. 5
( 38 .)


I have no reason to doubt my source, then an unopposed reason to believe
in the war generates, on its own, a requirement to believe it. But even
those who favor interpretation (b) are likely to say the following. If I do
have reasons to doubt the existence of a war, then the rumor of war I hear
from one lone source nevertheless provides some justi¬cation for believing
that there is a war; it is just an insuf¬cient justi¬cation.6 But if I hear more
reports from more and more sources, the justi¬cation will become great
enough that I no longer have suf¬cient justi¬cation to doubt the existence
of the war. At that point I am fully justi¬ed in believing “ and required to
believe “ that the war is going on. So both interpretations endorse the view
that suf¬cient justifying reasons will eventually yield requirement. This is
the essential point that this book challenges in the practical realm. That is, no
4 For an expression of this view in the practical realm, see Foot (1978a), p. 173.
5 In all these views, and in what follows, ˜required™ should be taken to mean ˜required, on
pain of acting or believing irrationally.™ The exception is for moral requirements, where
˜required™ should be taken to mean ˜required, on pain of acting immorally.™
6 If someone wants to maintain that such reasons provide no justi¬cation until they actually
require, then that person has simply identi¬ed justi¬cation with requirement. The arguments
of this chapter tell equally strongly against this extreme position.

Practical rationality and justi¬catory reasons

matter how strong the justi¬cation for some action becomes, it never follows,
simply in virtue of the strength of such a justi¬cation, that one is required
to do the action. Therefore, despite differences between interpretations
(a) and (b) that might be relevant in other contexts, both interpretations
are grouped together here as species of one view.7 This book neither
challenges, nor endorses, this view of justi¬cation and requirement in the
theoretical realm.
When one holds that justi¬cation and requirement are linked as the
above views link them, one needs to address the question of when it
is that justi¬cation becomes requirement. One might try to answer this
question from the agent™s own point of view, and hold that the point at
which one actually commits oneself to a belief or action is also the point at
which one takes one™s justi¬cation to have become a theoretical or prac-
tical requirement. Let us call this ˜the commitment view.™ If one holds
the commitment view, then one will also hold that if one takes oneself
to have equally good (or the very same) reasons for beliefs p and q, then
one must either believe both p and q, or neither p nor q.8 Alternately, it
is possible to hold that any nontrivial balance of justi¬cation generates a
requirement. Let us call this ˜the limiting view.™ On the limiting view, as
long as it is clear that there is any nontrivial reason for an action, and no
reasons against it, then the action is both justi¬ed and required. Philoso-
phers who hold a certain widely accepted version of practical reasons
internalism “ that any rational agent, simply in virtue of being rational,
will always be motivated to some degree by any practical reason relevant
to her choice of action “ are committed to something practically indistin-
guishable from the limiting view. This is because this version of reasons
internalism entails that all unopposed practical reasons, at least of a cer-
tain bare minimum strength, will produce action in a rational agent. In
other, more explicitly normative words, anyone who fails to act on such
an unopposed reason is, to some degree, irrational. And this means that all
reasons, at least of a certain minimum strength, provide prima facie rational

7 One interesting difference might be that the former view generates a class of possible
actions which are justi¬ed but not required: those possible actions for which the reasons
for and against are relatively close to balancing. But this class will in general not be very
extensive. On the other hand, the view of practical rationality advocated in this book allows
a very wide and interesting range of actions that are rationally justi¬ed but not required.
The range includes all morally required action, and also all self-interested action.
8 See Darwall (1983), p. 110, for an explicit endorsement of this view with regard to theo-
retical reasons, and the assumption that the same holds true of practical reasons.

Brute Rationality

requirements.9 This last claim is also a logical consequence of the limiting
view. The arguments in this chapter suggest that there are some practi-
cal reasons that do not provide prima facie rational requirements. Thus,
the arguments are directed not only against the common acceptance of an
overly strong parallel between practical and theoretical rationality, but they
also oppose an even more ubiquitous acceptance of practical reasons inter-
nalism.10 This opposition is articulated more explicitly in chapter 3, in the
course of responding to an objection that takes internalism for granted.
The general debate between internalists and externalists will be taken up
explicitly in chapter 8.
The argument of this chapter is primarily aimed at philosophers who
both take theoretical reasons to function roughly as the various above views
hold, and who also take theoretical rationality to provide a good model,
in the respects mentioned, for practical rationality. It may well be true that
reasons for belief function in one of the uniform ways endorsed by the
various views described above. That is, it may well be true that any reason
that justi¬es some belief would also require it, if it were stronger, or if there
were more reasons of the same sort, or in the absence of countervailing
reasons. Or it may be true (as the limiting view holds) that any reasons that
actually justify some belief also require it. Moreover, it seems perfectly
plausible that some practical reasons both prima facie justify and prima
facie require. This chapter does not attempt to call any of these claims into
question. But now consider the following claims about practical rationality.
Suppose that one could save forty children from severe malnutrition by
smuggling them food and medicine at high risk of injury and death to
oneself. In such a case, there is a very strong reason against smuggling the
supplies: that one risks injury and death. It would be seriously irrational
to risk injury and death in the absence of countervailing reasons.11 But

