Broome on Interpersonal Comparisons of Wellbeing

Campbell Brown and I, along with some BGSU graduate students, are currently reading through John Broome’s most recent book, Weighing Lives.  In our most recent meeting, we discussed Broome’s latest attempt to solve the problem of interpersonal comparisons of wellbeing (it differs from his previous attempts).  Here’s my rough reconstruction of his argument:

Defn: a person’s “wellbeing” = how good things are for that person.

1.   A person’s life (i.e., how things are for that person) contains “all the features of the world that can affect her wellbeing” (p. 94).

2.   Given 1, for any two people X and Y, if it’s possible that there’s a life such that X could have lived it and also that Y could have lived it, then “a person who lives a particular life has exactly the same wellbeing as any other person would have if she lived that life,” (p. 94), i.e., the goodness of lives is independent of who lives them.

3.   It is (conceptually and metaphysically) possible that, for any two people X and Y, there’s a life such that X could have lived it and also that Y could have lived it (there are a few technical details here I’m glossing that don’t affect the points made below).

4.    Thus, the goodness of lives is independent of who lives them.

5.   If the goodness of lives is independent of who lives them, a person’s individual cardinal scale of wellbeing (which we have available given Broome’s work in previous chapters) turns out to be a universal scale, “measuring the goodness of lives for everyone” (p. 95), i.e., everyone’s scale is the same and so the universal scale is cardinal.

6.    Thus, the universal scale is cardinal (i.e., it provides interpersonal comparisons of wellbeing).

This is a fascinating and clever argument.  The heavy lifting in the argument is obviously being done by premise 3, which Broome simply assumes to be the case, i.e., he assumes “that a person’s identity is not a part of what I call her life, or of how things are for the person.  I assume a person’s identity does not show up from her personal perspective” (p. 95).  Nevertheless, this is actually quite a controversial metaphysical assumption, viz., that selves are abstract bare particulars, that I could be abstracted away from all of my various individual physical and psychological properties (or at least those properties relevant to my personal good) and then be instantiated in the world with someone else’s individual properties (or, more precisely, I could be instantiated into a life with some specified set of properties that someone else could also be instantiated into as well).

The problem is that it’s difficult to make sense of the notion of a personal good, or personal wellbeing, on such a view, because it becomes difficult to make sense of the notion of the person conceived in this way.  On most contemporary accounts of agency, for instance, the self is simply identified with some sort of evaluational or noncognitive affective framework.  So on Gary Watson’s early account, the self is located in its valuational system, and while one can be dissociated from a certain set of ends and principles, this can only be done from the standpoint of another set: “In short, one cannot dissociate oneself from all normative judgements without forfeiting all standpoints and therewith one’s identity as an agent” (“Free Agency,” in Gary Watson, ed., Free Will, p. 347).  And on a Frankfurtian picture, one’s identity as an agent is located in certain decisive higher-order volitions, or (as I would have it) in one’s nexus of cares.  On these views, then, who I am is inextricably bound up in a certain subset of my psychological properties.

There’s a very plausible thought, then, that relates these accounts to goodness: something counts as good for me only if “for me” is short for “for me from my standpoint as _______,” where the blank is filled in with one of the above accounts, i.e., “for me in my standpoint as a judger of the value of X,” or “for me in my standpoint as a carer about X,” and so forth.  (Of course, there could be some objective goods, but surely not all goods are objective, i.e., are utterly independent of things I contingently value, desire, or care about.  And as along as some goods are subjective in this way, that’s all that’s needed to make the point about the connection between goodness and agency.)  Absent some specified standpoint, a standpoint with which the agent is identified, it’s hard to conceive of how something might be good for that agent.

But then if we strip the agent of all such standpoints, of all such filters through which things can even count as good, as Broome seems to do, we lose the capacity for even having a personal cardinal scale.  So there’s a nested dilemma: either we are abstract selves, in which case the notion of personal betterness is incoherent, or we aren’t, in which case we have a subdilemma: either the person who could have lived my life would have, in so doing, been me or he would not have been me.  If he would have been me (given that he took on my evaluative framework, say), then what we have is still just a personal cardinal scale of goodness (mine), and it’s not yet universal.  If he would have remained him, then any goods accrued would have been filtered through his evaluative framework, so we would still just have a personal cardinal scale of goodness (his).  In any case, though, Broome fails to get interpersonal comparisons of wellbeing.

