. 8
( 38 .)


even on the modi¬ed view, altruistic reasons would have very different
requiring and justifying strengths; any particular altruistic reason would
be able to justify much more than it would be able to require. Because
of this difference in justifying and requiring strength there would still be
many cases, like the job-interview case and the charity case, in which
one would not be rationally required to act on the stronger (justifying)

The criticism from internalism about
practical reasons

The previous chapter argued for the existence of a class of practical
reasons “ purely justi¬catory practical reasons “ that have no power to
make actions rationally required. It also suggested that altruistic reasons
form a signi¬cant part of this class. The strategy of that chapter was to point
out that a wide range of moral views grant, either explicitly or implicitly,
the existence of moral considerations that function in the same way. That
is, they grant the existence of considerations, the presence of which can
change an otherwise immoral action into a morally permissible one, and
yet that do not seem to be the sort of considerations that must weigh,
in a positive way, in the motivational economy of a virtuous person. On
the strength of some examples, and of further points of analogy between
morality and practical rationality, it was then suggested that practical ratio-
nality also includes considerations that play a similar normative role. But
if practical rationality does include reasons that function in this purely jus-
ti¬catory way, then it seems that there could be actions, favored by such
reasons, that even a rational agent might not be motivated to perform “
and this could be true even when the agent knows that there are no coun-
tervailing reasons. This implication offends against a dogma that is virtually
universally accepted by contemporary moral theorists: roughly, that ratio-
nal agents will be motivated to some degree by any practical reasons of
which they are aware, and which are relevant to their choice of action.1
Even those who balk at such a stark statement of the view typically nev-
ertheless accept the idea that in the absence of countervailing reasons, a
rational agent must act on any reason of which she is aware.2 It is the

1 See Cullity and Gaut (1997), p. 3. It is worth noting that not one contributor to the
Cullity and Gaut volume challenges the internalism requirement. See also Cohon (1986),
esp. p. 556; Smith (1994), pp. 60“62 and 151“77; Nagel (1970), pp. 66“67; Velleman
(1996), pp. 694“726; Darwall (1983), p. 52.
2 Joseph Raz, for example, offers some compelling arguments against the bolder view, but
accepts the weaker. See Raz (1999b), p. 99.

The criticism from internalism

purpose of this chapter to rebut an objection to the idea of purely justi¬-
catory reasons that emerges from an uncritical acceptance of such a view.
The strategy is borrowed from an argument that Christine Korsgaard has
recently used in rebutting an objection to the Kantian view that practical
reason can motivate action by itself. The objection she addresses comes
from Hume and neo-Humeans such as Bernard Williams.
Hume and Williams, among many others, argue, against Kantians, that
practical reason is incapable of generating motivation on its own.3 This
view is best interpreted as making the following normative claim: no par-
ticular basic motivations are, in themselves, rationally required or prohib-
ited. Christine Korsgaard calls this view ˜motivational skepticism,™ and in
“Skepticism about Practical Reason” she shows that, whether or not it is a
correct view, it should not be taken as fundamental.4 Rather, motivational
skepticism must be based on a prior view about what the principles of
rationality are. In particular, it must be based on a prior skepticism about
the existence of rational principles that can classify an action as irrational
based solely on a description of the action, and without additional infor-
mation about the contingent motivations of the agent who is performing
the action. This form of skepticism she calls ˜content skepticism.™ Another
way of characterizing content skepticism is the following: no action type
is simply irrational (or rational), regardless of what the agent wants. One
example of a principle denied by a content skeptic would be:

M It is irrational to refuse to take medicine that will save one
from lingering death, and restore one to perfect health.5

Another much more general such principle will be offered later in this

3 See Hume (1978), p. 415; Williams (1981), p. 105. Alfred Mele (1989) convincingly
advances a distinct but related causal view: that practical reasoning cannot produce motivation
underived from antecedent motivation. His view is compatible with the claim that, in order
to count as rational, an agent must have certain basic motivations that cannot necessarily be
reached by reasoning (such as aversions to pain, death, disability, etc.). Thus, as Mele rightly
notes, his position is neutral as between Humeans, such as Williams, and their critics, such
as Korsgaard. See Mele (1989), pp. 419, 432, and 436 n. 19.
4 Korsgaard (1996b).
5 See Williams (1981), p. 105, for a denial of this principle, based on a prior acceptance
of motivational skepticism. Williams would deny this principle even if it were expanded
to exclude the cases that make such a denial plausible: cases in which, for example, one™s
death produces great bene¬ts for others, or in which one anticipates that one™s rescue from
death will only lead to a life of great sorrow.

Brute Rationality

It would be more accurate to say that motivational skepticism must be
based on a prior view of what the principles of objective rationality are. But
Korsgaard does not seem to distinguish between objective and subjective
rationality. Indeed, because her ˜source of normativity™ is the agent™s own
practical identity, and because she takes reasons to be ˜endorsed impulses™
rather than facts, she seems to take the subjective notion of rationality to be
the fundamental notion.6 Nevertheless, this difference is not very impor-
tant here, for the dependence of motivational skepticism on an account
of subjective rationality can trivially be extended to a dependence on an
account of objective rationality, if one admits that objective rationality is
a distinct and more fundamental notion. The question of whether or not
certain basic motivations are rationally required is a question about how
a rational agent will act. A motivational skeptic would have to deny M,
even if it is understood as a principle of objective rationality, since if it
is objectively irrational to refuse the medication, then a fully informed
rational agent would be motivated to take it. Indeed, any rational agent
who believed that he was in a situation ruled out by M would be so moti-
vated. Thus even if M is understood as a principle of objective rationality,
it continues to entail that certain motives are rationally required. Because
of this, although the following discussion will often proceed in terms of
principles of subjective rationality “ for this is the way Korsgaard presents
her arguments against Hume and Williams “ it can easily be understood
in terms of objective rationality, simply by assuming that the agents who
¬gure in the examples and arguments possess all the relevant information
about their actions.
Korsgaard argues that until content skepticism has been established, it
remains an open question whether practical reason can generate moti-
vation without relying on contingent antecedent motivation. That is, it
remains an open question whether motivational skepticism is true. And
she does not think content skepticism has been established. This point “
that content skepticism has not been established “ coheres with the overall
picture of reasons and rationality that this book is advocating. But this
chapter argues that Korsgaard has not taken her arguments far enough,
and that their consistent application undermines the axiomatic status
of the dogma mentioned above: a dogma that Korsgaard herself, along
with virtually all contemporary moral theorists, also accepts. Following

6 See Korsgaard (1996a), pp. 94, 99 n. 8, 108.

The criticism from internalism

Korsgaard, we can call this dogma ˜the internalism requirement on practical
The internalism requirement is a motivational view about reasons that
Korsgaard clari¬es and endorses in the course of her argument. It holds
that for a consideration C to be a reason for agent A, it must succeed
in motivating A, given that A is rational, and that A is aware of C.8
That is, according to the internalism requirement, if a putative reason fails
to motivate an agent to whose action it is relevant, then either it isn™t
really a reason, or the agent is to some degree irrational. The internalism
requirement demands more of a reason than that it could motivate a rational
agent who had it. It requires that it would motivate a rational agent who
had it.9
This chapter argues that just as Hume™s and Williams™s motivational
skepticism depends upon a prior acceptance of content skepticism, the
internalism requirement depends upon the prior acceptance of the fol-
lowing view.
The requirement view: All practical reasons are prima facie rational requirements.
That is, if one acts against such a reason, then one is either acting irrationally,
or one is acting on other countervailing practical reasons of at least equivalent

The conclusion, similar to Korsgaard™s, will be that until the require-
ment view has been established, it remains an open question whether
the internalism requirement is valid. And the requirement view has not
been established. Indeed, though it is sometimes stated, one would be
hard-pressed to ¬nd any argument for it at all. One reason for this lack
of argument is that the internalism requirement, which follows from the
requirement view, has not been widely challenged, so that there is no felt
7 Recently, Sigr´ n Svavarsd´ ttir has provided compelling arguments for a similar conclusion
u o
regarding a different but related view: moral judgment internalism. That is, she undermines
attempts to use consonance with moral judgment internalism as an adequacy condition on
moral theories. See Svavarsd´ ttir (1999), pp. 218“19.
8 For the remainder of this chapter, the ˜awareness™ rider should be taken as understood.
9 There are also certain types of externalists who adhere to the internalism requirement as
it is here stated. The argument of this chapter is directed at such externalists as well. See,
e.g., Par¬t (1997), p. 101, and Brink (1986), p. 36. In this latter article, Brink argues for
the conceptual possibility that the recognition of a moral requirement might fail to give
an agent a reason to act, and this is partly what his externalism consists in. But Brink also
equates the question of whether the recognition of a moral obligation gives an agent a
reason for action with the question of whether an agent would be irrational subsequently
to fail to care about such moral requirements. This equation depends upon the internalism
requirement as it is here understood.

