. 22
( 38 .)


strength of a reason. It is true that it seems irrational to fail to take easy
means to save someone perishing at arm™s length, while it does not seem
irrational to fail to send money to Oxfam. But because distance does not
seem normatively signi¬cant, we should look elsewhere for the source of
the apparent irrationality of failing to take easy means help nearby peo-
ple who are in dire straits. Such a source is not far to seek, and it ¬ts
well with an ideal motive account that admits the existence of ranges of
acceptable degrees of motivation. Consider what would happen to a per-
son who failed to take easy means to save a nearby person from terrible
pain or death. For example, suppose that someone refused to let a gravely
injured person use the phone to call an ambulance. If such behavior were
to become publicly known, things would almost certainly go very hard
for the callous person. This provides an additional reason to give the aid.
It is this additional self-interested reason that is a plausible source of the
rational requirement to give help in such circumstances, since it appears to
be a reason that would produce a reasonably high degree of motivation
in an ideally rational agent.26 Note that when someone™s social or legal
position is such that they would not suffer any signi¬cant harm as the result
26 Of course this does not mean that these self-interested reasons always, or even usually,
provide the motivation for such acts. Almost certainly such acts are motivated by concern
for the person one is helping. But this does nothing to diminish the force of the argument

Brute Rationality

of public knowledge of their manifest indifference to (or even their delight
in) the suffering of others, then, often, they do not hesitate to show such
indifference. Hence the plausibility of the claim that power corrupts.27
Of course it is grossly immoral for such people to cause suffering, as it is
immoral to refuse to allow a gravely injured person to use one™s phone.
But the moral status of these actions is not the issue here. Rather, the issue
is their rational status. The conclusion we should draw from these examples
is that by itself, the prospect of saving someone from pain or death need
not provide a great deal of motivation, even to a perfectly rational agent.
In fact, for all practical purposes it seems that such altruistic reasons might
cause virtually no motivation in a rational agent. Those who are persuaded
by Kantian views will surely protest that this is unreasonable. But given
that we regard ourselves and our actions as generally rational, and given
that we eat nice dinners, and buy books and music, and go to movies,
and so on, while we could spare many people a tremendous amount of
needless suffering, this protestation is not very credible.28 This is, however,
consistent with the happier truth that a rational agent might be as strongly
motivated to save someone else from pain as she would be to save herself
from the same pain. That is, the Godfather was rational, but so was Mother

b e ne ¬ t s of re j e c t i ng th e as sum p t i on
This chapter has so far argued that it is a mistake for ideal motive theorists
to assume that a given consideration would provide a unique degree of
motivation to an ideally rational agent. This thesis has been defended by
example, and also by arguing that the notion of ideal or full rationality

here. The point is that the self-interested reason to give aid in such circumstances plausibly
makes it irrational to refuse to do so without a suf¬cient reason.
27 It is important here that the sense of corruption is moral. We do not have a corresponding
saying to the effect that power makes one irrational, for we do not generally regard
powerful and untouchable criminals as irrational, at least while they are running their
businesses ef¬ciently. Rather, our intuitions are that such people are breaking moral rules,
and that it is appropriate to punish them.
28 At this point, the notion of imperfect duties sometimes enters the discussion. But whether
or not the notion of imperfect duties is entirely coherent or credible, and whether or not
it bears on the rational (as opposed to the moral) status of actions, it is in any case not very
congenial to the view that the rational status of an action is determined by the strengths
of the reasons for and against it. And this view, which we might call ˜the suf¬ciency of
reasons,™ is a natural one for anyone who offers an account of practical reasons that grants
unique strength values to normative reasons.

The relation to other views

does not imply a determinate psychology, or a determinate set of desires.
The view defended so far implies that two different agents might both
be rational, and yet one might be, for example, quite altruistic, while the
other might not be. It also implies that a single agent might act ratio-
nally on two different occasions, even though on the ¬rst occasion he
acted altruistically, while on the second occasion he acted sel¬shly. This
is in itself a signi¬cant and, I think, welcome result, for it comports with
our actual judgments of the rationality of actions and agents much better
than do views that deny these claims. But the rejection of the uniqueness
assumption does more than merely allow ideal motive accounts to better
save the phenomena. It also allows them greater precision and variety in
their formal descriptions of reasons. This is because when they explicitly
reject the uniqueness assumption, they will formally characterize reasons
with two values rather than with only one. These two values are the min-
imum and maximum rationally permissible degrees of motivation that the
reason can produce. The current section explores some of the bene¬ts
of having two values with which to characterize the normative capacities
of a reason. The discussion, and the examples used, should make it clear
that the bene¬ts stem from the fact that ˜minimum rationally acceptable
degree of motivation™ functions in the same way as requiring strength, and
that ˜maximim rationally acceptable degree of motivation™ functions in the
same way as justifying strength.
Consider the following three situations and associated normative claims.
Assume in all cases that there are no other signi¬cant reasons bearing on
the case.

