[From Bill Powers (950323.0830 MST)]
Martin Taylor (950322 17:30)--
What you call "category" I might call a mapping between category
perceptions. There's a category of a kind we might call "symbol"
and a category of a kind that corresponds to a perception based on
something in the outer world. The label "red" is not a colour.
It's a word, and is of a category distinct from the word "smooth,"
though of the same kind.
A word is simply a configuration (visual) or an event (auditory).
Naming, I have conjectured, is the process of forming a category
consisting of the perceptions to be named and the perception of a word
(configuration or event). Once that perceptual function is formed, the
set of perceptions to be named can evoke a category signal, and the word
used as the name can evoke the same category signal. So higher systems
can control the category signal either by manipulating the nonverbal
perceptions which are perceived as instances of that category, or by
manipulating the word-configurations or events that also give rise to
the same sense of category. A category signal, as you note, is a single
structureless signal indicating THAT an exemplar of a category is
present in lower-level perceptions, but not itself being a member of
that category. Category signals are silent; they are not the words that
evoke them.
Thus some range of color-sensations is perceived as a particular
(unnamed) category, and eventually the three-letter configuration <red>
is made a member of that same category. This is the process of naming.
Once the name is included in the category-perception, the word can be
manipulated in lieu of manipulating color sensations, and as far as
systems above the category level are concerned the experience is the
same.
This concept requires, of course, that category formation be completely
arbitrary, so that any low-level perception can be used as a symbol for
any set of other low-level perceptions, via their common fate as inputs
to a single category-perceiver. This is how not only conventionalized
sounds and marks on paper can become words that evoke categories, but
how icons and hand gestures and squeezes (of affection or warning) and
throat-clearings and whistles and facial expressions and stamps of the
foot can evoke categorical perceptions.
When you see a category named "symbol", this category is formed from the
word <symbol> and a (large) set of familiar word-like perceptions. Once
this category has been formed, perception of it can be evoked either by
word-like perceptions or by the word-configuration <symbol>. This is the
order-reduction of which I spoke. What was formerly a category signal
evoked by any of a large set of words is now evoked by another word, and
that word can be manipulated, as a low-level perceptual object, in lieu
of manipulating the specific words perceived as examples of the
category. This process of verbal abstraction can, in principle, be
continued indefinitely. It does not involve higher and higher levels of
perception; it involves the same levels being used over and over.
In your convention for naming labels, you are unconsciously doing this
process of order-reduction. <Label> is a word, a configuration when
written. To say that <red> is a <label> is to say that <red> and <label>
evoke the same sense of category. The category signal itself indicates
that one exemplar is present; but the category signal is neither <red>
nor <label>. It is a nonverbal signal.
You seem to treat the higher logical operations as if they used the
labels of categories as arguments, rather than the category signals
themselves.
No, they use the labels, which themselves are low-level perceptions, as
the MEANS of manipulating category signals which are evoked by the
labels. The labels are inputs from lower levels to the category level.
The category signals are inputs to higher-level systems. When we "think
in words", we imagine labels at the lower levels, which evoke category
signals, which occur in sequences, which are inputs to logical reasoning
processes, which are inputs to principle-manipulating systems, and so
on.
Of course we are aware of all these levels of processes going on, so we
experience the words themselves along with the categories, and the
categories with the logic, and so forth, all collapsed into a single
experiential space. The words make all the noise; it is difficult to
experience the nonverbal processes that lie above them as something
other than the words, because those higher levels do not work in terms
of the words themselves. The words are noisy; the higher levels work
silently.
You don't accept that there exists a kind of perception that I call
"category," a perception that has nothing to do with words, though
much reinforced by verbal labelling.
Perhaps you can see now that this was an incorrect conclusion. In fact I
go farther than you have indicated: not only are categories directly
experienced as wordless signals, but sequences, logical rules,
principles, and system concepts are likewise experience wordlessly. We
manipulate words, which are low-level perceptions, as a means of
manipulating signals in these higher levels. I'm sure you can see that
this closed-loop approach eliminates the problem that seems inherent in
using words to talk about levels higher than language. In fact, it says
that the very same higher-level processes are used whether we evoke
experiences by using words or by manipulating the real world and our
present-time nonverbal perceptual inputs.
All of this conjecture is, of course, a different subject from that of
how categories are formed. The above is simply the best I have been able
to do in understanding how words and nonverbal experiences fit together
in a higherarchical PCT model. I don't expect it to make a lot of sense
to anyone else. You may see some similarity between what I call "order
reduction" and what you call "natural and derived categories."
···
-----------------------------------------------------------------------
Avery Andrews (950323) --
Ooops, somehow, spazzing out with my mailer, I managed to
rebroadcast Rick's `Re Relational Control message'. sorry!
You also seem to have broadcast to the wrong list???. Pitch accents??
Sanskrit??
-----------------------------------------------------------------------
Dennis McCracken (950322.2300 PST) –
You have my permission to reprint or cite from any of my writings on
CSG-L. Dag will tell you how to get those files. Probably the cheapest
way to get them would be to drive over to Dag's house in Valencia with
an armload of disks -- CSG-L generates over a megabyte per month of
typing.
Apropos of whatever: IOmega has just come out with their "Zip drive"
which will run off a SCSI port _or a parallel printer port_ (thus
becoming portable). It read/writes to 100 MB removable disks. The drive
is selling for $200 and the disks cost $20. I'm ordering a drive and
five disks today.
-----------------------------------------------------------------------
Best to all,
Bill P.