I want to be clear here. An agent is rational if his motivations can be rationally explicable whether we can deduce them or not, or whether she herself is aware of them or not.
Its a fundamental assumption to make. But I'm wondering philosophically, just how true this is. Can anyones motivations be so reduced?
It's the kind of assumption that one makes in physics, and it seems to have been spectacularly successful there. It seems to me that this supposition has been historically imported from that direction. Is this true?
To me, its more like a useful working hypothesis, to see what can be further deduced, than having any truth value itself