Now Broome’s much smarter than me, so what have I missed?

17 Replies to “Broome on Interpersonal Comparisons of Wellbeing

  1. I’m not sure how to interpret (1), which says: “A person’s life (i.e., how things are for that person) contains ‘all the features of the world that can affect her wellbeing’.” Suppose, for the sake of argument, that the desire-fulfillment theory is the correct theory of well-being. Would, then, P’s life include features such as P had a desire that Q? Or would it include only features such as P had an unfulfilled desire of a certain intensity and duration? Whether P had a desire that Q or a desire that R doesn’t necessarily have any affect on P’s well-being. What does necessarily have an affect on P’s well-being is whether P’s desires, whatever they may be, are fulfilled. Perhaps, then, Broome is not claiming that we can abstract selves from their contingent desires, values, cares, or what have you. He’s merely claiming that we can abstract “the features of the world that can affect her wellbeing” from her contingent desires, values, cares, or what have you.
    I’m not sure, though, where this would fit in with your complicated dilemma.

  2. I’m not sure how much fanciness we need in an account of personal identity to deny that anyone could live any given life. So I don’t think one has to take a stand on the agency inspired accounts of valuer identity to think that there is some limit on how different my life could be before it would not be mine. But I am entirely sympathetic to the general point that Broome’s argument (as reconstructed here – I’ve not read it for myself)seems to take on board heavy metaphysical baggage. One does not have to have an account of personal identity to think that identity must supervene on other features of a person, likely including features that make a difference to how that person’s life goes for him/her.
    I wonder about step 5. Can it be justified independently of step 3, given that it seems to presuppose that the individual’s cardinal ordering includes values for living the life of someone else? If that doesn’t make sense, why think that an individual must have preferences/views about such options?

  3. It’s not obvious to me that there’s any problem about essentialism or bare particulars here. I interpret premise 4 as saying something like this: If there are two possible lives, one lived by X and one lived by Y, and exactly the same things happen in each, then they have the same values for X and Y. (ie the fact that one is lived by X and one by Y makes no difference.) Suppose it happens to be the case that there are some lives that couldn’t have been mine. Then such lives cannot be paired with any life I could live to provide a counterexample to the premise. To get a counterexample, it has to be possible for X and Y to live the lives. Otherwise the antecedent is necessarily false and the conditional is trivially true.
    PS Is it really true that on most contemporary accounts of agency, a self is a framework? I don’t think I’m a framework.

  4. Ben’s comment makes me wonder whether the type-token distinction might be important here. Broome’s claim is not that two people might have the same life (token), but only that they might have lives of the same type. Perhaps this would make the metaphysical worries go away.
    But there would be a new difficulty. There are very many different ways of dividing lives up into types. We might divide them according to duration: short lives, long lives, etc. Or we might divide them according to wealth: rich lives, poor lives, etc. So Broome needs it to be the case that there’s some way of dividing lives into types such that: (1) all lives within a single type have the same wellbeing; and (2) different people might have lives of the same type. Whether this is possible may depend on one’s account of wellbeing.

  5. Ben,
    Is your disagreement with Ben’s statement of 3? Or is it that 4 is weaker than 3 and hence we can avoid any problematic features of 3 by just employing 4?
    You write:
    If there are two possible lives, one lived by X and one lived by Y, and exactly the same things happen in each, then they have the same values for X and Y.
    That claim is very weak, merely claiming that the value of a life supervenes on its other features. But surely the argument (again as presented above since I haven’t read the relevant Broome) needs something stronger. As presented 3 claims that for any two people there is at least one metaphysically possible life which is open to both hence which either of them could rank relative to the other lives open to them.
    And this seems to be needed to make the argument go through, since the goal is to generate a complete ordering of possible lives.
    Regarding Campbell’s comment on types and tokens, I don’t think it will help those of us who think that there is some limit to the changes you could make in the features of a person’s life without it no longer being possible that they themselves live it. For, at least as I understand it, the argument is premised on the idea that a person’s individual cardinal well-being scale includes all the lives possible for that person. If it already included more, I don’t see why step 3 is needed, since it is used to argue that the individual scale is universal (by which I take it he means it ranks all lives).
    Or am I missing how Broome runs the argument? (Again, I haven’t read the relevant Broome, so I could just be wrong on that.)

