. 16
( 38 .)


so forth, but that™s another matter.
Does it follow that you have no reason to avoid activities which are
unchaste, just because they are unchaste? The question may be skewed by
the controversial ethical status of chastity; so take the less tendentious case
of gratitude. Imagine that Tom simply has no sense of gratitude. It™s not
just that he subscribes to an ethical ideal which regards gratitude as a futile
emotion to be suppressed, in the way that Annabel does. He simply never
feels it, never expects it “ he just doesn™t see what this thing called gratitude
is about. So when Mary goes out of her way to help him, it™s not possible that
he should thank her for that reason, that is, simply and solely because he
sees for himself that gratitude is appropriate. (He may of course recognize
prudential reasons to observe the social conventions he™s been told about,

26 Williams (1995), pp. 38“39.
27 Williams (1995), pp. 37“38.
88 John Skorupski

etc.) Does it follow that Tom does not have that reason for thanking Mary “
that that particular fact is not a reason for him to thank her?
I think our response to this kind of question is interestingly uncertain.
In one mood we want to say “Certainly he has reason to thank her, whether
or not he can see that he has “ look at what she™s done for him!” However
in the remaining sections I want to argue that it™s also important to take
seriously the opposing answer, which says that he does not have reason to
thank her.28
This latter response can be seen as a kind of internalism. But before we
move on to examining it let™s consider whether it would help Williams to
make a case for internalism in his particular sense. If we say that Tom has
no reason to thank Mary, the thought that moves us is that a fact cannot
be a reason for an agent to • if it cannot be recognized as such by him.
This thought concerns not the agent™s knowledge of the reason-giving facts
but his ability to recognize them as reason-giving; it is not a question of
whether the agent has the information which enables him to know that the
reason-giving fact obtains, or that other facts obtain in virtue of which this
fact becomes a reason. The point is that the agent must have the ability to
recognize the reason-giving force of that fact (or combination of facts) were it
to obtain. He must be able to appreciate in his own right, or see for himself,
that that fact or combination of facts as such, were it to obtain, would indeed
be a reason to •. Let™s highlight this claim:

(III) X is a reason for A to • only if A has the ability to recognize that were
X to obtain, that would be a reason for A to •.

(III) is distinct from (I), even if we interpret (I) in a liberal, non-Humean, way
which allows that agents™ beliefs about reasons can motivate. Just because A
has the ability to recognize that X would be a reason to • it does not follow
that he actually believes that it is. He may not have thought about it, and
you can™t be motivated by a belief you don™t have. Also, as we™ve noted, an
existing capacity to recognize a reason for acting cannot be described as an
existing motive. So (III) does not sustain the view that A has a reason to •
only if •-ing would serve a motive in A™s S.29 In short (III) allows that in
Williams™ sense there can be external reasons.

Nonetheless, there is still a certain point in calling it “internalism.” For (III)
says that only considerations which the agent has the ability to recognize,

28 Trivially of course he does not have a reason to thank her in sense (iii) (§1); that is, he does
not believe that there is reason for him to thank her.
29 And of course it is quite consistent with denying the converse: that if •-ing would serve a
motive in A™s S then A has a reason to •.
Internal Reasons and the Scope of Blame

for him or herself, “from within,” as reasons, can be reasons for that agent.
Moreover it has a strong af¬nity to the requirement of effectiveness, which
Williams regards as crucial. If an agent simply lacks the ability to recog-
nize a type of consideration as a reason for •-ing, then it is not possible
that he should • for that reason, and so, by the requirement of effective-
ness, this cannot be a reason for that agent, even if we would regard it
as a reason. In contrast, if the agent can recognize the consideration as a
reason to •, then •-ing for that very reason opens up as an option for
him. It becomes possible that he should do so.30 Let™s call the constraint
on what it is for something to be a reason for someone, captured by (III),
cognitive internalism.31 Cognitive internalism is consistent with the view
that beliefs about reasons can themselves motivate. And Kant can certainly
be classed as a cognitive internalist: indeed, (III) is simply a corollary of
his central ideas about reason and autonomy. Autonomy, for Kant, is the
capacity to see reasons for yourself, or to “give yourself” reasons, and only
autonomous agents, who give themselves reasons, can be said to have rea-
sons. Moreover cognitive internalism has bite: as we shall see, it has many
of the implications for the scope of blame that Williams believes his inter-
nalism to have. If we view Williams™ critique of modern assumptions about
morality from this standpoint, it retains its interest even for those who are
unimpressed by Humeanism, psychological or normative, about practical


