I have a question regarding Buck-Passing about good — two questions, really — and Doug kindly gave me guest poster status so I could ask here. If this goes well, maybe I’ll leech off industrious readers contribute again soon.

Here is the Buck-Passing account of good that I got from Tim Scanlon’s What We Owe to Each Other:

For something to be good is for it to have properties that provide reasons for behaving in certain ways with regard to it.

Scanlon himself does not quite say this. He says almost this about value, and then proceeds as if he has given an account of both value and good. So, I’m assuming what I’ve written is the intended Buck-Passing account of good.
But if so, it has an obvious problem. Here’s a counterexample to illustrate.

Marsha is a corrupt, selfish, and sadistic government official who uses her position to enrich herself, punish her enemies, and cheat the public. However, she would never rat on a friend. Marsha therefore has a property — loyalty to friends —  that provides a reason for admiring her. But, she is not good.

Here’s my first question: has nobody noticed this difficulty? I have looked in a few obvious places but not scoured the literature.

The malady doesn’t look terribly serious. I can think of three ways to fix this particular problem with the Buck-Passing account of good.

  1. For a feature to be good is for it to provide reasons for behaving in certain ways with regard to its bearers.
  2. For something to be good-in-a-respect is for it to have properties that provide reasons for behaving in certain ways with regard to it.
  3. For something to be good is for it to have properties that provide sufficient reasons for behaving in certain ways with regard to it.

One could also go for an ‘ought’ formulation instead of a ‘reasons’ formulation, but that seems not to be in the Scanlonian spirit.

Here’s my second question: what fix (one of mine or some other) seems to you best?

35 Replies to “Help Jamie Fix Buck-Passing

  1. Off the top of my head, the first two seem quite promising to me (and equivalent) and the third still seems to have the problem you identified. By the way, the second formulation is the formulation that occurred to me before I finished reading, but now I don’t see how it differs in an important way from the first.

  2. I agree with David: the second occurred to me first, and the third seems like an obvious non-starter. Am I missing something attractive about it?
    In any case, perhaps the original formulation should have been in terms of a thing’s being ‘overall good’.

  3. “For something to be good is for it to have properties that provide reasons for behaving in certain ways with regard to it.”
    A problem with this claim, I think, is that the expression “provides reasons” is misleading. We tend to think that a reason has been provided only if a reason has been provided to someone. That is, we provide x with y, but it is never the case that x is just provided to nothing at all. But, if that is right, that would make goodness depend upon the existence of someone’s having a reason, instead of goodness merely depending upon there being a reason. That would be a bad consequence I think.
    Why does (1) say features are good, rather than things with features are good. Isn’t that against the spirit of the view?

  4. A couple of thoughts. The original formulation seems either obviously false, or so noncommittal that even the Marsha example is not a problem for it.
    On one way of reading the original formulation, it entails that an innocent person’s being tortured to death is good, since it has properties that give reasons to behave in certain ways towards it.
    That would be a pretty uncharitable reading of the view. It’s supposed to be saying that there’s some way of behaving towards something, such that when a thing is good, we should behave towards it in *that* way. But since it doesn’t say what that way is, we can’t say whether the Marsha case is a counterexample.
    The description of the Marsha case suggests that the appropriate way to respond to a good thing is with *admiration*. If so, the view is really: for something to be good is for it to have properties that provide reasons to admire it. If you wanted to get around the Marsha problem, I guess you’d say something like this: for something to be overall good is for it to have properties that provide reasons to have an all-things-considered admiration of it. I guess there’s no reason to have an all-things-considered admiration of Marsha, even if there’s reason to admire one of her features, or to admire her in a way.
    Obviously there are going to be cases where some pro-attitude other than admiration will be the appropriate attitude. I think there will be worries at this point but I’ll leave it there for now since this might get us far from Jamie’s worries. (I just realized that I didn’t really answer either of Jamie’s questions. Maybe the second one, a little. Sorry.)

