machine brains

Michael Edelman mje at mich.com
Mon Mar 1 09:05:07 EST 1999



dmb106 wrote:

> On Thu, 18 Feb 1999, Michael Edelman wrote:
> >
> > > ..."Thinking" occurs in mammals. It is the interpolation of extra synaptic
> > > events between sensory input and motor output. ....) Don't waste time talking
> > > about our
> > > awareness of the brain's thoughts. That belongs to religion.
> >
> > You're  defining the problem down to a much simpler, and much less interesting
> > one. Self-awareness is the heart of conciousness.
> >
> > We all know we're self-aware, and there's no need to invoke metaphysics to
> > realize that humans and many other animals can plan- which is to say they can
> > create internal, counterfactual models in their mind and experiment with them.
> >
> >  You seek to explain the brain as a robot with a strict mechanistic theory. But
> > if that's the case, who is writing this note?
>
> The robot is. You are nothing more that an intentional system. We
> ascribe beleifs and desires to you because that gives us most
> predictive power and understanding. Does that make the beleifs and
> desires we ascribe real?

If you are the intentional system, you are doing the writing. If you maintain that
the brain is doing the writing, you may as well say that the hand is doing the
writing. What is the prime mover?

>         I agree that self-awareness is the heart of conciousness, but
> if this is the case ask your self one more question...
> What is the difference between being self aware and behaving as though
> you are self aware?

To an outside observer, perhaps none. But our idea of the existence of self-awareness
is based on the notion that we all share the experience of self-awareness.

> ...A rat is aware that a red light means food, for example, and it
> behaves in such a way so as to eat. We do not, however, ascribe any
> internal life to the rat.

Though perhaps we should. It is not that we deny the rat an external life- only that
it is not necessary to ascribe this to explain most rat behavior.

Of course when we speak of conditioned behavior in lab rats we're talking descriptive
statistics and probabilities; we cannot with 100% accuracy predict the behavior of a
single rat. But that's also true of much human behavior. Macroeconomics is the
description of average behavior of large groups of humans; predicting individual
behavior is far from easy.

> - if you do then just think of a lower
> organism which could be similarly conditioned. We humans have the
> added ability of being aware that we are aware that a light means
> food, in some bizare experiment.

> It is my beleif that this infinate regress is what we call awareness.

Where do you get the infinity? We're aware that we're aware, and I suppose we could
conceivably be aware of *that*, but in general, we're not. At least, we don't, as a
rule, engage in infinite recursive metal descent. The problem with inifinities is
that they take an infinite amount of time to resolve...but we live finite lives and
make decisions in finite time.

> The fact that it is an infinate regress is not key, it just makes it
> intractable.  Just think about what it is like to be aware of
> somthing? i.e. aware that it is daytime, I dont think you have to be
> concious to do this. Then extend the idea and become aware that you
> are aware- why would you need to be concious at the next step? yet the
> effect of all this unconcious computing is what we call conciousness.

I don't think you're explained or described conciousness there. You've rather pushed
the problem back an infinite number of steps. Self-awareness is but one property of a
concious agent.

--
Michael Edelman     http://www.mich.com/~mje
Telescope guide:    http://www.mich.com/~mje/scope.html
Folding Kayaks:     http://www.mich.com/~mje/kayak.html





More information about the Neur-sci mailing list