I’m supposed to be writing a review of
Kieran Setiya’s book, Reasons
without Rationalism (Harvard UP, 2007). Even though
I disagreed with a lot of what he says (well, I’m an opinionated philosopher, so I would, wouldn’t I?), I found it a wonderful and
fascinating book.
In this post, I want to raise some problems
for a central principle that lies behind much of Setiya’s argument. (I’m sure
that PEA Soupers’ comments will help me with my review.) This is the principle
that he calls Reasons (p. 12):
Reasons: The fact that p is a
reason for A to φ just in case A has a collection of psychological
states, C, such that the disposition
to be moved to φ by C-and-the-belief-that-p is a good disposition of practical
thought, and C contains no false
beliefs.
I think this principle has the following
problems:
- Suppose that I have a
big collection of mental states C that includes no false beliefs, and also
includes the (true) belief that if I can φ,
then φ-ing would be an outstandingly admirable thing for me to do.
Presumably, the disposition to be moved to φ by this collection of mental
states C together with the belief that I
can φ would be a good disposition of practical thought. So Setiya’s principle implies that the fact that I can φ is in itself a reason for me to
φ. However, the fact that I
can φ is surely not itself a reason for me to φ. Intuitively, in this case, it is not
the fact that I can φ that counts
as a reason for me φ, but the fact that φ-ing would be such an admirable thing
for me to do. - Suppose that I have compelling evidence
that I have reason to φ. E.g., suppose that I get a message from my
exceptionally trustworthy adviser, Diotima, who urges me at all costs
to φ; unfortunately, Diotima also apologizes that she has no time right
now to explain why it is so important that I should φ. Surely
it would be a good disposition of practical thought for me to be moved
to
φ by the belief that Diotima has
advised me to φ (together with my true belief that she is such a
trustworthy adviser). But is the fact that Diotima has advised me to φ
itself a reason for me to φ? Surely it is merely compelling evidence that there is a reason for me to φ, not
a reason to φ all by itself. - Suppose that it is a fact that there is a compelling reason for me to φ. It would
surely also be a good disposition of practical thought for me to be moved to φ by
the belief that there is a
compelling reason for me to φ. So Setiya’s principle seems to entail that the fact that there is a
compelling reason for me to φ is itself a reason for me to φ. But can that be right? The fact that there is a compelling reason for me to φ doesn’t seem
to answer ‘Why?’ questions in the right way. If you ask, ‘Why should I do
that?’, it doesn’t seem a very satisfying reply to be told, ‘Because
there’s a compelling reason to, that’s why!’ - Setiya claims (p. 12–13) that his principle entails that “our
reasons supervene on our psychological states, together with the truth of
our beliefs”. As he says (p. 13), “If agents are psychological duplicates, and
their beliefs are similarly true or false, then their reasons must be the
same.” In fact, however, so far as I can see, Setiya’s principle does not
entail this. This is because his principle does not imply that whenever the fact
that p is a reason for A to φ, A must actually
believe that p. All that the
principle implies is that there must be such a fact as the fact that p.E.g., consider two worlds w and v; in both worlds, a certain
gambler (call him Alexei) has exactly the same mental states – so
Alexei in w and Alexei in v are perfect psychological
duplicates of each other. Moreover, in both worlds, the relevant collection of
Alexei’s mental states C contains no false beliefs.
However, in w a certain horse
(say, Bucephalus) will win the race, while in v this horse Bucephalus will in fact lose the race. (So, in
neither w nor v does Alexei believe either that
Bucephalus will lose the race, or that it will not lose the race.)Then, in v there is a
fact – the fact that Bucephalus will lose the race – such that
if Alexei believed that this fact obtains, the disposition to
be moved not to bet on Bucephalus by this belief together with this
collection of mental states C
would be a good disposition of practical thought. So in v, this fact is a reason for Alexei not to bet on Bucephalus. But there is no such
fact in w; so in w, Alexei does not have this reason
for not betting on Bucephalus. Thus, there can be a difference in Alexei’s
reasons between w and v, even though there is no
difference in Alexei’s mental states in the two worlds, and there is also
no difference between w and v with respect to which of Alexei’s
beliefs are true, and which are not — the only relevant difference between the two worlds is in whether or not Bucephalus wins the race!
Hi Ralph,
Glad to hear you will be sharing your thoughts on that very interesting book.
#1 and #2 look like solid points – you nicely bring out how his account runs roughshod over, e.g., the distinction between reasons to φ and reasons for deciding to φ.
