Re: death of the mind.

From: Glen Foy (spam33_at_butter.toast.net)
Date: 08/30/04 

Date: Mon, 30 Aug 2004 13:14:29 -0400

"Lester Zick" <lesterDELzick@worldnet.att.net> wrote in message
news:41334b60.35181257@netnews.att.net...
> On Sun, 29 Aug 2004 19:08:11 -0400, "Glen Foy"
> <spam33@butter.toast.net> in comp.ai.philosophy wrote:
>
> >
> >"Lester Zick" <lesterDELzick@worldnet.att.net> wrote in message
> >news:4131f1fe.18660890@netnews.att.net...
>
> [. . .]
>
> >> Glen strangely finds computation to be a metaphor but does not find
> >> behavior to be a metaphor. However, it would be a little more helpful
> >> to specify what kind of computer you think the brain is. Typically
> >> computationalists think the brain is a turing machine. Behaviorists
> >> think the brain is a behaviorcule. The problem reduces to what kind of
> >> machine the brain is and how it functions to produce the kind of
> >> effect it does. When behaviorists address the problem they just assume
> >> the brain is some kind of machine but are vague and inexact when asked
> >> what kind of machine it is. On the other hand, computationalists think
> >> the brain is a computer without being able to say what it computes or
> >> how.
> >>
> >> Regards - Lester
> >
> >That strikes me as a fair appraisal.
> >
> >As far as "what kind of computer you think the brain is", many
> >computationalists would say that at the lowest level it is something like
a
> >pulsed artificial neural network. But the brain is surely composed of
> >countless such networks, and all the really interesting questions, it
seems
> >to me, revolve around how these networks interact and influence each
other,
> >how the networks are layered, the highlevel architecture, not the
building
> >blocks. Being able to recognize the shapes of letters is lightyears away
> >from natural language understanding.
>
> Of course. The difficulty is that most computationalists naively
> assume that we're already able to recognize letters in cognitive
> terms. Certainly we can robotically, but it's by no means clear that
> all we have to do is architect high level robotic circuitry the same
> way to produce cognition. In other words, is cognition the same as
> computationalist robotics or is there some other mechanism at work
> which produces both robotics and cognition?
>
> The real problem is that as turing robotics currently stand, the field
> is the equivalent of neo Pythagoreanism where a finite binary number
> is simply assigned as the equivalent of cognition when in point of
> fact it is nothing more than a robotic equivalence. When this problem
> is recognized, addressed, and resolved, then we'll be able to concern
> ourselves with higher level design architecture and not before.
>
> Regards - Lester

>From the computationalist's point of view, I would argue that congnition is
a pattern of activation, involving many individual neural nets. The pattern
is hierarchical and highly structured, and is characterized by a high level
of unity and variety at every level in the hierarchy.

This is much more interesting than a binary number, but still may not be
what you're seeking.

Regards,
Glen