which is also eo ipso signi¬cantly irrational according to such views,
is not free or autonomous. Why then do we feel justi¬ed in holding the
perpetrators of such action morally responsible? Of course I am not the
¬rst to point out this inherent dif¬culty with Kantian moral theories. And
it is equally true that Kantians have many ways of attempting to meet it.
But it is a signi¬cant problem nonetheless. Kantian moral theories clearly
make the relation between rationality and moral responsibility hard to

17 One inadequate reply to this dismissal is ˜Whether you care or not, you are a rational being,
and so you are inescapably set up to act this way.™ For this response makes it impossible to
understand how irrational (or, hence, immoral) action is even a possibility.
18 One can think that there is an adequate reason for someone else to perform an action
that would be irrational for that person. This can happen when one is aware of relevant
facts of which the other person is ignorant. Obviously this cannot happen when one is
considering one™s own actions.

Conditions on an adequate theory

One way to avoid all the problems for contractualist and consequential-
ist moral theories, without giving up the intelligibility of the connection
between rationality and moral responsibility, is the following. One could
offer a theory of rationality that has suf¬cient latitude (a) to classify all
morally required action as rationally permissible, but also (b) to classify
much immoral action as rationally permissible as well. The ¬rst of these
features would allow one to avoid the problems that contractualist and
consequentialist moral theories encounter when they are offered in tan-
dem with maximizing views of rationality. The second would allow one to
escape the problems that Kantian moral theories encounter when dealing
with questions of moral responsibility. How might one build a theory of
rationality that incorporates this latitude? It will require a certain amount
of coordination between one™s moral theory and one™s theory of rationality.
In particular, it will require that when one is morally required to perform
an action that goes against one™s personal interests, there will always be
considerations that one can cite that are suf¬cient to rationally justify acting
against those interests. And this will require that there be considerations
one can cite to show that the action is not objectively irrational “ for if an
action is objectively irrational from the point of view of the agent on whom
morality is making its demand, then it is subjectively irrational. But for
this strategy to provide a solution, it is necessary that we understand these
justifying considerations as exactly that: rationally justifying considerations.
That is, the theory of rationality should hold that those considerations are
necessarily suf¬cient to make the action rationally permissible “ not that
they are necessarily suf¬cient to make the action rationally required. This
is the sense of ˜justify™ in which one might say that considerations of self-
defense can morally justify a person in killing someone who is attacking
her. The fact that one must kill the person in order to preserve one™s life
is a fact that makes it morally permissible to do so. But it by no means
makes it morally required. Rational justi¬cation, like moral justi¬cation,
is a matter of changing the status of an action from prohibited to per-
missible.19 In the case of moral justi¬cation, the change is from morally
prohibited to morally permissible. In the case of rational justi¬cation, the
change is from rationally prohibited to rationally permissible. A large part
of this book, including chapters 2 through 5, is devoted to clarifying this
concept of justi¬cation, and distinguishing it from the logically related but
19 The very same consideration that rationally (or morally) justi¬es an action may also make
the action rationally (or morally) required. But then the consideration is doing something
more than merely justifying the action.

Brute Rationality

distinct concept of requirement.20 The distinction between the justifying
and requiring roles of practical reasons is, in my view, one of the most
important features of such reasons, and also “ once one begins to look
at actual cases in a systematic way “ one of the most obvious. And yet it
has been completely overlooked by virtually every contemporary ethical
theorist, yielding the sorts of troubles detailed above.
If one™s accounts of rationality and morality are such that, when an
action is morally required, it will always be possible to cite considerations
that rationally justify it, then one has escaped all the problems described
above for contractualism, consequentialism, and Kantianism. Although no
particular account of morality will be offered in this book, the account
of rationality that is offered will make it easy to construct plausible moral
theories that necessarily include such rationally justifying features in all
morally required action. For the account of rationality will include the
claim that altruistic considerations provide reasons that can rationally justify
personal sacri¬ces: that is, altruistic considerations can make personal sac-
ri¬ces rationally permissible. This will mean that morally required action
will be rationally permissible just so long as those sacri¬ces are not blindly
offered at the altar of a dogmatic moral view that requires sacri¬ce even
when such sacri¬ce produces no bene¬ts “ or only insigni¬cant ones “
for others.

