Re: A troubling paradox...

Fred Mailhot <[EMAIL PROTECTED]> wrote in message news:<[EMAIL PROTECTED]>...
> Aside from the fact that I dislike the "what it's like" turn of phrase, I
> think there's a problem here with the notion that being "taught all of
> physics" really constitutes knowledge in the relevant sense...
> 
> 
> Anyway...does anyone have comments, or pointers to more discussion of this ?

Actually, I think the knowledge argument and variants do have merit in
pointing out to paradoxes which Russell was so fond of. We can dig
deeper.

The knowledge argument to my mind never showed the impossibility of
characterizing subjective experience. Rather, it showed the
categorical difference between a computation and a theory of
computation, or should I say the meta-language and object-language
distinction!?

Here, viewing the issue as distincton in mere human language leaves us
at the surface. Qualia of red is not readily expressible in human
language, *else* we would have gotten around the philosophy of
subjectivity much faster than the monumental mental fights that
involved some of the greatest minds in history of mankind!

As much as that is not a logical cause, it is convincing enough in my
view to dismiss the notion of a linguistic explanation for an adequate
theory of subjective color experience. I suspect that our language
does what it has been optimized for quite well, but when it comes to
expressing nature with the rigor of science it may fall short on a
huge problem such as subjectivity.

Note I am referring to a linguistic explanation in that the entities
in our theory of subjective experience themselves being linguistic,
not that our theory being denoted in human language which is somewhat
unavoidable as you may appreciate.

Incidentally, we should be able to construct analogies from hard
sciences that should serve us in clarifying the distinction in *types*
of knowledge. Let us first start with an unsuccessful analogy. Can a
theory of Newtonian motion lead a student to acquire knowledge of
motion in ways he has never observed or noticed? I believe the answer
is *yes*, in this case. A computer simulation could show him a very
accurate view of motion of spacecraft in orbital maneuvers. Surely
that is one thing not predictable *without* knowing laws of motion.
Note how this is a non-analogy for the knowledge argument. In this
case, the connection with the intuition was provided by a physically
accurate CGI animation based on laws of motion.

Can there be another example where the theory does not lead to a
change in the kind of subjective knowledge analogous to color
perception? I think so. Removing the mediation vehicle of computer
simulation in the previous attempt will not fix it for us. One could
argue that the student, by constructing *mental* images in his mind
according to theory could acquire an approximate sense of motion and
predictive power. Maybe he could make an animation of a simple
spacecraft maneuver himself.

Then, what shall be of interesting to us is how readily does the
theory provide the framework for achieving such mental imagery and to
what extent our brains are capable of such feats. I think not too
great. For instance, it is often cited that from the theory of general
relativity very few people have acquired an intuition of basic thought
experiments in space-time. This is in fact the capability to simulate
and thus predict the properties of tensors. I haven't had such a
"vision" and I find it unlikely that I might given my limited
knowledge of general relativity.

Still, there is a fine point in these physical theories in that they
provide for a *constructive* mode of thinking in which the agent may
construct from the mere cognitive substrate an imagery that didn't
occur to him before. Hence, there are two dimensions in which we can
get further away from theory-object mediation and these are complexity
and constructibility.

We can imagine many abstract theories that do not yield even an
*approximate* construction. That shall be troublesome for those who
wish to turn from the theory to the fine qualities of natural process.
Indeed, constrain one's attention to such high level psychological
terms such as desire and belief, and one shall find that the make of
these terms will never be revealed. Likewise, one can ponder how
effective a theory of color will it be for the agent when it has not
the means to mediate the theory to a simulation in its cognitive
space. Shall a too low level explanation in terms of neural
interactions be effective then? Perhaps that will be doable for an
agent that has such a flexible programmability that entirely new
universal machines may be defined. Is the human brain like that?
Perhaps so, but we do not know the means with which to program it. It
certainly isn't reading a dump of a neural network on a piece of paper
or holding a hard drive on top of one's skull.

These thoughts have exhausted me for the moment. We can carry on
another day hopefully. :)

Comments welcome,

Regards,

--
Eray Ozkural