[aato]

This kind of stuff is fun to think about. In my opinion identity is based on behaviour, not composition. So long as an entity matches a given behaviour profile, then for all intents and purposes that entity IS the thing that the behaviour is normally attributed to. Kind of like Duck Typing: "If it walks like a duck and it quacks like a duck, then it must be a duck."

So my response to your car example is that both vehicles are "the same car" since both get you around and both are owned by you.

In that sense, as your behaviour changes so does your identity. If you were able to 100% accurately encode all the actions you could ever take into a Markov Chain, then I would argue that Markov Chain is what makes you "you" at that point in time.

[chii]

So if an alien abducted you, created a clone to replace you while you are experimented upon, you'd still agree that you are living normally, and nothing happened to "you"?

[diminish]

Few years ago, I was in a PR show for a shopping center where Paris Hilton showed up. Everyone was interested to see her. It turns out she was not the true Paris Hilton but a doppelganger.

[Sacho]

Conversely, people usually go through profound behavioral transformations several times in their lives. Does that mean they are not "themselves" anymore?

[aato]

Very broadly speaking, I would say yes. Again this is just my take, but I argue that different sets of behaviours are indicative of different identities.

For example, at work I may keep to myself and be very quiet - in that setting my actions would define me as an introvert. With my friends I may be more outgoing and talkative, which is more characteristic of an extrovert. So I have two identities depending on my behaviour. I am not my introverted self when with friends, and not my extroverted self when at work.

[aato]

Interesting! I would say that I am living both normally and abnormally since "I", in this case, is referring to two entities which are indistinguishable in terms of their behaviour.