> So what do you think about replays in the causal model, where the replay
> has actual calculations occuring? Suppose you have a brain and you
> give it a Turing test, asking questions and so on, and are satisfied
> that it is conscious. Now you (somehow) reset the brain to the original
> state and run it through exactly the same Turing test again, giving it
> exactly the same inputs so that the brain runs through exactly the same
> calculations and produces exactly the same outputs. Everything is as
> it was for the first run. The brain is calculating and communicating
> and causing and everything else, but it is doing it exactly the same as
> it did the first time.
Two important points:
1) It is probably impossible, even in principle, to "reset" a human brain to a previous state. If there are subsystems that display chaotic behavior (and there probably are) this operation would require the simultaneous manipulation of every particle in the brain with infinite precision, and an equally impossible manipulation of its environment.
2) Even if you could reset the brain, there is no garuntee that it will give the same answers the second time around. It will probably give a very similar performance, but the combination of truly random natural processes and chaotic systems means that a repeat performance is not inevitable.
Billy Brown, MCSE+I