What the Neocortex Does

stephan anagnostaras stephan at ucla.edu
Fri Aug 4 12:42:17 EST 2000


In article <1eeoqcp.ad4vehlu3dxcN%herwin at gmu.edu>, herwin at gmu.edu 
(Harry Erwin) wrote:

> Bill Skaggs <skaggs at bns.pitt.edu> wrote:
> 
> > herwin at gmu.edu (Harry Erwin) writes:
> > 
> > > I'm beginning to suspect a primary function of the neocortex is the
> > > maintenance of a non-symbolic, dynamic model of the environment. The
> > > evidence seems fairly strong that bats live mostly in such a model
> > > (Griffin, 1958), and the difficulty we have training bats to use
> > > symbolic representations (yes/no tokens) in communicating with us,
> > > suggests that their internal model is non-symbolic, possibly like the
> > > analog models used in some contexts like wind tunnels and floatation
> > > tanks. 
> > 
> > Hi Harry!
> > 
> > I gather from this that you're having trouble training a bat to do a
> > task.
> 
> Hi, Bill!
> 
> Bats can be trained to signal 'yes/no', but it's a pain and a half. It
> doesn't come naturally to them.
> 
> > 
> > Anyway, you may be right but I don't think the argument is convincing.
> > Whatever sort of environment-model bats have, they certainly are
> > capable of using it to make at least some binary distinctions (e.g.,
> > moth/leaf).
> > 
> >       -- Bill
> 
> If I had a convincing argument, I'd do some sort of experimental test
> and publish a paper. I can e-mail you something speculative if you want.

I bet you they wouldn't have trouble telling the difference between a 
cave that contains a predator and one that doesn't. Or a poisonous 
insect from one that is palatable.  My point is, the bat probably 
doesn't care about the problems you are giving it, and doesn't care to 
represent them symbolically either, but this doesn't mean their internal 
model is not symbolic. 

Anyway,
Later,
Stephan






More information about the Neur-sci mailing list