From: Rowan Hamilton <rowan@arctic.org>

But it does seem to me that tossing out the old Copenhagen interpretation of the collapse of the wave function (which is what I was taught in college), and accepting the Everett interpretation (which I know a bit about) *does* seem to imply, at least in my naive perspective, the requirement of a self-aware observer. (Is there any other kind of observer?)

Hmm, maybe you could describe how. I thought Everett *got rid of* any special role for measurement, observer, etc. Instead what you get by with is correlation. So, two particles that haven't interacted, or a particle and a detector, start out having independent distributions of (phasey) states they can be in, but once they interact, their distributions are correlated. So you might start with particle spin up or spin down, crossed with detector detects it or not, so the world is spread over all four combinations: yes | ????? | ?????? detector ---+-------+-------- no | ????? | ?????? ---+-------+-------- | up | down particle But then they interact and it narrows down to (say): yes | ????? | detector ---+-------+-------- no | | ?????? ---+-------+-------- | up | down particle Which divides the distribution into two parts that aren't going to interact anymore. I am missing a lot here but the point is that this is just following the regular rules for the evolution of the wave function and yet the world has split into two separated worlds, one where the particle went one way and was detected, and one, where it went the other way and wasn't and won't be. And sure, if there are observers, in one world they'll see (the effects of) the event definitely happening, and in the other, definitely not, but any observers are latecomers to the action as it were, not particularly involved in or necessary to how it works. The world just really splits, and we inhabitants of the splinters just see (individual examples of) the results. --Steve