Reasons not to intend

The distinction between the 'right' and 'wrong' kind of reasons is taken to play at least three important roles: 'right' kind but not 'wrong' kinds of reasons contribute to standards of correctness, and in the case of reasons for attitudes, 'right' kind but not 'wrong' kinds of reasons can serve to make attitudes rational, and exhibit at least a strong asymmetry, in that it is at least substantially easier to believe or intend for the 'right' kinds of reasons, if not outright impossible to believe or intend for the 'wrong' kind.

Toxin-puzzle style considerations often lead philosophers to endorse the following thesis:

R is a RK-reason to intend to do A iff R is a reason to do A. 

This is because it seems difficult, if not impossible to intend to drink the toxin directly for the reason that so intending will result in a reward, and moreover that believing that so intending will result in a reward doesn't make intending to drink it rational, but only makes acting to get oneself to have that intention rational – corresponding to at least two of the earmarks of a 'wrong'-kind reason.

Proponents of the so-called 'state-given/object-given' distinction appear to go further than this, and claim that all RK-reasons which bear on intention are reasons for or against the object of that intention.

But this raises a puzzle.  Plausibly, given the 'right kind of reasons' approach, it is rational to intend to do A just in case the RK-reasons to intend to do A outweigh the RK-reasons to not intend to do A – or alternatively, just in case they outweigh the RK-reasons in favor of each of the alternatives to doing A.  But on the face of it, there is more than one alternative to intending to do A – or otherwise put, there is more than one way to not intend to do A.  Rather than intending to do A, you can intend to do ~A, true; but another way of not intending to do A is to have no intention one way or the other – to neither intend to do A nor intend to do ~A.

Reasons to do A constitute one kind of RK-reason bearing on the intention to do A, because they are RK-reasons to intend to do A.  Reasons to not do A – that is, to do ~A – constitute another kind of RK-reason bearing on the intention to do A, because they are RK-reasons to intend to do ~A – which is one of the alternatives to intending to do A.  (This follows merely by substituting '~A' for 'A' in the thesis above.)  But then what are the RK-reasons to have neither intention?

In principle, it could be that there are no RK-reasons to have neither intention.  But this seems like the view of desperation – if RK-reasons are required for rationality, then it would be a consequence of this option that not intending either way is never rationally the best option, which seems outright false – there are cases in which the most rational thing to do is to wait to make up one's mind.

But if there are RK-reasons to intend neither way, then what are those reasons?  The proponents of the so-called 'state-given'/'object-given' theory would have it that all RK-reasons which bear on the rationality of attitudes are reasons which bear on the objects of those attitudes.  But we've already used up all of the reasons which bear on doing A – the reasons for doing A are the RK-reasons to intend to do A, and the reasons against doing A – that is, the reasons to do ~A – are the RK-reasons to intend to do ~A.  So which reasons that bear on doing A are the RK-reasons to have neither intention, and in what way do they bear on A?  That is the question that any proponent of the 'state-given/object-given' distinction must answer, in order to avoid the result that not intending either way is never the most rational course.

I'm very pessimistic about how this question can be answered – am I wrong to be so pessimistic?  Or missing out on something else?