1) An agent ¬nds that she has a certain chance (say, 25 percent) of bringing
food and medicine to a group of forty children who need it and would
not otherwise get it. But there is also a high chance (say, 75 percent) that
the agent will perish painfully in the attempt. In such a case, the agent
would be acting in a rationally permissible (if unusually brave) manner,
if she decided to try to get the food and medicine to the children.
But such an agent would not be irrational to refuse to undertake such
a risk.
2) An agent ¬nds two hundred dollars in the street. The agent has recently
read in a reputable newspaper that a donation of this amount to Oxfam
has a roughly 25 percent chance of being used to provide food and
medicine to forty children who need it and would not otherwise get
it. In this case, it would be rationally permissible either to donate two

Brute Rationality

hundred dollars to Oxfam, or to refrain from making the donation. And
this is true even if the agent had no alternative plans for the money,
and would simply put it into her savings account for inde¬nite future
3) An agent with severe allergies ¬nds herself beginning to have a reaction.
She is without insurance, but she knows that she can get the required
treatment for roughly two hundred dollars at the local emergency room.
On the other hand, she also knows that she has a certain chance (say, 25
percent) of riding the reaction out without any permanent ill effects.
In this case the agent would certainly be irrational to refuse to pay two
hundred dollars to avoid the 75 percent chance of death or permanent
ill effects.
If one agrees with the normative claims in the above cases, and if one
holds that reasons have a speci¬c strength value and that this value deter-
mines, in con¬‚icts of reasons, what is rationally permissible to do, then, as
we saw in the previous chapter, it will be hard to deny the following three

1 ) The reason provided by a 25 percent chance of saving forty children
from the harms of severe malnutrition is of roughly the same strength
as the reason provided by the prospect of avoiding a 75 percent chance
of premature death.
2 ) The reason provided by a 25 percent chance of saving forty children
from the harms of severe malnutrition is of roughly the same strength
as the reason provided by the prospect of saving two hundred dollars
for inde¬nite future purposes.
3 ) The reason provided by the prospect of avoiding a 75 percent chance
of premature death is clearly stronger than the reason provided by
the prospect of saving two hundred dollars for inde¬nite future

As in the example in chapter 5, these three claims are in striking con-
¬‚ict with the expected transitivity of ˜of roughly the same strength as™
in the domain of practical reasons. But if practical reasons always have
one speci¬c strength value, we should expect such transitivity to hold,
except perhaps as the result of vagueness. Vagueness, however, does not
seem likely to be the explanation of this particular failure of transitivity,
since the two reasons involved in (3) are obviously of widely divergent

The relation to other views

On the other hand, if altruistic reasons that have comparatively high
maximum rationally acceptable degrees of motivation can also have
comparatively low minimum rationally acceptable degrees of motivation,
then our intuitions about the three cases are easily preserved. For to say
that an altruistic reason has a low minimum rationally acceptable degree
of motivation is to say that a rational agent need not be strongly motivated
by it. This is why it is rationally acceptable to keep two hundred dollars
for inde¬nite future purposes, even though one knows one might be able
to prevent a considerable amount of harm to others with the money. This
explains our intuitions about (2). But this is consistent with the claim that
such an altruistic reason has quite a high maximum rationally acceptable
degree of motivation: suf¬ciently high that it would be rationally permissi-
ble to be more strongly motivated by such a reason than by the prospect of
a high risk of painful death. This explains our intuitions about (1). Finally,
it is plausible to hold that the maximum rationally acceptable degree of
motivation provided by the prospect of saving two hundred dollars will
be signi¬cantly smaller than the minimum rationally acceptable degree of
motivation provided by the prospect of avoiding a high chance of painful
death. This explains our intuitions about (3).
Of course there are responses that can be made to defend the uniqueness
assumption. But given the arguments of pp. 120“28, which show how and
why the assumption might be false, and given the superior explanation of
the examples provided by a view that rejects the assumption, the burden
of proof should now clearly rest on the shoulders of those who want to
defend it. Perhaps the most tempting strategy for someone who wants
to preserve the uniqueness assumption is again to try to expand the role
of context in determining the content and strength of reasons, as was
attempted at p. 126. Following this strategy, one might try to claim that
the altruistic reason in (1) is different from and much stronger than the
super¬cially similar altruistic reason in (2), and that this is why the ¬rst
reason rationally justi¬es risking one™s life, while the second reason cannot
even require one to spend two hundred dollars. If this were true, then the
three cases would not present a failure in transitivity. But this strategy suffers
from the following problem. Unless one independently motivates some
rules by which context affects the strengths of reasons, one loses the ability
to argue against those who do not share one™s intuitions about particular
cases. And it is hard to see how one could motivate rules according to
which context would affect the strengths of reasons in ways that would
preserve our judgments about cases (1) through (3). Why, that is, should

