. 6
( 38 .)


to spend a hundred dollars on a bottle of good wine instead of donating
that amount to Oxfam, then one would also be violating the categorical
imperative and (therefore) acting irrationally. Admittedly, there are very
strong reasons to forego the wine and send the money off to Afghanistan or
Honduras. But despite agreement with this, virtually no nonphilosopher
would claim that buying the wine showed the slightest problem in practical
mental functioning, even in a fully informed agent. Again, it is not just that
such sel¬sh action is not insane. Smoking is not insane either, and yet many
ordinary people (even smokers) admit that it is ˜stupid™ to smoke, and that

14 One need not be a foundationalist to make this kind of point. Basicness may be context-
relative. See Heath (1997).

Practical rationality and justi¬catory reasons

the reasons they have for smoking are not really suf¬cient to justify the
health risks. On the other hand, those who buy wine, cars, new clothes,
CDs, and so on, do not need to provide a rational justi¬cation for doing
so. If the existence of people in extreme need is pointed out to them, they
may feel more or less pressure to provide a moral justi¬cation, but this is
a different matter. Even Kant does not start from the claim that immoral
action is irrational. Rather, his claim is the result of ingenious arguments
that start with a nonmoral conception of the nature of rational agency,
and lead to the claim that the actions we recognize as immoral would not
be performed by someone who was completely rational. This is not the
place to show exactly where these ingenious arguments go wrong. But it
is worth noting that when one is arguing for some practical conclusion
(such as the claim that it is irrational to lie, cheat, steal, etc.), one must
use normative words like ˜reason,™ ˜ought,™ ˜rational,™ and their cognates,
with the meanings one learned for them when one learned one™s native
language. There are no more basic normative notions or principles one can
appeal to.15 Thus, if it is clear that, according to the actual use of these
words, some reasons (say, altruistic ones) are only (rationally) justi¬catory,
and are not (rationally) requiring, then it is unclear how any argument
could ever show anything different. It is part of the point of this book to
make it easier to see that the normal usage of normative language does
indeed include such things as purely justi¬catory reasons, and that altruistic
reasons are among them.
Thus, despite Kantian objections, there are strong reasons to accept the
claims about the examples above: that it is not irrational to fail to donate
money, even though it need not be irrational to make great personal sacri-
¬ces to achieve the same ends. Nevertheless, in the face of these examples
and claims, one might point out that such a difference in the structures of
theoretical and practical rationality must be accounted for. In the absence
of such an account, it might be urged, the parallel should lead us to con-
clude that practical reasons stemming from the interest of others really do
require action, and that it is only human sel¬shness that blinds us to this
theoretically transparent fact.16 Or, like the egoist and desire-satisfaction
15 It is true that I am arguing for a fundamental normative notion “ objective rationality “
which is not represented in an obvious way in the language. Why cannot the Kantian
simply invent some similar notion? The answer is that any such notion must pass the two
tests given in chapter 1, and these tests make use of familiar normative words that must
be used in standard ways.
16 Because the idea of purely justi¬catory reasons is not discussed, this type of objection is
not found explicitly in the literature. Though it might not re¬‚ect their considered views,

Brute Rationality

theorists mentioned above, one might claim that such purported ˜altruistic
practical reasons™ are not really reasons at all, and can neither justify nor
require, if they do not engage the interests of the agent in some way. If
either of these claims is true, then all practical reasons may well have some
power to require. The following arguments attempt to motivate a differ-
ent parallel for practical rationality: a parallel with morality. The argument
draws attention to the plausibility and common acceptance of moral rea-
sons that violate a moral version of practical reasons internalism: moral
reasons that need not ¬nd a corresponding motivation even in a morally
good person. These are purely justi¬catory moral reasons. The analogy is
then intended to support the idea there are reasons that are purely justi¬-
catory with regard, not to morality, but to practical rationality. The initial
case for this claim emerged from the above discussion of some examples
of rational action. This discussion pointed out that it is generally consid-
ered rationally permissible to risk a great deal to save forty children from
malnutrition, but it is also generally considered rationally permissible not
to make even a small sacri¬ce to achieve the same end. If the following
point-by-point parallel between morality and this view of practical ratio-
nality is at all compelling, then objections to these claims that are based
on a putative parallel between theoretical and practical rationality should
lose much of their force.

th e paral le l w i th moral i ty
The parallel with morality includes the following features: ¬rst, a notion of
a prima facie immoral action; second, a notion of a moral justi¬cation for
such an action; third, the fact that the considerations that morally justify
are not identical with the considerations that make an action prima facie
morally required; fourth, the relevance of the distinction between persons;
and ¬fth, the fact that the existence of morally justifying conditions does
not imply that even a morally good person must be inclined to act in the
justi¬ed way. The following ¬ve subsections explain and defend each of
these ¬ve points in turn.