26 Replies to “Reasons not to intend

  1. Mark’s post argues against the following principle:
    RI: It is rational to intend to do A iff the reasons to intend to do A outweigh the reasons to intend to do ~A.
    (A side issue: I object to this use of “rational,” because I believe that rationality is always tied to the epistemic standpoint of the subject, from which the real reasons for action might be quite invisible. But I am willing to play along and understand the reasons on the right side as normative reasons and to substitute “S ought to intend to do A” for “It is rational to intend to do A.”)
    His idea is that one might sometimes have reason to intend neither to do A nor to do ~A: “there are cases in which the most rational thing to do is to wait to make up one’s mind.”
    Maybe Mark is correct that RI is false. Maybe one could have more reason to intend to A than to intend to not A, while one really ought to form neither intention. This sort of case seems pretty rare to me: it would have to be a case where forming neither intention was not tantamount to intending to not A. The best I can do is the following: Suppose there is a baseball outfielder who has been corrupted by the mafia. They want him to fail to catch balls that are hit to left field. He wants the payoff, but he doesn’t want to be fired. If he intends to catch the balls, he knows that he will catch them. If he intends not to catch the balls, he knows that it will be apparent to others that he is attempting to throw the game. Therefore, he ought to have neither of these two intentions. (God knows what intention he ought to have. Perhaps there are better examples … anyone?)
    But I think that a reformulated version of RI2 gets around this difficulty:
    RI*: It is rational for S to intend to do A iff the reasons for S to intend to do A outweigh S’s reasons to do anything else.
    Notice that this does not say, “the reasons to intend to do A outweigh the reasons to intend to do anything else.” This is to get around cases where the right thing to do is *to do nothing* or *to wait* (as opposed to *to intend to do nothing* or *to intend to wait*).
    This seems to me to solve the problem that Mark presented. Perhaps it has some other problem, so that it is not available to the theorist who favors using ‘right’ and ‘wrong’ kinds of reasons and intentions to solve the toxin puzzles.

  2. I’m not sympathetic to the view under attack, but how about a reason to deliberate further, or to get more information (bearing on whether it is better to do A or not)? We might think that once I intend to do A or not to do A, practical reasoning about whether to do A is closed off; so if it is best to continue to deliberate, then it is best not to form an intention.

  3. Hi, Daniel.
    Good ideas – I agree with all of that. But I didn’t ask what the reason was a reason to do – I asked what the reason to do it was, and in what sense it is a reason that bears on doing A, rather than whether to close off deliberation about whether to do A. It looks to me like the kind of reason that you are pointing out is a cost of forming an intention – that it closes off deliberation. But costs of forming intentions are paradigmatically state-given reasons – that is just what state-given reasons are: reasons associated with benefits or costs of being in a certain state.
    Jason –
    Agreed – RI* is preferable to RI. That wasn’t my problem; that was where I was hoping to start. My problem was that proponents of the ‘object-given/state-given’ account of the right kind of reasons can’t distinguish RI* from RI, because they can’t make room for right-kind reasons to not make up one’s mind.
    It is important to stress that this is not a problem for ‘the theorist who favors using ‘right’ and ‘wrong’ kinds of reasons and intentions to solve toxin puzzles’; it is a problem specifically for the ‘object-given/state-given’ theory of the ‘right’/’wrong’ distinction. It is an unfortunate fact about the current literature that so many authors use the object-given/state-given or equivalent vocabulary as if it just were the ‘right’/’wrong’ distinction, rather than a problematic theory about that distinction.
    However, please don’t substitute ‘ought to intend’ for ‘is rational to intend’. Please substitute instead understand reference to RK-reasons in the discussion of rationality as reference to subjective RK-reasons – ones that are in the subject’s ken.

  4. Mark – Agreed, I was too elliptical. It seems to me that if, for instance, relevant information (about whether doing A or not doing will turn out best) is still out there, and waiting for me to find out about it, that’s most directly a reason for waiting. It’s also a reason for not forming an intention, but that seems rather less obvious, which is why I thought I had to argue for it. Missing out on the information still to come in is a cost of acting, at least as much as a cost of deciding to act.
    But maybe you’re thinking that waiting is just a way of not doing A; it isn’t a third thing to do. If so, I’m worried that not forming an intention isn’t a third way of being either, at least once the option of doing A is under consideration. For if waiting counts as not doing A, then not forming an intention *as a deliberate policy* looks like a way of intending not to do A. More likely though, you weren’t thinking that, and we can agree that waiting is a genuine third option, which doesn’t preclude doing A eventually. In that case, I still want to say that the fruitfulness of further deliberation or evidence gathering is a reason both to wait and a reason not to form an intention either way.

