AI and cognitive psychology rant (getting more and more OT - tell me if I should shut up)

Anton Vredegoor anton at vredegoor.doge.nl
Mon Nov 3 20:15:58 CET 2003


bokr at oz.net (Bengt Richter) wrote:

>Now consider the experience of being "convinced" that a theory "is true." What does that mean?
>
>Does our experience-medium get shaped repeatedly through layers and layers of abstraction and
>re-representation to where some bit of memory is sensed to have a comfortable stability w.r.t.
>sensations derived from a series of experiemnts, and we are satisfied by some pleasure-connect
>with this state of brain? Is it pain and pleasure at bottom, tied to subtle internal state
>sensations? Do we form theories like patterns of sand form on a vibrating membrane at quiescent
>spots for a given mode of vibration? Do internal cognitive dissonances drive the "sand" away
>from untenable positions in patterns of theory? Is our conscious experience-medium passive in
>being shaped by the transducer, or does it have its own properties? E.g., is our sense of
>connected-ness of a broken line in our visual conscious experience due to shaping of a field?
>I.e., due to our very experience medium having field-nature and naturally taking some form
>across gaps between features due to its own properties?

>From what I know about neurons (the cells that transport information
using long wire like connections) there are only two states they can
be in: "firing" or "not firing". If they're firing a chemical
imbalance travels down the wire and causes a release of neuro
transmitters when it reaches a synapse (the connection with another
neurons receiving end). After firing there is an obligatory pause in
which the transmitters are reabsorbed and the chemical balance of the
transporting wire is restored. The neurons are not connected one to
one, but many neurons work together to make another one fire. And this
firing neuron sends *its* signal to a lot of others too.

Contrary to what one might expect after reading this, the neurons do
not use a binary information transmission scheme if looked at from a
somewhat higher level. Information is transported by changes in the
*frequency* in which neurons fire. We are convinced that a theory is
true if there are certain groups of neurons firing in a frequency
pattern that corresponds to the theory.

So essentially we are looking for harmony, which means that no single
observational fact is enough for determining truth. That's why we not
only need to hear someone say "yes" but we also need to see their face
while they're saying it and we also take into consideration countless
little details of the overall situation when forming the theory: "this
person is agreeing to something".

A practical application of this theory could also be possible.
Consider a kind of device that has countless little movable pins where
one could lay ones hands on in order to feel the activation patterns.
Like some braille device but bigger, and instead of just conveying
letters it would convey complete "meanings" by vibrating or
alternating in a certain way. 

Since this device would be close to the way our brain works (frequency
patterns activating the tactile cells in our hands) it would possibly
be more effective than a computer screen to transfer a certain
category of messages. I don't know which kind of messages though, but
I suspect there are things that cannot be conveyed any other way, like
there is no natural way to convey sound messages using a computer
screen. 

If it were a two way real time system one could have some weird
conversations with distant people "over the teletactile line"! Of
course in order to convey some more detailed information one would
have to program the thing using Python :-)

Anton












More information about the Python-list mailing list