I suggested that we don™t have a clear-cut response as to the truth of (III).
Let™s go back to Tom, and let™s suppose that he suffers from a psychological
syndrome that makes him incapable of experiencing or understanding feel-
ings like gratitude. Can we say that Tom has reason to thank Mary?32 One
might say “he has every reason to be grateful to Mary “ if only he could see
it.” However, does this mean that he does have reason to be grateful “ or

30 Note, however, that it may not follow from your having the ability to do something that it™s
possible for you to do it “ it depends on how we interpret “possible.” (You have the ability
to walk a tightrope, but I™m going to distract you whenever you try.) Williams™ requirement
of effectiveness should be so understood as not to fall to this kind of point.
31 (III), as noted, has an af¬nity with the requirement of effectiveness, but it™s narrower. If
you can see something is a reason, but that recognition has no motivating force for you
(perhaps given your psychology it could not have, then (III) still allows a consideration to
be a reason for you, whereas the requirement of effectiveness rules it out).
32 Exclude again the indirect reasons he may have, such as reasons of prudence to conform to
what he can see are the prevailing social conventions, for example.
90 John Skorupski

just that he would have if he could see it? Should we read it in the ¬rst way?
One also could say that the wounded bird has reason to thank the gardener
who looks after it with loving care “ if only it could see it. Or that the cat
has reason not to torture the mouse, if only it could see it. In saying that
we are not, I think, ascribing reasons to the bird or the cat.
What would count in favour of a practice of ascribing reasons to agents
who are quite impervious to such reasons? It may be thought that the
universality of reasons pushes in that direction.33 Surely anyone has reason
to thank a person who has helped them. True “ but does “anyone” include
Tom? If this seems evasive remember that “anyone” plainly does not include
the wounded bird, or the cat. Whether it includes Tom is precisely the
Reasons are universal in this sense: if the fact that p is a reason for A to
• in circumstance C then for any x, the fact that p is a reason for x to • in
circumstance C.34 But over whom or what does x range? Who or what is
included? The universality of reasons tells us that reasons of gratitude are
universal, but not over what domain they are. If any x has reason to thank a
person who has helped them, then every x has; but that does not include the
wounded bird, so the wounded bird does not fall within the scope of this
class of reasons, and thus within the range of “x.” The cognitive internalist
explains this by invoking the obvious fact that the bird is not an agent capable
of appreciating considerations of gratitude and their reason-giving force.
It falls outside the range of “x” because the range of “x” is constrained
by (III). On this explanation it follows that if Tom is really incapable of
understanding or feeling gratitude then he too falls outside the range of x,
as far as reasons of gratitude are concerned. If we ¬nd this disconcerting, it
is because we don™t want to believe that someone in other respects so like
us could be quite incapable of appreciating a class of reasons that we ¬nd
obvious. Or more generally, because we want to get all human beings into
the scope of all reasons. But now we need to explain why.
Reality is more complex than the stark example of Tom. There is wide
variation in, and a thick margin of unclarity about, the degree to which
people are able to appreciate all the variety of types of reason. Furthermore,

33 As suggested by Scanlon, (1998), p. 367, p. 372. (Scanlon allows that reasons can have what
he calls “subjective conditions.” The variably reason-giving force of ideals, discussed at the
end of this section, would be an example.)
34 Any occurrences of “A” in “p” and “C” must be replaced by “x.” This allows for references
back to the agent, that is, for “agent-relativity.” For example: if the fact that revising for
my exams will help me to pass is a reason for me to revise, given that passing is to my
advantage, then for any x the fact that revising for x™s exams will help x to pass is a reason
for x to revise, given that passing is to x™s advantage.
Internal Reasons and the Scope of Blame

