Fred Feldman (Pleasure and the Good Life) and Chris
Heathwood (“The Problem of Defective Desires”) point out the following paradox
for desire satisfaction theory, which seems to have been first suggested by Richard
Kraut. People sometimes desire to be
badly off. Desire satisfactionists say
that A’s desire to be badly off is satisfied iff A’s desires are on the whole
not satisfied. This leads to paradox, at least in certain cases. If having a desire satisfied is good for
you, then satisfying the desire to be badly off makes you better off; and in
some cases, the result will be that you are not badly off; which means that the
desire is not satisfied after all, so you are badly off. Paradox. (For a clearer formulation of the paradox, read Chris’ paper.)
Chris suggests in his paper that the paradox for desire satisfactionism
flows merely from a paradox about desire. Paradox arises just from the desire to have one’s desires frustrated –
whether or not desire satisfactionism is the right theory of welfare. If you
have a favorite way to resolve those paradoxes, the desire satisfactionist can
just employ your solution and save his theory of welfare.
I’m not sure Chris is right about this. Consider the sometimes paradoxical desire to
have one’s desires be, on the whole, mostly frustrated. Suppose that the solution to this paradox is
to say that this is an impossible desire. If desire satisfaction were true, this would entail that it is also
impossible to desire one’s life to go badly on the whole. (Or at least that desire could never be satisfied.) But that desire is not paradoxical, and it seems like it could be satisfied. So there’s a cost to desire satisfactionism
here – it entails that certain desires are paradoxical or unsatisfyable, while other theories of
welfare would not. (I realize this is a
bit sketchy, but this is after all just a blog post. Feel free to hammer away with de re/de dicto distinctions and
such in the comments.)
I think this sort of paradox creates problems for other
theories of welfare too, including Feldman’s “Truth-Adjusted Intrinsic
Attitudinal Hedonism.” (Feldman doesn’t seem to endorse TAIAH in the end, but he seems sort of sympathetic to it.) Very roughly,
the idea behind TAIAH is that pleasures are more valuable when taken in
propositions that are true. (Better to
be pleased that others like you when they actually do like you than when they
hate you, etc.) Just to make the view
sufficiently precise, let’s suppose that pleasures taken in truths (“true
pleasures”) are twice as valuable as similar pleasures taken in falsehoods (“false
pleasures”). Here’s a variant on one of
Chris’ examples. Suppose that A’s life
contains more pain than pleasure, so that his life has intrinsic value of –15 (pending
what else happens). A then takes
pleasure to degree 10 in the fact that he’s had a bad life. Call this pleasure P. Is P a true pleasure or a false pleasure? If it’s a true pleasure, then its intrinsic value
is +20, which means A’s life has intrinsic value of +5, which means P is not a
true pleasure after all. If P is a
false pleasure, then P has intrinsic value of +10, which means A’s life has
intrinsic value of –5, which means P was in fact a true pleasure. Paradox.
I think the desire-satisfaction account of well-being is not as clearly implicated in the worry here as people seem to be supposing. One central reason is that the most influential presentations of a desire-based account have generally held that 1) it is informed desires that are the relevant desires and that, in constructing a theory of well-being, 2) not all of an agent’s informed desires determine an agent’s well-being. I think Mill, Sidgwick, Railton, Griffin, and Parfit, among others, as noting that this is the best way to develop the view.
Now it is not always clear which (informed) desires are held to not be part of the subset that determines an agent’s well-being. Usually people exclude “moral” concerns, but I doubt that this move will be sufficient all by itself. Consider the person who cares that his group do well (when that group has no special moral standing) beyond the extent to which the group doing well benefits the agent. I have written on these issues so if someone wants to be directed to that, just let me know.
The general point is that it has always been a bit of a scandal that desire-satisfaction accounts have not adequately argued for a particular subset of the agent’s concerns as constituting an agent’s well-being. But they have noted that there is work to do here. So I see the general problems suggested in the post as not obviously giving the desire-theorist work to do (or problems) that they did not already acknowledge.
Now it would be a different matter if one could show that no principled distinction could be drawn by the subjectivist in a way so as to exclude the sort of concerns mentioned in the initial post. But until a case is made that this is so, I think the desire theorist should expect that if they can complete the work they announced they need to complete, then this problem will take care of itself.
Further, and likely less importantly, it would need to be shown that the sort of self-destructive desires mentioned can survive rational scrutiny.
Hi Ben,
I think this is a cool and underexplored topic. I agree that truth-adjusted hedonism faces the same paradox. (Incidentally, though I’ve read only the abstract, Matthew Cashen’s paper at last week’s APA might be defending the idea that true pleasure or happiness is better than false happiness. So this isn’t just a strawman.)
