On 3/11/2012 5:43 PM, acw wrote:
On 3/11/2012 21:44, R AM wrote:
This discussion has been long and sometimes I am confused about the whole
point of the exercise.

I think the idea is that if comp is true, then the future content of
subjective experience is indeterminated? Although comp might seem to entail
100% determinacy, just the contrary is the case. Is that correct?
3p indeterminacy in the form of the UD*, 1p determinacy from the perspective of those minds relative to bodies in the UD*.

However, I think that if comp is true, future experience is not only
indeterminate, but also arbitrary: our future experience could be anything
at all. But given that this is not the case, shouldn't we conclude that
comp is false?
You're basically presenting the "White Rabbit" problem here. I used to wonder if that is indeed the case, but after considering it further, it doesn't seem to be: your 1p is identified with some particular abstract machine

This identification, I think, is the identification that exists between a collection of bisimilar computations and a given "Virtual reality simulation" as the content of a first person experience.

- that part is mostly determinate and deterministic (or quasi-deterministic if you allow some leeway as to what constitutes persona identity) in its behavior, but below that substitution level, anything can change, as long as that machine is implemented correctly/consistently. If the level is low enough and most of the machines implementing the lower layers that eventually implement our mind correspond to one world (such as ours), that would imply reasonably stable experience and some MWI-like laws of physics - not white noise experiences. That is to say that if we don't experience white noise, statistically our experiences will be stable - this does not mean that we won't have really unusual "jumps" or changes in laws-of-physics or experience when our measure is greatly reduced (such as the current statistically winning machines no longer being able to implement your mind - 3p death from the point of view of others).

Also, one possible way of showing COMP false is to show that such stable implementations are impossible, however this seems not obvious to me. A more practical concern would be to consider the case of what would happen if the substitution level is chosen slightly wrong or too high - would it lead to too unstable 1p or merely just allow the SIM(Substrate Independent Mind) to more easily pick which lower-level machines implement it (there's another thought experiment which shows how this could be done, if a machine can find one of its own Godel-number).

I agree with ACW. The difficulty in understanding all of this is that we are effectively abstrating out all references to the physical world via a Platonic interpretation of the Church-Turing Theorem. Any time we talk of a "machine", we are really talking about an equivalence set of bisimilar computations in the sense that all of the computations in the set are equivalent in that they all act to generate the same output.

Onward!

Stephen

--
You received this message because you are subscribed to the Google Groups 
"Everything List" group.
To post to this group, send email to everything-list@googlegroups.com.
To unsubscribe from this group, send email to 
everything-list+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/everything-list?hl=en.

Reply via email to