  5. Scanlon might try to resist your counterexample by claiming that what we have reason to admire is not Martha herself, but her loyalty. So his view, on your original formulation, implies not that Martha is good, but only that her loyalty is.
    Here’s a related question. I’m inclined to think that better is the fundamental evaluative concept. So, I wonder, what would a buck-passing analysis of better look like? If you had such an analysis, and defined good in terms of better, this might solve your Martha problem.

  6. I think a slightly modified version of 2. is better:
    2*. For X to be good-in-a-respect R is for X to have properties (that is, some R) that provide pro tanto reasons for behaving in certain ways with regard to X. (Pro tanto meaning literally as far as R goes).
    As far as her loyalty goes, Marsha is good in that respect, and her loyalty gives a pro tanto reason to admire her, though evidently we have not overall reason to do so given her other traits, and so she’s not overall good.
    Of course there’s then a question of what we should make of pro tanto, defeated, reasons for admiration. Should these have an influence on our attitudes in the same way as pro tanto defeated reasons for action do?
    The formulation with “sufficient reason” (I think Skorupski adopts it somewhere) does not satisfy me, because it is conceivable that the value of something is so small (for instance, when taken in isolation from other parts with which it might form a whole), or so utterly unrelated to what we care about, that it still provides some reason to admire or desire it, but not sufficient reason to do so. Of course there’s a question of how can anything be sufficient to give a reason to admire it (as any value-maker must if Buck-passing is correct), but not able to give a sufficient reason to do so.

  7. Okay, thanks for all that help.
    So, nobody knows of any place this problem has been mentioned?
    1. Dave, Robert;
    The third may indeed suffer from the same problem. I was thinking maybe there isn’t sufficient reason to admire Martha, since, after all, she isn’t admirable.
    And Robert, the original formulation would have the same problem if it were in changed to be about ‘overall good’, wouldn’t it?
    2. Christian, I thought that was part of the point of Scanlon’s view. To be good is to have properties that provide everyone with a reason to admire, promote, etc. But yes, I think I agree with your second point.
    3. Ben, Scanlon’s view is that there are different ways to be good, corresponding to different positive propositional attitudes. That’s why the official formulation seems so sketchy.
    4. Campbell, maybe, but it seems obvious to me that there is a reason to admire Martha, only not a very good reason, considering. The stuff about ‘better’ is one of the things I’m ultimately after, but I want to get this more basic issue straight first.
    5. Francesco, yes, but now: what is it, then, for something to be good? That is, if we can have an acceptable Buck-Passing account of good-in-a-respect, how do we use that to construct an account of good? The natural way to proceed is by saying something about ‘sufficient reasons’, but nobody likes my (3) (properly so).

  8. It occurs to me that there may be an issue with predicative vs. attributive uses of ‘good’. The first option maintains the predicative use of ‘good’ but uses it on features, rather than persons. I suspect we could generate a new version of the same problem this way–i.e. Martha’s loyalty could be good in some ways and not in others.
    The second option still uses a predicate adjective but changes it to ‘good in a respect’. This struck me, as it did others, as the best option. But isn’t a still better option something like “For someone to be a good friend [or other F] is to have features that provide reasons for behaving in certain ways toward it”? I think this may be what the second option is getting at: when we say Martha is good in respect of being loyal to her friends, I think we mean that Martha is a good friend, or at least has some good-friend-making characteristics, though perhaps is not good in other ways.
    This comment, BTW, reflects my deep suspicions about predicative uses of ‘good’ generally.

  9. The example is not that different from ‘the Strike of the Demon’ cases in the Rabinowicz and Ronnow-Rasmussen Ethics paper. In fact, their counter-examples seem even worse as the Evil Demon does not seem to have any admirable properties even when he provides reasons to admire. There’s also nice replies by Stratton-Lake and Olson.
    I like many of the replies to your case so I won’t say much about that. I think the idea of keeping the contributory level and over-all level in mind does the trick.
    Anyway, I sketched a proposal for buck-passing in my defence of the view (in Ethical Theory and Moral Practice). That sketch is based on an idea in this quote from WWO:
    ‘Judgments about what is good or valuable generally express practical conclusions about what would, at least under right conditions, be reasons for acting or responding in a certain way (p. 96)’.
    I think the under right conditions is important. It suggests that value is understood in terms of properties that are disposed to provide reasons in different contexts. This means that value of a thing depends on how many reasons and how good reasons the thing provides for admiration, promotion, respect, and so on (Scanlon is pluralist about *valuing* – different valuable things are to be valued in different ways). I think this provides a ground for comparative value judgments. One thing is better than another if it provides more reasons, stronger reasons, in more contexts than the other thing (and these reasons are for whatever is involved in proper valuing of each object).