One interesting question this topic raises is, Are theories of welfare supposed to be analytic (as opposed to synthetic)? This is, of course, a big and deep issue. But if such theories are analytic, then I think the “piggy back” response I initially suggested will work. For if the true theory of welfare is analytically true, and the desire theory under discussion is true, then the desire to have one’s life go badly on the whole just is the desire to have one’s desires be, on the whole, mostly frustrated. However the paradox is solved for the latter desire, so too will the paradox be solved for the former desire, for they are one and the same desire.
What do you think, Is the true theory of welfare analytic?
Hi David,
You wrote:
I don’t think your first suggestion — moving to an informed desire account — helps avoid this paradox. Just consider a possible world with a person who is in fact fully informed and who in fact desires that his life go badly. For such people, the actual desire theory and the full-information desire theory are extensionally equivalent.
So I think only your second suggestion — ignoring some desires — may help. But, of course, and as you suggest, it is impossible to evaluate this proposal until the relevant subset of one’s desires is specified. I will note that one common specification — that we count only one’s self-regarding desires, or only one’s desires about one’s own life (Overvold, Parfit) — will obviously not work here, since the desire that one’s life go badly is surely self-regarding and about one’s own life.
Hi Chris,
Congrats on the new job! I hope you are enjoying your last few months in Northampton.
I don’t see how granting that the correct theory of welfare is analytic helps you. The point is that while it isn’t counterintuitive to say that the desire to have one’s own desires frustrated is impossible, it is counterintuitive to say that the desire to have one’s life go badly is impossible. So if the best solution to the paradox of desire leads us to say that such desires are impossible, this would be evidence that desire-satisfactionism is analytically false. That is, the grounds that we have for thinking that desire-satisfactionism is analytically true may very well depend upon whether such an account allows us to make intelligible the desire to have one’s life go badly.
Chris,
Yea, I myself would not want to put weight on the “informed” part of my reply, but I think others might. But of course, some cases in which a person has self-harming desires will involve a lack of full-info or vivid consideration, etc. So this will take care of some cases, but I too agree it will not take care of all cases.
One issue would be why the agent wants themselves to have less well-being. If they want this because they think they do not deserve to be happy for anything like moral reasons, then it would seem the standard “exclude moral desires” reply would work fine. So perhaps we will have to distinguish different cases where a person has self-thwarting desires and take the various cases one by one.
Chris – I would be hoping for a non-analytic theory of welfare. But if that’s impossible, then I say what Nishi says. If the choice is between (i) accepting desire satisfactionism and saying that you can’t desire that your life go badly, or (ii) rejecting desire satisfactionism, I reject desire satisfactionism.
I was at Cashen’s APA session, and I think Cashen does think what you suggest, though that’s not what he says in his abstract. In his abstract he says he’s going to show that false happiness (being happy about a falsehood) isn’t really happiness. He retracted that claim during the comments.
Oops, I retract that claim about Cashen. He doesn’t say that in his abstract. He did say in his session that you can genuinely be happy even if the object of your happiness is false, but he thinks that sort of happiness isn’t as good for you as “true” happiness.
Hey Nishi,
Thanks! Since we’re leaving, I’m feeling compelled to do all the sort of “touristy” things one can do around here.
On the philosophy: point taken. However, I don’t think the line of argument you suggest is necessarily decisive (and maybe you didn’t mean it to be). I agree that if the desire theory implies that it is impossible to desire that one’s life go badly, then that is a cost of the theory. For it sure seems like this desire is possible. But if we have lots of indepenedent reason to accept the desire theory, and we have independent reasons to think the true theory of welfare is analytic, then it may be that the correct conclusion to draw from the paradox is not that the theory is false, but that we have learned something surprising from the theory. (I suppose this may be a banal and uninteresting methodological point.)
Anyway, to make this surprising implication of the theory more palatable, we might point out that, if theories of welfare (and normative ethical theories generally) are analytic, they are certainly not obvious analyticities. We should not expect the theory to strike us as obvious analyticities (like “all bachelors are unmarried”) strike us. And we should expect the theory to entail some surprising analytic connections. One such connection is that the proposition that one’s life goes badly on the whole just is the proposition that one’s desires are, on the whole, mostly frustrated. Who knew?!
I think my point is that if theories of welfare are analytic, then, for any theory of welfare, we could find pairs of sentences that intuitively seem to mean different things, but that, according to the theory, mean the same thing. So perhaps we should approach arguments against a theory that exploit this counterintuitiveness with some caution. (All that said, I concede that your argument continues to have force. Maybe I have a hard time accepting the idea that normative theories are analytic.)
Ben says:
To be honest, I think I would too. But I also sort of want to be a naturalist. And I believe that, probably, if naturalism is true, then analytic naturalism is true. So if I deny analyticity, I think I deny naturalism. And although I do have a soft spot for non-naturalism, I don’t know if ultimately I can buy it.