  5. I wonder whether one could try to save the view by loosening the connection between RK-reasons and rationality. On many views, rationality is a matter of internal coherence rather than states being supported by RK-reasons. This would mean that not intending either to phi or to not phi would be rational whenever one hasn’t made a judgment that one has decisive RK-reasons to have either intention. This can be the case even when there really are decisive RK-reasons to phi (or not to phi).
    I guess the point is that the problem seems to follow from controversial premises about rationality rather than the RK-reasons view based on the state/object-given reasons. On this view, rationality is a matter of having RK-reasons rather than a matter of judging that one has RK-reasons. Not everyone accepts that.
    In thinking about this problem, it would also help to have an example that illustrates the problem. For one, that would help also to assess the solution that you mention is the desperate one.

  6. Following Jussi’s suggestion, here’s a simple case (the first of three). I’m considering whether to drive to Los Angeles on Monday morning. One reason to drive to Los Angeles then is that the ocean scenery along the way is pleasant. One reason not to drive to Los Angeles then is that it will be Memorial Day, so the traffic will be terrible. I’ll just stipulate that the fact that the Memorial Day traffic will be terrible outweighs the fact that the ocean scenery is pleasant. It’s now Saturday morning, and these are the only relevant reasons that I know about. But I’m expecting to get a phone call this afternoon from my brother, telling me whether he will be in Los Angeles on Monday. If he is, that will be enough, together with the pleasant scenery, to outweigh the bad traffic, but if he is not, that would be yet a further reason not to go to Los Angeles on Monday. Unfortunately, I don’t yet know which, so the only reasons that I have now weigh in favor of not going to Los Angeles.
    In this case I think it is irrational to form the intention to not go to Los Angeles on Monday, even though I have weightier reasons to not go than to go. The only rational course is to wait to decide.
    Here’s a second, different, case: I’m the Strategic Bomber, and I’ve just accepted a mission to bomb an ammunition factory. I find out that it is next door to an elementary school, which reminds me of all of the virtues of terror bombing – virtues which I find compelling (change the example if that makes it easier; I’m just going with a familiar one to make the point). Bombing the elementary school seems like a great idea to me – better than not bombing it (in contrast to the usual SB). I think that in cases like this it can be rational to not form an intention to bomb the elementary school, because I already know that I am going to do it anyway.
    There are also cases in which I don’t expect more information to come in, and I don’t already have an intention that I know will result in doing it, but I also trust myself to make up my mind later, in time before I actually have to act. For example, suppose that while I am deliberating about whether to drive into LA on Monday morning, I am interrupted over my Saturday breakfast by my fiancee, who wants to leave now for the farmer’s market. I think that among the rationally permissible options is putting off the decision about Los Angeles until later in the weekend. But my fiancee’s insisting that we go to the farmer’s market, even though it plays a role in making it rational for me to put off the decision and even though I am able to easily put off the decision for that very reason, does not itself bear on whether to drive to Los Angeles on Monday. So it’s not ‘object-given’.
    I suppose that Jussi’s solution would have it that in none of these cases do I judge that I have decisive RK-reasons to intend to drive to LA on Monday – for example, presumably in the first case I am waiting for my brother’s call to make up my mind about whether I do, and in the third case I put off making up my mind about whether I do to go to the farmer’s market first.
    I don’t know how this response would work in the second case, but more to the point, if you haven’t noticed, I can raise exactly this problem for the ‘state-given/object-given’ theory in terms of reasons for belief (or reasons for ‘judgment’, if you will). In my first case, for example, I currently have more evidence that there are decisive RK-reasons to not drive to LA on Monday than I do that there are decisive RK-reasons to drive to LA on Monday – that’s because the balance of reasons currently favors not driving to LA, and so as long as it’s sufficiently unlikely that my brother will be there, my evidence that I have decisive RK-reasons not to go can be pretty good. So why is it rational to wait to form a judgment about whether I have decisive RK-reasons to go? What makes it rational to wait isn’t itself evidence that I have decisve RK-reasons to go or evidence that I do not. It’s something else that the state-given/object-given distinction can’t capture.

