I like the hybrid approach, and discuss its implications for replaceability a bit here. (Shifting to the intrapersonal case: those of us who reject preference theories of well-being may still recognize reasons not to manipulate preferences, for example based on personal identity: the more you manipulate my values, the less the future person is me. To be a prudential benefit, then, the welfare gain has to outweigh the degree of identity loss. Moreover, itās plausible that extrinsic manipulations are typically more disruptive to oneās degree of psychological continuity than voluntary or otherwise ānaturalā character development.)
It seems worth flagging that some instances of replacement seem clearly good! Possible examples include:
Generational turnover
not blindly marrying the first person you fall in love with
helping children to develop new interests
I guess even preference-affecting views will support instrumental replacement, i.e. where the new desire results in oneās other desires being sufficiently better satisfied (even before counting any non-instrumental value from the new desire itself) to outweigh whatever was lost.
I agree that some instances of replacement seem good, but I suspect the ones Iād agree with are only good in (asymmetric) preference-affecting ways. On the specific cases you mention:
Generational turnover
Iād be inclined against it unless
itās actually on the whole preferred (e.g. aggregating attitudes) by the people being replaced, or
the future generations would have lesser regrets or negative attitudes towards aspects of their own lives, or suffer less (per year, say). Pummer (2024) resolves some non-identity cases this way, while avoiding antinatalism (although I am fairly sympathetic to antinatalism).
not blindly marrying the first person you fall in love with
people typically (almost always?) care or will care about their own well-being per se in some way, and blindly marrying the first person you fall in love with is risky for that
more generally, a bad marriage can be counterproductive for most of what you care or will care to achieve
future negative attitudes (e.g. suffering) from the marriage or for things to be different can count against it
helping children to develop new interests:
they do or will care about their well-being per se, and developing interests benefits that
developing interests can have instrumental value for other attitudes they hold or are likely to eventually hold either way, e.g. having common interests with others, making friends, not being bored
developing new interests is often (usually? almost always?) a case of discoveringdispositional attitudes they already have or would have had anyway. For example, thereās already a fact of the matter, based in a childās brain as it already is or will be either way, whether they would enjoy certain aspects of some activity.[1] So, we can just count unknown dispositional attitudes on preference-affecting views. Iām sympathetic to counting dispositional attitudes anyway for various reasons, and whether or not theyāre known doesnāt seem very morally significant in itself.
Plus, the things that get reinforced, and so may shift some of their attitudes, typically get reinforced because of such dispositional attitudes: we come to desire the things weāre already disposed to enjoy, with the experienced pleasure reinforcing our desires.
I think the hybrid view you discuss is in fact compatible with some versions of actualism (e.g. weak actualism), as entirely preference-affecting views (although maybe not exactly preference-affecting in the informal way I describe them in this post), so not necessarily hybrid in the way I meant it here.
Take the two outcomes of your example, assuming everyone would be well-off as long as they live, and Bob would rather continue to live than be replaced:
Bob continues to live.
Bob dies and Sally is born.
From the aggregated preferences or attitudes of the people in 1, 1 is best. From the aggregated preferences or attitudes of the people in 2, 2 is best, if Sally gains more than Bob loses between 1 and 2. So each outcome is best for the (would-be) actual people in it. So, we can go for either.
So, not all preference-affecting views even count against this kind of replaceability.
My next two pieces will mostly deal with actualist(-ish) views, because I think theyāre best at taking on the attitudes that matter and treating them the right way, or being radically empathetic.
I like the hybrid approach, and discuss its implications for replaceability a bit here. (Shifting to the intrapersonal case: those of us who reject preference theories of well-being may still recognize reasons not to manipulate preferences, for example based on personal identity: the more you manipulate my values, the less the future person is me. To be a prudential benefit, then, the welfare gain has to outweigh the degree of identity loss. Moreover, itās plausible that extrinsic manipulations are typically more disruptive to oneās degree of psychological continuity than voluntary or otherwise ānaturalā character development.)
It seems worth flagging that some instances of replacement seem clearly good! Possible examples include:
Generational turnover
not blindly marrying the first person you fall in love with
helping children to develop new interests
I guess even preference-affecting views will support instrumental replacement, i.e. where the new desire results in oneās other desires being sufficiently better satisfied (even before counting any non-instrumental value from the new desire itself) to outweigh whatever was lost.
I agree that some instances of replacement seem good, but I suspect the ones Iād agree with are only good in (asymmetric) preference-affecting ways. On the specific cases you mention:
Generational turnover
Iād be inclined against it unless
itās actually on the whole preferred (e.g. aggregating attitudes) by the people being replaced, or
the future generations would have lesser regrets or negative attitudes towards aspects of their own lives, or suffer less (per year, say). Pummer (2024) resolves some non-identity cases this way, while avoiding antinatalism (although I am fairly sympathetic to antinatalism).
not blindly marrying the first person you fall in love with
people typically (almost always?) care or will care about their own well-being per se in some way, and blindly marrying the first person you fall in love with is risky for that
more generally, a bad marriage can be counterproductive for most of what you care or will care to achieve
future negative attitudes (e.g. suffering) from the marriage or for things to be different can count against it
helping children to develop new interests:
they do or will care about their well-being per se, and developing interests benefits that
developing interests can have instrumental value for other attitudes they hold or are likely to eventually hold either way, e.g. having common interests with others, making friends, not being bored
developing new interests is often (usually? almost always?) a case of discovering dispositional attitudes they already have or would have had anyway. For example, thereās already a fact of the matter, based in a childās brain as it already is or will be either way, whether they would enjoy certain aspects of some activity.[1] So, we can just count unknown dispositional attitudes on preference-affecting views. Iām sympathetic to counting dispositional attitudes anyway for various reasons, and whether or not theyāre known doesnāt seem very morally significant in itself.
Plus, the things that get reinforced, and so may shift some of their attitudes, typically get reinforced because of such dispositional attitudes: we come to desire the things weāre already disposed to enjoy, with the experienced pleasure reinforcing our desires.
Good point about the degree of identity loss.
I think the hybrid view you discuss is in fact compatible with some versions of actualism (e.g. weak actualism), as entirely preference-affecting views (although maybe not exactly preference-affecting in the informal way I describe them in this post), so not necessarily hybrid in the way I meant it here.
Take the two outcomes of your example, assuming everyone would be well-off as long as they live, and Bob would rather continue to live than be replaced:
Bob continues to live.
Bob dies and Sally is born.
From the aggregated preferences or attitudes of the people in 1, 1 is best. From the aggregated preferences or attitudes of the people in 2, 2 is best, if Sally gains more than Bob loses between 1 and 2. So each outcome is best for the (would-be) actual people in it. So, we can go for either.
So, not all preference-affecting views even count against this kind of replaceability.
My next two pieces will mostly deal with actualist(-ish) views, because I think theyāre best at taking on the attitudes that matter and treating them the right way, or being radically empathetic.