Well, not quite. The survey showed that about one-third of the respondents shared my intuition that I (qua subject of the example) have no reason to purchase ITEM and that about two-thirds of respondents have the contrary intuition that I do have a reason to purchase ITEM. This doesn’t show that I’m wrong, but it does show that my intuition is not widely shared.
Now, here’s why this is important. In his excellent paper “Parfit’s Case against Subjectivism,” David Sobel argues that subjectivists – those who think that an agent’s reasons for action are all ultimately determined by the contingent pro and con attitudes that she would have under some procedurally specified conditions – can accept Parfit’s claim that we have current reasons to do what will prevent us from suffering future agony regardless of whether or not we have any current pro attitudes towards our avoiding future agony. Sobel argues that subjectivists will claim that anyone who will be in future agony will, in the future, necessarily have a future desire (when in agony) to get out of agony and so will, when in agony, have a reason to get out of agony. And he argues that, by appealing to this fact and what he calls the Reasons Transfer Principle, the subjectivist can hold not only that anyone who will be in future agony will have a future reason to get out of agony, but also that those who can avoid future agony have a present reason to avoid future agony. According to Sobel, the Reasons Transfer Principle (RTP) says: “If one will later have a reason to get O, then one now has a reason to facilitate the later getting of O.”
I have three main worries about Sobel’s argument.
(1) Suppose that I have no current desire to avoid future agony but that my X-ing will cause me to avoid future agony. That is, suppose that if I X, I will not suffer future agony, and if I do not X, I will suffer future agony. It’s unclear to me how an appeal to RTP will allow the subjectivist to accept the plausible claim that I have a present reason to do X. It seems that RTP allows the subjectivist to say only that if there is some act, Y, that will allow me to stop (as opposed to avoid) suffering some future agony, then I have a present reason to facilitate my later doing Y. But let us imagine that, despite my lack of any desire to avoid future agony, I’m going to X (moreover, let's assume that, as a matter of fact, I am moved to X even though I would not be motivated in the least to X or favor my X-ing were I under the subjectivist’s procedurally specified conditions), and let us imagine that my X-ing will prevent me from suffering any future agony. In that case, it seems that RTP doesn’t apply. Since I won’t be suffering any future agony (given that I will be performing X), it is not the case that I will have any reason in the future to get out of future agony (I can't have reason to get out of the future agony that I won't be experiencing). And since I would have no current pro attitude towards my avoiding future agony if I were in the subjectivist’s procedurally specified conditions, then I have no present reason at all to X. And I take it that this is absurd. Surely, I do have a reason to X givent that it will preven my future agony.
(2) One of the main motivations for adopting subjectivism, I take it, is that it is compatible with metaphysical naturalism – the view that the only facts and properties are naturalistic facts and properties. Thus, the subjectivist wants to identify the seemingly spooky normative property of one’s having a reason to do X with the perfectly naturalistic (and, thus, non-spooky) property of one’s being such that one would be motivated to do X if one were under some procedurally specified conditions. But, of course, anyone who appeals to RTP must deny that the property of one’s having a reason to do X is simply to be identified with the naturalistic property of one’s being such that one would be motivated to do X if one were under some procedurally specified conditions. So I wonder what naturalistic property the proponent of RTP is going to identify with the having-a-reason property referred to in the consequent of the principle.
(3) I wonder whether RTP is plausible. It’s seems to me that although what I will later have reason to do depends on what intentions I’m forming at present, what I have reason to do now does not depend on what intentions I’m forming at present. So if I’m now forming the intention to discard the anesthetizing pill and will follow through with this intention, then I will later have a reason to take the amputating pill. But I don’t think that I now have a reason to take the amputating pill or to facilitate my taking the amputating pill. For, at present, taking the anesthetizing pill is an option given that I need only respond appropriately to my reasons and thereby form the intention to take the anesthetizing pill. In any case, that was my thought. So the example was supposed to be a counterexample to RTP. But I take it that it wasn’t a very good one.
Hi Doug,
I wasn’t sure why naturalising the RTP would be a problem for the subjectivist. Any individual could be time-sliced. For each time-slice, there is the property of being such that that time-slice of the person would be motivated to do X if that time-slice were under some procedurally specified conditions. The person at the earlier time then can have the naturalistic property of being such that she has a (later) time-slice such that the time-slice would be motivated to do X if that time-slice were under procedurally specified conditions. I take it that the subjectivist who wants to accept RTP and remain a naturalist can use this natural property of the person to give a naturalist, reductivist account of having a reason even before one has the relevant desire.
Given my holistic inclinations, I’d think that RTP would be false as a universal absolute principle. But, presumably all Sobel needs is that it holds in the relevant cases because it is plausible generalisation that holds by default unless there are underminers.
Hi Jussi,
So take the normative property (norm-P) of S’s having a reason at t to do X at t’ (t’ being later than t). I want to know what naturalistic property (nat-P) that norm-P is to be identified with. I take you to be suggesting that the subjectivist could take it to be identical to the following naturalistic property: S is at t such that “she has a (later) time-slice such that the time-slice would be motivated to do X if that time-slice were under procedurally specified conditions.” But that can’t be it, because the subjectivist holds that S now has a reason to do X at t’ if S would now be motivated to do X if she were under under procedurally specified conditions. Thus, her having a (later) time-slice such that the time-slice would be motivated to do X if that time-slice were under procedurally specified conditions is not necessary for now having a reason to do X at t’.
Is the idea, then, that norm-P is to be identified with the following disjunctive natural property: being such that either she would now be motivated to do X at t’ if she were under under procedurally specified conditions or she now has a (later) time-slice such that the time-slice would be motivated to do X if that time-slice were under procedurally specified conditions?
That just doesn’t seem plausible to me.
Well, here’s a general theory of reasons:
Subject S has now at t a reason to do X at a time t’ which may be t or later (and facilitate Xing before t’) iff there is some time-slice of S such that it would be motivated to do X at t’ if it were under procedurally specified conditions.
There’s nothing disjunctive about this view; it is still recognisably subjectivist in spirit (even if, as you point out, it is not strictly identical to the old formulations of subjectivism); and it is a naturalistic reduction.
One advantage of this view is that you can derive from it both an account of the reasons for actions we have at the moment on the basis of our current desires (our current time-slice is one of the relevant time-slices that you can use in the formulation) and the norm-P of the reasons we have on the basis of future desires. This would be just instances of the same principle. In a sense, this view therefore explains why RTP might be true.
In effect this view of reasons could be used to give an account of prudence and prudential reasons that is akin to desire-satisfaction consequentialism. What the desires of each agent is for the consequentialist, the desires of different time-slices of an agent are to this view of the prudential reasons an agent has.
Now, the view might be implausible for other reasons and I would be interested to hear why, but it is a naturalistic reduction of reasons based on future desires.
Hi Jussi,
That’s good. I accept that what you’ve given is a naturalistic reduction and that it is recognizably subjectivist.
The question, then, is whether this is indeed the reduction that subjectivists, such as Sobel, would plunk for and, if it is, whether it’s plausible.