  7. Thanks for the examples Mark. They are really helpful for getting a grasp what kind of a problem this is. You are right that I would be inclined to use the judgment about reasons to deal with the first and the third cases. I know that this is a minority view, but, in the second case, my intuition is that if you know that you are going to do something, then you are intending to it in the sense relevant here. You might not intend to do it in the sense relevant for the DDE debates but that is a different story we discussed here earlier.
    This leaves with the last problem. I wonder if in this case one could say that there is a RK, object-given reason to intend to form the intention later on when there is more evidence available. This reason is given by the fact that you will know whether your brother will be arriving later on. This fact could be part of the facts about the intention formation based on evidence about reasons which you will get later. If there is this kind of an object-given reason to plan to form the intention later, it could make it rational to wait. After all, waiting would be carrying out that rational intention.

  8. What if we take the reasons in question to have as their object the decision, rather than A itself? So, it’s not that I have reasons to ‘neither A nor ~A’, but rather have reasons to ‘not decide whether to A’. In the first case, the fact that my decision would be a decision poorly-made were I not to wait for more information is a reason to postpone that decision. Of course, some story would have to be told about why decision-reasons “stall” the force of other action-reasons, but something like that seems plausible to me.

  9. Jussi – I don’t think the fact that I can also form an intention to form an intention later changes the fact that I am thereby not forming the intention now. And that’s what raises the problem.
    David – I think that move gives up the object-given/state-given game; decisions aren’t the objects of intentions.

  10. The decision itself (“I decide to φ”) is certainly not the object of an intention, but can’t the decision process (i.e., deliberation) be? I neither intend now to drive to LA nor to not drive to LA; rather I intend now to deliberate later about whether or not to drive to LA. Couldn’t there be object-given reasons for that intention?

  11. Mark,
    that was just the point. You don’t form an intention to either phi or not to phi. But, you do form an intention to form one of these intentions later when you get more information. I thought the challenge was to explain why having neither of the intentions was rational. The intention to form either one of the intentions later would be just the right thing to make lacking either one of the intentions rational. This is because not having either one of the intentions satisfies a plan for which there are object-given reasons (which are not reasons to either phi or not to phi).

  12. Forming the intention to drink the toxin also satisfies a plan for which there are object-given reasons – namely, the plan to intend to drink the toxin. If that’s an admissible move in the object-given/state-given game, then the game is up.
    My question is how a reason not to intend to do A could be object-given for A, not how it could follow from some reason that is object-given for some other intention. The moves that are being suggested look to me like they answer some other question, but not the one that I started out with.

  13. Mark,
    I think it’s likely that there can be no reasons merely not to intend to A that are object-given for A. Any reason that is object-given for A will (I would think) refer to positive or negative facts about A that are reasons either to A or to ~A. I just don’t see this as a worry, because I don’t think things ever seem otherwise. Rather, we expect that when confronted with the option to A, either the balance of reasons is for A or against A, or there are reasons to postpone or refrain from judgement that are object-given for deliberation. So, I don’t disagree with your position; I just don’t see it as a puzzle.

  14. I don’t see it as a puzzle, either; I see it as a refutation of what I understood the ‘object-given/state-given’ theory of the right kind of reasons to be. If there are reasons to postpone deciding whether to do A, and those reasons are reasons for which it is possible to not (for the meantime) intend to do A, and they make it rational to not (for the meantime) intend to do A, then they are right-kind reasons which bear on whether to intend to do A but are not object-given.
    The whole point of the ‘right-kind’/’wrong-kind’ distinction, as I began my post by pointing out, is to capture the distinction between reasons which bear on the rationality of a mental state and for which it is possible to be in that state, and reasons which do not bear on that state’s rationality and for which it is not possible, or at least not straightforwardly possible, to be in that state. If the ‘object-given’/’state-given’ distinction can’t capture such reasons, then it’s a bum account of the ‘right-kind’/’wrong-kind’ distinction.
    All I’ve been pointing out, here, is that the thesis that I started with:

    R is a RK-reason to intend to do A iff R is a reason to do A.

    is weaker than the thesis that:

    R is a RK-reason which bears on whether to do intend to do A iff R is a reason which bears on whether to do A.

    The former thesis only bears on reasons for intention, and the latter thesis, which is what I’ve been taking the object-given/state-given theory to be, bears also on reasons against it. That is what I have been suggesting makes it false. If you agree that there are legitimate reasons to delay decision about whether to do A, and that those reasons need not bear on whether to do A, then you are agreeing with everything that I want.

