reason in favor of A. For there was an undefeated reason in favor of B also.
The reason in favor of A does, admittedly, render the choice intelligible
to a certain degree. Nor does Raz claim that the appeal to the reason in
favor of A does render this choice intelligible, as a choice of A over B. But
still, we do want an answer to this further question: Why did the agent
chose A and not B? If choices are only rendered intelligible by reasons,
then in cases of underdetermination by reasons there is no resource that
could possibly render the choice intelligible. If, as Raz plausibly holds,
most choices are in fact underdetermined, this would seem to imply that
most human action is to some degree unintelligible. This conclusion is
false, and Raz does not endorse it. When a choice is underdetermined
by the available reasons, explanation is still possible. Such explanation
will appeal to our personality and character, among other things. As Raz
when we face conļ¬‚icting adequate reasons for action, the explanation of why we
followed the reasons we did will involve more than the invocation of our rationality.
It will allude to our tastes, predilections, and much else besides.50
But Raz holds that our tastes and predilections (saving in some odd
cases) do not provide reasons. How then do they help explain? How can
my personality make my choice intelligible? Raz does not say how. The
suggestion I would like to make is that personality and taste provide expla-
nations for action because they are names for certain brute regularities, and
because these sorts of brute regularities are explanatory. Imagine that two
people are in sufļ¬ciently similar situations that the same reasons apply to
their choice. Assume that these reasons underdetermine the choice, and
that among the rationally eligible choices are a rather selļ¬sh one, and a
49 Raz (1999b), pp. 46ā“66. Razā™s explanation for this fact is that there is widespread incom-
mensurability of reasons, which makes a determination of the balance of reasons impossi-
ble. The distinction between justifying and requiring strengths explains the same under-
50 Raz (1999b), p. 117. See also Raz (1999b), p. 66.
rather generous one. Further, suppose that one of the agents is quite selļ¬sh
while the other is quite generous. If the generous agent chooses the gen-
erous action, and someone asks why, we can provide a sufļ¬cient answer by
citing the generous character of that agent. Moreover, even if we admit that
the selļ¬sh action is rationally eligible, we will be puzzled if the generous
agent chooses the selļ¬sh action, and we will press for further explanation.
Yet there was presumably an adequate reason to act in this selļ¬sh way, since
it was also rationally eligible. This shows that merely citing an adequate
reason is not what makes the generous agentā™s generous action intelligible.
What we want explained is the agentā™s choosing the selļ¬sh action over the
generous one. This explanation cannot simply consist in citing the reason
that favors the selļ¬sh action.
Moreover, once the explanatory roles of personality and taste are rec-
ognized, it becomes much less plausible that desires and actions that are
determined by the available reasons need any implicit normative judg-
ments to make them intelligible. Rather, we should say that, as a matter
of brute fact, the overwhelming majority of human beings share certain
motivational characteristics, such as an aversion to pain, death, and injury,
and that human personalities also generally overlap in certain unsurprising
ways. We do not need two different forms of explanation: one, for desires
and actions that are determined by reasons, and another for those that are
not. A defender of Raz might say that though we do not need these two
forms of explanation, we nevertheless have them, and that a view such as
Razā™s helps explain the difference between them. That is, one might hold
that Raz is simply correct that in some cases our explanation of action
makes reference to reasons, and in other cases it makes reference to tastes
and personalities. But the point here is not to deny this. Rather, the point
is that while there are important differences between these two sorts of
explanation, there is also one extremely important similarity: both get their
explanatory force by appeal to brute regularity. Reasons-explanations are
explanatory because of the brute fact that human beings are almost exclu-
sively motivated by considerations that are, in fact, reasons ā“ whether they
are recognized as reasons or not. Personality-explanations are explanatory
because of the brute fact that this human being is typically motivated in
certain ways. In neither case is it necessary to postulate that the agent saw
a reason as a reason. He may simply have noticed that it was beginning to
rain, and opened his umbrella because he wanted to avoid getting wet.
Again, on both Razā™s view and on my own, typical intentional actions
are actions about which there is a story to tell. What commits Raz to
the judgment thesis is the further claim that this story āis of what the
agents took to be facts which show the action to be good, and which
therefore constitute a reason for its performance, making it eligible.ā51
Razā™s objector holds that there are some features of action that make them
intelligible objects of choice, but that are not good-making features, and
that do not even appear to be good-making to the agents who perform
those actions. Thus, much of Razā™s argument for the judgment thesis
consists in showing that the search for such features is doomed to failure.
