. 29
( 38 .)


Internalism and different kinds of reasons

Of course not all parties to the internalism/externalism debate are com-
mitted to the exclusive truth of only one of the views just outlined. Some,
for example, are Kantian internalists about reasons involving the needs or
objective interests of the agent, and yet hold that there are other reasons
that depend upon the agent™s contingent motivations.11 Such views are
more likely to approach the truth, and seem partly to have been shaped
by an implicit recognition of the justifying/requiring distinction. For sim-
plicity™s sake, I do not discuss these hybrid accounts. But it should be
clear that despite a more delicate feeling for the normative phenom-
ena, they are unlikely to be adequate unless they make their recognition
Many recent discussions of internalism and externalism are extremely
subtle, and it may appear that the present discussion is overlooking
important distinctions. But the distinctions between Humean internal-
ism, Kantian internalism, and externalism, as they have been drawn here,
are not intended to be subtle. The points that will be made here do not
depend in any way on the sorts of ¬ne distinctions that are made elsewhere.
Indeed, the reader will already have noticed that there is an important dis-
pute within the ranks of those whom I have described as ˜Kantian inter-
nalists,™ between those who hold that reasons stem from the existence of
objective goods or evils, and those who hold that reasons stem from some
imperative-producing feature of human rational processes.12 But in argu-
ing that Kantian internalism is the correct view of reasons in their requiring
role, there is no need to resolve this dispute. Signi¬cant work will have
been done if it is shown that neither Humean internalism, nor externalism,
is a viable view of requiring reasons. Here, then, is a summary diagram
explaining the relation between Humean internalism, Kantian internal-
ism, and externalism, as those notions are used in the remainder of this

11 For instance, David Copp™s ˜needs and values™ view of rationality places a similar division
in the center of his account of rationality. See Copp (1995), pp. 172“85. See also Foot
(1978b), pp. 148“56.
12 Garrett Cullity and Berys Gaut call the following two forms of Kantian internalism “the
recognitional view” and “legislative universalism,” respectively, attributing the former to
Aristotle and the latter to Kant. These two sorts of Kantian internalists share a commit-
ment to what Cullity and Gaut call “categorical reasons.” See Cullity and Gaut (1997),
pp. 3“5.

Brute Rationality

Q1: Can we have a reason to act in some
way regardless of our current motivational

Yes No

˜Humean™ internalism
Q2: Can we have a reason to act in some
way, and fail to be motivated by it, even if
we are fully rational?

Yes No

Externalism ˜Kantian™ internalism

In what follows, I will ¬rst present cases that seem to favor each version
of internalism over the other. Then I will argue that the plausibility of
each case depends very signi¬cantly upon whether we are considering the
justifying role of reasons, or the requiring role.

An example favoring the Humean interpretation over the Kantian
On January 13, 1982, Air Florida™s Flight 90 out of Washington™s National
Airport crashed into a bridge over the Potomac and landed in the freezing
water. An anonymous man, to whom a life-line was repeatedly given,
repeatedly passed it to others in the freezing water, until he himself ¬nally
succumbed to the cold and drowned. What are we to say about the rational
status of this man™s actions, and about the contributions made to their
rational status by the reasons that favor and oppose them? Certainly there
was a powerful reason against passing the life-line to another person: that
the man risked death in doing so. Were it not for the possibility of saving
other people, it would have been irrational to have acted against this reason.
That is, it would have been irrational to have passed the life-preserver to
someone else if no one else really needed it. But in the actual case, there
were also very powerful reasons in favor of the heroic actions: that several
lives could be saved. These reasons made it rationally permissible to act
against the other reason.
Let us now alter our description of the case. Suppose the heroic man
was in fact not motivated in his heroic actions by the thought of saving the