  15. I see. Insofar as your stronger thesis above captures the object/state view, I’m inclined to agree with you. And I must confess I don’t know this literature well, so perhaps there are good reasons for thinking that the view’s proponents are committed to that thesis. But why wouldn’t they want to just adopt your weaker thesis? After all, that seems more in line with how we think about these things: There are reasons to intend or intend not, all of which are reasons that stem from considerations that count in favor of or against particular actions. There are also reasons concerning deliberation. Such reasons are also, technically, reasons to not intend; but we certainly don’t think of them as such, especially since those reasons often have little or nothing to do with the particular intentions available to us (such as in the fiancee case). Mightn’t one think that precisely what we want is a view that clearly separates reasons to intend from reasons that merely “bear on” intention, noting that the latter have something other than intention as their (immediate) object?

  16. One could deny that there are reasons for not intending to A in these cases even if there are reasons to postpone deciding and for not doing the mental action of forming the intention to A. This is at least what Parfit used to think. This would mean that the reasons for postponing deciding would not be reasons not to intend to A, and thus the argument would not go through. The only question is is this enough to explain the rationality of not intending to A. It seems that it is as intending to A would require the action of forming an intention which one has no reason to do and which would be irrational to do.
    This matches the toxin-puzzle solution. There is an object-given reason for forming the intention even if there is no reason at all for intending to drink the toxin. I’m not sure I see why this is giving the came up.

  17. Interesting problem!
    I’m sympathetic to the view that there are no reasons for intentions at all, only reasons for actions…so I might not have a dog in this fight.
    Still, it seems important to note that if A-ing happens to be incompatible with B-ing, then if you have a reason to A, it might not automatically mean you have reason to not-B, even though A-ing implies not B-ing. Reasons govern intensional contexts. Your reason to A is not a reason to do everything that A-ing (combined with the present state of the world) entails, including omissions.
    I do think there are times when it makes sense to talk about reasons not to do something, such as the reason I have to not chop off my foot. But I don’t think my reason to eat bacon gives me a reason to not do anything incompatible with my eating bacon (like swimming).

  18. Jussi – the difference from the toxin puzzle case is that in the toxin puzzle case, we observe that it is difficult if not impossible to form the intention to drink the toxin without giving oneself an auxiliary reason, and we observe that the intention to drink the toxin does not appear to be a rational intention. Whereas in my kinds of cases, it is easy to withhold making up one’s mind without giving oneself an auxiliary reason, and so doing seems eminently rational. Think about it: ‘why hasn’t he decided not to go to Los Angeles on Monday yet?’ ‘because he’s waiting to hear if his brother will be in LA’. This looks straightforwardly like an ascription of the reason for which I haven’t decided – the reason for which I continue to not intend.
    In answer to David, that is why I think it matters whether the ‘object-given/state-given’ view encompasses my second thesis, or only my first. I was taking it that the ‘object-given/state-given’ framework was in the business of offering a theory which was supposed to account for certain things – the same things that the rest of us who have given accounts of the ‘right’/’wrong’ distinction were trying to account for.
    I also don’t think it’s any good claiming – if this is how Jussi is understanding Parfit – that there is an asymmetry between having an intention and lacking it, in that for having an intention to be rational, you must have sufficient RK-reason for it, but for lacking an intention to be rational, you need merely to lack sufficient RK-reason to have it, not to actually have sufficient RK-reason to lack it. That won’t work, because in my examples you do seem to have sufficient RK-reason to have the intention. At least, you have RK-reason to have the intention, and if it is insufficient, then we haven’t gotten any good explanation of why not – you certainly don’t have any better contrary reasons, unless there are RK-reasons not to intend.

