>and eventually you might just end up in physics-land again.

Sure, anything might turn out to be explicable by future physics (or indeed, future astrology, future phrenology, or any possible future knowledge). There are no knock-down arguments in this domain. The point is that the fact of your personal conscious identity is apparently not a physical fact, which is a prima facie argument against some of the stronger forms of physicalism.

>To be more serious: having prior notions of things is fine.

You can't use the physical definition to dispense with other notions of personal identity if you need those very notions to bootstrap the physical definition (and maintain it over time, as your 'set of atoms' is constantly changing). To point out that your informal notion of self is connected somehow with a loosely-defined lump of physical stuff isn't really to succeed in reducing anything to physics. It's just to say what pretty much everyone agrees with: that humans have physical bodies.

> You can't use the physical definition to dispense with other notions of personal identity if you need those very notions to bootstrap the physical definition

Now that you mention, I don't need to. "I" simply means "the person who is currently speaking" (or in this case, writing). This already determines, beyond reasonable doubt, which physical body we're talking about. No other definition needed, since you can't say "I" without speaking. QED. Sorry, I should have noticed this 4 comments earlier.

(Caveat: this definition doesn’t feel satisfying, which might be the source of our disagreement, but I can’t put my finger on why - perhaps you can)

> the fact of your personal conscious identity is apparently not a physical fact

The human brain is astonishingly good at BS-ing itself, so I wouldn't attribute much to how you intuitively feel about your own conscious self. To me, this smells like a more abstract and philosophical version of the puddle analogy.

> "I" simply means "the person who is currently speaking"

But on this definition you can raise the question of why the person who is currently speaking is the person who is currently speaking (which by alternating de re and de dicto interpretations can be understood in the same non-tautologous ways as e.g. "Why is the person who is currently the President the person who is currently the President?") So this linguistic analysis doesn't succeed in dissolving facts about personal identity into nothing. You still have the fact that you are who you are and not someone else. And the intuitive notion of personhood (on which your definition depends and which you haven't in any way reduced to physics) is intimately connected with the idea of a more-or-less uninterrupted stream of first-person conscious experience.

> I wouldn't attribute much to how you intuitively feel about your own conscious self.

If we set 'intuitive feelings' aside, then of course the whole problem of conscious experience disappears immediately.

> And the intuitive notion of personhood (on which your definition depends and which you haven't in any way reduced to physics) is intimately connected with the idea of a more-or-less uninterrupted stream of first-person conscious experience.

"person": a member of the genus Homo

"speaking": phenomenon wherein the previous object vibrates its own vocal chords and moves its own mouth to produce sound waves

"currently": a point in time in which the aforementioned act occurs

Notably missing are: any references to continuity of experience, conscious action, or even thought. Suggestion: I have, in fact, completely reduced it to physics.

> You still have the fact that you are who you are and not someone else

Please describe to me a hypothetical universe where that statement doesn't hold true. If there is none, then we must accept that this "fact" does not require any explanation, and therefore can be dismissed as trivial.

Your linguistic analysis of 'I' doesn't work if understood in purely physical terms. Lots of people are speaking at any given moment, but I don't refer to all of them if I say 'I'.

Consider the absurdity of the question “who am I?”, uttered by an amnesiac, if it means simply “Which lump of physical stuff is currently speaking?” Even if we ignore the problem mentioned in the first paragraph, the answer is trivial – the same lump that’s making the noises. But it’s difficult to accept that the amnesiac is confused over a triviality: they really don’t know who they are; they lack a key piece of non-trivial, non-tautologous knowledge.

>If there is none, then we must accept that this "fact" does not require any explanation

As I've tried to emphasise, I do not claim that the fact is puzzling or requires an explanation, only that it is a fact. By analogy, it is not puzzling that the sum of two and two is four, and no non-trivial explanation of this fact appears to be required, but it is still a fact. Now of course you can try to make such fact dissolve via various forms of philosophical therapy. For example, there are some philosophers of mathematics who would try to argue that all mathematical facts are tautologous. You are trying to do something similar with facts of personal identity. However, philosophers of mathematics who claim that all mathematical truths are disguised tautologies can point to lots of actual work that has been done to show that this position is more plausible than it might first appear.

> Consider the absurdity of the question “who am I?”, uttered by an amnesiac

In that context, that question would just mean "what was my life like prior to losing my memories?", wouldn't it? So in this instance, the missing non-trivial knowledge is their long-term memory, not a problem with understanding the definition of "I".

> For example, there are some philosophers of mathematics who would try to argue that all mathematical facts are tautologous. You are trying to do something similar with facts of personal identity.

I'm not arguing that the subjective experience of personal identity is tautologous, I'm arguing that it's a chemical reaction (EDIT: and that the underlying objective truth is tautologous). Any feeling about the intuitive concept of "myself" having some kind of esoteric component above the physical matter (which I also share, don't be mistaken) would, presumably, also be given voice to by a computer simulation of your brain. That would be your brain being caught red-handed in the act of lying to itself about how special it is. And if this turns out not to be the case, then we have rock solid evidence that something is indeed missing from our simulation, which gives us an opportunity to measure that something, which in turn makes that something physical.

> Your linguistic analysis of 'I' doesn't work if understood in purely physical terms. Lots of people are speaking at any given moment, but I don't refer to all of them if I say 'I'.

Valid point. Refinement: in any speech, "I" refers to the person making that speech.