Internalism and different kinds of reasons

freezing passengers. Rather, a woman with whom he had had one date, and
who worked at the local television station, had decided not to go out with
him again because, by his own admission, he never did anything heroic.
The crash gave him the opportunity to display some heroism, and he was
con¬dent that this woman, seeing his behavior in the editing room, would
reconsider her decision and give him another chance. Let us suppose that
he was reasonable in believing this about the woman. Let us also stipulate
that if he had not believed that she would see the footage of his actions,
and if he had not believed she would then go out on a second date with
him, he wouldn™t even have considered trying to save the passengers. This
is what is meant by saying that he was not motivated by the thought of
saving the passengers, but only by the prospect of a second date. And, let us
say, it also made no difference to him whether he thought the passengers
would really be saved by his actions “ he only cared that it looked as if
he cared. Let us even suppose that he almost ˜saved™ a passenger who was
already dead. What are we now to say about the rationality of this man™s
actions? The reasons that justi¬ed his risking his own life in the ¬rst, real-
life, case do not seem, in this ¬ctional case, to do so. That is, the fact that
the man knows that he is saving the lives of the passengers does nothing
to mitigate the irrationality of his actions. Despite the fact that he knows
that he is saving the lives of these people, his actions really demonstrate
only a pathological obsession with a woman whom he hardly knows.
This type of case suggests that the antecedent motivations of the agent
play a role in determining whether a given consideration is a normative
reason. For in a case in which an agent is strongly motivated to save the
lives of others at high personal risk, the fact that the action will proba-
bly save those lives seems obviously to count as a strong reason. But in
the case where the agent merely knows that his actions will save those
lives, but does not care about that fact, then the fact that the action will
probably save those lives does not seem to count as a reason at all. This
apparent dependence upon antecedent motivation argues in favor of the
Humean interpretation of internalism. Of course it does not provide a
deductive proof that the Kantian internalist is wrong, even in this case.
A committed Kantian internalist could still maintain that the heroically
strong motivation to save the freezing passengers would have been pro-
duced in any fully rational agent. This would probably imply that to save
one™s own life when one could have saved ¬ve others is in fact irrational.
Such a view, and the corresponding implication, seem false, but I will not
argue here that they are. At the end of this chapter the Kantian should

Brute Rationality

simply have less philosophical motivation to make such claims. For such
theorists should recognize, by that point, that they have been presenting
their arguments without realizing that reasons can play two distinct nor-
mative roles in determining the rational status of actions. In general their
arguments involve only the requiring role of reasons. And so they should
not assume that their arguments apply equally well when the primary role
of a reason is to justify.

An example favoring the Kantian interpretation over the Humean
Paradoxically, the following example comes from Bernard Williams, who
is himself a Humean internalist. The reader may or may not feel that
Williams has in fact provided a reductio of his own view, but the example
should at least show that any Humean as honest as Williams is will be
compelled to accept some counterintuitive consequences.
To set the stage for the example, Williams ¬rst concedes that “insofar
as there are determinately recognizable needs, there can be an agent who
lacks any interest in getting what he in fact needs.”13 The example itself
is that of a sick person who has no desire to take the medicine that will
restore his health. What should we say about the rationality of this sick
person™s refusal of medicine? Recall that the case is not one in which
the agent wants to die, perhaps as a release from suffering. Rather, this
sick person simply lacks a desire to be healthy. Williams™s point would
remain the same even if we stipulated that this agent knew that, if restored
to health, he would live a life of uninterrupted virtue and felicity. Most
nonphilosophers, including trained psychologists, would almost certainly
conclude that such a person was suffering from a mental illness (perhaps
depression), and that part of the very illness lay in a failure to be moved
by reasons that would move a more rational person.14 But Williams, and
other adherents to Humean internalism, cannot say this. Rather, Williams
is very explicit in endorsing the view that if such an agent:
really is uninterested in pursuing what he needs; and this is not the product of false
belief; and he could not reach any such motive from motives he has by the kind
of deliberative processes we have discussed; then I think we do have to say that in
the internal sense he indeed has no reason to pursue these things.15

13 14 See Deigh (1996), pp. 133“59.
Williams (1981), p. 105.
15 Williams (1981), p. 105. It is worth noting the phrase “I think we do have to say” here.
This is not the kind of phrase that precedes a conclusion one regards as independently