sum mary of ade quac y cond i t i on s
It may be worthwhile to summarize this chapter™s suggestions regarding
adequacy conditions on accounts of practical rationality. One should pro-
vide a fundamental principle that gives what I have called the objective
rationality of an action. This principle should pass the following two
T1 The question ˜Why should I always act that way?™ should
not make clear sense.
T2 It should not be possible to offer adequate reasons for acting
against the principle.
20 There are, of course, other related senses of ˜justi¬cation.™ For example, one of the impli-
cations of the view offered in this book is that actions that fall in a certain class do not
stand in need of rational justi¬cation. If one performs such an action, and is challenged
to provide a justi¬cation, it is possible to do so simply by showing that it belongs to
this class. Such a justi¬cation clearly neither changes the status of the action, nor cites a
consideration that does so. But it is plausible that this sense of ˜justi¬cation™ is derivative.

Conditions on an adequate theory

T3 The account should distinguish between rationality in the
fundamental normative sense “ the sense relevant to tests
T1 and T2 “ from rationality in the ˜mental functioning™
sense that is more directly relevant to questions of moral
responsibility, competence to give consent, and so on.
With regard to rationality in the ˜mental functioning™ sense:
T4 The account should explain the failures of attempts to de¬ne
rationality in this sense as being essentially the same as ratio-
nality in the fundamental sense, but relativized to some set
of beliefs of the agent (including beliefs that the agent should
T5 The account should explain the differential relevance of
desires with regard to self-interested and altruistic reasons.
T6 The account should yield the verdict of ˜irrational™ for the
kinds of actions that we really do take as counting against
moral responsibility, competence to give consent, etc. And
it should not yield the verdict of ˜irrational™ for the kinds of
actions that we do not take as indicating mental malfunc-
tioning of this sort. In particular, it should not automatically
yield a verdict of ˜irrational™ for all immoral action.
Further, with regard to morality:
T7 The account should be consistent with the claim that, for
agents who know all the relevant facts, no morally required
action is ever irrational in the fundamental sense.
T8 The account should be consistent with the claim that no
morally required action is necessarily irrational in the ˜men-
tal functioning™ sense.
This book will offer an account of practical rationality that meets all of these
conditions. Much of the work in providing such an account will depend
on the distinction between the justifying and requiring roles of normative
practical reasons, which is explained and defended in chapters 2 through
5. Because this distinction, though implicit in much commonsense and
philosophical reasoning, is not explicitly recognized by philosophers who
write about rationality and reasons, chapter 6 provides a way for many
contemporary philosophers to ¬t this distinction into their own accounts

Brute Rationality

of practical reasons. Chapter 7 is perhaps the heart of the book, and it is
where the actual account of objective rationality is offered, along with an
explanation of its relation to subjective rationality. Chapters 8 and 9 draw
out some implications of the view, with regard to some of the issues that
have been at the center of ethical theory for the past thirty years. One of
these is the so-called ˜internalism/externalism debate,™ which focuses on
the relation between the motives of agents, and the reasons those agents
have. The other concerns the role of normative judgments in the etiology
of intentional human action.
It may seem strange that the explicit account of objective and subjective
rationality comes so late in the book, especially since one of the points of
the book is to suggest that a failure to distinguish sharply between these
two concepts is the source of a great deal of confusion. But one of the rea-
sons why philosophers have been able to con¬‚ate objective and subjective
rationality is that, in most cases, any claim about the subjective rationality
of an action will imply the same claim about its objective rationality, and
vice versa. As was explained above, this is true whenever the context is
such that there is no harm in stipulating that the agent is aware of all the
relevant facts, and when it is not the speci¬c etiology of the action that
makes it subjectively irrational. Indeed, because of this connection, and
because we have stronger intuitions about subjective rationality than about
objective rationality, I will couch most of the early arguments in terms of
the subjective notion, and leave it to the reader to make the inferences
regarding the objective one. That such a strategy is unproblematic is of
course not a reason, in itself, for deferring an explanation of the relation
between objective and subjective rationality. But an easy and relatively
complete explanation of this relation is impossible until the distinction
between the justifying and requiring roles of practical reasons can be taken
for granted, and that is a reason for deferring discussion of the relation.
Once the justifying/requiring distinction is understood and appreciated,
the full account of rationality should be very easy to understand. It is my
hope that it will also seem compelling.