I have encountered this objection in conversations with Robert Audi, John Deigh, Cairin
Cronin, and Mylan Engel. Of course, the objection is natural if one holds the position
that there is a strong analogy between practical and theoretical rationality, or if one is
an internalist about practical reasons. But the possibility of purely justi¬catory reasons is
a direct challenge to these views, and cannot be denied simply based upon its obvious
con¬‚ict with them.

Practical rationality and justi¬catory reasons

Prima facie immoral actions
According to many standard moral theories, there is a notion of a prima
facie immoral action, and then there are certain types of reasons that can
justify such actions. Typically, for example, the prima facie immoral actions
are ones that harm other people. When, and only when, one believes (or
ought to believe) that one™s action will harm someone else, is one required
to have a moral justi¬cation for that action. Robinson Crusoe, skipping
stones in a lagoon, is not in need of any moral justi¬cation whatever for
doing so. To say he is morally justi¬ed in skipping stones for the following
reason “ that it harms no one “ is simply a very misleading way of saying that
actions that harm no one are not in need of moral justi¬cation. The fact
that an action harms no one is not a moral reason in any important sense.
It cannot ever make any otherwise immoral action morally acceptable,
as, for example, the fact that an action will alleviate someone else™s pain
can do.
According to these accounts, morality is primarily concerned with how
people are to act when their actions affect others. And since we, or those
we are concerned with, might be these very others, the types of acts we
are primarily concerned with are those that might harm other people.
This concern with how other people treat us is implicit in the reasoning,
for example, that leads to the choice of principles in the various versions
of Rawls™s original position that have appeared in contemporary moral
theories. The driving force behind the negotiations is a concern that one
not end up in a terrible position as the result of the actions of others. Even
Kant, whose morality seems as far removed from concern for self as any
moral philosopher™s, excludes some forms of behavior because an agent
could not will that other people would behave that way towards him.17
Hobbes, also, is clear that morality is a set of rules that we want other people
to obey, and that the reason one gives up one™s rights is to secure other
people™s compliance with moral rules.18 Versions of morality that try to
elaborate the Golden Rule, or that try to explain the force of the question
˜How would you like it if people did that to you™ also ¬t this pattern.19
They imply at least that a good heuristic for the content of moral rules is
whether one would want other people to follow them when one considers
that their actions may be directed at oneself. This range of views agree
that morality is primarily a set of rules by which people are to conduct
17 See Kant (1988), pp. 51“2, where he discusses an example of an imperfect duty to others.
18 19 See Strang (1995), pp. 378“85.
See Hobbes (1994), Pt II, ch. xvii.

Brute Rationality

themselves when their behavior will affect others. The point here is not
to deny this. Rather, the point is that it is an extremely common view
that some actions are morally acceptable not because there are suf¬cient
moral reasons in their favor, but because there is nothing morally to be said
against them.

Justi¬cations for prima facie immoral actions
Virtually all moral theories make it prima facie immoral to kill someone,
to cause someone pain, or to deprive them of freedom. And yet virtually
all moral theories also allow that in some cases it is morally permissible do
so. This characteristic cuts across the distinction between teleological and
deontological views.
First let us consider utilitarianism. For a utilitarian it is certainly true that
if all one knows about an action is that it will increase the risk of hurting
someone, then the presumption is that one should not do it; it is prima facie
immoral. And yet there is certainly no question that utilitarian views some-
times allow an action to be morally permissible even though it is virtually
certain to hurt someone. Indeed, it is often seen as a ¬‚aw in utilitarian
views “ especially act-utilitarian views “ that it is too easy to justify hurting
someone, since the avoidance of a slightly greater harm for someone else
(perhaps even for oneself) is generally suf¬cient to do so. Indeed, another
problem with utilitarian views is that there is no easy means to distinguish
such questionable justi¬cations from still more questionable requirements.
For if one is justi¬ed in causing an innocent stranger a great deal of pain
in virtue of that action™s being necessary to avoid the same pain for oneself
and one™s friend, this is in virtue of a certain utility calculation. And this
same utility calculation will also require one to hurt the stranger.
A classic deontological moral view is Kurt Baier™s. For Baier, it is cer-
tainly true that if all one knows about an action is that it will increase the
risk of killing or hurting someone, then the presumption is that one should
not do it; it is prima facie immoral. And yet, there is certainly no ques-
tion that Baier sometimes allows an action to be morally permissible even
though it is virtually certain to kill or hurt someone.20 His way of allowing
violations of moral rules is not, as in the case of utilitarianism, merely a
question of weighing the goods and the evils. Rather, it is a question of
there being a rule that allows such violations. In this way, Baier avoids an

20 See Baier (1965), pp. 99“108.

Practical rationality and justi¬catory reasons

unreasonably close connection between justi¬cation and requirement.21
For the rules that allow violations “ that is, the rules that provide moral
justi¬cation “ are not the same as the rules that make actions prima facie
immoral. The rules that allow violations do not require one to make those
violations. So Baier can explain the existence of a large class of actions that
are morally justi¬ed but not required.

