In an earlier post, Supererogation for Maximisers, I tried to reconcile two apparently irreconcilable claims: first, that maximising consequentialism is true; and, second, that supererogatory action is possible. Subsequently, the same topic has received significant attention in the comments to another post, Favourite Objections to Consequentialism, prompting me to revise my position. In this post I shall (1) briefly review my old position, (2) show why I’m no longer inclined to accept it, and (3) propose something new.
The problem of supererogation for maximisers lies in the inconsistency of the following three statements:
Max: An action X is permissible iff there exists no action Y such that (1) Y is an alternative to X, and (2) Y is better than X.
Super: An action X is supererogatory iff there exists some action Y such that (1) Y is an alternative to X, (2) X is better than Y, and (3) X and Y are both permissible.
Existence: There exists at least one supererogatory action.
The statement Max captures the core claim of maximising consequentialism. (For simplicity, I’ve chosen here to talk of actions as the objects of evaluation — i.e. as being better or worse than each other. Some might prefer to say that it is outcomes or consequences that are the proper objects of evaluation in a consequentialist approach. But I don’t see that rephrasing things in terms of outcomes or consequences would make much difference here. We can simply treat “action X is better than action Y” as convenient shorthand for “the outcome of X is better than that of Y” or whatever.) The statement Super is a natural analysis of the concept of supererogation, the basic idea being that a supererogatory action is one that does more good than was required for the agent to have acted permissibly — i.e. one that “goes beyond the call of duty” in the usual idiom. Finally, the statement Existence is just the fairly weak claim that some person has at some time acted in a supererogatory way.
However, as is easily seen, the three statements are jointly inconsistent. (Note: I assume that the relation “is an alternative to” is symmetric.) To some philosophers this has seemed a good reason to reject Max. Yet, for reasons I won’t bother you with here, I’m disinclined to follow them; I want to hang onto Max. And I’d also like to keep my Existence. So, for me, Super has got to go.
In my earlier post, I proposed replacing Super with the following:
Super*: An action X is supererogatory iff there exists some action Y such that (1) Y is an alternative to X, (2) X is better for others than Y, and (3) X and Y are both permissible.
Since one action might be better for others than another, even though the two are equally good simpliciter, there is no inconsistency in the conjunction of Max, Super*, and Existence. Maximising consequentialism allows for the possibility of actions that are supererogatory in the sense articulated by Super*.
However, now there is a new problem. As several commentators argued in their comments to the earlier posts (linked above), the revised analysis of supererogation results in maximisers running afoul of the following:
Better: For any actions X and Y, if (1) X is an alternative of Y, (2) X is supererogatory, and (3) Y is not supererogatory, then X is better than Y.
This states simply that, within a given situation of choice, a supererogatory action must be better than a non-supererogatory one. The problem is that the statements Max, Super*, Existence, and Better are jointly inconsistent. (Note: here I assume that the relation “is better for others than” is asymmetric.)
Earlier I was happy to reject Better, but now I’m not so sure. So, let me instead propose yet another analysis of supererogation:
Super**: An action X is supererogatory iff there exists some action Y such that (1) Y is an alternative to X, (2) X is better than Y, and (3) X and Y are both blameless.
Whereas Super* revised condition (2) in Super, our new analysis Super** revises condition (3). The key change here is that the notion of permissibility has been replaced by that of blamelessness. Obviously, then, Super** will avoid the problems of Super only if it is possible for permissibility and blamelessness to “come apart” in a certain way. That is, although there is no inconsistency in the conjunction of Max, Super**, Existence, and Better, that conjunction is inconsistent with the following:
Blameless: An action is blameless only if it is permissible.
In other words, maximisers can embrace Super** only if they’re willing to allow the possibility of blameless impermissible actions.
But this does not strike me as a bitter pill to swallow. I think we have independent reasons to reject Blameless. Some philosophers, perhaps most notably Derek Parfit, have already given compelling arguments for the possibility of “blameless wrongdoing”. But I want to suggest a slightly different argument, which appeals to a familiar distinction commonly drawn in law: namely, the distinction between a justification and an excuse. As I understand the distinction, a justification is some consideration that shows that the action is question was not wrong, whereas an excuse is a consideration that shows that, although the action was wrong, the wrongdoer ought not to be blamed for it. Unless lawyers are all guilty of hopeless conceptual confusion, the distinction between justifications and excuses suggests that there is sufficient conceptual room to drive a wedge between permissibility and blamelessness in exactly the way that Blameless denies.