I have some doubt about #3. I agree that, “If you ask, ‘Why should I do that?’, it doesn’t seem a very satisfying reply to be told, ‘Because there’s a compelling reason to, that’s why!’”
But I do not see why this shows that their being a compelling reason is not among the reasons in favor of performing the act.
An analogy: C can be among the causes of X even if citing C will not satisfy us after we ask ‘Why did X happen’.
Brad
Hello Professor Wedgwood,
Some of us here at the University of Washington are going to go through this book this quarter, so I’m glad you’ll be discussing it.
Regarding (1):
Suppose S believes both that if he can Φ, Φ-ing would be an admirable thing for him to do and that it is impossible for him to Φ now. It is unclear to me how, in such a case, the admirability of Φ-ing gives S a reason to Φ. It seems that in such a case the impossibility of S’s Φ-ing silences (from the point of view of S’s practical deliberations) the reason-giving force of the admirability of Φ-ing. That is, if S believes that it is impossible for him to Φ, the admirability of Φ-ing doesn’t provide any reason on its own for S to endeavor to Φ. But if the situation changes such that S can now Φ and S comes to believe that he can now Φ, then can’t this serve as a reason for him to Φ given his further belief in the admirability of Φ-ing?
Suppose I come across a child drowning in the local lake, and believe it is my duty to save him if I can but that I can’t just yet (perhaps I need to get my water-wings). Once I get my water wings, however, I can save him. Why can’t this very fact, that sufficient conditions have been met for me to do my duty, constitute a reason for me to do my duty? It may not be an ultimate reason for me to do my duty, and it certainly doesn’t explain why I have the duty at issue, but it seems to give me a reason to dive in.
I share your concern that, taken in isolation, the fact that one can Φ doesn’t seem to provide a reason for one to Φ. That would be crazy. But I didn’t think Setiya was claiming that that fact has its reason-giving force in isolation from one’s broader collection of psychological states.
Am I missing something here?
Cheers,
Bennett
Professor Wedgwood,
Regarding (2):
I admit to being confused by this point. Suppose we were talking about reasons for belief rather than reasons for action. If so, would you think that the testimony of a trusted and reliable advisor provided reasons for belief? Or, consonant with your objection, would you think that such testimony merely provided reasons for believing that there were other reasons for belief?
When I ask my friend the time, I take her testimony that it is noon to be a reason for believing that it is noon (with all the appropriate ceteris paribus clauses appended). Of course, I also take her testimony to provide me with reasons for believing that there are other reasons for believing that it is noon (e.g., that her watch reads noon). Given the appropriate elements in my C, my advisor testifying to its being noon may provide the crucial premise in a valid argument (one that could be traced by my deliberations) that concludes that it is overwhelming likely that it is noon. Of course, it is only in concert with other elements of my C that such testimony constitutes a reason, but that is not something that Setiya denies.
Why not say the same about practical reasons?
I guess my worry about your first two objections is this:
You have identified two sorts of facts that intuitively are not reason-giving when taken in isolation. The fact that one can Φ or that somebody says one ought to Φ do not, on their own, provide reasons to Φ. But Setiya takes the status of such facts as reasons for subjects to depend partially upon the broader constellations of psychological states of those subjects. And when this information is included, it seems much less clear that the facts at issue do not provide reasons.
Hi, Ralph.
I’m definitely with you on (2); Setiya’s condition looks like it gets into the same problems that epistemic, deductive, and evidential accounts of reasons get into, more generally. Not only: that Diotima told me so, but that: at least one person in this room ought to do it, it’s not Phil, and Phil and I are the only people in the room, etc. At the very least, so solve this problem, Kieran would need to distinguish between primarily and derivatively good dispositions of practical thought; I don’t have the book in front of me so I don’t remember if he actually offers such a refinement, but that’s the move I would make if I were going to try to stick with the view.
On (1), I’m with you in the principle that it’s important to distinguish reasons from their background conditions, and that I can looks like a paradigmatic background condition. I suspect that a solution to the general problem behind (2) would help Kieran out here, though.
I’m not with you on (3); I think that in general there are attractive reasons to think that existential facts about reasons should themselves count as reasons as well, though they shouldn’t carry extra weight, if we know about the reasons which make them true. I think (2) is a much better way of pressing the general problem.
