[From: Bruce Nevin ()Fri 921218 13:50:21]
(Bill Powers (921217.0930) to John Gabriel ) --
Don't we get something like discrete states in the latch
mechanism for steps in events and sequences? And indeed for
category perceptions and on up?
It seems to me that, as the strength of some category perception
grows, it becomes easier (more acceptable?) to fill in missing
category-attribute perceptions by imagination. At some
threshold, or seeming threshold, where imagined perceptions are
integrated with real-time perceptions, it is as though the
exemplar of the category is perceived as fully or truly present,
whereas before there were only unsupported signs or symptoms that
fostered a belief, readiness, or expectation.
There is an analogy to the relation of continua to discreta in
language. In language, pronunciations are continuous phenomena,
but speakers and hearers perceive words (morphemes) as (probably
event-level) sequences of discrete tokens, where the types are
sound contrasts established by social convention for their speech
community. Children probably learn a limited stock of words
first, and then learn the conventional contrasts and the
type-token relation of sounds to contrasts in words, which in
turn enables learning and recognizing a richer stock of words.
Language probably evolved in its first stages by such a route.
ยทยทยท
3. The idea of the rate of information transmission down a
discrete channel, being the the upper bound of the number of
binary decisions a recipent can make in a second, and first put
forth by Claude Shannon in the two 1949 papers in BSTJ.But all real neural systems work with continuous, not discrete,
variables. Neurons do not respond to incoming impulse streams by
making "decisions" but by harboring continuously-variable
chemical concentrations and potentials which in turn determine
the frequency of outgoing impulses. The computations done by a
neuron are analogue computations based on continuous internal
electrochemical variables. Certainly information theory could be
applied to these processes. But it doesn't help you model them.