Moreover, the distinction is supported by the kinds of things that ordinary folk are apt to say on occasion. Surely we sometimes hear such things as: “I can understand why you did what you did — you’re only human, after all, and I’m sure I would have acted in the same way if I were in your situation — and so I don’t blame you for acting as you did; still, I think that you really ought not to have acted differently.” Intuitively, at least, such statements seem quite intelligible. Of course, philosophical work remains to be done to explicate more clearly the distinct roles of permissibility and blamelessness. Nonetheless, I think, examples of this kind suggest that maximisers could do without Blameless, in which case the analysis given in Super** seems to constitute a plausible way in which they can accommodate the possibility of supererogation.
Independent of whether or not we should reject Blameless, Super** constitutes a plausible way for maxizers to accommodate supererogation only if Super** provides the correct analysis of the concept of supererogation. But, as you stated earlier on in your post, Super, not Super**, “is a natural analysis of the concept of supererogation, the basic idea being that a supererogatory action is one that does more good than was required for the agent to have acted permissibly — i.e. one that ‘goes beyond the call of duty’ in the usual idiom. [emphasis added]” What reason do we have to accept Super** over Super, then, as the correct analysis of the concept of supererogation? The fact that the former allows maximizers to accommodate supererogation is no reason to think that it’s the correct analysis of the concept of supererogation.
Campbell: If I read you correctly, your analysis implies that “merely” doing one’s duty where one had a supererogatory alternative available would be impermissible — wrong — but blameless. It seems to me, though, that the supererogation intuition is precisely the opposite: it involves going “above and beyond” the minimal requirements of duty, i.e., it starts with the thought that it’s *permissible* not to do what’s best. So yes, doing X where Y is supererogatory (and thus better) is blameless, but it’s blameless precisely because it’s not impermissible. Imagine someone failing to rush into a burning building to rescue a child being told, “Well, what you did was wrong, but I don’t blame you”!
Furthermore, there’s an ambiguity in the term “blameless.” On the one hand, it might mean “pointless to be blamed”; on the other hand, it might mean “not appropriately subject to blame (i.e., not blameworthy).” Your example (“I don’t blame you for acting as you did — I’d have done the same were I in your shoes”) strikes me as an example of the former sense — there’s no purpose to be served in *expressing* blame to you — whereas I suspect that you want the second sense here. But then I have a much harder time seeing the move as successful: blameworthiness strikes me as deeply conceptually committed/connected to wrongdoing, so while it may occasionally be *pointless* to blame you for wrongdoing, you’re always blame*worthy* for doing so. But of course my conceptual commitments may simply be a tenacious hangover from my religious upbringing.
Turns out Doug and posted comments nearly at the same time, so I didn’t see that my first comment was nearly duplicating his (although his was put a bit more succinctly than mine).
Doug and Dave,
You make good points. I guess I don’t want to insist that Super** is the “correct” analysis of the concept of supererogation. Presumably, the main standard of correctness here is how well the analysis coheres with the way in which ordinary folks use the term “supererogatory”. (Do ordinary folks use that term at all?) And I’m prepared to concede that, by such a standard, Super is the more correct analysis. Nonetheless, I stand by the claim that Super** provides a plausible way in which maximisers may accommodate supererogation. For it enables them to say “although supererogation is impossible, there is a closely related notion — call it supererogation** — that is possible.” Thus, maximisers may accommodate ordinary folks’ talk of supererogation, at least, by reinterpreting it in such a way that it makes sense.
To be sure, if Super is the correct analysis of supererogation (by the lights of ordinary usage), then maximising consequentialism implies that those who believe in the possibility of supererogation are conceptually confused. And that would indeed be a cost of the theory. But the cost may be reduced if we can find a surrogate for supererogation that maximisers can accept. My suggestion is that supererogation** is fit for such a role. Perhaps people are confused, but not irredeemably so. Clearing up the confusion would not take a great adjustment; we need only to make a fairly subtle substitution, replacing supererogation with supererogation**. This may still be a cost; but perhaps it is a cost that maximisers can bear.