  6. I’m not questioning the interpretation of Broome, because I haven’t got far enough in that book yet. I’m OK with 3, I think it’s probably true, but seems more questionable than 4. I agree that 4 is a pretty weak principle as I’ve interpreted it. If there’s a stronger interpretation, I want to know what it is.

  7. Doug’s suggestion — that perhaps Broome’s “merely claiming that we can abstract ‘the features of the world that can affect her wellbeing’ from her contingent desires, values, cares, or what have you — is interesting, but then I don’t see how Broome can get universal cardinal comparisons out of that. After all, the personal good I’d get from having a set of features of the world X affecting my wellbeing would likely be quite different were X true of your life (or some suitable third person in whose life you and I are both to be instantiated), given our different values, desires, beliefs, etc., so X couldn’t constitute the whole good in question (and thus couldn’t be what’s abstracted away to provide universal comparisons).
    I agree with Mark that we don’t need much by way of fancy theories of personal identity/agency to make the point here. I was simply appealing to a few otherwise opposed authorities to suggest the point that, *regardless* of your view of agency, it will be very unlikely to include the view that one can abstract away from all of one’s ‘good-contributing’ properties and still preserve agent identity.
    Mark then asks: “I wonder about step 5. Can it be justified independently of step 3, given that it seems to presuppose that the individual’s cardinal ordering includes values for living the life of someone else? If that doesn’t make sense, why think that an individual must have preferences/views about such options?” I guess I don’t quite see how it presupposes that the individual’s cardinal ordering includes values for living the life of someone else. Could you say more?
    As for Ben, I second Mark (I’m cheating, I know). But as I read Broome, it seems clear that he needs something more than the weaker claim you present.
    As for your other point, I don’t think I called the self a “framework.” Instead, I think it was a “standpoint,” and I merely said the self was, for these theorists, *identified with* a certain standpoint, whether that’s in judgments of value, higher order desires, or a nexus of cares. (And, in any case, I don’t know how seriously to take your assertion/argument that you don’t *think* you’re a framework.)

  8. “On most contemporary accounts of agency, for instance, the self is simply identified with some sort of evaluational or noncognitive affective framework.”
    OK, I’ll say it: I don’t just think I’m not a framework. I’m not one. If you really want an argument: I weigh too much to be a framework. (I’m also not a standpoint, which I think entails that I cannot be correctly identified with a standpoint.)
    So my interpretation of 4 is too weak to be what Broome needs. I agree it’s probably too weak to get anything interesting out of. So what is the correct interpretation? I apologize for being dense. I will read Broome for myself this afternoon.

  9. Imagine a person leading a life very much like yours, but just slightly different. Surely you could have lived that life. Call that person A.
    Now imagine a person leading a life just slightly different from A‘s, and call that person B. Surely A could have lived B‘s life.
    [long chunk of the argument omitted for obvious reasons.]
    … Surely, Y could have lived Z‘s life.
    Now it may be that you could not have lived Z‘s life, because anyone living her life would be too different from you to be you. That’s ok, it seems to me, for Broome’s purposes. The chain of commonly livable lives between you and Z establishes a common scale. For each neighboring pair of alphabet persons there is a common evaluative scale for their lives. So long as each scale is complete, so long as it ranks (cardinally, but this feature is independently secured by other arguments of Broome’s), a chain of comparability means universal comparability for all persons in the chain.
    Right?
    Warning: there may be some hidden and dubious assumptions in my argument.

  10. I don’t see where Broome endorses 3. I only see the weaker claim Jamie attributes. (p. 96) I think he *does* endorse the weak version of 4 I attributed to him. But the heavy lifting gets done by the overlapping chain claim, I think.

  11. David asks me to say more about my worry about 5 presupposing what I found implausible about 3. I now think Jamie’s chain connection suggestion pretty well obviates the worry I had, but I’ll explain anyway.
    It was just that the antecedent of 5 as stated was 4 which does not say what is in each person’s cardinal ordering. I took it that it was actually something like the thought behind 3 that allowed him to conclude that the individual orderings would include a ranking for at least one life livable by each other person.
    But Jamie’s version doesn’t even need that, so long as you can get overlapping chain connections of everyone to everyone.
    Do merely possible people get to be part of the chain too? Otherwise it looks to me like it might be a contingent matter whether we can get a chain from any person to any other. Perhaps I really ought just to pick up the Broome. I suspect that what he needs here depends on how he is using the idea of a cardinal universal ordering, and I don’t know that yet.

