This is an interesting point. I seem to recall a similar distinction being made by Jay Wallace in "Explanation, Deliberation, and Reasons" (reprinted as Ch. 3 of Normativity and the Will).

Thanks for the tip; I'll check it out.

Hi Jerry,
At least on the interpretation of Williams I advance in "The Obscurity of Internal Reasons" (Philosophers' Imprint 2009), this objection won't stick.  On this interpretation, Williams holds that the concept of a normative reason for S to do A is the concept of an explanation for S of why she would do A if she deliberated soundly.  Williams claims that such concepts do play a central role in actual deliberation.  The objection to external reasons is then that nothing can be an explanation of why an agent would act if it didn't connect with some motivational attitude or disposition that she has.

Stephen,
Thanks for the comments.  I have no basis for disputing your interpretation of Williams.  If I understand you aright:  nothing counts as a normative reason that does not connect with an individual's own motivations.  My problem is that -- prior to choice -- the deliberator has no way to determine what will in fact be motivated.  Imperatives like "Only do what you are motivated to do!" or "Ignore external reasons!" are literally of no help to someone e.g. confronted with a choice between A and B.  Which will motivate?  Which will turn out to be a "true" normative reason; and which will turn out to be merely external?  The deliberator cannot know until after the choice, A or B, has been made.  Therefore,  it cannot be a factor in making the choice.  But surely the deliberator should be able to tell what counts as a normative reason before choice has been made.  So how can the discrimination of internality be essential to identifying them?  

But it's possible I have missed the point; I haven't yet given your paper a proper read.  I reserve the right to change my mind :-).

Hi Jerry,
Distinguish two different kinds of objections to Williams.  The first (which I took you to be making in your original post) is that he wrongly privileges the post-choice perspective (of "motivating reasons"?), which leaves the deliberative role of normative reasons unaccounted for.  The second (which your reply seems to suggest) is that he cannot account for our ability to recognize in deliberation what normative reasons we have, because from that standpoint we can't identify our motivations or dispositions.  I don't think the first objection sticks, because (as I read Williams) he does in fact privilege the deliberative perspective: he just thinks that guidance by judgments about reasons is guidance by beliefs about what one would be motivated to do if one deliberated soundly.  But the second objection may stick: Williams may just be too optimistic about our ability to know what our dispositions are.
FWIW, I think Williams would say the following about the influence of normative reasons in the cases where our motivations are opaque to us from a deliberative perspective.  Suppose R is a reason for me to do A.  I might be motivated to do A by my belief that R, and thereby be motivated by my reason.  So I can be guided by reasons without knowing what is in my motivational set.  But in order to judge that R is a reason for me to do A (and thereby be motivated by this judgment), I must judge that R is an explanation why I would do A if I deliberated soundly, which requires identifying at least a relevant disposition.  (If this is right, then I think a good objection might be to say that if I am motivated through deliberation to do A by my judgment that R, then I must take R to be a reason to do A.  My Williams has to deny that.)
I hope this post isn't too confusing!

I don't see acceptance of an inability to discern one's normative reasons as a solution to your second objection.  I think we have to know our motivations, consciously or unconsciously, for them to affect our actions.  Instead, I find a problem with positing an "internalist condition" for normative reasons and motivations.  It is the supposition of that condition that seems to renders them opaque to the deliberator.  I claim the condition is incoherent because, in effect, it requires you to know what you're going to do as precondition of knowing what counts for you as a reason for doing it.

I would suggest, instead, that the concept of a normative reason for S to do A is the concept of an explanation for S of why she should do A if she deliberated soundly.  On this conception, a normative reason informs what one ought to do, viz. what action is justified.  It is perhaps one of many such reasons requiring deliberative reconciliation, but regardless something that can be determined prior to deciding which ought wins out.

This suggests explicitly replacing references to motivation with references to justification.  For example,

1.  R is a reason for A to do D only if R justifies (or helps to justify) A's doing D.

2.  R justifies A's doing D only if connects up in the right way with A's motivational set.

3.  If R is an external reason, one that does not connect up in the right way with A's motivational set, R does not justify A's doing D.

4.  Therefore, there are no external reasons.

On this account an external reason may lack justification but nonetheless motivate A, who may mistakenly believes it does connect up the right way with A's motivational set.  But there is no conflict between motivation and non-justification.  So this justificatory understanding of reasons is not afflicted with the problems that arise from a motivational view.

Hi Jerry, a few quick points in response:
(1) Your claim that "we have to know our motivations...for them to affect our actions" is very strong--implausibly so, it seems to me.  At least, the burden would be on you to prove it, not on Williams.
(2) Even granting this claim, I don't think it's right to say that this internalist condition "requires you to know what you're going to do as precondition for knowing what counts for you as a reason for doing it".  On this picture, a reason is something that connects in the right way with your motivations.  E.g. in virtue of your desiring milk in your cereal, the fact that there is no milk in your fridge is a reason to go to the store.  This account certainly doesn't require you to know that you're going to go to the store as a precondition for knowing what counts for you as a reason for going to the store.  (Neither does knowing you're going to (try to) put milk on your cereal entail that you know you're going to go to the store before you realize that you have a reason to do so.)
(3) The argument you propose looks like one of the standard readings of Williams' argument against external reasons.  (Puzzling: the conclusion of the argument is that there are no external reasons, but then you say that external reasons may nonetheless motivate?  Perhaps you don't mean to endorse P2 or P3?)
(4) I largely agree with your suggestion that the concept of normative reasons is the concept of an explanation of something normative.  (I don't endorse Williams' theory; I was simply defending it against your objection).  For analyses of normative reasons as explanations of 'should's, see also Toulmin 1950, Broome 2004.  I offered such an account in my 2001 PhD dissertation, but since have agreed with Raz 1999&Searle 2001 that reasons are usually explanations of goodness instead.  (Basic reason: "explanation" is factive, but we can have reasons to do things that we oughtn't do.)  Sorry for the additional references to my own work, and if any of this is overly argumentative.

