The trouble with the model is that modes of moral behaviour are not sufficiently nuanced to discriminate between persons. Let's say I have two people with alzheimer's . Let's say that neither remember who they are. And let's say they're both altruistic utilitarians. Now further, let's put them in a room where their life partners/spouses can't see them, but can write letters to each other on computer. Would either spouse be able to tell whether they're talking to their partner with alzheimer's, or not? I think not. So unless you can give a nuanced scale of what you mean by 'moral'... it's not going to discriminate between individuals. Individuation is, in my view, best done with genetics. However, memories aren't a bad model either. In my thought experiment here, spouse A can't tell if she's talking to partner B or partner A (with alzheimer's) unless either of them can remember *something* of their previous life, e.g. their partner's name. Obviously, SEEING their partner will individuate; but the physical form is genetic. Hence, genetics is the only true individuator.