Internalism and different kinds of reasons

In this case, the failure to have a desire for health seems to count against
the rationality of the agent. In general, if an agent is completely uncon-
cerned with things like the prospect of disease, death, injury, loss of free-
dom, and so on, this does not lead us to conclude that the agent has no
reason to avoid these things. Rather, such failures in motivation are all
the data we need in order to conclude that the agent is not completely
rational. Thus the example of the irrational sick person is consistent with
the Kantian interpretation of internalism, but argues against the Humean
interpretation of internalism. Of course it does not provide a deductive
proof that the Humean is wrong. When we confront a stubborn Humean
with a sick person who refuses medicine, that Humean can always dog-
matically assert that the sick person really does have some desire “ however
weak “ for health, or that the sick person really has no reason to take the
medicine. But for the Humean who takes the ¬rst of these routes, the
question of the strength of the reason will remain a problem, since it will
be hard to get a strong reason out of a weak motivation. And the second
route is not very attractive. In any case, at the end of this chapter the
Humean should simply have less philosophical motivation to make such
dogmatic claims. For such theorists should recognize, by that point, that
they have been presenting their arguments without realizing that reasons
can play two different normative roles in determining the rational status
of actions. In general their arguments involve only the justifying role of
reasons. And so they should not assume that their arguments apply equally
well when the primary role of a reason is to require.

th e p rop o sal
The overarching critical point of this chapter is that the ongoing debate
about practical reasons internalism has been hopelessly confused on
account of a failure to recognize that justifying and requiring are two
separate roles of normative reasons. The remainder of the chapter offers
arguments in favor of the following two more positive claims:
(1) Kantian internalism is true of requiring reasons.
(2) Externalism is true of justifying reasons.

The irrational sick person revisited
In the case of the sick person who refuses medicine the reason at issue “ that
the medicine will restore his health “ is relevant as a requirement on action,

Brute Rationality

and not as a justi¬cation for acting against any other reasons. This can be
seen by noting that in discussion of this example, what is at issue is whether
or not it would be irrational to ignore this reason, and to continue to refuse
the medicine, unless one had a powerful opposing reason that could justify
doing so. The Kantian internalist™s plausible position is that the agent would
be irrational, and that this is true regardless of the antecedent motives of
the agent. If the agent did not have the required corresponding antecedent
motivation, this would only show that the agent was to some degree irra-
tional. It would not show that the reason was not a reason. Here is the
argument for the Kantian view. Suppose that we have some reason, and
that it rationally requires an action. If this is so, then what is being required
by the reason is not only action, but also motivation itself. The point can
be put in the form of a rhetorical question: How could we be rationally
required to do anything in particular “ to avoid illness or pain, for example “
unless we were also rationally required to have the motives that would lead
us to do that particular thing? Suppose, for example, that one of our ratio-
nal requirements is the requirement to preserve our ability to reason. And
suppose that on a certain occasion it is necessary to take some medication
to comply with this requirement. If an agent takes this medication with
the false belief that it will destroy his ability to reason, is he nevertheless
immune to rational criticism? The above rhetorical question is, I think,
what stands behind the Kantian internalist™s intuition about the possibility
of rational assessment not only of actions, but of motives. Where require-
ments on action are at issue, there are implicit requirements on motives.
And this means that the motives cannot be antecedently necessary condi-
tions on the requirements, and, a fortiori, that requiring reasons do not
depend upon antecedent motivations. That is, the Humean interpretation
of internalism does not apply to requiring reasons. As a consequence, the
Kantian interpretation does apply, although this is not merely the result
of a simple argument by elimination. Rather, it is because the reasons at
issue are rational requirements that the Kantian view is correct. This is
because to say that these reasons have some requiring strength is simply
to say that failing to be motivated by these reasons is suf¬cient to convict
the agent of some degree of irrationality. And that is just what the Kantian
Derek Par¬t offers the same sort of argument in favor of Kantian inter-
nalism, and in favor of the idea that there are rational requirements on
motives. He argues as follows:

Internalism and different kinds of reasons

We have reason to try to achieve some aim when, and because, it is relevantly
worth achieving. Since these are reasons for being motivated, we would have these
reasons even if, when we were aware of them, that awareness did not motivate us.
But, if we are rational, it will.16


. 29
( 38 .)