Re: new paper on computationalism

From: <hal.domain.name.hidden>
Date: Mon, 8 Jan 2001 12:04:47 -0800

Jacques Mallah writes:
> Hello. I finally wrote my paper for the Plank Symposium (which I spoke
> at in October). In it I propose a new critereon for implementation of a
> computation, and mention the everything-list. I put the paper on my site at
> http://hammer.prohosting.com/~mathmind/100y.htm

It has many interesting ideas and covers a lot of ground but seems a little
unfocused. A few random comments:

The issue of what constitutes an implementation is complex and difficult
to understand unless you have read the relevant papers as background.
Hopefully your audience is familiar with the issue otherwise your paper
is hard to understand.

All these physically-oriented proposals have an ad hoc flavor, an
attempt to "cook the books" to get the answers we want. They do not
flow naturally from first principles.

I like your suggestion to pursue the question of whether one computation
simulates another. This gets away from the physical world and forces
us to focus on more fundamental issues. If you could come up with a
persuasive way of answering this question you might then be able to
leverage that and improve the analysis of implementation. For example,
if a clock implements every program, then a counting program simulates
every program. Find a way to reject the latter and you are on your way
to rejecting the former.

The overall program you are proposing is defined here:

   Suppose, hypothetically, that the problems in defining a unique and
   proper implementation criterion can be overcome. The next step in
   applying computationalism to quantum mechanics would then be to see
   if the appearance of the standard rules for probability could be
   derived using the wavefunction without collapse.

   Such probabilities would only be effective probabilities, because
   there is nothing non-deterministic in the model to be considered. An
   effective probability in this context is defined as the ratio of
   the number of observations of a given type to the total number of
   observations. In this case, this will be a ratio of the number of
   implementations of a computation of a given conscious type to the
   number of implementations summed over all conscious types.

As a practical matter, this plan has so many obstacles to overcome that
it looks hopeless. Not only do you have to resolve the issue of when
an implementation has occured (which is of doubtful success), you then
need to define what class of computations constitute conscious observers
suitable to your needs. This second problem is equally contentious and
far from resolution today.

Any such definition, if could ever be agreed upon, is going to be highly
complex. Now you have to go from this characterization of consciousness
to try to relate what constraints it puts on how the wave function should
be said to split. I don't see how you could possibly expect to come
up with meaningful statements about quantum theory which are contingent
upon whatever specific aspects of a computation render it conscious.

This struck me as funny:

   A disturbing possibility is that the measure of a brain (the # of
   ways in which it could be said to implement its regular computation)
   could depend on the size or structure of the brain. It is conceivable,
   for example, that a creature could be constructed which thinks
   like a human, but whose brain does not implement its computation,
   as mathematically defined, in nearly as many ways. Such a creature
   would be at great risk of enslavement by regular humans. The opposite
   scenario would also be possible. It is to be hoped that measure is
   not sensitive to the details of the construction of a brain.

Presumably your concern is that if people come to believe that a
particular creature's brain has less measure than others, we should care
less about the welfare of that creature, and so would feel comfortable
in enslaving it, since its suffering wouldn't matter. It seems absurd
to imagine that the majority of people would allow these untestable
philosophical musings to drive their ethical judgement on such an
important issue.

Anyway, even if it happened, by your own arguments, mistreatment would
be justified, hence there is no reason to hope for the contrary.

Overall I think there are interesting ideas here, but the problem
that it is working towards is so difficult that I don't think it can
succeed as a practical matter. Some QM approaches such as "many minds"
try to accomplish much the same task without hoping to solve all these
insurmountable philosophical problems. They look for wavefunction
splits that provide basic properties that would allow for causality,
consistency, memory, etc. These are simpler principles and more likely to
put practical and meaningful constraints on wave function decomposition.

Hal
Received on Mon Jan 08 2001 - 13:11:34 PST

This archive was generated by hypermail 2.3.0 : Fri Feb 16 2018 - 13:20:07 PST