  10. Also, instead of going for sufficient reasons could you say something like a thing is good overall if it provides more reasons to admire it than it provides against admiring (or for disliking)? In Martha’s case the opposite seems true if we look at the amount of her bad qualities she has so we might want to say that she is bad.

  11. Jamie,
    Maybe the original formulation should have been more complex and negatively formulated, like this:
    x is overall good iff x has no properties that provide reasons to deplore x, rid oneself of x, or…, and at least one property that provides a reason to pursue, admire, or … x.

  12. Robert,
    surely that cannot work. Can’t things that are good over-all have some bad features? Most good things I know of do.

  13. Heath,
    Yeah, that’s a second thing I was thinking, as it happens: maybe the primary account has to be an account of a good F. How much of the spirit of Buck-Passing could then be preserved?
    Hm, I don’t see the respect in which this example is similar to the Demon Strike. The reason for admiring Martha is object-given, after all. And I have a hard time with the ‘more reasons’ formulation, since I think reasons are very hard to count. (We’ve discussed this in another set of comments.)
    I’d expect that there are plenty of overall good things with properties that do provide at least one reason to rid oneself of the things that bear them. A trip to San Francisco, for instance, might be good even though it has the property of being time-consuming.

  14. I quess the similarity is that not all reasons overall need to make a thing good in the buck-passing account. I agree that that’s not much a similarity. The state/object distinction can probably be used in their cases even though they of course deny that.
    I don’t think you’ll need to start counting. In fact, I presume that you can formulate a similar account by just talking about ‘more reason’ without ‘s’ instead of ‘more reason*s*’ numerically. So, a more valuable thing provides more reason to value it in various and more diverse contexts than the less valuable one. And, a good thing overall provides more reason to admire it than reason to dislike it. This doesn’t require counting reasons just comparing their strength.

  15. Jamie,
    you ask “if we can have an acceptable Buck-Passing account of good-in-a-respect, how do we use that to construct an account of good?”. I think something close to Jussi’s proposal is the natural answer. X is overall good (that is, X has overall intrinsic positive value) iff X has properties that provide overall reasons to have a pro-attitude towards X. That is, X is overall good when the question “what to do with X” has somehow been settled. Two caveats to bear in mind: 1) once it is established that we ought to have a pro-attitude to X, it may still be that different pro-attitudes to X are appropriate (depending on the different things that make X good, and on the agents’ abilities), so that there may be a further question about which pro-attitude one ought to adopt; 2) that there is overall reason to favour X doesn’t of course entail that favouring X is what we overall have reason to do. There may be a Y that is better than X.

  16. I’m with both Campbell and Ben. I see two problems raised by Jamie’s example. One part of the problem looks like it arises from taking admiring to be an attitude that will work in a buck-passing account of good. I think Ben is wrong to think it is uncharitable to take Scanlon on this distributive reading, but that he is right that a good account should decide on which attitude figures in the account, rather than allowing lots of different attitudes to. Contra Scanlon, (and Jamie?), admirableness is not sufficient for goodness – so reasons to admire can account for admirableness, but aren’t going to be enough to account for goodness – and that’s why David and Robert seem right that Jamie’s third solution won’t turn the trick.
    But I think a second part of the problem has to do with the trade-off between different reasons. The bare existential claim in Scanlon’s original formulation wouldn’t suffice, no matter which attitude we supplied for it – for there could always be weightier reasons of the right kind on the other side. In principle, Jamie’s third answer looks like it is on the right track to address this problem, but (no suprirse – we discussed this in a recent thread) I’m with Campbell in thinking this will fall out, once we correctly focus on better than for the basic analysis.
    Jamie’s first two solutions (which do look different, to me, though I can see how they might have similar advantages and disadvantages) look like they don’t address the second part of the problem (so they would still need the third qualification anyway), and like they address the first part only obliquely. If we transform the buck-passing account of goodness into an account of goodness-of-a-feature or of goodness-in-a-respect, then how do we get from there to an account of goodness? So I’ve always favored Ben’s proposal that we need to limit the range of attitudes and find the right one to account for good, in response to the first part of the problem, and Campbell’s point in response to the second part.

