RE: Computability of consciousness

Nick Bostrom (
Thu, 1 Apr 1999 22:08:57 +0000

Hal writes:

But perhaps we would say that the causal
> relationship of these states is not the same, because *if* we had tweaked
> some aspect of Einstein's brain state at a given moment, it would have
> led to changes in the pattern of succeeding states. But making the
> corresponding change to the state of the couch would not have led to
> the same kind of changes in the succeeding couch states.

I think this is the idea behind David Chalmer's recent definition of what it means for something to be an implementation of a program (or computation). The causal structure of the implementation has to match the structure of the program. I don't have his book handy at the moment however.

Another example we may consider: Suppose we have a computer running an upload for 10 seconds. Suppose we computer has discrete time. Then we could in principle construct one computer (or one RAM chip) for each time step of the simulation; and we could load each of these RAM with the data that the RAM in the original computer had at the corresponding time. We can place all these RAM chips in one long row. In some sense, we then have what looks like the same structure instanciated in two places in spacetime -- once as a spatiotemporal process, and once as a spatial pattern. We presumably want to say that only the former gives rise to a consciousness. And the relevant difference seems to be that in the case of the process, the various states are causally connected, whereas with the spatial pattern that is not so.

Nick Bostrom Department of Philosophy, Logic and Scientific Method London School of Economics