9 Here and elsewhere I use ˜prima facie requirement™ to indicate not an apparent requirement,
but a requirement that persists until countervailing reasons remove it.
10 For an unargued acceptance of reasons internalism of this sort, see Cohon (1986), pp. 545“
56, esp. p. 556; Smith (1994), esp. pp. 60“2 and 151“77. Smith does indeed provide an
argument for moral reasons internalism, but it is based upon an undefended assumption
of the internalist requirement on reasons generally. The same is true in Nagel (1970).
In fact Nagel™s arguments only support a weaker view in any case, although he assumes
internalism about reasons at pp. 66“7 and elsewhere. See also Velleman (1996), pp. 700“4,
where Velleman asserts that it is “trivial” that “rationality is a disposition to be in¬‚uenced
by reasons,” and even takes the externalist to agree. Darwall (1983), p. 52 does the same.
11 This is true regardless of the etiology of the action. Therefore, since it would remain
subjectively irrational even if we stipulated that the agent was fully informed, we can
conclude that such an action is also objectively irrational. See p. 8.

Practical rationality and justi¬catory reasons

in the example there is of course also a very strong reason in favor of
smuggling the food and medicine: that doing so will save many children
from serious illness. In the example, the reason in favor of smuggling the
food clearly seems at least as strong as the reason against it. This is why one
is rationally justi¬ed in smuggling the food.12 But if all practical reasons
were comparable along one axis of strength, then it should be irrational not
to act on the reasons for smuggling the food, unless there are quite strong
reasons against it. But would it be seriously irrational, or irrational at all, to
fail to act so as to save forty children from serious malnutrition, if there were
no reasons, or only weak reasons, against saving them? For example, do we
regard ourselves as acting irrationally if, instead of preventing malnutrition
in forty children by (relatively) painlessly donating a hundred dollars to
Oxfam, we spend the money on a good bottle of wine, or on nothing
at all? If the answer is ˜No,™ then we have a case in which there are very
strong reasons in favor of donating the money (strong enough to justify
risking injury and death), and only weak reasons, or no reasons, against
donating it, and yet our action is not irrational. Callous or sel¬sh it may
be, but it is not irrational. The interests of the forty children provide a
very strong justi¬cation for actions that would otherwise be irrational, but
their interests do not provide any rational requirement to act.
It is the point of this chapter to make it plausible that the above descrip-
tion is correct: that there are some reasons, relevant to the rationality
of action, which can be very strong rational justi¬ers, but which do not
rationally require at all. That is, no matter how strong or numerous these
reasons become, it is never the case that we are irrational if we fail to act
on them. Let us call such reasons ˜purely justi¬catory reasons.™ One large
class of these reasons stems from the interests of others. It is not that purely
justi¬catory reasons have an upper limit on their possible strength, and
cannot ever become strong enough to require. For in one sense, to be
explained, the power of purely justi¬catory reasons may increase without
any practical limit. Rather, the point is that justi¬cation, as a function of
the normative reasons relevant to rationality, is a function logically distinct
from the function of requirement. Justi¬cation is a matter of making it
rationally permissible to do something that, without justi¬cation, would
be irrational. The arguments below will suggest that some reasons can
12 Again, since this remains true even when we stipulate that the agent is fully informed, this
means that such an action is objectively rational. In what follows I will generally omit to
point out implications of the sort mentioned here and in the previous note, unless there
is some special reason to do so.

Brute Rationality

justify without in any way tending to make it required to do the action
that they justify.
Some may take issue with this de¬nition of justi¬cation “ speci¬cally
with the idea that justi¬cation is to be understood as relative to something
that would otherwise be irrational or immoral. But in fact another way of
understanding the central point of this chapter is to see it as a defense
of this conception of justi¬cation, as against one that con¬‚ates the process
of justifying with the process of requiring. That is, the claim is that the
notion of justi¬cation gets its sense from contexts in which things stand
in need of justi¬cation. In the theoretical realm it is plausible to hold that
almost all beliefs would be irrational to hold (or at least that one ought
not hold them) in the absence of some reason to believe them. Thus,
it may be true that every belief stands in need of justi¬cation. This, together
with the fact that reasons for belief tend to require that for which they are
reasons, makes it very hard to distinguish justifying reasons from requiring
ones. And this makes it easier to con¬‚ate justi¬cation, as a process, from
requirement. In fact, it may be useful to distinguish justi¬cation from
requirement in the theoretical realm also, although I will not explore such
an idea in this book.
One might be tempted to think that the altruistic reason in the above
example “ that one can save forty children from malnutrition “ is a moral
reason, and not a ˜generic™ practical reason. Two things can be said about
this. First, it is a mistake to equate altruistic reasons with moral reasons.
Much of the most grossly immoral action is done completely sel¬‚essly,
for the sake of others: children, spouses, professional colleagues. But the
altruistic reasons in these cases generally provide as little moral justi¬cation
as would a corresponding self-interested reason, had the agent performed
the same sort of immoral action for her own sake. Second, a consideration
can be both a moral and a ˜generic™ practical reason. Very roughly, if a
consideration contributes systematically to the moral status of an action,
then it is a moral reason, and if it contributes systematically to the ratio-
nal status of an action, then it is a ˜generic™ practical reason. Altruistic
reasons are therefore ˜generic™ practical reasons, since they systematically
rationally justify actions that would otherwise be irrational.13 The notion
of ˜systematic contribution™ is discussed in more detail in chapter 4.

