Re: [Phys-l] disambiguating the terminology

[John Denker <jsd@av8n.com>]

On 02/16/2012 11:43 AM, curtis osterhoudt wrote:
> See this: http://www.youtube.com/watch?v=4b71rT9fU-I  , especially starting at the three-minute mark. 

As a further step in that direction:  Try explaining to an 
Italian person the difference between a ship and a sheep.
Or try explaining to a Chinese person the difference between
fly and fry.

There are actually good biophysics explanations for how this
happens ... and indeed why it *must* happen.

The key idea is that many calculations in the brain are done
using two-dimensional /maps/.  They're called cognitive maps.
Even language is represented this way.  

All this is related to the explanation of why the cortex is 
convoluted:  the convolutions serve to increase the amount
of two-dimensional surface area, i.e. to make more room for
cognitive maps.

This has some amazing consequences.  There is a theorem 
(Brouwer, 1911) that says you cannot map something from one 
dimension to another in a way that is one-to-one and continuous.
I call this the "flower pressing theorem".  So if you map
something that has three or four dimensions (the color signal
from the retina) onto the two-dimensional cortex, you are
absolutely guaranteed to have discontinuities, ambiguities,
or both.  Ditto for the representation of speech, which is
has at least six dimensions (pitch plus five formants).

Teuvo Kohonen has written a bunch of books about cognitive
maps.