From: Stathis Papaioannou (email@example.com)
Date: Thu Mar 27 2008 - 05:24:24 MDT
On 27/03/2008, Lee Corbin <firstname.lastname@example.org> wrote:
> > You can go up as high as you feel is necessary. You suggested that a
> > 1000x1000 patch of the Board that was looked up rather than calculated
> > might cause a small deficit in consciousness. If that's not enough
> > then imagine that half the Board is looked up and the other half
> > calculated: surely that should be noticeable?
> Noticeable to whom? Here is what I think is going on: *naturally*,
> the subject, the person who's supposed to be having experiences,
> can report nothing different than what he would report anyway.
> All the states that his brain would have reached (under what I call
> "totally authentic computation") would still be reached. So in a
> certain sense, he *doesn't* notice any difference.
> But in another sense he does. But first, our entire hypothesis depends
> on the agencies running the experiment to be perfectly aware of when
> his brain states are being looked up, and when they're being computed
> by the usual routes of causality and local reductionism.
"But in another sense he does". I can see how you might say this if
there is a period of looked up computation (LUC) alternating with a
period of totally authentic computation (TAC): the person behaves
normally during the LUC, and during the period of TAC recalls the LUC
experiences, but these are actually false memories, as there were no
actual experiences during the LUC.
However, the case I was considering was where the TAC and LUC are
occurring simultaneously, in the two halves of the GOL board. This
might result in, say, zombie vision but normal hearing, or zombie
aesthetic appreciation but normal arithmetical ability. If you were
100% zombie no external observer would know but you would (in a sense)
know, since you would not be having experiences. But is it possible
that your brain is partly zombified right now, with the normal,
conscious part systematically deluded into thinking that it can see
when in fact it is completely blind?
If you allow the possibility of partial zombification then you have to
allow that not only can an external observer never know a person's
subjective experiences, but the person himself can never be sure that
he is having the experiences he believes he is having. This seems a
high price to pay in order to maintain that a LUC cannot be conscious.
> Now, let me phrase my answer using an experiment so that there is
> no mistaking my meaning. Question: would you prefer
> (A) to be tortured for an hour in the old-fashioned way
> (B) for records of such an hour merely to be retrieved from
> a galaxy far, far away a long time ago in which you were
> tortured just the same, and merely the states for that hour
> interval brough to Earth and at the proper moment merely
> looked up?
I guess it depends on whether the looking up of the records repeats
the experience which is what we are debating. If the recording were
accurate down to the atomic level then yes, I think it would reproduce
the same experience as the original.
> > The only (naturalistic) way to avoid this conclusion is if the brain
> > contains fundamentally non-computable physics, and there is no
> > evidence that it does.
> Just so. I would be surprised if this were at all controversial between
> you and me, or between us and our favorite correspondents.
The argument above proves that a functionally perfect simulation of a
brain must be as conscious as the brain *unless* partial zombification
is possible. The possibility of mind uploading is something else you
might have to give up if a LUC cannot be conscious.
-- Stathis Papaioannou
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:01:02 MDT