[From Bill Powers (2003.09.27.0855 MDT)]
Bruce Nevin (2003.09.26.2249 EDT) --
As Rick correctly points out, you can't use the words fluently and
appropriately unless you have their meanings. Yes, meanings are also in
the realm of perception. But they are not identical with those perceptions
that we take to be the referents of words.
I think that people can use words fluently and appropriately even when the
only referents of words are other words, or even nothing at all but
familiar sounds. An excellent example is in a science-fiction story called
"The Gostak and the Doshes," by Miles J. Breuer, M.D. (Amazing Stories,
March 1930). Here is a sample, from the part where a man gets transported
to another world and is learning from its inhabitants. He has heard someone
saying, "The gostak distims the doshes", and asks a new friend
···
==========================================================================
"John, what's a gostak?"
...
"The gostak!" he exclaimed. "Hooray for the gostak!"
"But what _is_ a gostak?"
"Not _a_ gostak. _The_ gostak. The gostak is -- the distimmer of the doshes
-- see! He distims 'em, see?"
"Yes, yes. But what is distimming? How do you distim?"
"No, no! Only the gostak can distim. The gostak distims the doshes, see?"
"Ah, I see," I exclaimed. Indeed, I pride myself on my quick wit. "What are
doshes? Why, they are the stuff distimmed by the gostak. Very simple!"
"Good for you!" John slapped me on the back in huge enthusiasm. "I think it
is wonderful for you to understand so well after being here only a short
time. You are very patriotic."
Patriotic? There is much more like this in the story, which I must have
read when I was about 9 or 10 years old. It made an impression on me,
obviously. Since then, it's been evident to me that people use a lot of
words that don't actually mean anything -- there is no perception to which
they point, other than the words they might use to explain or define them.
I also learned about the "trip around the dictionary" some time later: look
up any word with a meaning that seems a little vague, then look up the
words that are used to define it, and so on. You won't get very far before
the definitions start using the word you began with. Sometimes you will run
into words with clear perceptual meanings, so even if the trip is circular
you can tie the whole collection of words to some kind of experience. But
sometimes, especially with very important-sounding words like "patriotic,"
or "intelligence," you can't.
This should be obvious for another reason: a referent of a word is a
singular perception; its meaning encompasses any possible referent of the word.
You seem to be saying that a referent is a lower-level perception, while a
meaning is a category in which many lower-level perceptions belong
together. But can that be true? What are the referents of the word "learn?"
One is memorizing; another is practicing to achieve competence; another is
reacting correctly to a stimulus. So are you saying that there is a single
meaning for the word "learn" which encompasses all of these very different
referents? I can't imagine what that meaning would be.
I think we have a clear example here of a word that everyone uses, thinking
it must have some meaning, when in fact its meaning is very fuzzy and
faint. That word is "meaning." I've read more than one usage of the term
that makes it sound like something sticky that adheres to a word so it is
carried by hearing or vision into the brain. Apparently, to many people it
is a property of a word, as if a word were like an object that gets passed
around and that has properties of its own, independently of any speaker or
hearer, like the squeak without the wheel. "Meaning" seems to be used as if
it were something in itself, rather than a name for whatever a word brings
to our minds. Of course these are all examples of a word that does indicate
perceptions, but the perceptions vary all over the place and none of them
is very clear. And the more we talk about what "meaning" means, the more it
just means a bunch of other words.
The blind people learn the meanings of vision words entirely from the ways
in which these words are combinable with other words, in their
conversations with sighted people. They couldn't learn all word meanings
in this way. There has to be some grounding of word meanings in nonverbal
perceptions.
But what have they learned when they learn these so-called meanings? It's
Searle's Chinese speaker paradox in a different context. I can learn to say
that John has a _green_ sweater, or that John has a _red_ sweater, but if I
am blind I have no more reason to use one adjective than the other unless
someone tells me John does not have a red sweater. Then I will say he has a
green sweater, but I have still not learned what green means. You could
have told me he has a perspective sweater and I would believe you and echo
your word. But I would not know the meaning of perspective, either. A
computer could know the likelihood of occurrance, or transition
probabilities, of words and thus construct sentences about English writers
that sounded as if they had meaning, but in the computer that constructed
them they would never be more than likelihoods or transition probabilities.
They could have no referents beside other words or symbols.
I think I decided some time ago not to take words too seriously if I
couldn't find any experience that they indicate (I decided also not to
assume I was a dummy for not understanding what people were trying to get
across by words like "holy"). Once having realized that people use many
words without any notion of an experience they might designate (high or low
in the hierarchy), I became much more suspicious of the whole game. I think
we could almost do without the term "meaning" -- it doesn't refer to
anything specific, and there are much more informative ways to talk about
the same phenomena.
