Re: consciousness based on information or computation?

From: Wei Dai <weidai.domain.name.hidden>
Date: Sat, 13 Feb 1999 17:21:47 -0800

On Sat, Feb 13, 1999 at 07:29:15PM -0500, Jacques M Mallah wrote:
> I do not believe your assumptions are simpler than mine. A string
> is less complicated than a computation, but it is a greater leap from a
> string to conciousness, and the illusions of dynamics and decisions, than
> it is from computations to consciousness.
>
> 0 --> strings ------------------------------------------> consciousness
> 0 ----------------------------> computations -----------> consciousness

That's not the way I see it, which is closer to:

0 --> strings = consciousness
0 ----------------------------> computations = consciousness

I don't understand what this "leap" is. Either strings can be concious, in
which case consciousness is just as complicated as strings are, or only
computations can be conscious, in which case consciousness is just as
complicated as computations are.

> > I assume you have read chapter 4 of Li and Vitanyi by now.
>
> I'm still on Ch. 1; I admit I should find the time to read it.
>
> > The answer to
> > your question is to ignore the time direction, or assume all directions
> > orthogonal to the tape direction are continuous. This may seem arbitrary,
>
> Indeed it does. And what happenned to your cloning scheme?

But my point, which you quoted below, is that it doesn't matter. You can
use the cloning scheme, or any other scheme you care to think of, as long
as the final distribution that is produced is recursive (see chapter 4 of
Li and Vitanyi) it will be close to this one (i.e., the one produced by
ignoring all directions orthogonal to the tape direction, because it
is the universal a priori distribution.)

> > but as the theorems in that chapter show, the details of the model don't
> > really matter, because the distribution you end up with it is universal,
> > which roughly means that no matter what model you use, as long as the
> > distribution it produces is enumerable, it will not be very different from
> > the universal a priori distribution.
>
> That doesn't seem to relate at all to the above question.

Do you see the connection now?
Received on Sat Feb 13 1999 - 17:22:56 PST