Anyway, enough autobiography. I need to think more about the original point: If theories of welfare are not analytic, then what to say about the original paradox?
Interesting stuff! I was discussing almost the exact same issue on my own blog not so long ago. A few highlights that seem relevant to the present discussion:
1) The impact of the paradox can be softened somewhat by appeal to contextual impossibility. A desire that most your desires be thwarted is perhaps only impossible in certain contexts (viz., those where a contradiction would otherwise result). The motivation for this view is the behaviour of analogous liar sentences such as “most statements on this page are false”. There are clearly contexts where this statement is meaningful and unproblematic (and even true). The same goes for ‘defective’ desires.
2) The paradoxes could have implications not only for welfare, but also morality (if consequentialist and grounded in welfare-aggregation).
3) I argue here that we could avoid the paradox by excising self-reference from the relevant desires. So even if one cannot desire that of all their desires, most are thwarted, at least “I desire that most of my other desires be thwarted” is not paradoxical. I further argue that the latter desire is so close to being about wellbeing (according to the desire-satisfaction theory) that it can fill in and do the required job. Whenever someone seemingly desires the be worse off, they really desire to be schworse off, where ‘schworse’ is defined by excluding the present desire from considerations of wellbeing.
(Actually, if you follow the link, my argumentative strategy in (3) was originally aimed at the ‘moral’ rather than ‘welfare’ paradox, but as demonstrated above it isn’t hard to adapt it to the present case.)
If either (1) or (3) works, then we can avoid the worst consequences of the paradox. If both work (so we only need to appeal to (3) in those rare paradoxical contexts), then it seems there is little problem at all for the desire-satisfaction theorist.
Richard – thanks for the link. About (1), you might be right, and the comparison with the analogous liar sentence is helpful. But it’s still a problem for the desire satisfactionist, because the desire to have one’s life go badly does not seem paradoxical in those circumstances.
(3) is an interesting idea, but it still seems like a cost to desire satisfactionism that other views don’t have to pay. Maybe it’s worth paying in the end.
“But it’s still a problem for the desire satisfactionist, because the desire to have one’s life go badly does not seem paradoxical in those circumstances.”
But is that actually saying anything more than “the desire to have one’s life go badly does not seem to be a desire that most of one’s desires be thwarted”? If we’re all already in agreement that no theory of welfare is obviously true, then I’m not sure that this objection really adds anything. Any substantive theory of welfare will make some non-obvious claims, and this just happens to be one of them for the desire-satisfaction theory. (I guess this is a similar point to what Chris was making above.)
I wonder if the desire satisfaction theorists could make some use of the distinction between first-order and higher order desires. They could try to say something like that only satisfaction of first order desires can make one better of. They could add to this that satisfaction of first order desires supported by higher order desires is more important from the point of view of well-being than the satisfaction of mere unsupported first order desires. Yet, they could deny that the satisfaction of higher order desires in itself improves well-being at all.
With this account, the paradox would not rise I believe. The desire to not get one’s desires satisfied seems for me to be the kind of higher order desire that would get excluded. And, in this way the desire satisfaction theorists could avoid another problem too. If the satisfaction of higher order desires counts towards well-being, one could get an infinite amount of well-being if one had just one first-order desire supported by an infinite chain of higher order desires satisfied. I know that the limitation above would need some kind of an argument, but the avoidance of these two problems seems like a start.
I realized something needs to be added to my suggestion above. One would need to say that higher order desires are any such desires in the content of which other desires of the agent feature. Only then would such desires include both the kind of desires for desire satisfaction or dissatisfaction discussed in the paradox and the traditional desires for desires in the Frankfurtian sense.
I think that it would also be an advantage of excluding the satisfaction of the latter kind of higher order desires from having an effect on the agent’s well-being. If this was not the case, and one knew about it, one could make one’s own life better by merely adopting suitable meshing desires. I could induce a desire to desire a random object and then a desire for that object, and this would suffice to make my life better because the higher order desire would get satisfied even if the first order would not. This sounds implausible. So, perhaps the desire satisfaction theorists of well-being would have grounds for excluding all higher order desires in the content of which other desires feature.
Jussi,
I am not sure that excluding all higher-order desires is necessarily the right way to go. Consider this example: Bob is a lazy slob who desires to watch TV all day. However, Bob wishes that he could be a more diligent person who desired to get some work done. It seems arbitrary to me to suggest that if Bob were to gain these new desires, his welfare would not be improved. Indeed, satisfying this desire would seem to be more important than the satisfaction of some first-order desires, because this second-order desire is a desire about the kind of person he wants to be.
This leaves preferentism open to the objection you mention at the end of your post. I suppose that a preferentist will have to respond to this objection in some way, so we can ask Chris what he has to say about it.