Brute Rationality

there be a stronger reason to save children from malnutrition when one
can only do so by taking a signi¬cant risk?29 The only answer seems to be
˜to preserve our intuitions and the uniqueness assumption.™
Another strategy for defending the uniqueness assumption from the
above argument is to suggest that the normative claims in (1) through
(3) cannot all be true for the same agent. Since ideal motive accounts can
easily make the strengths of reasons agent-relative, there would then be
no failure of transitivity. In order to use this strategy, it might be claimed
that for the normative claims in (1) to be true of an agent “ for it to
be genuinely rationally permissible for that agent to risk her life to save
a group of strangers “ that agent would have to be quite altruistic. And
for such an altruistic agent, it might be claimed that the normative claims
in (2) would be false. That is, it might be claimed that it would not be
rationally permissible for such an altruistic agent to keep the two hundred
dollars. This sort of objection, which appeals to the ¬ction of a stable
set of motivations that determine the strengths of an agent™s reasons, has
already been addressed above, at pp. 118“19. To repeat the point made
there, it is implausible to suggest that habitually mean and sel¬sh people
have no reason to do altruistic actions, or that for such people altruistic
actions would be irrational. Perhaps they have no (currently) motivating
reason to perform such actions. Perhaps, also, it is therefore very unlikely
that they will perform such actions. But the human mind is something
more complex and unpredictable even than a set of twenty dice. And just
as there is a chance that twenty dice, fairly tossed, will all come up six,
there is a chance that a sel¬sh person, even without the aid of a stroke or
a brain tumor, might see things in an unaccustomed light, and act in an
unsel¬sh way.30 When Sydney Carton does something far far better than he
has ever done before, this provides not even the hint of an argument that
29 Or: Why should a signi¬cant risk of death provide a weaker reason when one can do a
great deal of good by taking it? It is no answer to this question to point out that the reason
to avoid the risk is comparatively weaker, considered in relation to the reason to provide
food and medicine, than considered in relation to the reason to save two hundred dollars.
Comparative weakness of this sort is admittedly consistent with a unique ¬xed strength
value being assigned to the reason to avoid premature death. But if the strategy being
discussed here is to avoid failures of transitivity, then it requires that there be motivated
claims about changes in the absolute strength values of one or all of the reasons in the
30 One might be tempted to say that the correct description of what happens in such cases
is ˜the spontaneous and rationally inexplicable acquisition of an entirely new desire.™
Certainly, many cases of sudden and drastic changes in motivation may be the result of
irrationality. But in order for this sort of claim to affect the argument here, it must be
read as claiming that the only way in which any agent could possibly be capable of taking

The relation to other views

he has therefore done something irrational. And the same remarks can be
made about altruistic people acting in uncharacteristically sel¬sh ways.
If one rejects the uniqueness assumption, then cases like (1) through
(3) do not force one to abandon the hope of making theoretically useful
formal claims about something very much like the strength of reasons.
One can even preserve transitivity. All one needs to do is to identify the
minimum degree of motivation that a reason could produce in an ideal
agent with its ideal-motive-requiring strength (or IM-requiring strength, for
short). This label is appropriate, since a reason with a greater minimum
degree of rationally permissible motivation will of course rationally require
more of us than a reason with a smaller minimum. Similarly, we can
identify the maximum degree of motivation that a reason could produce
in an ideal agent with its IM-justifying strength. This is an appropriate label,
because a reason with a greater maximum degree of rationally permissible
motivation will rationally justify us in acting against more reasons than
would a reason with a lesser maximum. That is, a reason with greater
IM-justifying strength will make it rationally permissible to take greater
personal risks, and so on. Transitivity of IM-requiring strength and of
IM-justifying strength are both consistent with (1) through (3).
An important question is whether some reasons might have a minimum
rationally permissible degree of motivation of zero.31 Given the existence
of ranges of rationally permissible motivation, it seems rather arbitrary to
assert that the minimum for such ranges must always be nonzero. If there
were such reasons, they would provide a counterexample to the internalism


. 22
( 38 .)