On (4): I suspect (again, without remembering exactly what happens in the book) that you’re building more into what Kieran is claiming here than he is. The way I read the bits that you quote, he is not committed to saying that whenever P is a reason for A, A must actually believe P. He won’t get the kind of strong supervenience to which you’re providing a counterexample, but I’d like to see why he needs strong supervenience here for the purposes of his argument, since he still gets weak (intra-world) supervenience, and that’s consistent with the bit you quote. Of course, in that case he could leave out the bit about the truth of the beliefs part… that does make it puzzling.
It looks to me that even if Reasons were true (and I much doubt it is), it seems that it can be read in two ways. You could either take the good dispositions to be basic and use Reasons to give an account of reasons in terms of these dispositions. Or, you could take reasons to be basic and use Reasons to give an account of which dispositions are good practical dispositions in terms of these reasons. I know I would prefer the latter option – I much doubt whether we can give an account of good practical dispositions that would be independent and prior to the notion of reasons and our grasp of them. I’ll have to read the book though.
On Reasons, it could be the case that P&Q is a reason when neither P nor Q is a reason. Similarly, “if P then Q” (read as a material conditional) comes out as a reason when either Q or ~P is a reason. Also, reasons lose their intentional character (I think): we don’t distinguish between the reasons “Superman is coming!” and “Clark Kent is coming!” I don’t know how damning those observations are, but they strike me as odd.
Heath,
it could be the case that P&Q is a reason when neither P nor Q is a reason.
But that’s surely true. For instance, let P be that the lottery ticket is number 173604, and Q that ticket number 173604 will win. The conjunction is a reason to buy the ticket even though neither conjunct is.
Similarly, “if P then Q” (read as a material conditional) comes out as a reason when either Q or ~P is a reason.
I don’t think that does follow from Setiya’s formulation. Can you explain how it follows? (I agree that it would be a bad consequence — very bad.)
Jussi,
FYI, the main thesis of the book is (roughly) that we cannot do what you suggest – i.e. “take reasons to be basic and use Reasons to give an account of which dispositions are good practical dispositions in terms of these reasons”.
This is complicated by a distinction he makes between metaphysical and epistemic priority – he is only denying that reasons are *metaphysically* prior to good dispositions. That is the official line anyway.
Second, he does NOT assert that the metaphysical priority goes the other way (i.e., that dispositions are prior to reasons); instead, he claims there is metaphysical reciprocity between them.
Jamie,
My thought was just that if Q is a reason on Setiya’s view, then ~PvQ is a reason too. But now I am thinking that was wrong: what follows is that P&Q is a reason, for arbitrary P. That’s much less objectionable.
Thank you so much, PEA Soupers, for those terrific comments! (And thank you, Mark, for the discussion over lunch earlier today!)
I started trying to answer your comments individually. But then I thought that it would probably be more helpful for me first to make some general comments on exactly why I think Setiya goes wrong here. These comments grew longer and longer, and so I eventually decided that I would have to make them into a new post.
I’ll still try to answer your comments, though. So watch this space!
Ralph: I’m glad you liked the book! That means a lot to me. And thanks to everyone for the comments.
Before I turn to details, let me sketch the main idea behind “Reasons.” At the most abstract level, it attempts to capture the connection between reasons and good practical reasoning, a connection that surely exists. More strongly, the book assumes that there is an equivalence between the fact that something is a reason for A to φ and some fact about good practical reasoning. That might be true even if I got the details wrong, and it is what my later arguments require. More specifically still: in my view, reasons are simply premises of sound practical reasoning. (Joseph Raz adopts a similar view, though his conception of practical reasoning is very different from mine.) It is this idea that “Reasons” purports to make at least tolerably precise.
With this background, I can state my general response to examples of the sort that appear in #1, namely that the claims implied by “Reasons” are correct and the oddness of asserting them is merely pragmatic. Since the fact that I can do something is a premise of so much practical reasoning – perhaps all of it – mentioning that reason will be terribly uninformative. It will say almost nothing about the specific content of the corresponding practical thought. Still, there may be contexts in which this fact is salient, and then it will seem less odd to mention it as a reason. Imagine we agree that those in need should be offered help whenever possible and that A is in dire need; but so far we can’t find any way to help him. You and I are lamenting this possibility when I realize that I can help A by doing so-and-so. It would not then seem peculiar for me to say that, in the present context, the fact that I can do so-and-so is a reason to do it. Of course, it does not follow that in other contexts the fact that I can do so-and-so is a reason to do anything. (This is, I think, very close to the response that Bennett Barr provides on my behalf.)