Scott,
that’s a nice example. I had a hunch that my suggestion might face something like this. However, I do not actually think that it is much of bullet to bite compared to the original paradox and the other problems. I would in fact say that Bob’s life is not yet improved by the appearance of the first order desires. Consider the case in which Bob, as a lazy slob he is, finds that all the new first order desires are just too weak to move him anywhere from the couch. There is a pull, which just lacks the ummpph. If this is the case, we can ask whether Bob’s life in the end was improved at all with these new desires. I might want to say that no – it wasn’t at all. This would fit the idea that satisfaction of higher order desires does not improve well-being.
Yet, I can agree that the appearance of them is important for Bob. They can be seen as an enabling condition for his future potential well-being. Once they come to the scene of his mental economy, Bob has first-order desires to be satisfied – and thus it is possible that through this his lazy life improves.
I don’t think the paradox is so bad. If the desire-satisfaction theory of welfare is true, then a desire that one’s life go badly is unsatisfiable. That doesn’t mean you can’t have the desire. A parallel would be the desire to drink H2O without drinking water (if welfare-theory is synthetic) or the desire to trisect an angle with ruler and compass (if welfare-theory is analytic but very hard).
Heath – Suppose that I am miserable, for the vast majority of my desires have been thwarted. The one speck of light in all this gloom is that I have a second-order desire that my life be miserable in just this way. In this case, that desire is fulfilled, so it cannot be impossible as you suggest. (It’s only paradoxical/impossible in certain contexts.) In this case, my life does go badly on balance, despite having this *one* fulfilled desire.
Jussi – You defined second order desires as “higher order desires are any such desires in the content of which other desires of the agent feature.”
An interesting consequence of this is that it leaves us open to interpersonal desire paradoxes. Suppose Amy desires that Bob’s life go badly on balance; but Bob desires that Amy’s life go well on balance. Each of them is such that, apart from the desires mentioned above, they are exactly on the borderline between being well off and poorly off. This should give rise to a paradox as before: each is well-off iff they are not.
We could avoid this by removing the italicized section of your definition. But anyway, as Scott pointed out, we often consider second-order desires to be relevant to wellbeing, so…
Richard,
Sorry, this is perhaps a failure in my understanding, but I totally fail to understand the interpersonal desire paradox. I would like to hear more about it though. You say that both Amy and Bob are first on the borderline of being well-off and badly-off. Thus, both of their desires that are directed towards the well-being of others are frustrated. On the balance, this makes their lives a little bit (one desire’s dissatisfaction worth of) worse. This then satisfies Amy’s desire for Bob’s negative well-being. She moves back to the original borderline with one desire re-satisfied. And, Bob stays below the point, because Ann still fails to be well-off on the balance like he hopes. And, there they stay. So, I really do not get what the paradox is supposed to be or what ‘each is well-off iff they are not’. I take it that the paradox is created by the self-referentiality of the relevant higher-order desires and this is lacking in the two person case.
And, I no-where denied that second-order desires are relevant for wellbeing. I accepted that in some cases their satisfaction is a necessary enabling condition for the future well-being created by the satisfaction of the first-order desires. What we need is an example where it is clear that *mere satisfaction of them* creates plausibly new well-being, where this well-being is not brought about with the satisfaction the first-order desires that the second-order desires create.
Perhaps the paradox might be solved by adopting a form of desire-satisfaction that includes only first order desires as counting toward a person’s welfare. On such a view, the desire that one be badly off is the desire that one’s first-order desires not be satisfied, on the whole. But that desire itself is a second order desire. So, no paradox.
Jussi, I was thinking that when Amy’s nasty desire was satisfied, that would push her above the borderline, since she is on the borderline when that desire is entirely excluded from consideration. (If a thwarted desire makes one slightly worse off than having no desire at all, then changing a thwarted desire into a fulfilled one should add *two* increments of wellbeing.)
But then, since Amy is now well-off, it turns out Bob’s desire is satisfied rather than thwarted, so he is now well-off too. So Amy is poorly off. So Bob is. And so on and so forth.
You see, in the two person case we get an indirect self-referentiality. Amy’s desire refers to those of Bob’s, one of which in turn refers back to Amy’s desires (including the original one).
It’s really just analogous to a two-person liar paradox, e.g. where Amy says “Everything Bob says is false”, and Bob replies “Everything Amy says is true”! (I admit the paradox is much easier to see when statements, rather than desires, are involved!)
Fair point in your final paragraph though. Perhaps excluding second-order desires from direct calculations of wellbeing (an idea I see Campbell has also taken up) wouldn’t be so bad after all.
But then, I think the really interesting issue here is not about ‘wellbeing’, but the foundation of the paradox in a desire that most of one’s desires be thwarted. Is it impossible to have such a desire? Or merely contextually impossible, as I suggested earlier? Are there any other ways to deal with it?