I question its plausibility, for it doesn’t seem to account for my having a reason to do X where X will prevent me from suffering any future agony in the sort of case that I described in my worry (1). Suppose that X at t’ is taking the anesthetizing pill before the agony would otherwise initiate. And assume there is no time-slice of me such that I would be motivated to do X if I were under procedurally specified conditions. (This is compatible with my wanting to take an anesthetizing pill after the agony has initiated — a distinct token act — and it’s compatible with my wanting the agony to stop and wanting to facilitate its stopping by doing Y.)
Given that, in this example, it is not the case that there is some time-slice of me such that I would be motivated to do X at t’ if I were under procedurally specified conditions, the proposed principle implies that it is not the case that I have a reason to do X, nor a reason to facilitate my doing X. That seems quite implausible to me given that my doing X would prevent me from suffering any future agony.
Hi Jussi,
Let me add: I take it to be Sobel’s aim to accommodate Parfit’s (and many other people’s) intuition that S has a reason to do X at t in those instances in which doing X at t would prevent herself from suffering future agony. So whatever naturalistic reduction is proposed, it had better be one that vindicates this intuition.
As I am understanding it, Doug disagrees that S necessarily has a reason to do X at t in those instances in which doing X at t would prevent herself from suffering future agony. Doug, as I am understanding him, would say that is not so in cases in which I could but will not take action at t that would have completely avoided the agony. In such cases, he claims, I lack a reason at t to implement any back-up plan that takes into account my acting as I myself think I ought not. Thus I think Doug’s real fight is with Parfit, not me. I was taking on board the plausible thought that we do necessarily have a reason now to avoid future agony. That still seems quite plausible to me. But if it is false it is bad news for Parfit’s argument which is what I was out to show was not telling against subjectivism. If Doug is right, Parfit would need to say that there are exceptions to the general thought that I necessarily have reason now to take action to avoid foreseeable future agony. In that case, I assume I can just take on board such exceptions in my dispute with Parfit. In other words, I don’t see the danger to my position posed by the very interesting kind of cases Doug is pointing us to.
Likely I am reading this too quickly but Jussi’s formulation looks a lot like what I had in mind. However I tried to argue that there likely are principled grounds for not worrying about past procedurally correct desires. Suppose one thinks that only the satisfaction of now-for-now desires matter. In that case, I cannot now help facilitate the satisfaction of desires that I only had in the past unless I have a time machine. So my RTP, with these assumptions, would not say I had any reason to facilitate such past desires because I cannot. There are also very tricky cases where I will have a very serious change in stance. Perhaps I will later desire to torture whereas now the idea is repugnant to me. Do I have a reason to help along my future self regardless of his desires? I allowed that such cases are difficult but also not needed to deal with Parfit’s agony case.
Doug’s cases are quite interesting. I have not thought enough about what I would like to say about them. It feels to me like an advisor would be torn about what to advise. They would want to say that you should do the wise thing but that given that one is very likely not to, one should plan now for the case where you do not now do the most wise thing. Suppose the thing a heroin addict has most reason to do now is to quit and they know it and they are capable of doing it. But they also know they are unlikely to do so. It seems that it would be wise of such a person to invest now in clean needles. It feels like these thoughts are perhaps, as Jan would say, just relative to different sets of information. That is, I would be tempted to find a solution that says there are different kinds of reasons at play here, such as the difference between objective and subjective reasons, rather than claim that there is no sense in which one ought not buy the needles or no sense in which one ought to.
As to Doug’s 1, I guess I was thinking that the subjectivist can say that the ideal procedure would consider my motivations in possible futures (not just the actual future). If in certain possible futures I would very much mind being in agony, then I have reason to avoid those possible futures.
Hi David,
Well, if your position is only that Parfit’s agony argument as originally constructed by him (where it assumes that S necessarily has a reason to do X at t in those instances in which doing X at t would prevent herself from suffering future agony) doesn’t pose a threat to subjectivism, then you’re absolutely right that there is no danger to your position. Because my putative counterexample to RTP is as much a counterexample to Parfit’s claim that S necessarily has a reason to do X at t in those instances in which doing X at t would prevent herself from suffering future agony as it is to your RTP. But I think that I have a real fight (I would say disagreement) with you insofar as you are a subjectivist and someone who thinks both that RTP is true and that RTP can be used by the subjectivists to ward off close cousin’s to Parfit’s agony argument. So here’s my version of the agony argument, which is, I take it, an argument against your subjectivism.
(P1) Necessarily, if a subject, S, can avoid suffering agony at T10 by, and only by, performing a costless act, X, at t1 (where t0 is the present and X is, say, taking the anesthetizing pill), then S has at t0 a reason to perform X at t1.
(P2) If either RTP is false or RTP fails to imply S has at t0 a reason to perform X at t1 in the sorts of cases described in my worry 1, then subjectivism cannot vindicate P1 and is therefore false.
(P3) RTP is false — see my putative counterexample, which I assume that you and many others will not take to be a counterexample at all.
(P4) RTP fails to imply S has at t0 a reason to perform X at t1 in the sorts of cases described in my worry 1.
(P5) RTP doesn’t help subjectivism to vindicate P1 without giving up the primary motivation for the view, because the primary motivation for subjectivism is a commitment to metaphysical naturalism and its unclear what naturalistic reduction of the normative property (norm-P) of S’s having a reason at t to do X at t’ that the advocate of RTP can be given by the subjectivist to vindicate P1.
(C) Therefore, subjectivism is false and/or subjectivism can vindicate P1 only by abandoning the primary motivation for adopting in the first place.
So I know that you reject P3 and you’ll have many people on your side in doing so. But, if you have the time, I wonder what you would say about P4 and P5.
Hi David,
I see that our comments crossed in the either.
You write: “As to Doug’s 1, I guess I was thinking that the subjectivist can say that the ideal procedure would consider my motivations in possible futures (not just the actual future). If in certain possible futures I would very much mind being in agony, then I have reason to avoid those possible futures.”
I take it, then, that you concede that RTP as quoted from your article won’t do. So what would the revision be? Would it be this:
If one will in some possible future later have a reason to get O, then one now has a reason to facilitate the later getting of O.
That doesn’t seem plausible. But what would the revision be, then, exactly.
I meant crossed in the ether, not crossed in the either.
Hi Doug and David
I just wanted to say that I agree with David even if the dialectic is quite complicated and beyond me. It’s true that your last comment requires reformulate the proposal I gave (I didn’t quite get the case (1) first), but I don’t think that it requires taking into account merely possible motivations. Here’s another go:
Subject S has now at t a reason to do X at a time t’ which may be t or later (and facilitate Xing before t’) iff there is a sound deliberative route from the motivations of all S’s timeslices to wanting to X at t’.
Doug, it seems like you need to grant to us that the future time-slices can have a desire to avoid their present agony even if they have no desire to X because the earlier slices already actually have done so. If we later asked the person ‘would you want to be in agony right now – the kind of agony you avoided by Xing?’ she would presumably say ‘No!’. So the future slices have some desire not to be in agony at that moment (which they are not given that the person Xed).
