A personal constraint is a constraint on action that arises from certain associative relations, such as kinship, friendship, etc. Typically, they are injunctions to treat one's personal relations with a certain form of priority over strangers even if, for instance, not doing so would promote more good overall. One could construe this as a constraint on rational action, viz., that any action that disobeyed such a constraint is all-things-considered irrational, or as a moral constraint, that any action that disobeyed such a constraint is all-things-considered immoral. (I'll leave this question open.) Personal constraints differ from personal options. I might have an option to treat my wife with a certain priority in the face of a greater good elsewhere, but I need not do so--in a conflict case I could also be justified in acting for the greater good. Personal constraints deny that one can be justified in refusing to act with priority to one's associates. Do such constraints exist? I want very briefly to run an argument up the flagpole that they do not.
As I've just started exploring this topic in depth, it's possible that this argument has been given a bajillion times already (that would be helpful to know!). Anyway, here goes:
1. Personal constraints depend on and are regulated by a certain "associative relation", call this Relation A.
2. As I bear Relation A to a person to a lesser degree, my associative duties to that person are weakened. (by 1)
3. The person to whom I bear Relation A to the strongest degree is my future self.
4. Hence my strongest associative obligation is to my future self. (by 2 and 3)
5. Though I have the option not to do so, I have no constraint against sacrificing my future self for the greater good.
6. Hence, there is no person I am not justified in sacrificing for the greater good, though I may have options not to do so. (By 4 and 5.)
(6) is the denial of personal constraints. As far as I can tell, the argument (roughly) valid, give or take. Are the premises true?
1. You might deny (1) if you believe that personal constraints are, as Samuel Scheffler calls them, "non-reductive". In other words, that they don't depend on any particular interaction, say, they only depend on the fact of one's personal relationship in a certain way. (That he's my father, versus that we've spent all this time together, etc.) But this is compatible with the argument here. Relation A need not be reductive. The only thing that the argument requires for Relation A to be plausibly construed as scalar, and as diminishing the "further" one is from the agent. And it seems to me that even a non-reductive Relation A has these features. That x is my father entails greater associative obligations than that y is second cousin.
3. This is, to my mind, the crucial premise. But it is hard to see how, on either a non-reductive or reductive account of Relation A, (3) could be false. Non-reductively, "being my future self" seems to be at least as strong as any relationship one could bear to someone else (say, "being my father"). If the account of Relation A is reductive, it is hard to see how one could bear the proper sort of "interaction and influence" (to quote David Brink) to my future self any less than to others close to me. Brink, for example, suggests that the right sort of "interaction and influence" just is psychological continuity and connectedness. But under this conception of "interaction and influence," (3) obviously follows. I guess I'm skeptical that any plausible Relation A could both (a) establish personal constraints, (b) mark a strong distinction between one's self and one's associates. Anyway, it would be helpful to know of plausible ways of denying (3).
5. Could one hold that one behaves immorally/irrationally by sacrificing one's self for a greater good? Perhaps this might be plausible if the greater good was substantially insignificant, for instance, if sacrificing one's self would increase the good by a net of precisely one enjoyed lollypop, say. For what it's worth my intuition is that it remains permissible to sacrifice one's self under those conditions. But even if it's true that one is blocked from sacrificing one's self when the overall benefit is small, because one's constraints get weaker as one gets "further away" along the dimension of Relation A, this will lead to extremely narrow constraints. To put this another way, let's say that there are certain "permissibility conditions" under which I might sacrifice my child's interests for the impersonal good (such as there is a substantial amount of good to be achieved by doing so). But the "permissibility conditions" get wider as Relation A diminishes. But if the circumstances of permissibility for the person to whom I bear Relation A to the strongest degree, my future self, are wide to begin with, the set of permissibility conditions for the sacrifice of others will be so wide as to wonder whether any constraints exist at all.
I'm a bit wary of this argument at a few points; mostly the argument for (3) relies on simply denying that there are plausible denials of (3). I may have just not run across them yet. It would be especially helpful to be made known of such arguments, if they exist.