Stephen,
(1) I believe that, insofar as deliberation is a rational process, a motivation must be known for it to be taken into account.  You seem to attribute a similar sentiment to Williams in your paper: "Williams has argued, you’ll recall, that in order for R to be a reason for A, A has to be able to recognize it as a reason through deliberation" (p. 17).  (I'm not certain what the "recall" references, so maybe this is off-topic).)  Anyway, it seems to me:  something is outside the deliberative process to the extent that it is unknown; and anything outside the deliberative process cannot be reflected in what the person chooses and does = the person's actions.

(2) Many accounts of normative reason specify it with reference to something like "what one would be motivated to do, were one to deliberate rationally or soundly".  This seems to imply that identifying a normative reason requires first deliberating rationally or soundly, in order to discover what one would then be motivated to choose and do.  Which seems to imply you need to choose in order to find out what counts as a reason for choice.  But I may be mistaken in focusing on identifying normative reasons, because Williams at any rate is usually unconcerned with identifying some particular reason R (pace your Standard Argument), but instead queries simply whether "A has a reason to do D".

Perhaps the relevant distinction is not pre-choice/post-choice but rather 1st person/3rd person.  From a 3rd person perspective values (motives, desires, pro-attitudinal things) are determined by action -- i.e., it is from what a person does (plus plausible assumptions about what alternatives the person thought were available, &c.) that we infer what he or she values.  From the 1st person perspective, of course, it is the exact reverse:  it is a person's values that determine which alternative is chosen, which action gets done.

(3) (External reasons are reasons the way subjective truths are truths.  I should have said simply "an apparent reason".  "P2 or P3"?)

(4) Thanks for the references, and for the arguments as well.

Distinguish between (i) taking a motivation "into account", and (ii) being influenced by it.  Granting what you say about the former, I think it's hard to deny that we can be influenced by motives we're not aware we have (playing a "background" role).  The gap in your argument I'm suggesting is that Williams can allow that we can be motivated by things that actually are reasons for us in virtue of motives that we're not aware we have--even if we don't realize that they are reasons for us.  But to be clear, I'm not disagreeing with you that Williams makes deliberation about our reasons too dependent on our beliefs about our motivations.  I think that criticism is right.

In the passage you quote, I'd emphasize the words "has to be able".  The claim isn't that (for Williams) something is only a reason if it is actually recognized, but simply that it must be possible for the agent to recognize it (because otherwise it wouldn't be an explanation for the agent.)

Regarding (2), I think Williams' view is that we can often have justified beliefs about what we would be motivated to do if we deliberated soundly, without actually deliberating soundly.  (I find that plausible.  For example, a trusted authority--perhaps in possession of the relevant information for "sound" deliberation--might tell us).  In these cases, the belief can come before the motivation.

Finally, I suspect there might be an equivocation in your use of "determined".  From a third person perspective, an agent's motives are indicated by her actions, but not caused by her actions, right?  Whereas from a first person perspective, one's actions are caused but not indicated by one's motives.  So there isn't really any contrast?

I do deny we can be influenced by motives we're not aware we have.  But in this I am using a sense of "aware" that includes both conscious and unconscious levels.  The mind is a bigger place than the part that talks to itself.  Insofar as you infer a person's values from their behavior, you are imputing agency.  Agency essentially involves doing A rather than not because of value V, in the belief that A facilitates realizing V in a way that not-A does not.  But you can't have beliefs about V if you are totally unaware of V.  Still, I would agree (with Fingarette) that anything unconscious must be capable in principle of becoming conscious, which may ameliorate our differences.

You are right to point out that Williams shows a firm appreciation of the deliberative perspective.  And I agree with Williams that it is a kind of conceptual truth that anything an agent does must be explicable in terms of his or her values.  Where he goes wrong, in my view, is in supposing that the internal/external distinction has significance for the agent, i.e. that deliberation should or could take into account the "internalist condition".  Then he puzzles over what it would be for a person to believe they have a reason when that reason is external.  But, according to his own lights, it is conceptually impossible, whatever reason is considered and whatever decision is made, for a person to run afoul of the internalist condition.  So I reiterate its irrelevance for the deliberator.  

I prefer to regard my use of "determined" as generously broad rather than equivocal.  The 1st and 3rd person perspectives contrast but do not contradict; in one case the determination is causal, in the other epistemic.  But, as Parfit notes, there is a significant support for the notion that normativity need prove itself by actually moving someone to do something rather than not.  What may have encouraged that notion is the fact that such movement is indeed necessary from the 3rd person perspective; for there it is only from actions (ultimately) that motivations can be inferred.  The mistake, it says here, lies in overlooking the fact that from the 1st person perspective, in stark contrast, the appreciation of normativity and value does not require manifestation in action.

Okay, thanks for all the explanations!  I look forward to seeing your objection in print.

Thanks for personally clarifying issues, pointing out some of the more important concerns in this debate, and indulging my minor-league speculations.  I suspect it will take a while for me to absorb all of this properly, let alone review the literature you've suggested. 

It's been a rare treat to get your input. 

Elijah Millgram has a related thought about commensurability of reasons when he claims that commensurability of reasons/considerations is a result and not a precondition of deliberation. (See "Incommensurability and Practical Reason" Reprinted as chapter 9 of Ethics Done Right). I believe the point is relevant to Millgram's critique of Williams in his "Williams' Argument Against External Reasons"  30(2), 197-220 June 1996.

Thanks for the references!