  17. Jamie, if being good provides everyone with a reason to admire something, then the ancient Greeks had a reason to admire, promote, etc. David Lewis for being a good philosopher. But they couldn’t even have had thoughts about David Lewis. So, I suggest they had no reason to admire him, which implies they were provided with no reason to admire, promote, etc. him even though he was a good F; a good philosopher in this case.

  18. Ok. Then why not:
    x is overall good iff x has on balance few properties that provide reasons to deplore x, rid oneself of x, or…, and on balance enough properties to provide sufficient reason to pursue, admire, or … x.
    No doubt, refinement would be required. But something down this road seems the right way to construe Scanlon.

  19. Hi Jamie,
    Maybe this came up already, but there seems to be a different worry about the Scanlon’s S.
    S. For something to be good is for it to have properties that provide reasons for behaving in certain ways with regard to it.
    S isn’t normed to those with typical dispositions. If the larger community is composed of do-gooders atypically (relative to the actual larger community) disposed to act favorably (I mean, justly) toward free-riders and foul-dealers, say, then the property of being a foul-dealer/free-rider confers value/goodness on those that possess it. That seems mistaken. But maybe I’m misreading S. It’s actually not so obvious to me how to read S.

  20. Robert; so are you basically on board with the ‘sufficient reason’ formulation, then?
    Okay, Jussi, I’ll go with ‘more reason’ (which implies no counting, but only weighing). Indeed, let me now combine that thought with some others and suggest:

    For A to be better than B is for A and B to have characteristics that make it that case that one ought to prefer A to B.

    So this combines the ‘most reason’ idea (here expressed by ‘ought’), the ‘please choose one attitude’ idea, and the ‘better before good‘ idea.
    Christian: my understanding of Scanlon is that he thinks there is a reason for the Ancient Greeks to admire Lewis, although they weren’t aware of it (understandably). It’s a bit like Bernard Williams’ guy who has a reason to pour the contents of his glass down the sink (it’s petrol), but doesn’t realize that he has this reason.
    Mike: Scanlon doesn’t think those dispositions are in any way relevant to what people have reason to do. (I think they are, but I’m trying to follow Scanlon as far as possible for the moment.)

  21. Mike,
    Why would that be? The BPA is about reasons to behave in certain ways whereas your antecedent is about how people are disposed to behave. Even if the community is disposed to treat the freeriders in a certain way, it is an open question whether they have reasons to treat them in that way. If these people really deserve such treatment and if they have properties that provide reasons to treat them favourably, then BPA gives the conclusion that at least they have some value. This cannot be read of from the actual dispositions though.

  22. Jamie,
    I had something like that in mind. But, things have to get more complicated. Whether we prefer A to B depends a lot on the context where we are. So, if I am stranded in the mountains I will easily prefer a warm jacket to hearing Mozart. I would not want to conclude on the basis of this that the jacket is overall better than Mozart’s music. The comparative value judgments must therefore be based on dispositions to provide reasons more generally in different contexts. I’m not sure whether all of this can be put into a formal necessary and sufficient conditions for better than but it does strike me as the right idea.

  23. Jussi,
    How’s about the following revision of Jamie’s proposal to deal with your latest worry:
    For S1 to be better than S2 is for S1 and S2 to have characteristics that make it the case that everyone ought to prefer S1’s obtaining to S2’s obtaining, where S1 and S2 are states of affairs.
    It seems that we all ought to prefer the state of affairs where you keep from freezing while not getting to listen to Mozart to the state of affairs where you get to listen to Mozart while freezing.
    Perhaps, we also need to change ‘characteristics’ to ‘intrinsic properties’ in order to avoid problematic cases where, given certain relational properties, I ought to prefer my child is saved and you ought to prefer that your child is saved.