the degree and extent to which a person has the ability can vary greatly
depending on the circumstances. People can suffer a temporary blockage
on their ability to see a reason, or they can have an ability that has not
yet developed. We can then talk about the reasons they have, even though
they™re unable to see them, or are not yet able see them “ because we
take as our benchmark their normal ability or the ability that they have a
determinate potential to develop.35
Williams considers a case of the ¬rst kind in Ethics and the Limits of
Philosophy.36 A despairing teenager, Susan, attempts to commit suicide. Even
though she can see, in a way, that things will be ¬ne in three months™ time,
she doesn™t care. Williams™ discussion centres on whether we can say that it™s
in Susan™s real interest to stop her; our concern is with the related question:
does Susan have reason not to commit suicide? Let™s assume that for the
moment she just can™t see the fact that things will be better in three months
time as a reason not to commit suicide. Her inability is caused by the very
depth of her despair. In this case one can truly say “Look, there really is
reason for you not to do this. You will feel much better in three months,
and that really is a reason. You™re not in a state to appreciate that just now,
but believe me it™s true, and you™ll agree with me later.” We reconcile that
ascription of reasons to Susan, even in her suicidal state, with (III) by relying
on her ability to appreciate these reasons when she is in her normal state.
Or consider some little boys playing a game of running across a railway
track at the very last moment in front of an oncoming train. It™s not that
they don™t appreciate the danger “ on the contrary, the danger is the whole
point. Rather, they don™t value the bene¬ts of the life before them above the
bene¬ts of the glory and respect they gain from their gang right now. Don™t
consider a little boy who fully sees the imprudence of the risk but is seduced
by the desire to belong; consider rather a dashing one who really subscribes
to the ideal of bravery and cool, and regards prudential considerations as
beneath him. Can we say that this little boy has more reason to avoid play-
ing this game than he appreciates? We may think that we can; in which
case we may reconcile our response with (III) by appealing to the assess-
ment he will make when his capacity to appreciate and weigh reasons has

35 “determinate potential” raises tricky questions, of course. How determinate? If there™s
reason to think capital punishment is wrong is there reason for a two-year-old with strong
moral potential to think so? If not, what development of the potential is required? Note
also the difference between realizing the potential to grasp a reason and being merely
indoctrinated into counting it as one.
36 Williams (1985), pp. 41“43.
92 John Skorupski

But suppose that this kind of prudence is just not in his nature. He
will always rate glamour and cool above everything else. Can we in this
case say that he has more reason to avoid the game than he appreciates? He
appreciates all the facts; he just cannot accept that they generate the balance
of reasons that we think they do. We and he appreciate both reasons of
glamour and reasons of prudence. But he, even “in the full maturity of his
faculties,” gives the former a degree of strength relative to the latter which
we think to be misguided. Yet surely if reasons universalise, so too does the
strength of reasons.
However, in talking about glamour we are talking about ideals. The
reason-giving force of ideals depends, at least within limits, on what mat-
ters to a person “ what comes home to that person as worthy of pursuit.
In Scanlon™s terms it depends in part on subjective conditions.37 So if the
ideal of glory is more important to a worldly hero than to an otherworldly
ascetic, there is more reason for the hero to follow the risky path of glory
than the ascetic. That is consistent with the universality of reasons, because
what ideals matter to a person (stably, without self-deception, etc.) is written
into the facts which generate the reasons. True “ a difference of ideals is the
very thing that™s most commonly experienced as a difference about what™s
important, and it™s natural to put this as a difference about what the balance
of reasons “really” is. Natural, but indefensible: for as between competing,
universally intelligible, ideals it can happen that there is no “real,” interper-
sonally invariant, balance of reasons. What ideals have reason-giving force
for you depends on your nature; in the case of ideals we expect to ¬nd a
variety of human natures.
In contrast, however, we do not think that how much reason you have to
take moral considerations into account depends on your particular human
nature. Moral considerations have a reason-giving force that does not vary
with the particularities that differentiate one human nature from another.
One can™t just say “I can see that doing your duty is admirable, and why
there™s reason for some people to do it “ but I™m not that kind of person.”
It seems, then, that with moral obligations, as against ideals, we have uni-
versality without subjective conditions. Nonetheless, in both cases “ the
reason-giving force of ideals and the reason-giving force of moral obliga-
tions “ (III) applies. It™s just that with moral obligations, as against ideals, you
cannot consistently accept that a moral obligation has reason-giving force
for others without also recognizing its reason-giving force in your own case.

37 See n. 33.
Internal Reasons and the Scope of Blame

But now what if you just can™t see that some putative moral considerations
have reason-giving force at all, for anyone “ it™s not in your nature to see
it?38 In that case (III) says that these considerations give you no reason
to act.
In the ¬nal section we shall consider the signi¬cance of this for our moral
practice of blame. But ¬rst we must examine more closely the connection
between moral obligations and reasons.


Moral obligation is linked to reasons for action by a crucial implication that
Williams notes, for example, in the following passage:

Blame rests, in part, on a ¬ction; the idea that ethical reasons, in particu-
lar the special kind of ethical reasons that are obligations, must, really, be
available to the blamed agent. . . . He ought to have done it, as moral blame
uses that phrase, implies there was reason for him to have done it, and this
certainly intends more than the thought that we had a reason to want him
to do it. It hopes to say, rather, that he had a reason to do it. But this may
well be untrue: it was not in fact a reason for him, or at least not enough of
a reason. Under this ¬ction, a continuous attempt is made to recruit people


. 16
( 38 .)