Correction: in the 2-person liar paradox, Bob should instead reply, “What Amy *just said* is true”. (Oops.)
And Amy should have said “What Bob *next* says is false”. (Man, I really screwed that one up! Hopefully the general idea I was getting at is clear enough, anyway.)
Richard,
thanks a lot. Now I think I see what you mean. I am a bit slow sometimes – sorry. I guess it now turns onto how big of a difference satisfying and frustrating one’s desires makes to one’s well-being and the time sequence. But, you can be right.
I had a reason for adding the proviso that the content of the excluded higher-order desires need to include desires of the agent herself. This was that I thought (ok, this is a bit rude) that my life would be improved if my desire that some persons desired me was satisfied by their first-order desires. But, if there is a interpersonal paradox too, perhaps one might have to say that it is not the satisfaction of this desire that makes my life better, but the consequences of those desires (blushing…).
Campbell,
yes, that’s the suggestion we are discussing at the moment.
Anyway, I have tried here to defend the desire satisfaction view by trying to make it safe from the paradox. Now, I want to put my cards down for what it’s worth if anyone’s interested. My own main worry with the view comes from a kind of Aristotelian direction. I would be interested in hearing what the defenders of the view have got to say about these matters. What I have in mind is that the view seems too formal to be a view about human wellbeing. At times our desires can be pointed towards strange things – turning on radios for no further reason, eating mud, moving a block of wood around on the floor, etc. etc. Now, it seems strange to me that satisfying such desires somehow is relevant at all for one’s wellbeing. For this reason, I might want to say that satisfying some desires makes one’s life better, but this is because of the things that are in the content of those desires and their reason-providingness and role as goods in human life. Just a thought.
Ooops. It appears that in my haste to post my clever new idea I neglected to read through the earlier comments and notice that the idea was already under discussion. That’s a little embarrassing. But perhaps I can redeem myself by offering something genuinely new.
Consider the following two principles, where Dp means “the agent desires that p”:
(P1) (Dp & (p -> q)) -> Dq
(P2) Dp -> ~D~p
P1 says that the agent’s desires are closed under logical implication, so the agent desires everything that is implied by something she desires. And P2 says that the agent cannot desire both that p and that not-p. These principles seem fairly plausible. But together they imply that the agent cannot both (i) have a certain desire, and (ii) desire that that desire not be satisfied. For suppose that the agent desires that p, so we have Dp, and that she desires that this desire not be satisfied. It’s a necessary condition of that desire’s not being satisfied that not-p. Hence, by P1 we have D~p. But that gives us Dp & D~p, violating P2.
This tends to suggest, I think, that the kind of second-order desire needed to make sense of the agent’s desiring to be badly off, on a desire-satisfactionist approach, is impossible.
Campbell,
I don’t think either of your principles is true.
(P1): I gather from your gloss that the internal “->” is logical entailment. Surely the second one is something weaker — material implication, say. I’ll assume so. But even so, it seems completely implausible. Nobody desires that 746 is greater than 691, even though it’s entailed by everything everyone desires. And if Bertrand desires that arithmetic be finitely axiomatizable, it does not follow that he desires every proposition, even though every proposition follows from what he desires.
(P2): I often desire that I finish writing a lecture and at the same time desire that I not finish writing the lecture. That’s called having conflicting desires, and it’s commonplace. It is an interesting and telling difference between belief and desire that there is something irrational, or inconsistent, in believing two contradictory things but nothing irrational in desiring both. (There’s a recent discussion on Brian Weatherson’s blog about whether hope goes with belief or with desire in this respect; I think it plainly goes with belief, myself.)
Hey, is the clock on this blog set to Greenwich Mean Time?
Jamie,
1. I think you’re right to reject P1. But perhaps there’s a more plausible principle in the ballpark — for example, (Dp & B(p->q))->Dq, where B is for belief. That would avoid the Bertrand case at least. More to the point, I think that something as weak as D(p & q) -> (Dp & Dq) would be sufficient for the argument I suggested.
2. Even if P2 doesn’t hold for all desires, as you claim, I wonder if there’s a species of desire for which P2 does hold — i.e. “all things considered desire” or “overall desire” or some such thing.
3. The analogy with belief is helpful. We can all agree, I hope, that there’s something very wrong with saying “I believe that p, but I believe that that belief is false.” My thought was there’s something similarly gone awry in “I desire that p, but I desire that that desire is not satisfied.” I’d be interested to hear people’s intuitions on this. (I should say that I’m not worried about sentences such as “I desire that p, but I desire that I don’t have that desire”, which seem to be a different case.)
(Warning, this post contains shameless self-promotion.)
