. 19
( 38 .)


stopping to help much less signi¬cant (for example, if the buses came with
much greater frequency) the woman would be unreasonable not to stop
and help. Exclusionary permissions do not seem to allow us this addi-
tional claim. For if there is a permission to exclude the altruistic reason
in the ¬rst case, surely it does not disappear or cease to apply when the
strength of the opposing self-interested reason is decreased. One could of
26 Raz (1999a), p. 90. In Raz (1975), there is an argument for certain exclusionary per-
missions based on incommensurability. But this cannot be the general form of such an
argument, since such a form would never allow the exclusion of only one type of reason “
say, altruistic ones. Rather, arguments based on incommensurability will always allow the
exclusion of reasons on ˜both sides™ of the incommensurability. Moreover, the notion of
incommensurability itself is problematic.

Accounting for our normative judgments

course say that it does disappear or cease to apply. But at that point the
theory of exclusionary permissions seems to lose a good deal of its appeal,
especially when compared with the justifying/requiring distinction. One
could also argue that exclusionary permissions are sometimes permissions
to take certain reasons as having less strength than they actually have, rather
than being permissions to exclude certain reasons completely. But again,
this move seems in need of a great deal of defense, whereas the justify-
ing/requiring distinction really ought to be taken as the default position.
That is, chapter 4 showed that, whatever one™s view of rationality, the
roles of justifying and requiring are logically distinct. Thus the concepts
of requiring strength and justifying strength are also distinct. And it is a
much stronger thesis to hold that these strengths are always the same, for
any given reason, than that they are not.
The justifying/requiring distinction could explain the above case in the
following way. The altruistic reason to help the man with the map has
a certain minimal requiring strength, and a relatively greater justifying
strength. Its requiring strength is suf¬ciently great that the agent is not
rationally permitted to act against it when the opposing reason is merely
that she will have to wait ¬ve minutes for the next bus. But the justifying
strength of the altruistic reason is still suf¬ciently great that it can make
it rationally permissible for the agent to risk even the extreme annoyance
of missing a very infrequent bus. Of course one may disagree with the
normative judgments expressed in the discussion of this example. In fact,
I myself do disagree with them, since I claim that altruistic reasons have
no requiring strength, rather than the minimal requiring strength to which
the explanation appeals. But, as has been mentioned a number of times
in previous chapters, nothing of great signi¬cance rests on this particular
claim. Moreover, and as Raz himself insightfully claims in his presenta-
tion of exclusionary permissions, the point here is only to explain how
someone who makes the above normative judgments can be interpreted
in a coherent way.27 It is neither here nor there whether such a person is
correct in her particular normative judgments.

conc lu s i on
Making a distinction between the requiring and justifying strengths of
practical reasons, and holding that altruistic reasons have far less requiring

27 Raz (1999a), pp. 91, 93.

Brute Rationality

than justifying strength, allows us to explain why it is rationally permissible
(if rather stingy) for middle-class people to donate nothing to charity, but
also why it is rationally permissible for them to donate quite a lot. The
distinction allows that being sel¬sh and mean, even to the point of violating
moral requirements, need not involve any irrationality, but also that it is
rationally permissible to devote one™s life to the relief of the suffering of
others. And yet the distinction, and the associated claim about altruistic
reasons, does not make ˜everything permitted.™ For many reasons have
requiring strength: those that involve nontrivial harms to the agent, and,
perhaps, those that involve great harms to others.
The distinction between justifying and requiring gives a sensible sense
to the idea that the reasons in favor of donating two hundred dollars to
charity are far stronger than the reasons against doing so: they have far more
justi¬catory strength. This is not a special technical sense of ˜stronger.™ Many
ordinary people would agree that saving forty children from malnutrition
is, in some relatively straightforward sense, more important than saving
two hundred dollars or getting a new winter coat. The distinction between
requiring and justifying strength, and the claim that altruistic reasons have
more justi¬catory than requiring strength, gives sense to the claim that
despite this, one is not rationally required to donate the money. That is, one
is not always rationally required to act on a reason that is, in an important
and intuitive sense, clearly the stronger of the two primary reasons bearing
on one™s choice. While this claim may sound paradoxical in the abstract,
the theory that stands behind it provides a better explanation for many of
our normative judgments than do the notions of incommensurability or
exclusionary permissions.