Part of the problem for Razā™s objector is that Raz forces him to assume
that the notion of a reason is only used in one way, either as an explanatory
notion, or as a normative one. Call this view ā˜the univocality of reasons.ā™
Thus, the objector makes the following claim:
Once we draw a clear distinction between features which show an act to be an
intelligible object of choice and ones which show it to be good or of value we
will see . . . that reasons belong with the ļ¬rst, and not with the second.52
The univocality of reasons will certainly land the objector in difļ¬culties
if we can produce claims in which reasons are functioning in an unam-
biguously normative way. But why cannot some reasons be explanatory
without being (even perceived as) normative, just as some reasons can be
normative without being explanatory? Because the objector assumes the
univocality of reasons, he is forced into an implausible view according
to which explanatory reasons have features that are plausibly held only
by normative reasons: universality, for example, and the ability to justify
actions. This is why the objector wrongly concedes that if the fact that
an action will hurt another person is a reason that makes it intelligible,
then everyone would have such a reason.53 It is true that at one point the
objector tries to deny that reasons must pro tanto justify actions. But Razā™s
response is that restricting the role of reasons in this way fails to preserve
the normativity of reasons, and this ā“ surprisingly ā“ is enough to silence
the objector. But if some of the things we call reasons are not normative,
this is no problem. The objector never raises this point. It is true that
Raz is not making these assumptions without realizing it. He explains
that it is ācentral to [his] approach that the same concept is crucial both
for intelligibility and to justiļ¬cation (and therefore also to evaluation).ā54
This thought is expressed more clearly a few sentences later, when Raz
51 52 Raz (1999b), p. 25.
Raz (1999b), p. 24.
53 54 Raz (1999b), p. 31.
Raz (1999b), pp. 26, 28.
explicitly gives the following test for a consideration possibly making an
action an intelligible object of choice:
the same consideration must also show that the act is justiļ¬ed or at least that there is
something to be said in its justiļ¬cation, something that in the absence of contrary
considerations makes it justiļ¬ed.
But this is exactly the point of dispute, and Raz provides no defense for it.
The view offered in this chapter is that we humans are imperfectly rational
in the following sense: we are motivated by a wide range of considerations,
most but not all of which are normative reasons. There is no need to hold
that a consideration can only belong to the former class if it is regarded by
the agent as belonging to the latter.
Of the four philosophers considered in this chapter, Scanlon is perhaps the
most explicit in his endorsement of the judgment thesis. For example, he
[i]n order for a consideration to be an operative reason for me, I have to believe
it. In addition, I have to take it to be a reason for the attitude in question.55
The only source of motivation lies in my taking certain considerations ā“ such as
the pleasures of drinking, of eating, of hearing from a friend ā“ as reasons.56
[what are usually called desires] are not a matter of preconceptual appetite but
involve at least a vague appeal to some evaluative category.57
In fact, Scanlonā™s judgment thesis concerns not only actions, desires, and
intentions, but beliefs, anger, and so on. Scanlon characterizes these all as
ā˜judgment-sensitive attitudes,ā™ by which he means that an ideally rational
person would have or lose these sorts of attitudes depending upon their
judgments of the sufļ¬ciency of reasons.58
Unlike Raz, Scanlon is happy to appeal to regularity in order to make
claims about what sorts of things explain human actions and other attitudes.
One central way in which Scanlon argues for the judgment thesis is by
55 Scanlon (1998), p. 56. These particular claims helpfully distinguish two senses in which
judgment might be necessary for a reason to motivate an agent: judgment about a matter
of fact, and judgment about a normative matter. Again, this chapter is only challenging
the necessity of the latter.
56 57 Scanlon (1998), p. 65.
Scanlon (1998), p. 35.
58 Scanlon (1998), p. 21. Actions are included because they are so closely linked to intentions,
which are attitudes.
appeal to a regular connection between normative judgments and attitudes.