The new proposals constructs a hypothetical set of motivations for the agent when she is choosing whether to X or not. It seems like, in your case (1), instrumental reasoning from that set can lead to a motivation to X from the future desire not to be in agony at that future time. This would give a reason to X in your case. You might use the Smith-styled advisor here too but it seems that we still have a naturalistic reduction and we can generate the reason in (1). So, this does seem to enable the subjectivist view to support the basic Parfit intuition.
In many cases, this principle will support RTP but not in all depending on what happens during the sound deliberation. Of course, as David points out, sound deliberation would mean also more information about one’s future actions.
Finally, I would very much recommend Michael Smith’s paper in recent Ratio entitled ‘Deontological Moral Obligations and Non-Welfarist Agent-Relative Values’. Despite its title, there is a response in it to this very objection to subjectivism which is very interesting.
Hi Jussi,
Are you talking about S’s actual or possible time-slices? Because I don’t see that there is a sound deliberative route from the motivations of all S’s actual past, present, and future time-slices to wanting to X at t’ in the sorts of cases that I’m imagining (cases in which I won’t experience any future agony given that I am going to perform X at t’).
If, by contrast, you mean to be talking about S’s actual (?) past, actual (?) present, and possible (?) future time-slices, then can you explain to me what the deliberative process is that gets me to wanting to do X at t’, but that hopefully doesn’t imply that I have a reason to facilitate my not getting chewed up by sharks by buying at present some very expensive shark suit (see http://www.neptunic.com/sharksuits_csuit.htm) because there is some possible future in which I suddenly become very stupid and decide to swim in shark invested waters. Surely, I don’t have a reason now to purchase a very expensive shark suit.
You write: “The new proposals constructs a hypothetical set of motivations for the agent when she is choosing whether to X or not.”
In the case that I’m imagining the agent is not choosing whether to X or not. Assume that she has already decided to do X. We can still ask whether she has a reason to do what she decided to do. Or if she has already performed X, we ask whether she had a reason to perform X.
Hi Jussi,
To be clear, I don’t think that I have to grant that necessarily future time-slices of any agent will have a desire to avoid being in agony. What Parfit and I would concede is only that “my state of mind would not be agony unless I had a strong desire not to be in this
state.” (Quoting Parfit.) But I don’t think that this implies that necessarily an agent who is not in agony must have a desire not to be in agony. That seems like an entirely contingent matter to me. Why do you think it necessary?
Hi,
only actual present and actual future time-slices. We know three things of them:
1. Not even one of these time-slices experiences any agony given that one of them decided to X.
2. Every one of them strongly prefers itself not to be in agony to being agony.
3. Not even one of them has any preferences of the agony of the other time-slices.
Now, we ask, given that the agent has already performed X, did she have a reason to decide to X and then X? Well, on this reductivist view, this is very roughly the same question of whether the decision to X and the act of Xing satisfied more of all the aggregated preferences of the time-slices than the other the decisions/acts available to the agent at the time. The answer of course is yes. So we get the right result.
However, part of me wants to take the potential time-slices into account. If it is likely that I will come into contact with sharks, a suit I must save for. If it’s unlikely that this will happen, then I shouldn’t. Because we are taking into account all the possibilities in the possibility space, some of the time-slice agents and outcomes will be likelier than others (there will be more of them in the possibility space) and therefore there will be more time-slices with the relevant preferences. So, this kind of expected utility preference satisfaction prudentialism will work in the same way as expected utility preference satisfaction consequentialism does.
Hi Jussi,
So you think that
necessarily: For any actual present and actual future time-slices of any agent, every one of them strongly prefers itself not to be in agony to being agony.
If so, I disagree. I accept that those time-slices that are in agony will strongly prefer not be in agony, but whether a time-slice that is not in agony prefers not to be in agony to being agony seems like a contingent matter. By the way, I’m assuming that for S to prefer A to B is for S to be disposed to bring about A as opposed to B when other things are equal and S has the option to bring about one or the other, but not both.
Moreover, I wonder if you’re now going beyond subjectivism. As Sobel defines it, it’s the view that an agent’s reasons for action are all ultimately determined by the contingent pro and con attitudes that she would have under some procedurally specified conditions. But now you seem to be saying that the relevant preferences (the ones that are source of reasons to avoid future agony) are not contingent.
Hi,
no. I didn’t say anything about necessity, did I? You asked does the agent happen to have a reason and I explained why she does. So, it seems like you are moving the goal posts here.
In any case, it seems plausible that most beings desire not to be in agony at present (good evolutionary reasons for this…). If a being didn’t have such desires, then I’m not sure I have a strong intuition that it would have a reason to avoid agony so that’s not a problem. So, it’s not clear to me that beings necessarily have reasons to avoid future agony.
I’d want to also say that what is to be in agony is to be in a state you don’t want to be in. So, beings that can experience agony, need to desire not to be in that state. If we accepted this view of agony, then we would get necessarily reasons to avoid future agony.
In any case, the question was, if we assume that a person doesn’t care about future agony even if she cares about present agony and continues to do so, will she have a reason to act now to avoid future agony on the subjectivist view? And, the answer is yes. True she won’t have such a reason if she stops caring about present agony. But this isn’t any worse than not having a reason to get rid of agony if you have it if you don’t want to. So, it’s not a further problem.
Thus, my view satisfies Sobel’s definition unless we accept a desire-based account of agony. All it assumes are contingent desires at a time and a procedure to aggregate them.
Finally, I don’t quite see what the grounds for the asymmetry would be about what the agony-experiencing and non-agony-experiencing beings’ desires. Of both of them, it’s true that if they were experiencing agony, they would try to get away from that state. This is all that’s required of them to have the disposition to bring about… I don’t see how there could be a difference in these dispositions in the way you think.
You guys are hard to keep up with but I will hope to set aside some time soon to figure out what I might say about some of these tricky issues.
Hi Jussi,
You said: “We know…Every one of them strongly prefers itself not to be in agony to being agony.” If whether or not a given time-slice of a given agent prefers not to be in agony is merely contingent, then how do you know that all the time-slices of the agent in my example have this preference? If it’s contingent, can’t I just stipulate that the agent has at no time a preference for not being in agony? I and Parfit would still claim that even though the agent doesn’t at any present or future time have preference not to be in agony, the agent does have a reason to do now what will prevent herself from suffering future agony. From what you’ve said, I gather that you don’t accept this claim.
In any case, Parfit’s agony argument assumes that “We all have a reason to want to avoid, and to try to avoid, all future agony.” And my revised version of it assumes that “Necessarily, if a subject, S, can avoid suffering agony at T10 by, and only by, performing a costless act, X, at t1 (where t0 is the present and X is, say, taking the anesthetizing pill), then S has at t0 a reason to perform X at t1.”