  24. On ‘better than’:
    This comes to mind as a possible problem case, at least for Doug’s revision..
    A full house (or having a full house) is better than a high pair (having a high pair), but it is not the case that everyone ought to prefer the one to the other.
    First, it is not the case that I ought to prefer Bob having a full house to his having a high pair when I am playing against him.
    Second, only those committed to winning the poker game ought to prefer their having one to the other (there are additional restrictions in the same vein of course).
    To deal with these cases, I guess you could make a distinction between better-for-a-purpose or better-of-a-kind and better-than-from-the-POV-of-the-universe?

  25. Jussi,
    Even if the community is disposed to treat the freeriders in a certain way, it is an open question whether they have reasons to treat them in that way. If these people really deserve such treatment and if they have properties that provide reasons to treat them favourably, then BPA gives the conclusion that at least they have some value.
    Let me see if I have this. I thought the Scanlon excerpt was stating that the dispositions of the community to believe/act–whatever they happen to be– was what determined whether those properties provide reasons to treat them favorably.
    But what you seem to be saying here is that the community’s dispositions to believe/act do not determine whether those properties provide them with reasons to act in certain ways. Rather it must be true independently of what the community is disposed to believe and do that the properties of these people (viz., the free riders/foul-dealers) are such that (quoting from above) “. . .these people, [the free-riders and foul-dealers], really deserve such treatment”. If they do not “really deserve such treatment” then, presumably, quite apart from what the community is disposed to believe, the properties of these people do not give the community reason to act favorably toward them. Is that pretty much what you’re saying?

  26. Mike,
    I think you got me right even though I’m not alltogether certain to which properties you are referring – the properties of the freeriders I presume.
    I don’t think Scanlon accepts the inference:
    P1. Community C believes that property F of X is a reason to phi.
    C. Therefore, that X has the property F is a reason to phi.
    He’s got quite elaborate story about the truth of reason-judgments. He doesn’t want to sat that their truth-makers are REASON-FACTS. Rather, the reason-judgments have to pass certain epistemic deliberation-procedures. That a community has certain beliefs about reasons does not guarantee this.
    That might go for the value of states of affairs but I guess you (or, I, I should say) want to have an account of value that also goes for the value of objects and qualities – i.e., for all things that are bearers of value.
    I’m also worried about the Strike of Demon cases with these formulations and relatedly whether there are any states of affairs which everyone ought to prefer. So, in your proposal S1 is better than S2 if everyone ought to prefer S1 over S2. However, it might be that a Demon threatens someone that unless they prefer S2 they will be killed. In that case, they ought to prefer S2 even when most people ought to prefer S1. The formulatation needs to say something about the source of the oughts or reasons that matches whatever it is the value of which we are assessing.

  27. Jamie, you said: “My understanding of Scanlon is that he thinks there is a reason for the Ancient Greeks to admire Lewis, although they weren’t aware of it (understandably).”
    I’m not objecting to the idea that there is a reason for the ancient Greeks to admire Lewis. He was, after all, a good philosopher. I’m objecting to the claim that Lewis’s being a good philosopher or having some other properties “provides” the ancient Greeks with a reason to admire Lewis. I want to distinguish between (a) the existence of reasons and (b) the possession of reasons. I accept, for the sake of the argument, that (a) is true, that something’s being good generates the existence of reasons, but I’m denying (b). I deny that (a) implies (b) and I’m suggesting awareness of reasons, or something in the neighborhood of awareness of reasons, is necessary for some thing to be provided with a reason.

  28. Christian,
    Ah, I see. There is a reason, but it hasn’t been provided to those who don’t know about it — they don’t have it, so to speak. I suspect that ‘provides’ might just be a mistake, then. I personally think Buck-Passers should have a special editor that removes all occurrences of ‘provide’ and cognates from their papers… but that’s another story.