The meaning of a new noun can be learned contextually before any referent
is experienced.
But doesn't this ignore the fact that "context" is not just the other words
used, but the experiences referred to by the contextual terms, the other
words? If you have no contextual terms with perceptual significance, you're
back to "the gostack distims the doshes." The contextual terms evoke
perceptions which are put into relationships with each other, for instance,
and you can then tell someone the word for that perceived relationship. At
the configuration level, I learned that "a person with his elbows sticking
out and his hands on his hips" (get the picture?) is standing with his arms
"akimbo." Even if you've never seen a person with his arms akimbo, the
description ("hand ON hips") uses specific perceptual referents and words
describing relationships, so that akimbo becomes becomes a pointer to the
perceived relationships, not to the words. If you're smart, you can figure
out a little of what "akimbo" means from less direct descriptions: "He was
standing with his arms akimbo, so when John spoke, he turned and hit John
in the chest with his elbow." So how would he have to be standing for that
to happen? One plausible imaginary picture is the right one.
The learning of the meanings of operator words over these nouns
(adjectives, prepositions, relational nouns, intransitive and transitive
verbs of the simpler sort that only allow concrete nouns in their
argument) depends as much on context as on example referents.
I would interpret "context" as "direct perceptual examples." I'm sure
imagination plays a large part in this way of learning referents for words,
and I know from personal experience that this can lead to misunderstandings
when you don't imagine the right experiences. I may have told the story
about my father but it's worth another hearing: when he saw the printed
word "misled", he had always imagined something like a missile flying
off-track instead of where it was aimed, so he pronounced it "missile-d".
Of course that perceptual referent for the term worked well enough in all
contexts he encountered, so he didn't realize his error until he was a
well-known scientist past 65. Knowing how my dad was, I would guess that
when he heard someone say "mis-led," he just smiled to himself at the poor
fellow's lack of education. When the realization finally dawned on him, I
happened to be present and I'm afraid I was delighted in a mean sort of way.
The meanings of higher-order operator words (adjectives like true, verbs
like doubt, indeed, adjectives like fair) can hardly be learned
ostensively, in part because even if a referent can be singled out it can
hardly be sufficiently an exemplar.
We shouldn't think of perceptual referents only as concrete things. They
can be as complex as perceptions can be, and can be of any level of
perception from intensities to system concepts. I don't see, model-wise,
how a word or phrase (a configuration or event perception) can come to
evoke a perception of a system concept, but that whole area in the model
needs work, and we can be sure that "Catholic Church" is a printed
configuration that definitely evokes a system concept, whether we can
explain that or not. And don't tell me it's an association -- you'd only
have to explain how association works, and nobody knows that, either.
Realizing this, it actually is not a surprise that the blind people
learn the meanings of words for which they have no perceptual referents.
That itself, however, may be a surprise.
I think the (latent) surprise is that they think they know what the words
indicate to other people, when they don't and can't know. Francis (?)
Galton, who discovered color-blindness, wrote about how he learned to use
color words as others used them, and used them in such a convincing way
that neither his friends nor _he_ realized that he did not see colors as
they did. He was so good at deducing the usages of color words from context
that his color blindness was completely concealed even from himself. I have
conversed with more than one blind person for a while without realizing the
person was blind, because the person had learned the language of the
sighted, both verbal and body, with exquisite thoroughness. Yet the
giveaway always came when the perceptual significance of a word became
critical: "I see that you're skeptical," said when I was nodding in
agreement, for example. "I see" cannot "mean" the same thing to a blind
person as to a sighted one.
The other side of this is that perceptions above a certain level are
independent of sensory modality: "fairness" does not refer to vision or
sound or touch or smell, but to relationships among people and strategies
by which they deal with each other in any modality. We perceive these
relationships, categories, sequences, strategies, principles, and so on,
and -- to be a little more germane -- use a word like "fairness" to refer
to certain experiences at those levels -- I can't be more specific. The
experiences come first; when we hear people talking about "fairness" or
"unfairness" when those experiences are happening, we can acquire a direct
perceptual meaning for the term. Or we might acquire a perception of a
logical proposition: "A is being fair to B and C if A punishes or rewards B
and C exactly the same way for the same behaviors."
But then we may fail to recognized the same term when used with other
experiences: A 13-year-old girl forbidden by her parents to dress like a
prostitute may well complain of "unfairness," when the only perception she
refers to by this term is that of not being allowed to do whatever she
wants to do. It makes no difference that her parents apply the same rule to
all the girls in the family: it's still not "fair." People often use words
simply to give themselves any advantage they can. Consider the "Clean Air Act."
I think that "meaning," like many other words we casually use, is not very
useful, and that we can deal with the same phenomena in other and clearer ways.
Best,
Bill P.