Linguistics (was Re: Zombies and PCT)

[From Rick Marken (01.02.13.1500)]

Bruce Nevin (01.02.13 15:44 EST)]

The structures of language and the use of those structures
("language behavior") are not the same thing.

Yes. I think the structures of language are perceptions, some
of which are controlled ("used") and some of which are
irrelevant side effects of control.

How those structures themselves are perceived and how those
perceptions are controlled has not been determined, for the
most part, but it is clear that language is structured

Sure. As is all behavior, when seen by a person who can perceive
"structure". But that structure, like the rings formed by the
individuals in the CROWD program, is probably an irrelevant side
effect of the fact that people are controlling perceptions that
have nothing to do with that structure. The individuals in the
CROWD program, for example, are controlling for proximity, not
for being in a ring. The ring structure you observe is an
irrelevant (to the individuals themselves) side effect of
controlling for proximity. Similarly, the grammatical and other
structures (such as subject/verb phrase order) we see when we
look at spoken language are very likely to be (like the rings)
irrelevant side effects of the fact that speakers are
controlling for perceptions that have nothing to do with
these structures.

Data from various forms of testing for controlled variables, plus
methods for identifying dependencies among those controlled variables,
tests and methods that have been in use in descriptive linguistics...
for the past 50 years.

I find that very hard to believe. I studied some linguistics in
college and it seems to me that linguists are rather attached to
the notion that people "generate" grammatical utterances rather
than control perceptual variables. It's hard for me to imagine
how anyone could systematically test for controlled variables
if they assume that language is a generated output.

Some references:

I don't think I'll have much chance to read these. But I would
like to know what you think constitutes a linguistic test for a
controlled variable. It would be nice if you could give a brief
description of the testing for controlled linguistic variables
that is described in these references.

Best regards

Rick

···

--
Richard S. Marken Phone or Fax: 310 474-0313
MindReadings.com mailto: marken@mindreadings.com
www.mindreadings.com

[From Bruce Nevin (01.02.13 19:40 EST)]

Rick Marken (01.02.13.1500)--

I find that very hard to believe. I studied some linguistics in
college and it seems to me that linguists are rather attached to
the notion that people "generate" grammatical utterances rather
than control perceptual variables. It's hard for me to imagine
how anyone could systematically test for controlled variables
if they assume that language is a generated output.

I have said many times that Generativist Linguistics is a crock, and I said I was talking about descriptive linguistics.

I don't think I'll have much chance to read these.

Then you don't have time to find out what I'm talking about. That's all right. But if you don't find out what I'm talking about, I don't think you should tell me what I'm talking about.

Even in this exchange, you are not concerned with what I am talking about. You are omitting the point that a side effect that is irrelevant for control of proximity can be relevant for control of being perceived by others in an unwanted way (not wanting to stand out, or be singled out, or the like). Consequences that are unintended with respect to control of one variable can disturb one's control of another variable. We may even come to control those consequences by means of the manner in which one controls the first variable while still controlling it much as before. Language is full of this kind of thing (think of dialect, or shifting from a formal register when talking to an interviewer at the door to an informal register when the phone rings and you answer a friend) and indeed all aspects of human culture.

But I would
like to know what you think constitutes a linguistic test for a
controlled variable. It would be nice if you could give a brief
description of the testing for controlled linguistic variables
that is described in these references.

I've done this before. You could search for pair test in the archives.

Select two utterances A and B, preferably very short, with one speaker of the language pronouncing a number of his repetitions of each utterance, randomly intermixed, while another speaker of the language indicates which of these pronunciations are repetitions of each other. In most cases the hearer's and speaker's judgements of repetition will agree either in close to 100% of the cases or else in about 50%. In all cases, if the hearer's list of what were the repetitions has a close correlation with the speaker's list, then the two given sets of repetitions are phonemically distinct (e.g. heart vs. hearth). Otherwise there is no phonemic distinction between them (e.g. heart and hart). By experimental substitution of small portions from one member of the pair to the other(which can be done by the speaker, or digitally now rather than by tape-cutting), one can isolate the particular portion in which the phonemic contrast is located. This gives a first approximation to the most basic elements of a language (the contrasts) as represented by sounds, and a segmentation of utterances into those elements. Disturb the sound of a given segment enough and a different word is perceived, or a non-word (which could be a word but isn't, e.g. vary the last segment of harp and get harb). This first approximation is refined by later steps of experiment and analysis.

There are further issues of course, and more is done to identify just what the controlled variables are (a difficult task because a great many are controlled at once, with redundancy among them which enables sometimes very "sloppy" control to be adequate where other times very precise control of the same variable is necessary) but this is an example, as you requested.

         Bruce Nevin

···

At 03:06 PM 02/13/2001 -0800, Richard S. Marken wrote: