. 9
( 38 .)


psychological processes by which one discovers ef¬cient means, and by which motivation
to take those means is generated, are rational processes. For an excellent discussion of
the relation between the processes of practical reasoning and the principles of practical

Brute Rationality

after restricting the nature of reason, Hume goes on to demonstrate that it
cannot produce motivation out of whole cloth. With this description of
Hume™s argument, Korsgaard has at least illustrated her point: she has shown
us a philosopher whose motivational skepticism is clearly dependent on
an antecedent content skepticism.
But there is another way in which Korsgaard exploits Hume, to reach
a more general conclusion. Because of his view that rational principles
can only judge of truth and falsity, Hume believes they have nothing to
say about motives or actions, since neither motives nor actions can be
true or false. Motives, and actions based on them, cannot, for Hume,
be straightforwardly irrational. They can only be derivatively irrational.
They are derivatively irrational when they are based on false beliefs about
objects or about the ef¬cacy of means. But Korsgaard points out that
there is a way in which reason might be seen to bear on action in a less
derivative way, even on a view very close to Hume™s. True irrationality,
as she calls it, would be to choose inef¬cient means in full knowledge:
without any mistaken beliefs. Hume explicitly denies the existence of
true irrationality.13 That is, he denies that the principle ˜choose ef¬cient
means to your ends™ is a normative principle, for he denies it one of the
necessary characteristics of a normative principle: the possibility of not
being followed. By allowing the possibility of true irrationality, Korsgaard
turns the principle into something normative.
Now, even if true irrationality is granted, by taking the principle ˜choose
ef¬cient means to your ends™ as a normative principle of reason, it need
not be granted that reason can generate motivation without the existence
of some antecedent end. Certainly the particular instrumental principle
under discussion is incapable of doing it, simply because it is instrumental,
and therefore depends upon contingent ends to supply it with content.
But Korsgaard argues that once true irrationality is granted, our attitude
towards other putative principles of reason should become more liberal. In
particular, our attitude might become so liberal as to allow principles that
do bear directly on particular actions. It is this possibility that Korsgaard
wants to argue for, and so she spends some time arguing for the existence
of true irrationality.

rationality, see Mele (1989). In light of Mele™s paper, it should be borne in mind that the
processes of practical reasoning may be a subset of rational processes. That is, the processes
by which one acts rationally, and the failures of which can explain irrational behavior,
may include more than merely processes of reasoning.
13 Hume (1978), p. 416.

The criticism from internalism

True irrationality
Korsgaard argues that it is possible that a person could engage in ¬‚awless
means/end reasoning, and yet fail to be motivated to take recognized
means to her acknowledged end. Her extremely plausible explanation for
this is simply that there might be interference in the transmission of motive
force from acknowledged ends to recognized means. There are things
that prevent us from acting rationally: rage, depression, drugs, arrogance,
aneurysm. Here the case of theoretical irrationality does provide a useful
analogy.14 We all admit that there might be theoretical reasons decisively in
favor of believing some claim, and yet I might be unconvinced by them.
This would not make them any the less reasons. It might only be evidence
that I am irrational. So Korsgaard thinks Hume is wrong, as he surely is, in
failing to acknowledge the existence of true practical irrationality.
Once it is admitted that it is possible to be truly irrational, Korsgaard
points out that Hume™s limitation of rational processes to mathematics
and means/end reasoning is less compelling. She illustrates this point with
the example of prudence.15 Hume takes prudence to come from a pas-
sion that, should it disappear, would take with it both the motive and
the reason for prudential behavior. In this respect the ˜prudential™ pas-
sion is for Hume just like any other contingent passion. It possesses no
special rational authority. One can have it, or lack it, and one™s lacking
it has no bearing on one™s rational status, or the rational status of one™s
actions. But if one admits, as Hume does not, that one can sometimes
simply fail to be responsive to rational considerations “ if one admits the
possibility of true irrationality “ then when one fails to take means to
one™s greater good, there are two ways of describing what has happened.
The ¬rst is Hume™s explanation: one simply, and in a rationally accept-
able manner, did not have one™s greater good as one™s end. The second
is Korsgaard™s explanation: one was irrational for failing to have had one™s

14 In fact, the analogy with morality supports the same point. There may be moral consider-
ations that count decisively in favor of some action: that I fail to perform the action does
not alter that fact; it only means that I am morally imperfect.
15 This choice of illustration by Korsgaard shows that she takes prudential motivation as a
potential rational process. And yet the reasoning involved in being prudent may be of only
the broadly instrumentalist and constitutive sort countenanced by motivational skeptics.
This suggests that Korsgaard may agree that rational processes need not all be processes of
reasoning. See note 12 above. Such a view allows one to reject the internalism requirement
while still holding the view that Alfred Mele calls ˜motivational internalism™: roughly the
view that arguments will generally only have a practical effect on an agent if that agent
has the appropriate antecedent motivations. See Mele (1989), pp. 422“29.

Brute Rationality

greater good as one™s end. One is precluded from taking the second expla-
nation if one has, like Hume, limited practical reasoning to mathematics
and means/end reasoning, but that limit needs defense. By admitting the
possibility of true irrationality, one can admit that prudence has rational
authority even though it sometimes fails to motivate. That is, the obvi-
ous fact that many people are not prudent no longer rules out, or even
argues against, prudence as a rational principle. One™s decision to admit
prudence as a rational requirement will depend on arguments that cannot
be refuted simply by showing that prudence sometimes fails to motivate.
This way of dealing with putative rational principles on which people
sometimes fail to act widens the ¬eld of principles for which philosophers
can argue.

It is of course possible to acknowledge a very wide ¬eld of rational
principles “ to doubt that Hume™s notion of rational processes is even
remotely complete “ and yet still maintain that motivation cannot come
from reason alone. That is, one can maintain motivational skepticism even
if one grants that rational principles involve much more than the princi-
ple that one should choose ef¬cient means to one™s acknowledged ends.
This is what Bernard Williams does.16 And it is Williams who is Kors-
gaard™s next target. Williams is arguing for the view that reason-claims
must imply a motive. That is, he claims that if I say that you have a reason
to , I must be taken to mean you have some desire or goal that would
be served by your -ing, or that you adhere to some principle that speaks
in its favor, or that some other similarly motivational claim is true of you.
Of course, you might have some other goals that yield reasons against
-ing. Neither Williams nor Korsgaard are committed to the implausible
view that one can only be said to have a reason to do an action if that
action turns out, all things considered, to be favored by reason. Indeed,
the notion of being so favored, all things considered, presupposes that
there may be reasons pulling in different directions. Any given action will
probably have reasons both for it and against it, and hence not all reasons
will, or even could be, acted on, even by a perfectly rational agent. That is
why the discussion is always cast in terms of motivation, and not of action
or intention. It is at least plausible that all reasons provide motivation to

16 Williams (1981).

The criticism from internalism

rational agents who have them. It is not plausible that all reasons provide
such agents with suf¬cient motivation to prompt actual action. To restrict
talk of reasons in such a way that they always do provide rational agents
with suf¬cient motivation to act is to distort the notion of a reason beyond
Williams argues that a reason-claim must imply a motive, because oth-
erwise we could not use the reason to explain the action for which it was
claimed to have been a reason. Therefore, the argument continues, unless
a consideration would motivate an agent if that agent were fully rational, it
cannot be a reason for that agent. This, again, is what Korsgaard calls ˜the
internalism requirement.™ Though Korsgaard is arguing against Williams,
she does think that the internalism requirement is “clearly correct.”17
Korsgaard admits that different considerations have the capacity to moti-
vate different individuals. It might seem therefore that the internalism
requirement entails that we can only determine whether or not a consid-
eration is a reason for an agent by looking at the particular motivational
capacities of that agent.18 This is in fact Williams™s position. He claims that
reasons are always relative to what he calls the agent™s ˜subjective moti-
vational set™ “ a collection of motivational entities that includes desires,
but that can also include principles one adheres to, projects one has, and
other things not resembling desires except in being sources of motiva-
tion. Williams assumes that rational processes must start from something
that can motivate: that they must start from something in one™s subjective
motivational set. But Williams does not take means/end reasoning to be
the only rational process by which motivation for particular actions can
be teased out of one™s subjective motivational set. He even goes so far as
to allow that imagination might be such a process: one imagines what it
would be like to achieve some end, and a desire for that end might be
created. But even in this case Williams still claims that the capacity for
imagination to engender a desire is dependent on the contents of one™s
subjective motivational set.
But, Korsgaard argues, once we, like Williams, have abandoned the
Humean view that logic and means/end reasoning are the sole ratio-
nal processes, there may turn out to be processes of practical reason that
can yield motivation on their own. Korsgaard™s point is that until ratio-
nal processes are exhaustively inventoried, it is unsettled whether they
can motivate us on their own. For example, if all rational people could,

17 18
Korsgaard (1996b), p. 329. Korsgaard (1996b), p. 325.

Brute Rationality

in principle, be convinced to accept some particular practical principle
by some yet-to-be-discovered argument, then, even on Williams™s view,
that principle would yields reasons for everyone. According to Korsgaard,
Williams simply denies that there could be an argument like this. But
this shows that Williams™s motivational skepticism about practical rea-
son depends on a denial that there are any substantive rational princi-
ples that could be shown, by argument, to be valid. Therefore Williams
cannot be taken as showing this to be so. It only looks like Williams
limits the principles of practical reason by means of the internalism
But, Korsgaard rightly argues, the internalism requirement cannot, by
itself, limit the content of principles of practical reason. This is because
of the possibility of true irrationality relative to any proposed principle.
When someone fails to be motivated by the reasons that some putative
principle supplies, we can always preserve the validity of the internalism
requirement by claiming that this failure is suf¬cient to show that the
person is acting irrationally. Therefore the internalism requirement cannot
limit the content of principles of practical reason to ones that only generate
reasons out of an agent™s antecedent motivations. For example, consider
the principle ˜Death is to be avoided at whatever cost.™ Though such a
simplistic principle is certainly not valid, it is not ruled out merely by the
internalism requirement. For we can consistently maintain both that it is a
rational principle, and that the internalism requirement is true, by claiming
that anyone who does not avoid death at all costs is acting irrationally.19
Of course, our ability to make this claim does nothing to favor the position
that the principle ˜Avoid death at whatever cost™ will provide motivation
to every rational agent.20 Rather, our ability to make this claim shows
only that citing the internalism requirement cannot, by itself, rule out that

19 Elsewhere Korsgaard denies that she defends the internalism requirement in this way. See
Korsgaard (1997), p. 219 n. 11. There she seems to advocate the position that even after
one has shown that some principle presents an unconditional normative requirement, one
must still show that rational agents will be motivated to act in accord with it. This suggests
that, at least at that point, she is taking the notion of rationality, as applied to agents, as
a purely descriptive notion. At the very least, it suggests that she is using the notion of
rationality here in a way that is conceptually independent of the idea of complying with
unconditional normative requirements.
20 See Dreier (1990), pp. 12“13. Dreier points out that this way of preserving the internalism
requirement, because it is available to the advocate of any rational principle, cannot be
used to argue in favor of any particular rational principle.

The criticism from internalism

con se que nc e s of th e arg um e nt : th e i nte rnal i sm
re qu i re m e nt i s too st rong
According to Korsgaard, the question is open as to whether there might
be some processes of practical reason that can motivate all rational agents
regardless of their contingent antecedent motivations. This opening of
the question can of course be done without bringing the internalism
requirement itself into question: this is in fact what Korsgaard does. But
the possibility of such principles of reason drastically changes the way we
can explain the truth of the internalism requirement. Consequently it
suggests a different understanding of the requirement itself.
Here is how someone like Williams might understand why the inter-
nalism requirement is true. Adherents of views like Williams™s believe a
reason-claim is true in virtue of the existence of some antecedent moti-
vation from which rational processes could produce a motivation to do
some particular action. The picture this creates is one of a sort of reservoir
of motivational fuel, from which rational processes, like the processes of
internal combustion and transmission, extract and direct energy in speci¬c
directions. Now, if an automobile has a supply of fuel, and is in gear and
running, but does not have its wheels move, then there is no question but
that there is something wrong with the transmission mechanism. And in
the same way, if it is true that a person has a reason, and therefore has
˜motivational fuel,™ and is awake and aware of the reason, and yet that
person is not motivated by that reason, it is also tempting to say that there
must be a failure in rationality. This is a natural analogy when rational
principles are conceived of as principles that essentially govern only the
transmission of motivation. Thus, when one takes one™s own ends as the
ultimate source of all reasons, as Williams does, it is natural to claim that
it is irrational not to be motivated, to some degree, by every reason one
has.21 That is, it is natural to adhere to the internalism requirement.
So, if one grants Williams his motivational skepticism, he might seem
to provide an explanation of the truth of the internalism requirement. He
might be seen to do this by saying that all reasons stem from antecedent
motivation, so that an appropriate motivation always exists, at least in a
21 One might point out that in some cases incompatible means and are available. Then it
would be plausible to say that even a rational agent either would not be motivated to do
or would not be motivated to do . But in this case it is still true that the same reason favors
both and . That reason is therefore providing motivation, no matter which option the
agent picks. So this kind of situation (which is the typical one) does not undermine the
temptation of the internalism requirement.

Brute Rationality


. 9
( 38 .)