Re: minimal theory of consciousness

From: <hal.domain.name.hidden>
Date: Fri, 16 Jul 1999 22:58:57 -0700

Wei Dai, <weidai.domain.name.hidden>, writes:
> We are still pretty far away from such a theory, so perhaps we should ask
> instead what is the minimal theory of consciousness that we need in order
> to make practical decisions? I argue that all we need is a theory that for
> each object or process, give the person using the theory the probability
> that the object or process gives rise to that person's consciousness. In
> other words, it should tell me the probability that being some object or
> process feels exactly like the way I feel right now.
> [...]
>
> The minimal theory of consciousness is the link between a theory of
> everything and decision theory. In order to derive useful information from
> our subjective experiences, we need to know the conditional probability
> that we feel the way we do under various assumptions. Given a set of
> assumptions, a theory of everything would produce a probability measure P
> over the set of all possible objects or processes. Now to get the
> probability that I feel the way I do, I compute sum_over_all_x P(x)Q(x),
> where Q(x) is the probability that being x feels the way I do.

Do you think this variant would work. Suppose that there are multiple
possible distinct universes, forming a set U of all possible universes,
and a probability measure P() defined over elements of U, which tells
how much contribution that universe makes. I think that is the direction
many of us start from.

Now suppose we need only answer the question, does anything in each
universe from U give rise to my consciousness. Q(x) where x is selected
from U is the probability that universe x instantiates my consciousness
(I'm not sure whether we'd get probabilities other than 0 and 1 though).

My motivation is to avoid the replay problem as it seems that so many
of these paradoxes involve replays. I only have to decide whether my
consciousness is ever instantiated in a universe, I don't have to decide
whether each replay is separately conscious.

I can then apply your formula, letting x vary over all universes in U,
computing sum over x of P(x)Q(x). I don't fully understand the meaning
of the result, "the probability that I feel the way I do", but I wonder
if this would be a valid alternative way of getting to it.

Hal
Received on Fri Jul 16 1999 - 23:07:04 PDT