Jussi,
You wrote:
I discuss this kind of case in section V.2 (starts on p. 20) of my paper “The Problem of Defective Desires”. (You might have to read some of what comes earlier to understand some of the terminology and some of the references.) In a nutshell, I say that satisfying such desires is relevant to welfare, but it may come at the expense of diminishing the excellence of a life (since these activities lack excellence). Excellence, I say, is a scale of evaluation of lives that we care about, but one different from the welfare scale.
(Self-promotion over, for now.)
Another way to go that doesn’t totally abandon the desire approach is a hybrid account (suggested by Parfit and others): what’s good for a person is satisfying desires for things on the “objective list.” You’d have to specify what’s required for entry on the objective list, of course, but surely eating mud wouldn’t be on it. Although appealing to such a list does constitute a serious departure from the anti-paternalistic spirit of the desire theory, desires are still relevant on this view.
Dear Chris,
no need warn about self-promotion. Your take on the matter sounds interesting and I am more than happy to check it out.
Yes, I have always found the hybrid views most intuitive.
You are right that at least pure objective list theories face the problem with paternalism. Agent may be able to ‘tick all the boxes’ of the good life ingredients list and still feel strongly that her life is not going well at all. It seems offending to say at this point that she must be mistaken, that actually her life is going well. There seems to be at least some first person authority to knowledge about goodness of one’s own life (even if not complete). However, do you think that desire satisfaction views fare any better on this account? Doesn’t it seem conceivable that agent’s all (or most and important ones too) desires are satisfied and she is feeling that her life is still going very badly? Must the agent be in this situation necessarily mistaken?
Anyway, here is something I find deeply puzzling in the hybrid views I still like the most. If I remember right, the way Parfit formulates his view is just like you put it – satisfaction of desires improves agent’s life on the condition that the desires are relevantly connected to the items on the objective list. Now, one can formulate a converse view (I am sure I’ve seen this somewhere, but my memory fails me here) according to which the satisfaction of the things on the list makes one’s life better on the condition that these are things which the one desires. One could rig this view to a full objective list view by trying to put some kind of an autonomy item on the list, but I’ve never found it plausible that this solves the paternalism problem. Anyway, my problem is that with these two views, which seem necessarily co-extensive, it is hard to think of a way in which one could argue for one or another. Thus, I am left wondering… Of course, one could go with Lewis and Jackson and say that necessarily coextensive properties must be one, but there seems to be a significant, substantial difference between the views.
Sorry everyone, I’ve gone more and more off-topic into my personal ramblings. I’ll try stop here, and perhaps put some thoughts on paper about all of this if I have the time. Big thanks for everyone too about this discussion.
Jussi,
For my own part, I don’t think we should worry too much about ensuring that people cannot (or even cannot easily) be mistaken about how well their lives are going. And I don’t think failing to ensure this is offensively paternalistic. Not even an unrestricted actual desire theory ensures this, and surely such theories don’t count as very paternalistic.
While we are on the topic, could someone explain to me what the “paternalism objection” is supposed to be? I assume it’s not this:
Necessarily, no one is in a better epistemic position than myself with respect to claims about how my life is going… therefore, theory x, which implies that someone can be in a better epistemic position than myself with respect to claims about how my life is going, is false.
(Because desire-satisfactionism is a view that obviously implies that someone can be in a better epistemic position than myself with respect to claims about how my life is going. This is Chris’s point, right?)
And it’s not this:
Necessarily, nothing can be good for someone unless that person wants that thing, so theory X, which implies that something can be good for something even if that person doesn’t want that thing, is false.
(Because on the desire-satisfactionist view, what’s good for you is the satisfaction of your desires, but, as I think Chris has also pointed out, you may not want these desires to be satisfied– and you may even regret having the desires themselves. So the desire-satisfactionist should also reject the claim that necessarily, nothing can be good for someone unless that person wants it.)
I assume it’s not this either:
Necessarily, if something will make my life go better, then it is up to me whether that thing will make my life go better. So theory X, which implies that things can be good for me even if it is not up to me whether that thing will make my life go better, is false.
(This is a little unclear. The idea is this: there are some propositions such that it is up me whether they are true or false. For example, the proposition that I am writing this email right now is such that it is up to me whether it is true or false. According to this version of the paternalism objection, propositions about which things are good for me have to be up to me.
This is also a silly objection, and if it works, it refutes desire-satisfactionism as well, for it’s not up to me whether desire-satisfactionism is the correct theory.)
Sorry this long-winded, possibly confused, and off-topic. Can someone help me?
Kris,
I think that you have raised an interesting question. I am not sure if I can answer it in a way that will satisfy you, but here is my shot.
Non-paternalistic theories claim that whether some state of affairs is good for a person ultimately depends on some attitude or other that the person takes towards that thing, while paternalistic theories deny this dependence claim.