You’ve defined Max in terms of permissions. I thought Max was at least a position on requirements, that one ought to perform the act that is best from amongst one’s alternatives. If that’s right, then shouldn’t any act one can perform be either required or prohibited, and hence, the only acts that are permissible will be required. And then, if the required act is the only act that is a candidate for being permissible, then it will also be the only candidate for the supererogatory. But the required act will be the best one, so it will turn out both that the alleged supererogatory act is both required and best, not permissible and less than best, which is what defenders of supererogation have supposed.
Also, Super** and Blameless together seem the same as Super. I don’t see what the necessary condition in Blameless adds to Super.
As far as I can tell, a maximizer ought to reject Supererogation. They should tell a story like the one you are telling, but a little different. Supererogatory acts are acts, required or prohibited, the performance of which makes the actor praiseworthy. All I mean by this is that the act is tough to perform, required a bit of sacrifice perhaps. Super acts teach us something about each other, that one places the interests of others before oneself, etc. That captures enough of the rhetoric surrounding supererogation and its something the maximizer can happily endorse.
Christian,
Thanks for your comments. A few replies:
1. You write:
I don’t think that follows. Suppose that two alternatives are tied for first place — i.e. equally good and better than all the rest. Then maximising consequentialism, on its most natural construal, will say that neither of these two alternatives is either required or prohibited. Of course, MC requires that the agent perform either one or the other alternative (and it will prohibit the agent’s performing neither of them), but that’s not to say that either action is such that MC requires the agent’s performing that particular action. More generally, as I like to think of things, an action is required, or obligatory, just in case it is uniquely permissible among the alternatives. Hence, MC implies that an action is required just in case it is uniquely best.
2. Super is equivalent to the conjunction of Super** and the following biconditional: an action is blameless if and only if it is permissible. Is that a problem for my argument? If so, I don’t see why.
3. I take your positive proposal to be largely in same spirit as mine. Plausibly, we don’t blame people for acting wrongly when their acting rightly would have been very difficult for them — involving significant personal sacrifice and so on.
Hi Campbell,
I wonder if you could say more about actions and motives. In particular, consider the following:
You go into the pond to save the drowning child out of love, or a sense of duty, etc. Basically – out of a morally appropriate motive.
I go into the pond to save the drowning child because I’m hoping for a reward, to impress the crowd around me, or because I want to kidnap the child, etc.
On your view, are we both performing the same action (better: actions of the same type), but with different motives, or do we have two different actions here (actions of different types)?
Also, does ‘Blameless’ concern particular action tokens, or does it concern types?
Campbell,
I agree with your first point. I think ties are possible. But I am not sure that makes much of a difference. I am suggesting that MC tells us we are required to perform the unique best act open to us, or some act from the unique disjunction of best acts open to us. Either way, it tells us something about requirements, not just permissions.
You wrote: “Super is equivalent to the conjunction of Super** and the following biconditional: an action is blameless if and only if it is permissible. Is that a problem for my argument? If so, I don’t see why.”
If Super** is correct, then an action is supererogatory only if it is permissible. But that is somewhat questionable. I think that swimming alonside the lifeboat that would sink if I were on it (cold water, the ship just sank, many people on the lifeboat…), is supererogatory, but one could easily describe a case in which doing so would not lead to the best consequences. Hitler is on the boat, unbeknowst to me, etc.
So, permissibility is not a requirement of the supererogatory. So, Super** won’t work. Now, spelling out a plausible counterexample will depend upon whether one is considering objective or subjective versions of MC, I just suggested a counterexample to the objective version.
I do think that our accounts are similar as well. MC can account for the super… acts by giving a theory of praiseworthiness that is independent of act appraisals. Reasons for acting, when they are of the right kind, can account for intuitions about supererogation.
Christian,
Regarding your first point, I don’t think there’s any substantive disagreement between us. I certainly agree that MC tells us something about requirements. However, I think that what it says can be expressed without using the word “requirement” (or “required” etc). I take it to be true by definition that any sentence of the form “S is required to do X” is equivalent to “S is not permitted to not do X”. (In this respect “required” and “permissible” are like “necessary” and “possible”; logically speaking, they are duals.) And I also take it to be true by definition that an action is required only if it is permissible. It follows that an action is required if and only if it is uniquely permissible. Hence, although the statement Max, as I wrote it, does not use the word “required”, it does tell us something about when an action is required: specifically, it tells us that an action is required if and only if it is uniquely best.