So, if you’re allowing that we don’t all have a reason to act so as to avoid future agony, then you’re just rejecting Parfit’s initial assumption. But Sobel seems to want to grant (at least for the sake of argument) that we all have a reason to act so as to avoid future agony and to argue that subjectivism can vindicate this claim. But it seems that what you’re now saying isn’t compatible with Sobel’s aim.
Hi Doug,
yes – because I was stipulating that case. If I stipulate a person’s contingent desires, then I know what the person desires. If it’s my thought-experiment, then I make the facts. I just wanted to show that, in the light of this minimal assumption, that people want to avoid present agony we can generate reasons to avoid future agony on a subjectivist view.
yes – you too can stipulate that a person doesn’t desire not to be in agony at any moment. I’m not sure what to say about this case, but here few things:
1. It’s not clear to me intuitively that such a person does have a reason to avoid agony he doesn’t want to. To say this is not to reject Parfit’s assumption. It’s not a modal claim though your revision is. Funnily enough, elsewhere in the book Parfit admits that normative theories do not need to apply to outlandish scenarios. This should go here too.
2. It seems to me that being in agony *means* being in a state you want not to be in. If someone lacks a desire not to be in a state, I don’t see how it could count as agony. This view would get the strong modal version of your revised version of Parfit’s view. You might reject this view of agony, but then I would like to see a more plausible alternative.
3. We ascribe beliefs and desires to make sense of other people’s behaviour on the basis of something like the principle charity. So, it’s constitutive of understanding other people that we make them rational, believers of the true and lovers of the good. For this reason, I would want to see more substantial examples of cases in which we have agents’ whose behaviour we can make sense of without needing to ascribe to them a desire to avoid agony whilst simultanously there being a reason on the light of the agent’s other ends to do so. It seems to me that understanding people requires assuming that they care about their agony even if they might care about other things more.
We should ask Sobel about this, but I don’t think his aim was to argue that we all have a reason to avoid future agony even if we don’t want to avoid agony in the future.
Finally, the Smith article mentioned above fairly convincingly argues that procedural rationality, including instrumental rationality, requires certain constitutive substantial desires from agents. If this argument works, then no matter what an agent’s starting-point desire-set is, if the agent is capable of instrumental reasoning, then she will end up with a desire to avoid future agony.
Hi Jussi,
We’re talking past one another. I thought that we were talking about my case, the one that I proposed in (1) of the original post and the one that I reiterated in my objection to your initial proposed naturalistic reduction. I thought this because you said: “It’s true that your last comment requires reformulate the proposal I gave (I didn’t quite get the case (1) first).”
You write: “It seems to me that being in agony *means* being in a state you want not to be in. If someone lacks a desire not to be in a state, I don’t see how it could count as agony.” Parfit and I are willing to concede this for the sake of argument. But from this, it does not follow that someone who is not in agony must desire to not be in agony.
You also write: “We should ask Sobel about this, but I don’t think his aim was to argue that we all have a reason to [act now so as to] avoid future agony even if we don’t [now] want to avoid agony in the future.” This is precisely what I take to be one of his main aims.
I’ll have to check out Smith. But right now I’m wondering if Sobel’s appeal to RTF will allow him to avoid all plausible versions of the agony argument.
I think backing up a bit and trying to explain what I wanted to argue in my paper might be helpful.
Parfit argued that subjective theories of normative reasons are false because they cannot vindicate certain obvious truths. The obvious truth he focused on was that we have reasons now to take steps to avoid future agony. Parfit claimed subjectivists cannot guarantee that result. He was clearly focused on cases where it is clear that we do have such reasons. Doug describes cases where that is disputable. I think to consider the threat Parfit’s argument poses to subjectivism, we should ignore the cases where it is not clear that we now necessarily have reason to avoid future agony.
Parfit does not put it this way, but I take his challenge to subjectivism to amount to the claim that subjectivists cannot ensure that future reasons we have transfer to our current situation and give us reasons now. Parfit claimed, without argument, that subjectivists must say that my current reasons are determined by my current subjective motivational set after it is exposed to procedurally correct deliberation.
I offered a variety of responses but the one we are focused on here is the claim that subjectivists can include it in their notion of procedurally correct deliberation that one be motivated by the subjective concerns one will (and I should have said, or would) have in the future. That is, subjectivists can say that you have to care about what you will (or would) care about. It is not procedurally rational, I was trying out saying, to now not care about the fact that you will soon care very much to get out of agony.
So I crucially did want to say, as Doug suggests, that I necessarily do now have a reason to facilitate the latter getting of what I will later have subjective reason to get (setting aside Doug’s interesting cases where anyone, subjectivist or not, may reasonably doubt that we have current reasons to further future reasons).
I don’t really feel that I am much invested in the literal truth of the RTP. I thought I was until Doug showed me that there are grounds that might appeal to anyone to doubt that one has a current reason to further future reasons. I still doubt that, but if he was right I do not think that would affect my case against Parfit. We could then just focus on the cases where Parfit is right that we necessarily have such reasons and wonder whether subjectivism can account for that. I am afraid that my own interest in the RTP is really just in how effective it is in blunting Parfit’s argument. I think Doug’s issues about it outside of that context are very interesting issues but I don’t see them yet as central to the fate of subjectivism which was my main concern.
I should have also said I do not see the issue about naturalism. The typical subjectivist says that it is certain kinds of desires that carry normative authority, for example informed one’s. There is no getting around the thought that the subjectivist must favor some kinds of desires or some kinds of procedures over others. If that costs them their naturalism, then that is a cost that the mainstream view must pay. But I don’t see the new threat to our subjectivist’s naturalist credentials that she has a non-standard conception of the relevant procedural test.
Hi David,
Thanks. That’s very helpful. Can I infer that it’s not clear to you that Smith has a reason to do X in the following sort of case?
The Pill: Smith will suffer horrible agony from t10-t15 unless he takes the anesthetizing pill at t1 (call this token act “X”), and he knows this. It’s now t0. Smith has at t0 no pro-attitude towards his not being in agony from t10-t15. Moreover, he will not at any future time have a pro-attitude towards his not being in agony at t10-t15. He does, however, have at t0 a pro-attitude towards his skin turning purple at t10, and he falsely believes that his doing X is a necessary means to his skin turning purple at t10. Thus, he does X and doesn’t experience any agony from t10-t15. Nevertheless, had he not done X, he would, as I said, suffered terrible agony from t10-t15. And if he had suffered any agony at t10, he would have then had a very intense pro-attitude toward his not continuing to suffer agony until t15.
This case (unlike the case for which I conducted the survey) seems like a clear case in which Smith has a reason to do X. Am I right in thinking that you disagree? If so, then I now understand better one of main points of disagreement. For I think that the following is a powerful and sound argument against subjectivism:
(P1) Smith has a reason to perform X (that is, to take the anesthetizing pill at t1).
(P2) Subjectivism (even if supplemented with RTP) cannot vindicate P1.
(C) Therefore, subjectivism is false.
This argument seems to me to be absolutely central to the fate of subjectivism, although I take it that your response would be to deny P1.