13 It may be worth mentioning here that purely justi¬catory ˜generic™ practical reasons that
also happen to be moral reasons need not be purely justi¬catory moral reasons. This is
why some morally required behavior might not be rationally required.

Practical rationality and justi¬catory reasons

It is important to keep in mind the distinction between altruistic and
moral reasons, and the possibility that one and the same consideration
might be both a moral reason and a ˜generic™ practical reason, because the
˜generic™ practical reasons that this chapter argues are purely justi¬catory
also happen to be altruistic ones. If one con¬‚ates moral and altruistic
reasons, and fails to see that moral reasons can also be ˜generic™ practical
reasons, then one may wrongly conclude that this chapter succeeds only
in showing that moral reasons can be purely justi¬catory. Of course egoists
and desire-satisfaction theorists might deny that altruistic practical reasons
exist, or that they systematically contribute to the rational status of action.
This is not the place to argue against either of these views, but one example
may at least suggest why they are both inadequate: if some habitually sel¬sh
and mean person suddenly decides to turn over a new leaf, and begins
helping other people, no one would call such behavior irrational. The
idea that we have a ¬xed set of desires, and that it makes sense to try to
maximize their satisfaction, is as much of a ¬ction as the idea that we are
only (rationally) motivated by our own interests. The rational options open
to an agent are those that would be judged rational if they were chosen.
And, for all we can actually know about any particular agent, altruistic
sacri¬ce is always something that could be chosen.
It may seem that there are more uncontroversial cases of practical rea-
sons that justify without requiring, even when they are strengthened. For
example, suppose that I am in a stuffy room, and know that I could get fresh
air either by opening one of the three windows, turning on the central
air-conditioning, or opening the door to the porch. Its being stuffy seems
to be a reason that would justify performing any of these actions. But just
in virtue of their being so many justi¬ed options, none of them could be
required. And of course, even as the reason stemming from the stuf¬ness
becomes quite strong, none of the available actions become required. So
this seems to be an example of a reason that could be inde¬nitely strength-
ened without yielding a requirement to do any of those particular actions
that it justi¬es. But these sorts of ˜multiple option™ examples do not illus-
trate the same point as the examples given above. The belief that they
do illustrate the same point rests on a con¬‚ation of basic and derivative
reasons, and on using a wrong level of description. The practical reason
supplied by the fact that it is stuffy, which justi¬es my either opening a
window or the door, is dependent upon the fact that stuf¬ness is unpleas-
ant and that fresh air reduces stuf¬ness. Barring other circumstances, there
would be no reason to open the window or door, if it were not the case

Brute Rationality

that I ¬nd stuf¬ness unpleasant and that fresh air reduces stuf¬ness. But
nothing similar can be said about the reason ˜because it will make me
feel less uncomfortable.™ It is basic.14 ˜Because it will make me feel less
uncomfortable™ is also the sort of reason that can require action, given a
description of the required action that is at the same level of generality as
the basic reason. If one is feeling extremely uncomfortable, and can do
something about it rather easily then one is required to do something to
make oneself less uncomfortable. If, as in the stuffy room example, there are
a variety of ways of doing it, one is of course not required to do each of
them. But one is nevertheless required to do something to make oneself less
uncomfortable. If one has to do something that is itself unpleasant, or risky,
then this reason also justi¬es doing that thing (at least, if the risk is not too
great, etc.). Unless one realizes that talk of practical requirement demands
appropriate levels of description, it will never be true that any action is
required, since there will always be irrelevant features of the action one is
contemplating. That is, to take a standard moral example, I am required
to save the baby, but not to save the baby by diving into the water, as opposed
to saving the baby by jumping into the water.
Because of the above considerations, it is important to take as motivating
examples those of the sort that are offered above, in which saving other people
from death and pain justi¬es but does not require one to risk pain, injury, and
death, and also justi¬es but does not require one to forego a small pleasure.
For these cases of justi¬cation without requirement do not depend upon
a description being at too high a level of detail. They are not ˜multiple
option™ examples. Of course Kantians may simply reject these types of
motivating examples. It is a standard Kantian claim that immoral action
is also irrational. And some Kantians might also claim that if one chose


. 5
( 38 .)