  12. Just so everyone who wants to can follow along, here’s the relevant passage from Broome (pp. 94-95):
    “The question is not whether two people could live the same life simultaneously. It is whether there is a life such that one person could live it, and also such that another person could live it. I assume that is possible. It is not physically possible, but conceptually or metaphysically possible. This means I assume that a person’s identity is not a part of what I call her life, or of how things are for the person. I assume a person’s identity does not show up from her personal perspective. I do not assume that every possible life could be lived by every possible person. Other properties of a person besides her bare identity might prevent her from living every possible life. Perhaps I could not have lived Cleopatra’s life, for example. I assume only that some lives could be lived by more than one person….”
    The overlapping chain argument is introduced later, and it was one of the “technical details” I was glossing in premise 3 of my reconstruction of Broome. Ben suggests that premise 3 thus overstates the case, but the claim I’m objecting to — that some lives can be lived by more than one person — is still part and parcel of the argument, even on a weaker reading of 3. In other words, I’m objecting to the claim that every person “belongs to at least one overlapping pair” (p. 96) of lives. So I’m objecting to the first move in Jamie’s chain, that “surely” I could have lived the life of someone else who’s only slightly different.
    Now I’m not saying that *I* don’t have a variety of possible lives: things could have been different *for me*, I agree. I’m simply disputing the move from there to the claim that one of my possible lives could have been yours. I’m disputing, in other words, the claim that “a person’s identity is not a part of what [Broome] call[s] her life” (p. 95).
    Back to the framework that is Ben. Oops! I guess I did call you a framework! No matter. The argument that you weigh more than a framework, while quippy (there’s my neologism for the day), trades on an ambiguity in the “I.” Occasionally “I” denotes a body, of course. But sometimes, as in the case of ownership with respect to certain questions of agency, it doesn’t. So when the agency theorists I cited talk about what makes some motivationally efficacious desire *mine*, say, they have in mind a relation it bears to some other psychic elements (or framework!), and they identify the self with these elements (or framework). Surely, then, if we’re talking about personal goods, we’re talking about the ownership of those goods — what makes them *mine* — and so I was suggesting it’s most plausible to think of the ownership of such goods in the same way we think of the ownership of various motivationally efficacious desires.

  13. David,
    This is very interesting.
    Just to be clear: you think that there are no overlapping lives, right? Let me be more precise.
    Lives here are individuated by “all the features of the world that can affect her wellbeing”. I think you are denying this:

    There are two people, x and y, and one possible life, L, such that x could have lived L and y could have lived L.

    That seems hard to deny, but maybe I’m missing something.

  14. Jamie: sorry for the delay. Yes, that’s what I’m denying (or, more precisely, what I’m merely questioning). To say that I could’ve lived L *and* you could’ve lived L (not necessarily simultaneously, of course) seems to require the “I” to be detachible from its wellbeing-affecting properties in a way that strikes me as implausible. Here’s one possible story. Suppose the self consists simply in a bundle of psychological properties, and the identity of the self is a matter of its origins in combination with a causal dependency chain running from its origins to the current time. I can thus imagine living a different life, given that the same origins could have (supposing indeterminism) produced a different causal chain than the one currently instantiated. And I can imagine the same for you. What I can’t imagine, however, is one of those alternatives being lives we *both* could’ve led, given our different origins. And if we alter one or both of our different origins to make them identical, this makes the life either mine, yours, or someone else’s, but not one that both you and I (or some third person) could’ve lived.
    (This is different than the argument I gave before, I recognize, but it’s drawn from the same set of intuitions.)

  15. I guess I share David’s intuitions. Given the account of lives I like, the only way in which I could have lived your life is if I could have been you.
    I guess this where I get uncomfortable. Some friends of counterpart theory, such as David Lewis, are willing to say that I *could have been* identical with someone that I actually am not identical with. If you are willing to say this, then I guess you should be willing to say that you could have lived someone else’s life. I’m not sure I want to say that I could have been identical with someone I actually ain’t.

  16. i m a student of psychology department in karachi university of pakistan n i want a interpersonal value measuring scale without any cost can u send me on my e-mail address please i will be very thankfull to u

Comments are closed.