Fitting the view into the contemporary

The claim that some reasons have greater justi¬catory strength than requir-
ing strength entails a number of further claims that are at odds with a good
deal of current philosophical dogma. For example, it entails that one need
not, rationally, always act on the stronger of two opposed reasons, even
in the absence of other relevant considerations. And it holds that this is
true whether one takes ˜stronger™ to mean ˜stronger in the requiring role™
or ˜stronger in the justifying role.™1 The of¬cial view advocated in this
book also denies the internalism requirement on practical reasons, for it
holds that it is not irrational to be completely unmoved by altruistic rea-
sons. Given these con¬‚icts with contemporary views, and given also what
appears to be a signi¬cant structural difference between the view advocated
here and other views “ two strength values, as opposed to only one “ some
readers may have begun to suspect that the notions of practical rationality
and reasons for action that form the subject of this book, while inter-
esting and signi¬cant, are simply different notions than those of concern
to other contemporary philosophers who use the same lexicographical
terms. At the very beginning of chapter 1 I explained why this suspicion
is unfounded: we are all engaged in the same project of trying to pro-
duce an account of the fundamental normative notion relevant to action.
Moreover, we all take the relation between this notion and the ˜mental
functioning™ interpretation of rationality to be suf¬ciently close that a fully
informed agent, acting in a subjectively rational way, must also be acting
in an objectively rational way “ a way that the fundamental normative
principle allows.
Nevertheless, it will serve a number of purposes to show how the view
developed here is related to certain accounts of the same subject: ideal
1 The ¬rst of these claims is true because the reason with less requiring strength may never-
theless have suf¬cient justifying strength to make it rationally permissible to act on it. And
the second is true because greater justifying strength does nothing to generate a rational

Brute Rationality

motive accounts of normative reasons. One of these purposes is to show
that these accounts, when one rids them of an assumption that is, inde-
pendently, quite implausible, yield views that include two values for any
given normative reason. These values function in a way that is isomorphic
to justifying and requiring strength. All the arguments, therefore, in favor
of distinguishing between justifying and requiring strength can be added
to the arguments in favor of modifying ideal motive accounts by removing
the implausible assumption. And all the arguments against the implausible
assumption can be added to the arguments in favor of regarding reasons “
the kind of reasons under discussion in this book, and the kind of reasons
under discussion in the work of other ethical theorists “ as having two
independent dimensions of strength.

mot ivat i ng i deal mot ive account s
Normative reasons are reasons of the sort that are involved in claims such
as ˜One reason for Jones to take a vacation is that it would make him feel
more relaxed™ or ˜Gates ought to donate more to charity for the following
reason: it would help a lot of people in a signi¬cant way.™ Even when such
reasons are altruistic, they are not speci¬cally moral. Indeed, one purpose
for which philosophers develop accounts of normative reasons is to provide
a morally neutral foundation for subsequent arguments that seek to show
why one should be moral. Such reasons get their normative signi¬cance
in virtue of the way they contribute to claims about whether our actions
are rationally permissible or irrational. For to call an action ˜irrational™
in the relevant sense is, among other things, to claim that it ought not
be performed. Running parallel to this normative sense of ˜reason,™ there
is another, nonnormative sense, which we might call ˜motivational™ or
˜explanatory.™2 Motivational or explanatory reasons are involved in claims
such as ˜Smith spat on Jones for the following reason: he (Smith) hated
people of Jones™s religion.™ As this example suggests, motivational reasons
need not provide any normative support for the actions they help to explain.
For ordinary agents, there is always the possibility that some of the moti-
vational reasons involved in the explanation of a particular action will not
be normative reasons. And there is also always the possibility that some
2 For one clear presentation of the distinction between normative and motivating reasons,
see Smith (1994), pp. 94“98. Smith holds that an agent must regard his motivating reasons
as normative reasons. The ¬nal chapter of this book argues that this is false.

The relation to other views

normative reasons applicable to an agent™s choice might completely fail
to be motivational reasons for that agent. There is, however, an unde-
niable appeal to the view that when normative and motivational reasons
come apart in either of these ways, it must be the result of some failure
on the agent™s part. Thus, it seems plausible to claim that in an agent
who was ideal in relevant respects, all normative reasons would have some
motivational pull, and that all motivating reasons would also be normative
Motivated by the plausibility of this last claim above, one popular and
well-known type of account of normative practical reasons associates such
reasons with the motives that an agent would have under certain ideal
conditions.3 For example, on such an account the question of whether an
agent has a reason to have a cup of coffee is reduced to the question of
whether, under certain ideal conditions, that agent would be motivated
to some degree to have a cup of coffee.4 And the question of the strength
of a reason is reduced to the question of the strength of the associated
hypothetical motivation. It is because of the ideal element in the conditions
under which it is claimed that the agent would be motivated, that such
accounts are sometimes referred to as ˜ideal motive accounts.™ All such
accounts take the following essential form:
p is a reason for S to do A if, and only if, were S to consider p
in the right way he would be given some motivation to do A.5