Brute Rationality

need to defend it. Of course there have been principled defenses of other
forms of internalism, but these often rest on an unargued assumption of
the internalism requirement on practical reasons, or of the requirement
If, against the requirement view, some reasons are not prima facie ratio-
nal requirements, then the internalism requirement will in fact be false. To
make the parallel with Korsgaard™s argument against Hume and Williams
more obvious, it is useful to note that the internalism requirement is a
motivational thesis about reasons, and that the requirement view is a the-
sis about what the content of a reason-claim is. In a sense, the argument
offered here is the same argument as Korsgaard™s. The difference is only
that the present argument recognizes a wider variety of potential principles
of rationality than Korsgaard does, and, hence, the potential for a more
complex relation between reason-claims and claims about the rationality
of actions.

kor s ga ard ™s arg um e nt : mot ivat i onal ske p t i c i sm
de pe nd s on conte nt ske p t i c i sm
Although Korsgaard argues that motivational skepticism about practical
reason always depends upon a prior acceptance of content skepticism, she
does not directly attack content skepticism itself. And so she does not
directly attack motivational skepticism either. Rather, she is concerned
with the link between the two forms of skepticism. Her point is that
philosophers who wish to argue that reason is unable, on its own, to
motivate action, must ¬rst argue about which principles are to be admitted
as rational principles. And, if she is right, they cannot reasonably use
motivational skepticism as a premise in such an argument. Korsgaard uses
Hume as an example of a philosopher who is sometimes interpreted in a
way that commits him to this mistake. But she shows, convincingly, that
his actual argument respects the priority of content skepticism.

10 See, e.g., Smith (1994) for a defense of moral judgment internalism based on an unde-
fended assumption of the internalism requirement on practical reasons. See Foot (1978b),
p. 152 for an implicit argument for the internalism requirement on practical reasons, based
on an undefended assumption of the requirement view. Derek Par¬t makes the same move
in Par¬t (1997), pp. 101, 130. See also Broome (1999), pp. 400“1, for a relatively clear,
but undefended statement of the requirement view. Broome states the position in terms
of ˜ought,™ rather than ˜is rationally required to,™ but the position is essentially the same,
since for Broome ought-claims are ˜strict demands,™ rather than pro tanto or prima facie

The criticism from internalism

On Hume™s view reason only helps us choose ef¬cient means to ends.
For Hume, normative standards for the choice of those ends come from
another source. Reason cannot even rank ends, or determine that we
should satisfy the ends we regard as ˜our greatest and most valuable enjoy-
ments.™11 Hume™s argument is that reason is concerned only with abstract
relations of ideas and relations between objects. When reason is con-
cerned with the ¬rst of these types of relations, it is doing mathematics,
which cannot give rise to motivation. And when reason is concerned
with the second type of relation, it is involved in causal reasoning, which
gives rise to motivation only from pre-existing motivation. Thus, by sur-
veying the types of rational processes, we can see that neither of them
can generate motivation on its own. As Korsgaard notes, this particu-
lar argument depends in an obvious way on presuppositions about what
processes count as rational processes. That is, if we go so far with Hume
as to grant that reason is answerable only to principles of mathematics
and causality, then it is indeed quite plausible that reason cannot, on its
own, direct an agent towards one action rather than another, indepen-
dently of some contingent and antecedent motivation. If Hume could
defend his notion of the content of practical reason “ its limitation to
math and science “ he would therefore be able to defend his motivational
What Korsgaard is rightly pointing out here is that it only looks as if
Hume is arguing from motivational skepticism to a restriction on what
counts as a rational process. But he is not. Rather, the argument goes in
the other direction. Hume argues that reason™s essential function is to judge
truth and falsity. It does this by making determinations of the accuracy of
representations. And he claims that actions and passions do not represent
anything, and therefore cannot be true or false. Thus reason has nothing
to say about them. For Hume, no rational principle rules in, or rules
out, any particular action, and, correspondingly, no rational process, on its
own, necessarily leads to a motivation to any particular action.12 That is,

11 Hume (1978), p. 416.
12 The logical relation between rational principles and rational processes is a complex one,
and no explicit theory of that relation will be offered in this book. But it should be clear
that if it is a rational principle that one should take ef¬cient means to one™s ends, then the


. 8
( 38 .)