Regarding your second point, I do assume that an action is supererogatory only if it is permissible. (Though I don’t see how this is implied by Super**.) But I’m not persuaded by your counterexample. If we’re going to distinguish objective and subjective permissibility, then I think we should also distinguish objective and subjective supererogation. I doubt you’ll find an example of an action that is supererogatory and impermissible, when these are either both understood in the subjective sense, or both understood in the objective sense.
Hey Campbell,
Right. There isn’t anything substantive we are disagreeing about. I was pushing the point that Max should be stated in terms of requirements because I want to leave open the possibility that there are “mere” permissions, actions we are permitted, but not required to perform, and that ‘permission’ is ambiguous between mere permissions and requirements. I am thinking of someone who believes (1) all permissions are mere permissions and (2) there are no requirements (a nihilist). Such a person, with a weird metaphysics, would be an MCer given the way you characterize it and I wouldn’t want them to count as one.
You wrote: “I do assume that an action is supererogatory only if it is permissible.”
I think this is bad for a few reasons. First, I think supererogation is impossible on an MCer’s view given the way they are standardly characterized. However, praiseworthiness is possible even on an MCers view. Attributions of praise capture enough of the talk so that an MCer denying supererogation and accepting praiseworthiness doesn’t look so bad. Here I just echo points made by Kris on the “Objections to Consequentialism post”. Praiseworthiness is logically independent of an acts being required, so the “only if” above suggests a logical relation that I don’t think we should accept. The reasons one acts from determine whether the performance of an act is supererogatory, but I doubt whether the fact that the act performed is better or worse than alternatives determines whether it is supererogatory.
And we should want this because doesn’t it just seem like unforseen consequences of an action are just irrelevant to one’s deserving praise for performing it?
This raises the last point. I don’t think there is room for objective supererogation because I think the class of the “revisionary” supererogatory just is the class of acts the performance of which would make the performer deserving praise. But this is just to say the act is done from certain reasons, which is to make the status of the act determined by subjective features. I’m not sure how one could make a case for objective supererogation.
And that leaves open the question of how an MCer that is a subjectivist will make sense of supererogation. I suggest she do so in the same way as the objectivist. It might turn out though, on this view, that only strictly required acts are supererogatory. It is hard to see how one could reasonably believe some act leads to the best results, intentionally refrain from performing it, and yet still deserve praise for doing so. I’m not sure about this though.
Christian,
1. How do you propose to formulate MC in such a way that the nihilist you describe must reject it? Perhaps you have in mind the following:
Max*: For any act X, (1) X is required if and only if X is better than every alternative; and (2) X is merely permitted if and only if X is equally as good as some alternative but worse than none.
But your nihilist can accept all of this. Moreover, Max* is simply equivalent to Max. So you must have some other formulation in mind. But I don’t see what it could be.
2. I don’t understand your motivation for denying that “supererogatory” implies “permissible”. I take it that, in your view, the best way for a proponent of MC to accommodate talk of supererogation is to tell a certain story about praiseworthiness. But I don’t see how accepting that “supererogatory” implies “permissible” could prevent the telling of such a story.
3. I think we should distinguish the supererogatory and the stupid. Making great sacrifices in order to benefit others may be supererogatory, but making great sacrifices in order to benefit no one is just plain dumb. This makes room for a distinction between subjective and objective supererogation in the following way. When the agent makes great sacrifices in the belief that his doing so will benefit others, his action is subjectively supererogatory. But when the agent makes great sacrifices and, as a matter of fact, his doing so will benefit others, his action is objectively supererogatory. Suppose I give all my money to an apparently reputable charity in the belief that doing so will benefit starving children in third-world country, when in fact the charity is so ineffciciently run than no one benefits at all. My action is supererogatory in the subjective sense, but not in the objective sense.
Christian,
Please ignore 2 and 3 above. I think I now see your point. I agree that, if we’re going to explain supererogation in terms of praise and blame in the way that you, I, and Kris are partial to, then objective supererogation doesn’t make much sense.
I also see now that, given a few plausible assumptions, the conjuntion of Max and Super** implies that an action may be both supererogatory and impermissible. Suppose there are three alternatives X, Y, and Z such that (1) X is better than Y, and Y better than Z; and (2) all three are blameless. Then Max implies that Y is impermissible, and Super** implies that Y is supererogatory. That worries me.