Regarding naturalism, my point was only that I fear that one would have to reject metaphysical naturalism in order to endorse RTP or anything like it. So insofar as one would be attracted to subjectivism in the first place on the grounds that it is compatible with metaphysical naturalism, then one would not be attracted to your response to the agony argument.
Now, you don’t see any threat to your naturalistic credentials. If so, then could you tell me what naturalististic reduction of the normative property of S having at t at a reason to do X at t’ you would propose?
I see that t0 (t-zero) looks a lot like the word ‘to’. I’m sorry about that. It’s looks different as I type but not once posted.
Doug,
On the naturalism issue, my point was that I see no reason to think that if the other mainstream subjectivist views are ok with naturalism, that my proposal is not. I have no worked out reduction to offer you. But I am not worried unless you show me that this addition I have made itself looks problematic from the point of view of naturalism and I don’t yet see that you have made any such case.
If I am understanding the key thought in the Smith case, you are still urging the thought that the RTP only applies when I will actually come to have certain concerns in the future (as opposed to it being the case that I would have come to certain concerns had certain things happened). In my recent longer post I allowed that I should have said that the latter sort of cases are relevant too.
Hi David,
Can you explain to me how RTP (as quoted from your article) implies that, in The Pill, Smith now has a reason to do X?
Or if you concede that RTP (as quoted from your article) doesn’t imply that, in The Pill, Smith now has a reason to do X, can you tell me how exactly you would reformulate RTP. I suggested one implausible revision above. And you never commented on whether you accepted that revision or some other.
Hi David,
So, on the naturalism issue, is the thought that Smith would, if in the appropriate procedurally specified conditions, have at present a pro-attitude towards his not being in pain from t10-t15?
Hi David,
You stated RTP as follows: “If one will later have a reason to get O, then one now has a reason to facilitate the later getting of O.”
I take it that, in The Pill, ‘O’ would stand for ‘the state of affairs where Smith doesn’t suffer agony from t10-t15’. And what I’m not seeing is how S *will* later have a reason to get the state of affairs where Smith doesn’t suffer agony from t10-t15 given that Smith will not later be having any pro-attitude towards this state of affairs.
It would certainly be complicated to work out exactly how the precise proposal would go that allows that counterfactual concerns are relevant. Perhaps if you told me what you are worried about in my vague proposal we could get to more fruitful issues more quickly. My thought generally is that subjectivists should think that a temporally extended agent should not concern herself only with her current concerns but must look out for her ability to further the concerns she may have in the future. To my way of thinking, the other path looks to artificially favor the current time slice’s concerns over concerns of other time slices that are (or would be) just as much those of the agent whose reasons we are considering. Similarly a non-subjectivist story might say that I should only look out for the non-subjective interests of my current time slice–but that seems to give priority to a part of me over other parts of me for no obvious reason. The subjectivism I favor would say that one has reason now to see to it that one’s subjective concerns (that one will have or that one would have had) are served in the future. If that is not subjectivism, I do not mind not being a subjectivist. Nonetheless, on that view all of my ultimate reasons are determined by what will or would further things that I just happen to care about. And rationality is just a matter of wisely promoting what I have (or would have) such reasons to care about throughout my life.
I think the main issue is in danger of getting lost.
1. On naturalism: David seems to me to be right about this. Being a reason will have some naturalistic supervenience base, according to everyone (including all sorts of subjectivists but also all objectivists). The question of naturalism is the question of whether being a reason is identical to that base. This is just a separate question, and if a simplistic subjectivist can answer Yes, I don’t see why David’s more complicated subjectivist can’t say the same.
2. But, David, you haven’t really explained what the updated version is — the one according to which a person’s merely possible future desires could provide the person with actual present reasons. I agree with Doug that this looks pretty dubious, but it’s hard to say unless you tell us what the view is.
I can briefly say why the Merely Possible Future Desires thing looks dubious. There are two different views that are easy to entangle. One is a desire-based view of reasons, a la Williams, etc. The other is a kind of egoism accompanied by a desire-satisfaction conception of welfare. It seems to me that allowing future desires to produce present reasons already is a step away from the first kind of view toward the second. Allowing merely possible future desires to enter the picture is another step in that same direction.
Since I find the egoistic view extremely implausible, each of the two steps makes me fear for your Humean soul.
Hi David,
You write: “Perhaps if you told me what you are worried about in my vague proposal we could get to more fruitful issues more quickly. My thought generally is that subjectivists should think that a temporally extended agent should not concern herself only with her current concerns but must look out for her ability to further the concerns she may have in the future.”
I thought that I had, although maybe some of this was in response to Jussi. So you might have missed it. But, to reiterate, my worry is that such a revised version of RTP will be subject to the too-many reasons worry. Consider that I may in the future (for I could stupidly decide in the future to swim in shark invested waters) have a concern to be protected from shark bites with an expensive shark suit. But if I’m not going to stupidly decide in the future to swim in shark invested waters, then it seems false to say that I have a reason now to buy a shark suit because it’s on sale and thus would facilitate my achieving ends that I may later have.
In any case, do you admit that RTP as it stands won’t do? And if so, isn’t the burden on you to propose some way for the subjectivist to vindicate the extremely plausible claim that, in The Pill, Smith now has a reason to do X?
Doug,
The way you insert [now] to this quote is revealing:
“You also write: “We should ask Sobel about this, but I don’t think his aim was to argue that we all have a reason to [act now so as to] avoid future agony even if we don’t [now] want to avoid agony in the future.” This is precisely what I take to be one of his main aims.
I accept that this is one of Sobel’s aims. And, the theory I gave which Sobel seemed to like satisfies this aim easily.
However, that quoted sentences was in response to your demand that the subjectivist must be able to give an account that entails that an agent necessarily has reasons to X *even if her future time-slices will not desire to avoid agony when they exist*. This is what you said here:
“I and Parfit would still claim that even though the agent doesn’t at any present or future time have preference not to be in agony, the agent does have a reason to do now what will prevent herself from suffering future agony. From what you’ve said, I gather that you don’t accept this claim.”
In the quoted sentence, I wanted to deny that establishing this is Sobel’s aim. This is why the correct reading of the quote is:
“We should ask Sobel about this, but I don’t think his aim was to argue that we all have a reason to [act now so as to] avoid future agony even if we don’t [in the future] want to avoid agony in the future.
And, that this is not his aim is exactly what Sobel confirms in his response above:
“Parfit does not put it this way, but I take his challenge to subjectivism to amount to the claim that subjectivists cannot ensure that future reasons we have transfer to our current situation and give us reasons now…
I offered a variety of responses but the one we are focused on here is the claim that subjectivists can include it in their notion of procedurally correct deliberation that one be motivated by the subjective concerns one will (and I should have said, or would) have in the future. That is, subjectivists can say that you have to care about what you will (or would) care about”.
Note how this response, which Sobel seems to still accept, is to the challenge which assumes that all we need to do to response to Parfit is to explain how the reasons an agent has in the future because of his desires then will transfer to the agent’s reasons in the current situation where he lacks any future directed desires. So, he seems to share my view about what is required to respond to Parfit – we can assume that the future time-slices care about their own agony and we then need to explain how the reasons based on these desires transmit to the earlier time-slices through instrumental rationality.