The example in #1 belongs to a family further illustrated by Jonathan Dancy in Ethics without Principles, 38ff., where he distinguishes reasons from enabling conditions. Now, I agree that some necessary conditions of a fact’s being a reason for me to φ are not reasons for me to φ since they are not premises of any pattern of sound practical reasoning that would motivate me to φ. They are mere enabling conditions. But I deny that there is a significant distinction between reasons and enablers among the premises of sound practical reasoning.
As far I understand them, I find Dancy’s arguments for this distinction unconvincing. In his example, my reasoning is that I promised to φ, this promise was not given under duress, I am able to φ…so I will. (We are assuming that if the promise had been given under duress, there would be no reason to keep it.) Dancy notes that it seems wrong to say, about the fact of no duress, that it is an additional reason to φ, along with the fact that I promised. And so it does. But what follows? Only that the conjunctive fact that I promised and there was no duress is not a stronger reason to φ, in the circumstance, than the fact that I promised. Not that the fact of no duress is not a reason in that circumstance, too. Similar considerations apply to the fact I am able to φ.
The implicit argument I am disputing here rests on a bad model of balancing reasons: it assumes that, if the fact that p is a reason for A to φ, and the fact that q is a reason for A to φ, the conjunctive fact that p & q must be a stronger reason than either of them by itself. But when the relevant facts are premises of the same sound reasoning, as in the present example, that will not be so.
About #2, it is not clear that it is a good disposition of practical thought to be moved as Ralph suggests, rather that being moved by the inferred belief that I should φ, as in #3, but if it is, I can’t see why it would be wrong to say that the fact that Diotima advised me to φ is a reason to φ. About #3, I agree with Mark that the fact that there is a reason to φ is a reason to φ. As with Dancy’s argument, this can be made to seem absurd only in conjunction with views about weighing reasons that I reject. It is not as though, when the fact that p is a reason to φ, the consequent fact that there is a reason to φ strengthens the case for doing φ, which is further strengthened by the fact that the fact there is a reason to φ is a reason to φ, and so on!
Finally, the claim that Ralph reports in #4 was a mistake on my part. What I meant to say was that, if the fact that p is a reason for A to φ and B has the same psychological states and her beliefs are similarly true or false, then the fact that p – if it is a fact – is a reason for B to φ. In other words, the status of a fact as a reason for an agent is fixed by her psychology and true/false beliefs, though whether it is a fact in her circumstance, and thus whether there is a reason at all, is not.
Thanks again for the comments and objections!
Thank you so much, Kieran, for those illuminating replies to my comments!
I guess that you’re right that biting the bullet with respect to my points (1), (2) and (3) may not be a very high cost for you to pay. I’ll say a bit more about this in response to your comments on my other post (“The Diagnosis”), especially with respect to my third point (3).
Let me just clarify one point that is relevant to the comments that Heath and Jamie have made here. On Kieran’s view, *every* truth that could be the content of one of the “premises” of a piece of good practical reasoning that results in my being moved to φ is itself a reason for me to φ. So, if there is any such good piece of practical reasoning that starts out from beliefs in the true propositions {p, q, r, …}, then the fact that p is a reason for me to φ, and so is the fact that q, and the fact that r as well.
Apart from this point, however, there doesn’t seem to be any obvious way of arguing that Kieran’s view entails any general claim about the logical structure of reasons. E.g., it doesn’t entail that reasons are closed under entailment (in the sense that whenever p entails q, and the fact that p is a reason for A to φ, the fact that q is also a reason to φ); it also doesn’t (so far as I can see) entail that reasons agglomerate (in the sense that whenever the fact that p is a reason and the fact that q is a reason, the fact that p & q is also a reason).
However, Heath is right at least to this extent, that Kieran’s view does entail that reasons can’t be individuated any more finely than facts. So if the fact that Hesperus is shining is the same fact as the fact that Phosphorus is shining, then whenever the fact that Hesperus is shining is a reason for A to φ, the fact that Phosphorus is shining is also a reason for A to φ (and vice versa). (If Kieran thinks that’s too a high price to pay, he could always insist on individuating facts more finely.)
Anyway, this has been a very rewarding PEA Soup thread. I hope that it makes it clear that if you haven’t done so yet, you should all read Kieran’s book!
About the individuation of facts: reasons are premises of sound practical reasoning, so they need to be individuated as finely as the contents of the beliefs that figure in such reasoning. Beyond that, I don’t have a view!