I think there might still be a confusion somewhere. First, Max* is what I am getting at. I would include ‘distinct’ modifying ‘alternative’ in clause (2) though. Is Max* different from Max? Well, Max* says when there are no ties, the best act is required. Max says, when there are no ties, the best act is permitted. If by permitted in Max you just mean required, then they seem the same to me unless I am missing something. But, Max* has the virtue of making explicit that when there is a unique best act, that act is not merely permitted, an ambiguity that remains in Max.
More importantly, you wrote: “I also see now that, given a few plausible assumptions, the conjuntion of Max and Super** implies that an action may be both supererogatory and impermissible. Suppose there are three alternatives X, Y, and Z such that (1) X is better than Y, and Y better than Z; and (2) all three are blameless. Then Max implies that Y is impermissible, and Super** implies that Y is supererogatory. That worries me.”
What worries me is not that there could be an action that is supererogatory and impermissible, but rather, that there couldn’t (assuming a praiseworthiness account of the supererogatory). Max implies Y and Z are impermissible. But if Super** applies to X or Y, then X or Y is permissible because X or Y is blameless. But, that’s just a logical falsehood. So, Super** does not apply to X or Y.
At this point, I just don’t see what’s wrong with Max* and the view that an act is supererogatory just in case the act is performed in, say, the interests of others at the sacrifice to oneself. Sure, this view would have to spelled out, but it seems plausible to me.
I suspect that people hesitate to think that an act can be both impermissible and the agent praiseworthy for performing it because they are, in some sense, considering subjective and objective versions of Max together. If objective Max is true and far away consequences make some act open to me better than another, then surely we can deserve praise for doing what is less than best and hence, impermissible. Otherwise, it’s likely that nobody would actually deserve praise. Yet, if subjective Max is true and only consequences the agent reasonably believes will follow from his act make an act better than an alternative, then it isn’t obvious that there will be cases in which an act is both impermissible and one for which an agent would deserve praise for performing.
Christian,
You suggest that the sentence “X is permissible” is ambiguous between “X is merely permissible” and “X is required”. But that is not so. To say that an act is permissible is not to say that it’s merely permissible; nor is it to say that it’s required. These are three distinct claims. True, if an act is permissible, then it must be either merely permissible or required, and the mere fact that the act is permissible does not decide between these two possibilities. But that is not a sign of any ambiguity. If it were, we would have to conclude that the sentence “X is possible” is ambiguous between “X is merely possible” and “X is necessary”, which surely it is not.
Perhaps an analogy will be helpful. Suppose that a number of coloured balls are placed in a bag from which one ball is then chosen at random. Consider the following two statements:
(A) For any colour X, it is possible that the chosen ball is X if and only if at least one of the balls placed in the bag is X.
(B) For any colour X, (1) it is necessary that the chosen ball is X if and only if every ball placed in the bag is X; and (2) it is merely possible that the chosen ball is X if and only if at least one of the balls placed in the bag is X and at least one is not X.
Would you say that there’s some ambiguity in (A) that is eliminated by (B)? I certainly wouldn’t. It seems to me that (B) is just a needlessly convoluted way of saying exactly the same thing as (A). But this case is directly analogous to that of Max and Max*.
You also write: “But if Super** applies to X or Y, then X or Y is permissible because X or Y is blameless.” Thus, you seem to assume that an act is blameless only if it is permissible. But, as I said in my original post, that assumption — which I labelled “Blameless” — must be rejected by anyone who accepts both Max and Super**.
Hey Campbell,
As I think more about how to state Max, I don’t know exactly why I wanted it stated in terms of requirements. Maybe I just like the word ‘required’. I wasn’t intending to make a semantic thesis about ambiguity, though. I kind of like your statement of Max now.
You said: “But, as I said in my original post, that assumption — which I labelled “Blameless” — must be rejected by anyone who accepts both Max and Super**.”
What’s odd about this is that Blameless is a clause in Super**. So, it isn’t clear how one can accept Super** and reject Blameless. Perhaps all you mean is that one can accept Super** by accepting its first two conditions and denying its third.
But what are your thoughts on my earlier proposal on praiseworthiness?
Christian,
Good — we’re making progress.
You’ll have to explain what you mean by saying that “Blameless is a clause in Super**”. I’m just not getting it.
I’m sympathetic to your praiseworthiness proposal. I was thinking along similar lines at one point. But, for some reason, which I don’t now clearly recall, I decided to state my proposal in terms of blame, rather than praise. In order to assess your proposal, it would be helpful to have a more concrete formulation to work with.