[Martin Taylor 950327 11:10]
I'm here for a couple of hours today, and a bit of Friday, and then not
until April 18, so this will probably be my last contribution to the category
discussion, and it has to be short.
Bill Powers (950324.0905 MST)
Your description of what constitutes a category perception is very close to
mine except for some points that are probably significant.
the
same configuration, BANK, can be used to refer to two very different
things, the bank of a river and the bank in which you keep money. To
those we can add the bank of a turning airplane and the actions we call
banking a fire or banking a billiard shot off the side-cushion. As we
normally use the word BANK, it seems to participate in the thing we are
talking about; it almost IS the thing we are talking about. It is as
much a part of the main subject as the ball's shape is part of the
bouncing.
There are those who think in words, and those who cannot imagine (literally:-)
how that can be done. I know from talking to other people that many of
them claim to think in words. I do not know what that could feel like.
You arrogate the word into the same category as the concept to which it
refers. I don't. I take the word to be a category, different from all
other words, and the various kinds of "bank" to be categories on their
own, categories that would be the same categories whether I knew a word
for them or not. The word-category /bank/ is _associated_ with the
concept-category /place of money/, and with /a type of billiard shot/, and
with /land beside a river/ and with ... The word-category is not part of
those concept categories, for me, and as far as I can introspect, the
word-category does not participate in any thoughts I have when dealing
logically with those concept-categories.
All the chairs seemed similar,
even though each appeared at a different visual angle and each was
different from the others -- at the configuration level. The category
perception was momentarily visible by itself, in that short space
between noticing the similarity and the time when I thought "chairs."
I think you are describing a category perception of the same kind as I
try to describe, followed by a description of the dynamic operation of
the associative feedback loop that both gives you the perception of the
word category /"chair"/ and reinforces the perception of the concept
category /chair/.
Only what went on in my thoughts was not that literal
sentence, "I'll have to tell Martin...". It was just a sort of "Ah."
Yes. That's about as closely as I could put these kinds of perception into
words. But isn't that perception itself a kind of category perception? A
category of perceived kinds of event? You didn't know how you were going
to tell me--but you perceived the imagined successful telling. Any of
many ways would have served the purpose, but none were part of the perception
itself, I assume. And you had not associated any words such as "I'll have
to tell..." with the perception of /having to tell/.
At the moment, you feel (or felt when you wrote) that a word _is_ a
category. My claim is that a word is simply a low-level perception that
_is used to point to_ a category perception, and that the category
perception itself is not a word.
As you may see, I agree with the second sentence, which is the opposite of
what I previously had construed you to be thinking. But I disagree with
the first sentence. A word IS a category, in just the same sense as /chair/
is a category before a word "chair" is perceived to belong with it.
So far as I can see, that's one point of disagreement. The other is whether
these derived categories (as I call them) such as /chair/ or /bank/ can
be (or are) formed without being based on what I have called "natural
categories." I think this is a discussion that has to wait until I have
more time; for now, I think my recent postings have said all I can say without
going into more detail than I can now afford.
As you can see, I agree with most of your comment to Bill Leach. One thing
I would do, though, is to assert that your diagram refers to derived
categories, and in it the "OR" function should be "Logical function" and
should not (necessarily) have "word configuration" as an input. So here's
my version of your diagram (I added the output side not because I thought
you forgot it, but because it is often ignored in the discussion):
> > category reference signal
> >
> category signal (wordless)
>------------------O--------
> >
=========== Output
>> Boolean || function
>> function || / / | \ \
=========== To both analogue
/ / / and category control
/ / / system reference inputs
perceptions / / / | | | \ \ \
being
categorized
Basically all I'm trying to do is model how words can come to
designate categories.
Yes, I can accept that your model is one way. But then, so is the associative
connection, and that type of connection also accounts for the widespread
branching of associations that occur for any category perception, whether
it be a verbal category or a non-verbal one. (As I remember, Alice's
comment on Jabberwocky was that it filled her head with pictures, but she
was not sure of what--the nonsense words did not have associated concept
categories, but had other perhaps verbal associations that did).
We'll bring the Looking Glass back next month.
Martin