In 'The Limits of Well-Being' (in The Good Life and the Human Good), Shelly Kagan claims that a person is nothing more than a body and mind, and then adds in a footnote (fn#7, p.182):
In contrast, it is not nearly as plausible to assert that a person's life is comprised solely of facts about that person's body and mind. This raises the intriguing -- and generally over-looked -- possibility that it might be one thing for a person to be well-off and quite another for that person's life to go well.
Now, it looks like 'success theory' and the intuitions supporting it, such as provided by my 'Molly the Mathematician' thought experiment, are really about what it is for a person's life to go well. If that really is a distinct matter from a person's welfare, then this severely undermines success theory as a theory of welfare.
Kagan goes on to claim (p.186, original italics, link added):
If something is to be of genuine (ultimate) benefit to a person, it must affect the person; it must make a difference in the person. That is, it must affect the person's intrinsic properties. Changes in merely relational properties cannot be what is of ultimate value for the person... What benefits the person must make some intrinsic difference in the person. Otherwise there would be nothing in it for him.
So, being genuinely successful, or having desires fulfilled in actual fact, might constitute the good life. But none of that benefits the person except insofar as it directly impacts them (e.g. through the happiness we feel when we believe our desires have been satisfied). Molly's life goes better in case #1, but she's better off in case #2.
I don't know what to think of all this, so I'd love to hear some other opinions. Can we really separate a person from their life in such way? Is it really true that benefits must be internal to the person? Can't I be made better off by things that happen in the world (that, say, improve the worth of my life)?
I'm reluctant to give up welfare externalism so easily. But Kagan notes that we may do so without actually giving up as much as it seems (p.189):
If this is right, then the importance of well-being might be less than we often take it to be. In many cases, the pursuit of external personal goods will be far more important than the pursuit of the internal goods that happen to comprise well-being. The more narrowly we understand well-being, the more likely that this is the case. If well-being is limited in its extent, then it may also be limited in its significance.
Should we really want to impoverish our concept of welfare in such a way? It won't really affect my moral views in any case. If I adopted this view, my consequentialism would change from welfare-maximization to, well, the maximization of what I used to call 'welfare', i.e. internal and external personal goods. What the point of such a semantic revision would be, I'm not quite sure.
Update: I now think Kagan's argument equivocates on different senses of 'affecting' someone. See here for more detail.