[Martin Taylor 991022 23:02]
[From Bill Powers (991022.1606 MDT)]
... My hangup is
that I can't see how the assignment of symbols happens -- that is, why and
how the signal indicating presence of a category somehow gets associated
with visual or auditory configurations, transitions, and events. The
mechanics of naming eludes me.
That was the point of what I called the "grand flip-flop" or some
such name. Association, and in particular the association of names
with perceptions (and of perceptions with names) falls out as a matter
of course. It doesn't have to be forced. You didn't like it when I
described the system before, so I didn't want to bring it up again
and obscure the main point of the architecture--which is that if one
allows the category level (and _only_ the category level) to get its
perceptual inputs from every level below, many of your misgivings that
you expressed about the hierarchy vanish in the wind.
A side-benefit of seeing the category level as a category interface
rather than as an interposed level is that it permits the construction
of higher analogue levels. Sequence, for example, can be analogue
and at the same time, on the other side of the category interface,
symbolic/digital, without having to see sequence as being either above
or below category, or as there being two distinct sequence levels
(one of your concerns was where the sequence level fits--above or below
category).
I'm not trying to force consideration of this architecture on you, or
to substitute it for the "standard model." It was just that you mentioned
some problems that had been bothering you, all of which are solved by
this architecture, so I thought it might be worthwhile bringing it up
again. If you prefer, I'll just drop it again.
Martin