3 For the canonical versions of this type of account, see Darwall (1983), pp. 41, 81; Brandt
(1979), pp. 10“15; Smith (1994), pp. 150“81. Brandt™s account is cast in terms of rational
desires, rather than reasons. But rational desires function for Brandt in exactly the same
way that reasons function for other ideal motive theorists. One might say that Brandt so
identi¬es reasons with rational desires that he has no need for the additional term ˜reason.™
And indeed the word ˜reason™ hardly occurs in Brandt (1979). See also Railton (1986).
4 The phrase ˜can be reduced to™ is meant to be neutral as between a number of different
views about what such accounts are trying to provide: meaning analyses, truth-conditions,
or something else. Perhaps least controversially, ˜Question X can be reduced to question
Y ™ may be taken to mean ˜X can be straightforwardly answered by answering Y.™
5 This way of characterizing such views is taken from Darwall (1990), p. 262. Technically,
this characterization excludes accounts such as Michael Smith™s. For on Smith™s account,
the hypothetical motivation is not a motivation that the ideal S would himself have, to
perform A. Rather, the hypothetical motivation is, for Smith, the desire, on the part of
the ideal S, that the actual S perform A. Since this distinction is irrelevant to the point of
this chapter, I group Smith with Darwall and Brandt. Smith™s modi¬cation is an attempt to
avoid what has been called ˜the conditional fallacy.™ See Johnson (1999). In J. Gert (2002b)
I propose a different solution, based on the idea of an appropriate level of description for
action types and desires: descriptions in terms of the basic ends of the agent. Nothing of

Brute Rationality

Among ideal motive accounts, there are both naturalistic versions, and
versions that make use of ineliminably normative terminology. The only
difference between naturalistic versions and nonnaturalistic ones is that the
phrase ˜in the right way™ in the above general characterization is replaced,
in naturalistic accounts, by ideal conditions that can be speci¬ed in purely
naturalistic terms, while for nonnaturalistic accounts the ideal conditions
have an ineliminably normative aspect. This difference, though important
in other contexts, will not be relevant to the purposes of this chapter. Nor
is it important, for current purposes, whether such accounts are taken as
providing an analysis of the meaning of reason-claims, or a statement of
their truth-conditions, or something else. Moreover, although the plau-
sibility and usefulness of any ideal motive account will of course depend
on exactly how it speci¬es the relevant ideal conditions, the arguments of
this chapter will not depend on a detailed assessment of the merits of any
of these speci¬cations.
It is not the primary purpose of this chapter to attack the general strat-
egy of associating practical reasons with idealized motives. Rather, the
purpose is to point out an assumption common to existing versions of
such accounts, and to examine the consequences of rejecting this assump-
tion. The assumption is that any given consideration would generate a
unique degree of motivation in any given agent, if that agent were ideal
in relevant respects. We can call this ˜the uniqueness assumption.™ On
an ideal motive account the uniqueness assumption entails that we can
always describe the strength of a reason with a single scalar value. If one
rejects the uniqueness assumption then one holds that, within a theoret-
ically signi¬cant range, a given consideration might generate any degree
of motivation, even in an ideal agent. As a result, one will need two val-
ues in order to characterize the normative capacities of any given reason.
These two values will correspond to the maximum and minimum of the
range of acceptable degrees of motivation. The point of this chapter is to
show that, with the simple rejection of one unnecessary assumption, ideal
motive accounts become much more similar to the account of normative
reasons advocated in this book. For the minimum and maximum of the
range of acceptable degrees of motivation correspond, respectively, to the
requiring and justifying strengths explained in previous chapters.

substance would have to be altered in the present argument in order to accommodate my
suggested solution to Johnson™s problem. Consequently, that problem, and my solution to
it, are not presented here.

The relation to other views

th e as sum p t i on
The uniqueness assumption is sometimes smuggled into ideal motive
accounts concealed in the word ˜the.™ For example, Stephen Darwall makes
the uniqueness assumption on behalf of naturalistic ideal motive accounts
when he explains that on such accounts:
the normative force of reasons is fully constituted by the motivational pull a con-
sideration exerts when considered in light of knowledge and experience.6


. 19
( 38 .)