[From: Bruce Nevin (Tue 93119 10:43:12 EST)]
Rick Marken (931108.0900) --
This way of saying it has got across to you part of what I have been
trying to communicate, and that's good, but there's another piece of it
that I haven't yet put in a way that speaks to you.
pronouncing the phonemes in a word does not depend upon
perceiving (even in imagination) phonemically different words and
controlling the contrast between the desired word and the alternatives.Bingo! Now you can see what our problem was. It sounded like you
were claiming that a contrast was a controlled perceptual variable.
It COULD be one -- but it seems highly unlikely that a contrast is
ALWAYS what is controlled during regular speech.
What I said was that contrast is not controlled in speech. I said that
contrast is controlled in establishing the hyperarticulated targets for
speech--the values of the reference perceptions in the elementary control
systems (ECSs) that we now refer to rightly or wrongly as phoneme
detectors.
Contrast in this sense is the maximal differentiation of pronunciation
targets within constraints that we as observers can describe as a
multidimensional space (see next paragraph). As such, contrast in this
sense is not varied during the course of speaking. The time scale for
variation of pronunciation norms is much longer than that for control of
words and sentences, and the time scale for maximizing the
differentiation of pronunciation targets within the pronunciation space
is correspondingly long also.
This pronunciation space is defined by physiological limits of opening
the mouth, moving the tongue, etc., by acoustical limits as to which
articulatory differences make some acoustic difference (and which are
acoustically alike despite articulatory differences), and by socially
established limits as to which of these available means is actually
exploited in the given language and dialect (e.g. front rounded vowels as
an accident of coarticulation, but not as means for distinguishing
utterances).
I said (Fri 93115 15:36:24 EST):
4. The contrast observed from the outside in (2) might be a result of
control of the contrast between targets, as in the first proposal that
you paraphrase above. However, this control would not be effected by
varying pronunciations directly, but by varying the reference settings or
targets for pronunciations (with an indirect effect on pronunciations).
This was in context of Bill's question, which of two notions about the
control of contrast is more likely to be right. When I say "control of
contrast" here, I refer to an observed phenomenon--norms exist and they
are maintained. I do not mean to pre-judge how that is accomplished.
One notion is that this control (as an observed phenomenon) is indeed due
to perceptual control by the language learner in the course of learning
the language and in the course of accomodating to different dialects.
The other notion is that it is a byproduct of controlling perceptions
such as a perception of understanding, a perception of being understood,
and a perception of being accepted by others as normal, reliable,
accessible, friendly, etc. The former seems prima facie more plausible
to me, because of cross-dialect communication, etc., but what do I know.
Now, some consideration of the sources of confusion--at the risk perhaps
of muddling things more.
I probably invited the confusion between contrast and what we might call
distinctness. Contrast is the optimization of differences between
targets in the physiologically, acoustically, and socially defined
pronunciation space. Distinctness in the course of speaking is a
function of gain on the speaker's control of pronunciation with respect
to the targets (reference perceptions). Distinctness may vary a lot even
in the course of speaking a single word (as when emphasizing a single
phoneme in it). Contrast is a function of establishing the pronunciation
targets themselves. Whether the process of establishing the targets
involves the learner's control of perceptions of contrast or is a
byproduct only of controlling perceptions of social relations with others
is the question that Bill mooted. My answer was I don't know, but I
suspect the former.
Probably where I invited the confusion was in discussion of language
learning. I have tried to distinguish language learning (and the related
methodology of linguistics) from language origins in evolution, and both
of these from language as perceptual control during speaking. These
three perspectives have nevertheless got repeatedly muddled. As an
infant learns a language, it starts with many different perceptions of
what is said. Some lumps of talking are isolated, especially in the
baby-talk of encouraging elders. But the child does not know the
pronunciation targets that speakers of the language have as their
internally maintained reference perceptions. Only a small repertoire of
utterances can be controlled in a socially recognized way by such means.
Later, when the child begins to control the pronunciation targets, much
falls rapicly into place. Variations of pronunciation are normalized
because the reasons for failure to reach the actual targets are obvious
to the child by introspection into her own pronunciation process.
Crucially, the targets (the phonemes) become available as means for
differentiating utterances from one another, for identifying words by
these points of contrast between them, and for recognizing parts of words
that are reusable in the making up of other words. The phonemes are
means for representing words efficiently by their points of contrast with
one another. They locate the points of contrast within words and
identify them with perceptual targets. In the process of learning, the
perception that words are different comes first, and wherein they are
(conventionally) different comes later, whereupon learning of new
vocabulary and of how to use words takes off at a much faster rate.
Your manifest interest is in language as perceptual control during
speaking. We have to consider these other kinds of issues as well. Very
different sorts of perceptual control are possible before and after the
establishment of phonemic contrasts.
Bruce Nevin
bn@bbn.com