Practical rationality, morality, and purely
justi¬catory reasons

Because the normative notions of practical and theoretical rationality seem,
due to their respective names, to be species of one genus, it is often assumed
that there should be a very strong parallel between the two notions.1 In
particular, it is often assumed that for practical rationality, the business of
normative reasons is to count in favor of (or against) doing something,
and that for theoretical rationality, the business of normative reasons is
to count in favor of (or against) believing something.2 And in both cases
it is assumed that reasons do this by providing justi¬cation which either
is requirement, or which would tend, if the reasons became stronger or
more numerous, to mount in strength and become requirement.3 A closely

1 Because theoretical rationality is relative to the epistemic situation of the agent, the notion
under discussion here should be taken to be subjective rather than objective practical
rationality. Moreover, morality also exhibits the kind of relativity to the agent™s epistemic
situation that is missing in the case of objective practical rationality. Making this explicit
removes one potential source of confusion when we are comparing the analogy between
practical rationality and theoretical rationality with the analogy between practical rationality
and morality.
2 In the remainder of this book the quali¬cation ˜normative™ will generally be dropped
when talking about reasons. But it is always to be understood that the reasons being
discussed are normative, and not (necessarily) explanatory. Explanatory reasons are causal
or psychological entities that explain my actions or my beliefs. But such reasons may well
fail to justify or require them in a normative sense: they may fail to be relevant to ˜ought™
claims about those actions or beliefs.
3 This view is so widespread that many theorists do not seem to recognize that there is a
position opposed to it. As a result, it is not often clearly stated. Nevertheless, for relatively
clear endorsements, see Darwall (1983), pp. 19, 54; Korsgaard (1996a), pp. 225“26; Audi
(1997), pp. 146“47; Scanlon (1998), pp. 18“23; Copp (1995), p. 42; Velleman (1996),
pp. 705ff; Edgley (1965), pp. 182“88. Foley (1991), pp. 365“66 also favors a uni¬ed
treatment of rationality, and expresses something very much like the above view, according
to which “reasonability is a matter of the relative strength of your reasons” and “[t]he
rational is that which is suf¬ciently reasonable.” See also Foley (1992), p. 111. A slightly
different analogy between practical and theoretical reason was, according to Darwall (1999),
p. 9, popular with rationalists like Balguy and Clarke. See Balguy (1978), p. 45; Clarke
(1978), p. 614. Their analogy, translated from faculty-language to norm-language, is equally
undermined by the analogy presented in this chapter.

Brute Rationality

related position holds that if a belief is held for no reason, or if an action
is done for no reason, then the respective belief or action is unjusti¬ed
and irrational.4 As more theoretical reasons are found for the belief, or
as more practical reasons are found for the action, or as existing reasons
become stronger, the belief or the action becomes increasingly justi¬ed. If
the justi¬cation becomes strong enough, then the belief or the action is
The above position, that suf¬cient justifying reasons will always yield
requirement, is consistent with two interpretations. The ¬rst interpreta-
tion, (a), allows some actions and beliefs to be justi¬ed but not required.
The second interpretation, (b), is one on which ˜increasingly justi¬ed™
means only ˜closer to justi¬ed,™ and on which any action or belief that
is actually justi¬ed is also required. The essential point, shared by both
(a) and (b), is that any reason that can justify can also require, if it is instanti-
ated strongly enough or in suf¬cient numbers, or if countervailing reasons
are weakened or removed. For example, those who adopt interpretation
(a) may hold that though I am not required to believe a rumor of war heard
from one source, I might be justi¬ed in believing it. But if I begin to hear
the same reports from a great many sources of equal reliability, and if I
have no reason to doubt that there is a war going on, eventually I will be
required to believe it. Those who favor interpretation (b) may hold that if


. 4
( 38 .)