To repeat, as far as I see, we have done this part satisfactorily [well, I think we need to add something like what Smith says]. We haven’t explained how you could have reasons to avoid future agony even when you don’t want to avoid agony in the future. But, as far as my and Sobel’s understanding of the challenge goes, this was never at issues. That’s a further challenge for the subjectivist which you set – even if it’s not any harder than the challenge of whether the subjectivist can create a reason to avoid present pain if one doesn’t want to do so.
Finally, here’s what Parfit says about agony on page 74:
“This case might be claimed to be impossible, because my state of mind would not be agony unless I had a strong desire not to be in this state. But this objection overlooks the difference between our attitudes to present and future agony. Though I know that, when I am later in agony, I shall have a strong desire not to be in this state, I might have no desire now to avoid this future agony”.
Note how Parfit seems responds to this thought by drawing the distinction between our attitudes at present and in the future and not by drawing the distinction between the desires of a person in agony and a counterpart who is not in agony at the same moment. So, he seems to grant that in the future we have desires not to be in agony then.
Finally, it’s clear from 74-75 that his crux with the view we are discussing is merely that it’s not subjectivist enough. You seemed to accept above already that this view still is recognisably subjectivist.
Sorry about the long post. I’ll go back to marking essays…
Hi Jussi,
You write: “We haven’t explained how you could have reasons to avoid future agony even when you don’t want to avoid agony in the future. But, as far as my and Sobel’s understanding of the challenge goes, this was never at issues. That’s a further challenge for the subjectivist which you set”
Let’s grant that I’m posing a further challenge in light of Sobel’s response to Parfit’s original challenge. I’m curious how Sobel or you propose to meet that challenge. Does Sobel reject my premise that, in The Pill, Smith has a reason to perform X (that is, to take the anesthetizing pill at t1)? Or does he reject my premise that subjectivism (even if supplemented with RTP) cannot vindicate P1. I take that Sobel denies this second premise. But I don’t see what the plausible alternative to RTP is.
I think I’m with Jamie – I’m worried I’m losing the big picture. I’m wondering if it will help to accept a Sobel-like treatment of Parfit’s case and contrast it with another case where that treatment seems inapt (like Doug’s strategy, but pared down). The goal is to for the subjectivist to get the right results in a non ad-hoc way, and I don’t see how they do that yet. So here’s my contrast case.
I’m in a chamber with 4 other people. If I pop a cork out of the wall by pulling a string at t1 the room will flood. If I retrieve the cork at t2 and proceed to put it back in, which I can do by tn (>t2), but no sooner (it’s hard and I have to wait for the waters to rise high enough for me to reach the hole), 2 people will have died from the flood waters, but I’ll save the other 3 of us. If I don’t put the cork back in, we’ll all die.
I will pop the cork out at t1. (Dammit!)
Assume that I have a (most?) desire-based reason to retrieve the cork at t2 (you know me – I want to save people). The trick is to make this (ir)relavant at t1. By RTP, I have reason to facilitate that at t1; i.e., I have reason to pop the cork out, so it is retrievable later. But that’s exactly what got us into this mess! I think I have *no* reason to pop the cork out at t1. Let us all live. (This is consistent with having some reason to not pop the cork out at t1, even desire-based. One can assert that considered judgments about the absence of reasons are no good, but I don’ think this is the best case for that move.)
This places pressure on the combination of subjectivism and RTP. David indicates that he is not wedded to RTP. I agree with Jamie that modalizing it is not attractive. In any event, this is a clear case where we all agree on what I will have reason to do at t2. So the challenge is for subjectivists to distinguish these cases. Otherwise the patch for the agony case gets the wrong results on other cases, and the enterprise starts to look like an ad hoc way of trying to get the same results as an object given theory of reasons.
General comment on Doug’s initial case: I gave Doug’s answer. But I’m generally wary of holding motivational and behavioral profiles fixed while asking for normative judgments. This is because I think whether one ought to/has reason to/etc. phi can influence his motivation to phi/behavior vis-a-vis phiing. (Our judgment about the case would not be exogenous to it, as it were.) But that is a bigger issue, I think.
Hi Matt,
Thanks for this example. It makes the same point as the one from my survey, but is, as you kindly put it, pared down.
I keep falling further behind on this thing.
On the Jamie stuff: The cases that Doug is worried about, I think, are cases where if I do Z, I will later get lots of agony. For bad reasons, I do not do Z. So I never feel the agony. We want to say that I had good reasons to not do Z based in the fact that it would have caused me agony but as I never in fact feel the agony, we cannot say this reason is generated by actual future reasons I have to get out of the agony. Thus my earlier formulation that says reasons I have latter transfer and provide reasons to facilitate the avoiding of what I will latter have reason to avoid, is incomplete. What I need here is the thought that because Z would have caused me to be in a situation where I had reason get out of agony, that gives my present self a reason to not Z. I would not want to say that the fact that I might be hit on the head and then want an apple gives me a reason now to facilitate the latter getting of an apple. This situation does not have the feel to me of a deep problem but rather something that needs a technical fix. Admittedly that is easy to say, but wonder if others have the feel that there is a serious problem looming here.
Matt: I like that case. I am not sure what my intuition about whether I have any (even quite outweighted) reason to do the dumb thing at t1. I do have something like a consideration in favor of doing so that someone who could not replace the cork lacks, but I am not sure that is enough to count as having a reason to do it as opposed to just having less reason to not do it.
The cases I was focused on are cases where it is allowed to be obvious that because I will have a reason later, I have a reason now. There may be many cases that are not of this form. But it is cases of that form that are being supposed to be the problem cases for subjectivism. I was offering a fix for those supposedly problematic cases. You may say that is ad hoc. I would say that we all need to distinguish cases where I necessarily have reason now because I will have reason later from these tricky cases you guys are interested in. Providing such a distinction is a job for subjectivists and non-subjectivists alike. Suppose we have in hand such a distinction. The threat is that in the transfer cases, subjectivism cannot accommodate the fact that our reasons transfer in such cases. That is what I was trying to say is not the case. I need help seeing how the existence of non-transfer cases is supposed to be a problem for my reply to Parfit.
Jussi,
Yes, Parfit does not rely on the thought that a person might, while in agony, have no desire to get out of it. At least for the sake of argument he allows that that is necessarily the case. That is why he focuses on the case of future agony rather than present agony. So in the context of replying to Parfit I help myself to the thought that there is necessarily a subjectivist friendly rationale for saying that the agent in agony necessarily has a reason to get out of it while they are in it. The problem, Parfit suggests, is that the subjectivist has no way of ensuring that I now care about the reasons I will have once I am in future agony.
David, I take it that Parfit needs no transfer principle of the kind you offer. He need not derive present reasons from future ones (actual or counterfactual). In the agony case, he’ll just say that the fact that my phiing now will avoid future agony is a reason to phi now (or something like this).