  19. One more suggestion: maybe it helps to distinguish between objective and subjective reasons here. It looks like the troublesome case (waiting for the phone call) is one where there is a subjective RK-reason for not forming an intention to A, but no reason for not doing A. But there isn’t an objective RK-reason for not forming an intention to A, because the subjective RK-reason depends on ignorance. (Notice that if we asked the agent whether it was best to do A or not, he would say “I don’t know yet.”)
    So I wonder whether the original thesis works better as one concerning objective reasons:
    R is an RK objective reason to intend to do A iff R is an objective reason to do A.
    Note that in the toxin case the WK-reason is both objective and subjective, so the restricted thesis still gives us the right answer about the toxin case. The diagnosis would be that reasons concerning the intention to do A can diverge from reasons concerning doing A in two ways (i) because the former reasons are WK; (ii) because the former reasons are subjective, and depend on a lack of information.
    To counterexample this, we’d need a case where there is an RK objective reason concerning the intention to do A, and no corresponding objective reason concerning doing A. Are there any such cases?

  20. Daniel – In my first case, I believe that my brother will call later to let me know whether he will be in Los Angeles. It’s because I believe that, that I have a reason to delay my decision – a subjective reason, as you say. But it might even be true. In which case I have an objective reason to delay my decision. And my second and third cases didn’t turn on a lack of present information at all.

  21. Mark – It’s not (just) the belief that your brother will call that is relevant; it’s the fact that you don’t know whether he will be in LA. The objective reasons depend on whether he will be in LA; you wait because there is the prospect of aligning your reasoning more closely with those reasons.
    I thought that the second case didn’t involve an RK reason not to intend (or at least it’s open to the defender of the thesis to say that it’s WK). The third case arguably fits my analysis, since the idea seems to be that you will deliberate better later on (when there is more time). Better deliberation will (usually) bring your reasoning better in line with the objective reasons, and so again we can explain not forming an intention by saying that doing so would involve relying on subjective reasons that are epistemically suboptimal (they are based on reasoning that you expect to improve on later).

  22. Hmm, sorry if this is repetitive – I haven’t read all the comments – but suppose we say the following (using the case of belief):
    One is rational in believing P iff one has sufficient object-given reasons to believe P.
    Thus: One is rational in neither believing P nor believing NOT P iff one has neither sufficient object-given reasons to believe P nor sufficient object-given reasons to believe NOT P.
    As such, there are in fact two sorts of reasons not to believe P:
    (1) Object-given reasons to believe NOT P.
    (2) Facts about the insufficiency of one’s object-given reasons to believe P.
    Given this, although there are never object-given reasons to neither believe P nor believe NOT P, it can sometimes be rational to neither believe P nor believe NOT P – namely, when reasons of the second sort rule out both beliefs.
    Now, the existence of both these sorts of reasons *not* to believe P is compatible with the claim that R is an RK-reason to believe P iff R is an object-given reason to believe P. But it does mean that there are sometimes reasons that are relevant to the rationality of believing P which are neither object-given reasons for P nor object-given reasons for NOT P…

  23. Hi, Karl.
    I think this lines up with one of Jussi’s earlier suggestions. For the record, I think there there are two problems with the idea that this can solve my problem.
    The first is a problem about why the object-given reasons to believe P are insufficient in the right cases. If your reasons to go to the store are at least as good as your reasons to not go to the store, then that makes your reasons to go to the store sufficient. But if your reasons to believe that p are at least as good as your reasons to believe that ~p, that doesn’t make your reasons to believe that p sufficient. Why not? Is it a sui generis fact about your reasons that they are insufficient? Or, as in the case of reasons to go to the store, are they insufficient because they are outweighed by reasons for the alternative?
    According to my view, what makes the reasons insufficient is that they are outweighed. That gives us an elegant, unified treatment of what is going on in all cases of sufficiency, which doesn’t require an ad hoc answer as to why reasons that are not outweighted are sufficient in some cases (in the case of reasons for action, for example), but not sufficient in other cases (in the cases of reasons for belief or for intention, for example), and which can explain how the bar for sufficiency goes up and down in different sorts of cases – it goes up or down accordingly as there are better reasons for the alternative.
    If you accept this view, then saying that the reasons to intend are insufficient in my cases is not enough; you have to say what the reasons not to intend are, that make them insufficient. Contrapositively, if you want to avoid saying what the reasons not to intend are, you need a sui generis account of their insufficiency, which is what I think is problematic.
    The second problem is one that I think I noted above; it’s that we can say why someone withheld belief – for example, because more evidence was going to come in. This shows that it is perfectly possible to withhold belief for the reason that more evidence is going to come in – which is one of the earmarks of it being a RK-reason to withhold belief – i.e., to not believe. Similarly, we can say that what makes it rational for her to withhold is that she believes that more evidence is going to come in – which is another of the earmarks of the fact that more evidence is going to come in being a RK-reason to withhold – i.e., to not believe.