For example, consider the unrestricted desire theory. According to this theory, a state of affairs obtaining is good for me only if I take the appropriate attitude towards it (namely, desiring that the state of affairs obtains). Now consider the objective list theory. According to it, some things just are good for me, whether I want them or not. This is the paradigmatic paternalistic view, for our attitudes towards a certain state of affairs does not determine the value of that state of affairs for us.
Some theories are harder to classify. Consider Feldman’s version of Intrinsic Attitudinal Hedonism. According to this view (which I will briefly summarize and hopefully won’t mangle too badly: Chris can correct my mistakes if I make too many) pleasures are not sensations but are instead constituted by an attitude we can take to various objects. I can be pleased that I am in my office writing these comments, and this counts as a pleasure. Whether or not it counts as a pleasure depends on the attitude I take towards the object. So, according to this view, whether or not something counts as a pleasure or not depends on my attitude I take towards that thing. This seems to suggest that the theory is, in this regard, non-paternalistic. There is a sense, however, in which it is paternalistic. Consider the following example:
Bob would receive an awful lot of intrinsic attitudinal pleasure if he did action X. However, he would prefer, all things considered, to do action Y instead because he believes that Y would be better for him than X (even though he knows that action Y will result in less overall intrinsic attitudinal pleasure for him than will X). As I understand him, Feldman would say that action X would contribute more to Bob’s welfare than action Y will because it will lead to more intrinsic attitudinal pleasure than will action Y. In other words, as long as desire and intrinsic attitudinal pleasure can come apart, it would seem that someone could always desire something other than pleasure more than he desires pleasure. In these cases, it would seem paternalistic to claim that the pleasure is better for the person than what the person wants for himself.
This problem is even worse for traditional, sensual versions of hedonism. According to these views (which I am not sure anyone really accepts) a pleasure is a simple sensation and is the only thing that directly contributes (positively) to your welfare. In that case, it would seem that a certain course of action would lead to more pleasure than another, but could fail to be the action that you most want to perform. It seems paternalistic to claim that the pleasure just is better for you, regardless of what you want.
Like I said, I am not sure I can answer your question fully, for I am not sure I really understand how these terms are operating. But if I were forced to give an answer, I suppose I would say that the mark of a non-paternalistic theory is that it makes the welfare value of any state of affairs depend on the attitude you take to the thing, whereas a paternalistic theory does not require this connection.
(Some of the things that Sumner says may be relevant to this question, for he seems to object quite strongly to any paternalistic feature of a theory of welfare at all. See, for example, his Welfare, Happiness, and Ethics especially section 2.3, “The Case against Objectivism”.)
Lots of good comments since I last checked in here.
I suggested that desire satisfactionists have this problem: the desire to have a bad life does not seem paradoxical, but the desire to have one’s desires thwarted is (sometimes) paradoxical. Richard responded:
“But is that actually saying anything more than ‘the desire to have one’s life go badly does not seem to be a desire that most of one’s desires be thwarted’? If we’re all already in agreement that no theory of welfare is obviously true, then I’m not sure that this objection really adds anything. Any substantive theory of welfare will make some non-obvious claims, and this just happens to be one of them for the desire-satisfaction theory.”
I think that’s probably the best response to my point. But I want to make a distinction between different ways a theory can “surprise” us. Some theories surprise us by giving answers to questions we didn’t already have answers to. That’s OK. Other theories surprise us by telling us that something we were pretty sure was true is actually false. In such cases, opponents of the theory say there’s a counterexample, and proponents say there’s just a surprising result. I take this paradox to be a counterexample. The claim that the desire to be badly off is paradoxical is not merely non-obvious. It seems wrong.
Kris,
I think the intuitive idea is this. Consider a form of the Objective List Theory according to which the following things are good for a person:
And the following things are bad:
Suppose someone loves rock ‘n roll and hates Bach, loves cheeseburgers and hates caviar, loves basketball and hates golf. And suppose he does what he loves. This theory implies he’s got a terrible life, even though he says it’s going great. It’s as if the theory says, paternalistically, “You don’t know what’s good for you, but I do; you’d be much better off if you did what I said you should do.”
This sounds crazy. Lots of people want a theory that doesn’t have this feature (or has it to a lesser degree somehow). True, it’s possible for a person not to want any of his first-order desires to be satisfied. Desire satisfactionism may seem paternalistic to such a person. But since such cases are so rare and odd, desire satisfactionism seems “less” paternalistic to many people (or to me, at least).
Scott,
FTR, what you say seems right to me: that a theory’s appeal to the subject’s attitude makes it less paternalistic. (People also cite this as the mark between “subjective” and “objective” theories of welfare.) But I do think paternalism is going to have to be a matter of degree. No theory (or no theory anyone has ever defended) will be completely anti-paternalistic. But some theories are more paternalistic than others.
Jamie – I think the clock is set to west coast time.