  29. Jamie,
    Yes, sufficient reasons view, but with the understanding of ‘overall good’. So perhaps it should just be ‘x is overal good iff x possesses properties that overall are sufficient reasons…’
    Ok, overall weight.

  30. Chiming in late, it does seem to me that with good Jamie’s suggestion of preference seems to have much to recommend it, especially since it makes the connection with better relatively straightforward. It would also accommodate Mark S’s reasonable thought that we need to focus on some attitude or other to make the account precise enough to evaluate. I think this is consistent with thinking there are other positive evaluative concepts that might link with different responses — I suspect being admirable is one of those.
    But I do have a worry with preference, which has to do with my thinking that the appropriateness of attitudes is circumstance relative. I think you ought to prefer better things to worse things when you are in a position to choose between them, and probably also that you ought to be so disposed most of the time. But since I think we can distinguish between being disposed to prefer and preferring itself, there is no real problem with someone who is only disposed to prefer the better in conditions where she has the choice.
    I’m not sure whether that shows she doesn’t have a reason to prefer the better in conditions where she is not in a position to choose between the better and the worse. So it might not be a problem. I guess I think it does give her a reason to prefer, but this is a reason that rationalizes preferring but does not necessarily require preferring in those conditions. (For example, when a person is really angry I don’t think there is much wrong with preferring some bads to some goods, and maybe even that the cause of the anger gives that person a reason for that preference, but that this is so only when s/he is not actually in a position to choose the bad over the good. When she’s in such conditions she has reason to change her preferences.)
    This may be a bit idiosyncratic to me in that I worry that fitting attitude accounts can require to much of people.

  31. Hi, Mark – I actually suspect that your worry is kith and kin to the wrong kind of reasons problem. Suppose that angry people (or perhaps people who are rightly angry, to head off some objections) have reasons to prefer some bad things to some good things. Or suppose that parents have reasons to prefer things that benefit their children to things that benefit more other children instead and thus are better. Those reasons look like they bear on what the people in these situations ought to prefer, but I think, at least, that they don’t seem to affect which is the correct preference. So those reasons look like the wrong kind to affect correctness.
    My idea is that in activities that are governed by norms of correctness, like chess, parking regulations, and perhaps preference, correctness doesn’t vary from person to person – it is an agentless or ‘agent-neutral’ concept. Examples like yours rely on variations between the reasons that different agents have – idiosyncracies of their reasons. But if correctness is to be an agentless concept, it can’t depend on such idiosyncratic reasons. That’s why I think that buck-passing accounts need to appeal to the notion of correct preferences, and why I think that the most general solution to the wrong kind of reasons problem is going to have to work by ruling out idiosyncratic reasons as not mattering for correctness.

  32. Hi Mark,
    I’m a bit of two minds about this. Partly that’s because I’m open to thinking of some features of an agent’s psychology as part of the situation the agent is in, rather than just as an idiosyncrasy. This would be especially tempting to me if the person in question is rightly angry. Then it seems to me there is nothing wrong with the preferences that make sense given the anger.
    But I suppose we can say that, and still say that these are the wrong kind of preferences to determine betterness. I agree with that. And I have no trouble with talking about correct and incorrect preferences. I’m just very tempted to try to limit the conditions in which certain preferences are correct in the sense that they are required of us. They may still, for all that, be the ones that are correct for determining betterness.
    Since I’m a non-consequentialist who thinks that right action is partly determined by what is better (reading better in an agent-neutral way) and partly also by considerations that don’t reduce to betterness of what is chosen, I suspect I’m committed to a more complex relationship between goodness and preference even at the time of action or choice. I think that people should most prefer that their own friend (say) get a certain job, even when they don’t think that it would be better if the friend get the job — so long as the friend’s getting the job would not be much worse impartially than someone else getting it. Since I don’t like agent-relative teleology as a way of capturing this view, I don’t have an easy way to define goodness in terms of what I think I ought to prefer, though I think there should be a way to work backwards to something that would do the trick by delineating how other considerations might interact with betterness to determine overall preference. And that might also handle the worries I voiced above.

Comments are closed.