  24. Thanks, Mark, that’s very helpful. (And sorry, Jussi, for repeating some of your suggestions!)
    Still, I think bringing in the case of action actually confuses the debate here. It seems to be a simpler case, to be sure, but I think those appearances are misleading.
    Let me try to say a bit about why. In the case of belief, as you note, one sort of case in which suspension of judgment seems rational looks like the following:
    (1) I have some evidence for P.
    (2) I have some evidence for NOT P.
    (3) I have good higher-order evidence that much better evidence concerning P is on its way.
    In this situation, all other things being equal, the rational thing to do is to suspend judgment.
    Similarly, in the case of intention we might have the following situation:
    (1′) I have access to some object-given reasons to intend to A.
    (2′) I have access to some object-given reasons to intend to NOT A.
    (3′) I have good evidence that I will soon have access to much more powerful object-given reasons concerning these intentions
    Once again, in this situation, all other things being equal, the rational thing to do is nothing.
    Ok, now the case of action. Here a parallel case would look like the following:
    (1”) I have access to some object-given reasons to A.
    (2”) I have access to some object-given reasons to NOT A.
    (3”) I have good evidence that I will soon have access to much more powerful object-given reasons concerning A’ing.
    The problem here is that, supposing A is an action at a particular time, the considerations mentioned in 3” will already be captured within one’s object-given reasons to NOT A – i.e. the evidence mentioned in 3” will support the claim that A’ing (now) would be overly hasty and that I should wait, etc. So, in the case of action, we don’t need to consider evidence that more information concerning A’ing is on its way *separately* – because these facts will be a subset of our object-given reasons to NOT A.
    This is what allows one to say that sufficiency is just a matter of outweighing in the case of action. But to say this is misleading – for, in fact, the very same substantive account of when it is rational to end deliberation is present in all three cases. It’s just that in the case of action, this account is being “hidden” within the set of object-given reasons to NOT A – whereas in the case of belief/intention it must be kept separate.
    Anyway, there are two basic points here. First, there is nothing ad hoc about the treatment of these cases I’ve suggested. Rather, the very same account of when stopping deliberation is rational will apply to belief, intention, and action. The only difference will be that in the case of action we can capture these considerations within the framework of outweighing. But this, it seems to me, is a relatively superficial feature of the action-case – one that is simply a product of the fact that the actions we are considering are actions at a particular time, but the belief/intentions we are considering are not beliefs/intentions at a time.
    And second, given all this, it’s unfair to complain to me that I need to give you a substantive account of sufficiency. After all, my account of sufficiency will be the same as the account of when it is rational to stop deliberation that is part of your account of the action case. So it’s not like I’m helping myself to anything here that you aren’t helping yourself to as well.

  25. Karl –
    I’ll grant that it looks like different views about what it is for some reasons to be sufficient are going to lead us to different levels of satisfaction with the idea that believing that my brother will tell me later this afternoon whether he will be in LA on Monday can make my reasons to intend to not go to LA on Monday insufficient, even though it doesn’t give me any RK-reason to not intend to not go to LA on Monday.
    But here’s how I look at it: here’s something, believing which affects the rationality of something (intending to not go to LA) – if I didn’t believe it, then my (subjective) RK-reasons to intend to not go to LA on Monday would be sufficient. It behaves in every way like a reason not to intend to not go to LA on Monday – I can not have the intention on those grounds, it can make it rational to not have the intention, and it can make reasons that would otherwise be sufficient reasons to have that intention turn out to be insufficient reasons to have the intention. I say: if it walks like a duck and talks like a duck, it’s a duck. Or whatever ducks do.

Comments are closed.