Douglas,

This argument had parallels in a debate that popped up repeatedly in the 1980's in the fields of psychology and economics, over whether or not it was actually possible for a person to be selfless.

On one end (represented primarily by Robert Cialdini, at least in the field of psychology), researchers contended that since compassionate acts almost always benefited the self in some way (i.e. through building relationships, improving ones status, confirming ones moral views, or just plain making oneself feel good) then compassionate acts weren't truly altruistic.

The opposite side (represented largely by C. Daniel Batson) pointed out that all of those views assumed the primacy of rationality--that is, that humans were first and foremost a rational actor, and that all of their decisions were calculated choices. With those as priors, any benefit to the self became a calculated benefit and one could argue that it was selfish.

Remove the idea that rationality is primary (which Hume did) and you open yourself up to another, much kinder possibility--that humans have a set of irrational (or, perhaps better, pre-rational) impulses that lead them to prefer kindness and compassion in some situations. Those pre-rational impulses are what we usually refer to when we say someone is acting in a selfless way. And evolution seems to have fashioned them over time so that they tend to be beneficial to people on both sides of the emotion.

I'm sure that the field of philosophy has more parallels to this, since much of psychology is shaped by philosophical ideals (usually in an impoverished form). We have an entire branch of our field--one of the largest--which was formed based off of Wittgenstein (and I was delighted when I found that out).

J