That is, he repeatedly points out regularities of the following sort:
[A] rational person who judges there to be compelling reason to do A normally
forms the intention to do A.59
Judgments of there being good reason will correlate with action in a rational
In context, these claims are intended to do two things. The ļ¬rst is to
form part of an argument against the idea that brute desires are the pri-
mary source of intentional action in humans. That is, Scanlon is pointing
out that there is a regularity between something cognitive on the one
hand, and action on the other. Scanlon might admit that desires ļ¬gure in
action-explanation somehow; but if they do, these desires are responses
to something else. Thus, an action is best explained by reference to this
something else, and not by reference to antecedent desires. The second
thing these claims are intended to do is to suggest that the ā˜something elseā™
is a normative judgment about the sufļ¬ciency of reasons. That is, these
claims are part of an argument that intentional action is normally the result
of the agent making certain normative judgments. Scanlon expresses this
particular interpretation of the correlations by characterizing intentions as
ā˜judgment-sensitive attitudes,ā™ and not, for example, ā˜judgment-producing
attitudesā™ or ā˜judgment-constituting attitudes.ā™61
Let us grant that desire is normally explained by something else, so that
this something else ought to be regarded as the real source of intentional
action. Why should we believe that this ā˜something elseā™ is a normative
judgment? Scanlonā™s correlations not only do not settle the issue, they do
not support the judgment thesis at all. If the correlations Scanlon asserts
really obtain, this is just as likely to be because both the action and the
judgment are explained by a third something: the agentā™s having noticed
the facts about the situation that provide the reasons that favor it. For
example, the agent notices that someone has fallen and needs help getting
up. That the agent notices this nonnormative fact (which happens to
provide a reason) could explain both the subsequent normative judgment
ā˜I have a reason to help this person get up,ā™ and the motivation to do
so. One reason to favor this explanation over Scanlonā™s is that normative
judgments are in fact very rarely part of the phenomenology of desire,
intention, or action.
59 60 Scanlon (1998), p. 61.
Scanlon (1998), pp. 33ā“34. See also Scanlon (1998), p. 66.
61 See also Scanlon (1998), pp. 23ā“24 for the assumption of this direction of explanation.
Why does Scanlon choose normative judgments as the ā˜something else,ā™
instead of mere awareness of facts that are themselves reasons? One possible
explanation might be termed ā˜the philosophical fallacy.ā™ This fallacy occurs
when a philosopher tries to become clear on the nature of some general
phenomenon, but examines it primarily as it occurs in the context of
philosophical discussion or reļ¬‚ection. Thus, when Scanlon examines his
own desire for ice cream, he notices that there is something cognitive
behind his desire: a belief about the taste and sensations of the ice cream.
But because he is also doing philosophy, he also notes that this fact seems
to provide a reason for eating the ice cream, and that he himself makes
this very judgment.62 The same thing occurs when he considers other
cases. Thus, in all of these cases the one constant factor is a normative
judgment that he has a reason to do the action. It may be because of
this philosophically induced regularity that Scanlon then concludes that
it is judgments of this sort that are doing the motivational work. To a
philosopher this unifying thesis is more attractive than the claim that in
one case it is the thought about the taste that motivates, while in another
it is the thought that a person needs help, and in another it is some other
thought. But since Scanlon recognizes a plurality of values and reasons in
any case, this increased unity is really an illusion.
Scanlon may also have fallen into the philosophical fallacy when he
introduced the notion of normative reasons by claiming that they are
answers to ā˜why shouldā™ questions.63 Suppose, for example, that an agent
is in a situation in which she has the following reason to : by -ing,
she will make her sister happy. If, in this context, the agent actually asks
a ā˜why shouldā™ question, there will not only be this reason, but there
will be the nonnormative belief ā˜ -ing will make my sister happy,ā™ and
the further normative judgment ā˜making my sister happy provides me
with a reason.ā™ In this reļ¬‚ective context it will become more difļ¬cult to
determine whether it is the reasons themselves, or beliefs about facts that
happen to be reasons, or normative judgments, that are doing the relevant
motivational work. But in typical cases no ā˜why shouldā™ question precedes
action, decision, or desire. In these typical cases there are nevertheless
still reasons for decisions, actions, and desires, and reasons why we in fact
decide, act, and desire. If one examines reasons in a more typical and less
philosophically reļ¬‚ective context, it is easier to avoid holding that it is
judgments about reasons that get one to decide, act, or desire.
Scanlon (1998), p. 35. Scanlon (1998), p. 18.
It is true that Scanlon admits that ā[j]udgment-sensitive attitudes can
arise spontaneously, without judgment or reļ¬‚ection.ā But he also thinks
that in such cases āthe formation of these attitudes is generally constrained
by general standing judgments about the adequacy of reasons.ā64 This
qualiļ¬cation, taken together with the bulk of the claims cited above, sug-
gests that normative judgments, whether conscious or unconscious, play
a role in the formation of almost all of our desires and intentions. Never-
theless, it may be safest to attribute a slightly more modest version of the
judgment thesis to Scanlon ā“ a version according to which desires and
intentions merely involve or encompass normative judgments ā“ for at one
point he almost explicitly denies the claim that āall desires arise from prior