I worry about this focus more generally: “The cases I was focused on are cases where it is allowed to be obvious that because I will have a reason later, I have a reason now.” I’m not sure that there are cases where it is obvious that I have my present reason *because* I will have some future reason. In the agony case, it isn’t obvious to me I have reason to move my hand (is that the right body part?) *because* later I have reason to do something (move my hand, do something to make the pain stop, continue on without pain (in the case where I did move my hand)). I don’t think Parfit is thinking of the case that way. Maybe we have different views about Parfit’s analysis of the case?
Even if there is a principle that captures cases where this ‘because’ holds, I think pairing it with subjectivism is a dangerous combination, and calls for special attention to cases like my cork one, and for reasons similar to the one you mentioned in earlier in the thread – I become a Tea Partier in my golden years, e.g. The more nuanced the transfer principle gets to carve the cases aright, the more worried I am that these are ad hoc moves to mirror the objectivist’s verdicts.
Matt,
just for what it’s worth, I think there’s a subjectivist non ad-hoc way of dealing with your case. What this seems to show in the end is that RTP is not an absolute principle by the lights of subjectivism.
Here’s a sketch:
At t1, I’ve got one set of options – to pop the cork or not to pop it and either facilitate putting the cork back on or not to do so.
Let’s say we an advisor, Jussi+, at t1 who has the desires of all my future-time slices (and we make these coherent, unified, and informed). Let’s say that we reduce my reasons to what Jussi+ wants me to do. He’ll want me to not pop the cork and if we wants me to do that he will not want me to facilitate putting it back on later on. So, in this case, future reasons don’t transfer backwards but this is an outcome of the theory.
At t2, I have another option set, to put the cork back on or not do so. Here Jussi+ aggregates another set of desires from all the time-slices of Jussi from t2 onwards. Here presumably he wants me to put the cork back on.
And, in the original agony case, even if I don’t desire to avoid future agony, my future slices will desire not to be in agony. Here, when Jussi+ aggregates the time-sliced desires, he will want me to take steps to avoid future agony. So here RTP holds locally but that’s again a motivated outcome of the view.
There is a question of whether the subjectivist should motivate the caring of the future desires based on the current desires. i guess the point in the Smith paper is is that whatever current desires the current self has if she is capable of instrumental reasoning she must have substantial desire to care about the future desires.
Doug,
I think the best responses to that further challenge are in Mark Schroeder’s Slaves of Passions and the Smith paper I’ve going on about. I should emphasise that I’m not subjectivist myself.
Hi David,
You’re correct, although I would tweak things slightly: the cases are ones in which if I don’t do X, I will later get lots of agony. For bad reasons, I do X. So I never feel any agony. What you need, then, is the thought that because refraining from X-ing would have caused me to be in a situation in which I would have had a reason to get out of agony, that gives my present self a reason to do whatever would have facilitated my later getting out of agony had it been the case I was going to be in agony.
But I worry that any such thought would imply that even though I eat lots of fiber and am not the least bit constipated, I now have a reason to buy a very expensive and powerful laxative because my refraining from eating any fiber would have caused me to be in a situation in which I would have had a reason to get out of a state of severe constipation, for, on your thought, this gives my present self a reason to do whatever would have facilitated my later getting out of a state of constipation if it had been the case that I was going to be constipated.
Do you think that I should go out and buy expensive and powerful laxatives even though I don’t and am not going to have any problems with constipation?
(I hope you appreciate the opportunity for making all sorts of jokes.)
[Jussi+ will want Jussi] to not pop the cork and if we wants me to do that he will not want me to facilitate putting it back on later on.
I don’t see this. Jussi+ wants you not to pop the cork. But how does it follow that he won’t want you to facilitate putting it back on later? He knows that you will pop the cork. So (I think) he wants you to put it back on later. So he wants you facilitate putting it back on later.
Are you sticking with Matt’s example, or have you changed the example?
[I propose that we ignore Doug’s most recent example, by the way.]
Jamie,
I’m sorry if my latest example hits too close to home.
Matt,
I am not sure I need to put much weight on the because. It would be fine, I think, for my purposes if we just talked about cases where necessarily if I later have a reason to get X, I have an earlier reason to facilitate getting X. But I don’t see that the because is doing any harm. I would have thought Parfit should say that it is because I have a reason to get X that I have a reason to take the means to X, for example. I don’t see anything odd in saying that because I have a reason to fly someplace tomorrow I have a reason today to buy a ticket. Maybe you can help me more understand the worry.
Hi Jamie,
I’m sticking to Matt’s example. Say the theory is that my reasons at t1 are reduced to whichever plan of action Jussi+ wants me to follow most at t1 on the basis of aggregating the desires of my present and future time-slices. There are four plans available for me:
1. Not pop the cork and not facilitate putting it back.
2. Not pop the cork and facilitate putting it back.
3. Pop the cork and put it back at t2.
4. pop the cork and not put it back at t2.
Seems to me that plan 1 satisfies most of the present and future time-slices. 2 is close but requires extra effort for no gain. So, as the view is defined, Jussi+’s knowledge of my actual future actions is irrelevant for what plan he wants me to follow. Only his knowledge of what plans I could follow counts.
Of course, after t1 when I have popped the cork things look different. The plans available to me are updated at that point and so which plan Jussi+ wants me to then follow is different. So, at that point, knowledge about what I have done becomes relevant.
Now, you might think that this view is ad hoc just to get around the problem. I would want to motivate it by the Kantian thought that our analyses of normative concepts must reflect the fact that we use them under the assumption of freedom.
Jussi, I agree, obviously, that plan 1 is the one Jussi+ wants you to follow. Even so, he wants you to replace the plug. He wants you to follow this plan: replace the plug if it is out. And he knows you are going to pull it out. So, I think, he wants you to (follow the plan in which you) replace the plug.
Try this example. Suppose Doug is going to pull out the cork, and this is well-known. There is some day-glo paint he can right now daub on the cork, so he’ll be able to find it in the dim Arizona light after he pulls it out. If he doesn’t brush a little paint on the cork now, there’s a good chance he will not be able to find it when the time comes.
I hope the good version of the subjectivist view says that Doug has a reason to put some paint on the cork, even though that is no part of Doug+’s favorite plan.
Hi Jamie,
wMatt and Doug gave the challenge of formulating a non ad-hoc theory according to which the subject has *no reason whatsoever* to facilitate putting the cork back on before or at the time of whether or not the cork is popped. The theory that I gave that reduced reasons to which plan the ideal version of me wants to most follow does this – as you seem to accept.
You seem to give a conflicting desideratum which is that the correct subjectivist theory has to generate a reason to facilitate putting the cork back in (this is what Matt and Doug denied). Fortunately, you show yourself how we can generate a subjectivist theory that provides this reason. We just add to the information on the basis of which the ideal version of the agent wants the real agent to act. Or, maybe, that it is settled that I am going to pull the plug removes 1 from the plan-options which the advisor compares.
The question then is whether Matt and Doug or you are right about the reasons in the situation. But, however that debate goes, there seems to be a plausible subjectivist theory to understand the reasons in the situation (maybe this is a version of the consequentialising debate).