The set of justifying considerations is different from
the set of requiring considerations
So far the argument has been concerned with prima facie immoral actions
and justi¬cations for them. A concept very closely related to prima facie
immoral action is prima facie morally required action. For whenever one
action is prima facie immoral, there is another action (namely refraining
from the ¬rst action) that is prima facie required. And whenever an action
is prima facie required, there is another action (again, refraining from that
action) that is prima facie immoral. Therefore, many moral views claim
that certain considerations place an action into the category of prima facie
morally required; in fact these are the same considerations which place a
different action into the category of prima facie immoral. Examples of such
considerations, again, might be that an action will avoid causing pain to
someone other than the agent, or that one promised to do some action.
It is prima facie morally required to do actions when these considerations
apply to them, and it is prima facie immoral to refrain from such actions.
The previous section pointed out that many moral theories also claim
that certain considerations can provide justi¬cation for prima facie immoral
actions. For many moral theories the set of justifying considerations is dif-
ferent from and wider than the set of considerations that make action prima
facie required. That is, for many moral theories there is some consider-
ation that can justify an action, but which cannot make an action prima
facie required.
The above feature does not apply to as wide a range of moral views
as the previous two features. For utilitarianism, it is in fact precisely the
21 However, it is unclear whether Baier avoids an unreasonably close connection between
immorality and irrationality. In Baier (1978), pp. 248“50 he attempts to de¬‚ect this charge
by distinguishing irrational actions from actions which are merely ˜contrary to reason.™
But in fact his notion of ˜contrary to reason™ plays the same fundamental normative role as
does the notion of objective irrationality in this book. Since there cannot be a suf¬cient
justi¬cation for doing an action which is contrary to reason, on Baier™s view, there also
cannot be any such justi¬cation for doing something immoral.

Brute Rationality

considerations that make an action prima facie required that also serve to
justify action. That is, if an action will avoid causing someone pain, it is
prima facie required by utilitarianism. But now suppose that an action
is prima facie immoral for some reason. It is also true, on a utilitarian
view, that such an action might nevertheless be justi¬ed by the fact that
the action will avoid causing someone pain. Thus utilitarianism does not
have the characteristic asymmetry described in this section. But in fact it is
also precisely this symmetry in utilitarianism that makes it objectionable.
More particularly, if the way justi¬cation works is merely by a sort of
balancing of pros and cons in a homogeneous ¬eld of considerations,
then morality becomes amazingly demanding, and the distinction between
persons is lost.22 These are precisely the problems that have forced some
to reject at least act-utilitarian views.23 Whether or not one agrees with
these criticisms, the point here is only that when a moral theory makes no
distinction between what can justify and what can require, this is a notable
departure from commonsense intuition.
For other moral views, the fact that an action will hurt someone other
than the agent is a consideration that makes avoiding it prima facie required.
And what makes it justi¬ed to hurt someone else is not merely that one™s
action will avoid hurting yet another person. Rather, it is often that the
person has consented, in virtue of some bene¬t for that very person. For
example, a plastic surgeon is justi¬ed in this way when she in¬‚icts pain
on her patients. Or, it might be that the action will prevent a far greater
harm for someone. For example, I am justi¬ed in taking someone™s car
keys away from her if she is so drunk that if she drives she signi¬cantly
increases the likelihood of seriously injuring or killing someone. Both of
22 In fact, the latter objection only applies to consequentialist views according to which
bene¬ts and harms are always bene¬ts and harms for a particular agent. Some distribution-
sensitive consequentialist views may avoid this objection by ranking states of affairs partly
based upon some index of the number of people who are independently pursuing their
plans of life, and how successfully they are doing so. See Schef¬‚er (1995), pp. 26“30 for
a description of some of these ranking principles.
23 See Rawls (1971), p. 27; Sen and Williams (1982), pp. 4“5; Smart (1991), p. 110. Smart
gives an exceptionally clear statement of the second objection on behalf of nonutilitarians.
This objection is different from the claim that utilitarianism is quite demanding, though it
does lead very quickly to this objection. It is only the former objection that is peculiar to
utilitarianism, for a Kantian view could easily be comparably demanding. But because of


. 6
( 38 .)