On the subject of the clock, I think we should now set it to Singapore time, in honor of Kyle. Or maybe New Zealand time. Isn’t it tomorrow already over there? That will make it seem like we’re getting things done earlier.
Hi Scott and Chris,
Ok, this is sort of helpful.
If I may, I’d like to raise another question about paternalism and theories of welfare. The question is this: why should anyone think that the paternalistic objection has any force at all?
Analogues of the paternalist objection to other kinds of normative theories do not seem to me to have any force. Consider for example a “paternalistic” objection to a kind of Rossian pluralism.
According to this view, killing people, breaking promises, and cheating on ethics exams are prima facie wrong, whereas helping old ladies cross the street, feeding the poor, and calling her mom on mother’s day are prima facie obligatory.
Suppose someone loves to kill people and doesn’t like to call his mother, loves to break promises and hates feeding the poor, loves to cheat on ethics exams and hates helping old ladies. And suppose he does what he loves. This theory implies he behaves terribly immorally, even though he may like how he is behaving. It’s as if the theory says, paternalistically, “You don’t know what you should be doing, but I do; you’d be have much better if you did what I said you should do.”
I take it that this sort of objection has no force. If that’s right, then the paternalistic-type of objections don’t work against normative theories.
So why the disanalogy? Why is this objection supposed to work against theories of welfare?
But the paternalist objection is taken by many
Kris,
That’s a good question. But one thing is clear: it’s not at all implausible to suppose that the true theory of moral obligation is paternalistic (since ethical egoism is not at all plausible), but it is implausible to suppose that the true theory of welfare is (very) paternalistic.
Kris,
This is where Sumner might be helpful. I must admit that I have always found his argument to be pretty close to question-begging, but he is at pains to show that it is not. Still, I’m not sure….
Anyway, his point is this. Theories of welfare are supposed to tell us what is good for me from my own perspective. He says that they are essentially subject-relative or perspectival. By this, he does not just mean that they are supposed to describe a value for a particualr person. Rather, they are perspectival in a deeper sense than that.
The main point is that we must be careful to distinguish welfare value from other kinds of value, such as perfectionist value, moral value, or aesthetic value. For example, the life of King Lear would make quite an aesthetically pleasing story, but that does not mean that it was a good life for Lear to live. We must, therefore, find a way to distinguish welfare value from other sorts of value. For Sumner, the perspectival nature of welfare value is what sets it apart from all other values.
Once we recognize this, though, we will see that no objective theory will work. Objective theories do not appeal to the agent’s perspective and so cannot be descripively adequate, according to Sumner. Although I am not sure if the objective/subjective distinction coincides directly with the paternalistic/non-paternalistic distinction, this much is true: objective theories are the paradigmatic objective theory. So if his argument is correct, it follows that any pure objective theory will not be descriptively adequate.
The difference between this and moral value is that moral value is not perspectival. It is (on mnost accounts) universal in some sense.
Like I said, I’m not entirely convinced by Sumner’s argument. It would seem to rule out any theory that had any paternalistic element, and I think Chris is correct to suggest that there is a coninuum rather than a sharp divide between paternalistic and non-paternalistic theories. That would mean that all but the most radically non-paternalistic theory could be correct, but I’m not ready to go there yet.
Ben,
The time that actually shows up on the postings is California time. The time marked on the PREVIEW seems to be GMT, or something like that. It’s four hours later than my (EDT) clock.
I see that I wrote something rather stupid in my last post. I said:
“objective theories are the paradigmatic objective theory”
Of course this is true, but not very exciting or informative. What I meant to say was that objective theories are the paradigmatic paternalistic theories.
Kris,
I suspect what might be puzzling you about the paternalist objection is a kind of category mistake. Often, when the paternalist objection is raised, for example by capabilities theorists like Sen and Nussbaum, they raise it as an objection to a theory of welfare used for *political* purposes.
The rough idea is this. Politics should be in the business of promoting access to certain capabilities necessary for a good life, and not (for a variety of reasons) in the business of promoting the actual exercise of those capabilities, at least not when it comes to non-basic requirements for living. But this isn’t to say that there aren’t other senses in which a person who is “well off” in the sense that is most useful for political purposes could not be “poorly off” in other sense. Both Sen and Nussbaum, I suspect, would readily admit this. So, for example, Rawls’ obsessive counter of blades of grass could be fairly well off by their standards, if he is provided with all kinds of rights and resources (albeit unexercised ones), and yet in some sense not living a life nearly as good as that of someone living with few rights and resources under a repressive regime, if the latter person at least uses what little she has to enjoy a higher quality of life.
But if you are looking for good reasons to worry about paternalism in theories of welfare for non-political purposes (“prudential purposes”?), then Scott Wilson is right to point to Sumner’s argument from the perspectival nature of welfare. I think I’m basically in agreement with Scott about the overly broad implications of that argument, though.
Justin Tiwald