For what’s worth, maybe both Matt and Doug and you are right about what reasons the agent has. Maybe – as Sobel wrote earlier – we use the ‘term’ reasons in a context-sensitive way so that the context of utterance determines a set of information relative to which the reasons-claims are true. Maybe Matt and Doug were in a context where the information about the actual future actions is not relevant, whereas you are in a context in which it is. On the subjectivist view, this would be reflected by the idea that the context determines on the basis of what information the ideal advisor wants me to act or what plan-options he compares.
So, I’m not yet seeing a case for which there would not be a non ad-hoc subjectivist account.
No, I don’t disagree with what Matt has said. I have (very strong) intuitions that there is no reason to pull the plug out in Matt’s example, but there is a reason to put paint on the plug in my example. So I think we want the theory to deliver different verdicts in the two cases, despite their similarity (each includes the information that the agent is going to do something he ought not to do, each is about another act that facilitates replacing the plug later).
I do not agree that the view you gave says there is no reason to facilitate replacing the plug, because I think the contingency plan, replace the plug if it has been removed, is part of the plan that Jussi+ most wants you to follow. I gave the paint example to point out that it isn’t going to be easy to fix this problem, since we want to get the result that you do not have a reason to pull out the plug, but Doug does have a reason to paint it.
Ummmmm….
Oh, got it.
I don’t think the current problem is specific to subjectivist views. As far as I can tell, Parfit has the same problem. This is a different issue.
Hi Jamie,
ahh, I’m very sorry. I see that I got confused with the details of the examples. You are right that in Matt’s case the act of facilitating is the pulling the plug itself whereas in your case this is not true – the relevant act is a separate act of painting the plug. Can we just combine the cases in to one? If so, we want:
1. No reason to pull the plug out in order to facilitate putting it back later.
2. A reason to paint the plug so that we can locate it later and a reason to replace the plug later.
And so, we need the advisor not to want me to pull the plug (so as to get 1) and yet want me to paint the plug bright and put it back (so as to get 2).
It seems to me that contextualism can do the work here. Claim 1 is made with the presupposition that both pulling the plug and not doing so are options. So, its truth depends on what the advisor wants me to do in the contingency in which it is not settled what I will do, and here he wants me not to pull the plug. Claim 2 is made with the presupposition (or even knowledge) that I will pull the plug in any case. So, its truth depends on what the advisor wants me to do in the contingency in which it is settled that I will pull the plug (and here he wants me to paint the plug and put it back on).
Everyone,
I thought that it might be useful to make clear that there seems to be three separate issues currently under discussion.
Issue 1: Is RTP false? Matt’s example concerning the cork (and my much more convoluted example from the original survey) suggest that it may in fact be false.
Issue 2: Can RTP be employed by the subjectivist to account for all cases in which an agent has a reason to act so as to avoid future agony. My example entitled “The Pill” suggests that it can’t. Sobel has suggested that this is merely a technical problem for which there is likely to be some technical fix. And I’ve been questioning whether there is likely to be some technical fix given that the thought that Sobel needs for the fix has counter-intuitive implications.
Issue 3: Regardless of whether Sobel’s RTP response works, can the subjectivist successfully respond to various challenges (like Parfit’s Agony Argument). Jussi has been arguing that the subjectivist can — in part by appealing to the work of others such as Michael Smith.
Contextualism sounds a little fishy to me, Jussi. Do you mean that one person could say, “Jussi has no reason to pull the plug in his actual circumstance,” and another person could say, “Jussi has a reason to pull the plug in his actual circumstance,” and both speak truly?
I think the examples do threaten RTP, Doug. My point was really that at least some features of RTP are not specific to subjectivism — a point that Dave agrees with, of course.
Hi,
contextualism is always a bit fishy but given that we can put all of this in terms of oughts too I was thinking along the lines of Wedgwood’s contextualism which seems fairly plausible to me.
I would assume that ‘Jussi has no reason to pull the plug in his actual circumstances’ comes out false in both contexts given that it’s never the case that the advisor wants me to do pull the plug. I was more thinking that ‘Jussi has reasons to paint the plug bright’ and ‘Jussi has a reason to put the plug back in later’ come out true in some contexts (that presuppose that Jussi is pulling the plug out) and false in others (in which this isn’t presupposed).
I agree that RTP is threated but I wonder whether Sobel was suggesting as an exceptionless principle in the first place (and I think he said no). Also, I agree that the cases are a problem for anyone who accepts RTP whatever metaethical view of reasons they held.
I should say that as a result of this discussion and previous chats with Doug, I now see that the Reasons Transfer Principle as I envisioned it in the paper is quite suspect and likely false. Matt’s examples seem to me telling. I think I intended the RTP in the paper to be a completely general thesis. I would no longer make such a claim. I continue to suspect that something like it will adequately serve to show that Parfit’s Agony Argument does not show that subjectivism is false. But Doug has shown me that I do not know yet how to formulate a version of the view that would accomplish this. I see that as work that remains to be done. I don’t feel Doug’s worries in my bones, but perhaps that is just because I have not yet wrestled with the issues enough. I am up to my neck in new issues so I will have to come back later to these issues.
I have started to vaguely worry about what a subjectivist should say about whether there is a tighter connection between the (procedurally correct) concerns I have at t and my reasons at t than there is between my concerns at t+1 and t. I suspect the subjectivist can vindicate some extra tightness of connection here, perhaps by claiming that it is the satisfaction of now-for-now desires that give ultimate reasons, and perhaps stressing the uncertainty about what one will still want when t+1 arrives. But I suspect that in cases of radical ideology changes, our intuitions are that I have little or no current reason to further my future self’s goals which I currently find quite alien. In a creature that switched back and forth between ideologies, such a thought would, I think, be sure to be directly collectively self-defeating between the different ideological perspectives I will have and from the point of view of the goals of the different time slices. This has the feel to me of a situation where something big must give. If folks know of good stuff to read on such issues, I am all ears.
I think the real worry for RTF is that reasons may be time centered. The worry is more serious if the RTF is supposed to transfer the full strength of reasons (or perhaps transfer strengths proportionally). If it only says that the reason is transferred, allowing that it may lose all but some tiny amount of its strength in the transfer, then the principle may be harmless.
I personally think it is very plausible that there are time-centered reasons. For instance, it is pretty plausible that reasons are future biased, in that at any given time t we have reason to prefer that a certain bout of pain be over at t. In Parfit’s Operation example, we might on Wednesday have reason to prefer that our own operation is the Tuesday one, but no reason on Monday to prefer that our own operation is the Tuesday one. (Tom Dougherty’s Operation 2, discussed here eight months ago, with the full example spelled out in Tom’s Ethics paper, might make real trouble for this view, so maybe I should not assert so confidently that it is plausible.) I think there are other, even more compelling examples, but they are a bit complicated to present and this is already a long comment.
If only the satisfaction of now for now desires have authority we may be able to generate something that looks like future bias on the cheap.