Why is Setiya’s
principle vulnerable to the problems that I listed in my previous post? What exactly is the diagnosis?
that the diagnosis is that there are in fact two sorts of reasons. These two sorts of reasons do not always
coincide; and Setiya is implicitly conflating these two sorts of reasons here.
1. One way of thinking of “reasons” insists
on a close connection between reasons and reasoning.
Basically, the idea is that a reason for φ-ing is a starting point for a bit of reasoning
that terminates in φ-ing. (We can understand “reasoning” very broadly here, as any process of thought
that results in your being motivated to act intentionally in a certain way or to
change your beliefs or intentions or other attitudes, etc..)
Even if we think of reasons in this first way,
we could still distinguish between motivating reasons and normative reasons. We
could say that your motivating reason
for φ-ing is whatever actually
motivated you to φ, while a normative
reason for you to φ is something that could
rationally move you to φ. But it
still seems to me that if we think of reasons in this first way, then all of
your reasons – both normative and motivating reasons – must be facts
about your mental states, since (pace
Jonathan Dancy) it seems to me that it is only ever facts about your mental states
that could possibly motivate you to act.
Of course, if you were asked what your
reasons are, you would normally cite an external-world fact. (E.g. you might
say, “Why did I fly via Iceland? Well, I just had to get to my aunt’s funeral, and all the direct flights were ridiculously expensive.”) But if you really are motivated by this external-world fact, that must be because you believe that this external-world fact
obtains. So, if we continue thinking of reasons in this first way, but also
identify reasons with these external-world facts, we run the risk of seriously
exaggerating the role that belief
plays in our reasoning (and correspondingly underestimating the role that is
played in our reasoning by other mental states besides belief). As a result, it seems
to me that a cleaner way of conceiving of reasons of this first kind is to
think of them as always consisting in facts about the relevant agent’s mental
2. A second way of thinking of a normative reason
for you to φ is as a fact that plays a role in a certain sort of normative explanation – e.g., an
explanation of why your φ-ing has the positive normative status that it has. So, e.g., a paradigmatic reason of this second kind would be some fact that explains why you ought to φ, or why it is permissible for you to φ (or in other words, the reason would be the reason why you ought to φ, or why it is permissible for you to φ). Reasons of this second
kind are certainly not restricted to facts about the agent’s mental states: the
explanation of why you ought to φ, or of why it is permissible for you to φ, will
very often consist of some external-world fact about your situation at the relevant time.
Now, it will certainly sometimes happen that a reason of the first sort is (a true belief whose
content is) a fact that also explains
why the action has the positive normative status that it has. But this need not
always be the case. It is rational for you to be moved to φ by your rational
belief that Diotima has advised you to φ; but the fact that Diotima has advised
you to φ does not in itself explain why it is a good idea for you to φ. On the
contrary, Diotima’s advice is merely evidence
that it is a good idea for you to φ, not an explanation
of why it is good idea for you to φ.
So, why do I think that Setiya’s principle
implicitly conflates these two ways of thinking about reasons?
- On the one hand, there are certain elements in his principle
that are clearly suitable for the first way of thinking about reasons. The principle centrally
deploys the idea of a good disposition
of practical thought. Such dispositions involve responding to a “collection
of mental states” by being “moved” to φ. So talking about a “good
disposition of practical thought” is in effect just a way of talking about
- On the other hand, there are certain other elements that are
really only suitable for the second way of thinking. Reasons are
identified with facts –
normally facts about the external world; and it is also required that the collection
of mental states that (along with belief in the fact that constitutes the reason)
must be capable of rationally moving the agent must contain no false beliefs. These elements are, it seems to me,
out of keeping with the first approach, since we can certainly be
rationally moved to act by rational beliefs that happen to be false.
The first three problems that I raised in
my previous post — (1), (2), and (3) — all highlight implications of Setiya’s principle
that would all be perfectly reasonable if we are thinking of reasons in the first way,
but seem quite wrong if we are thinking of reasons in the second way (i.e., if we require reasons to play an explanatory role in
The fourth problem that I raised (4) highlights
a claim that Setiya makes that would indeed be true of reasons of the first
sort (since reasons of this sort do seem to supervene on facts about the
relevant agent’s mental states), but which is in fact not true of his own
principle, on account of the element that he incorporates into his own
principle in order to make it cover reasons of the second sort.