From Andy Papanicolaou and Tom Bourbon [920604 13:50]
Response to Bill Powers [920603.1700], Re: Habits
The hope behind the selection of this specific topic (acquisition
of, or learning of the skill or the habit of producing correctly
and reliably new speech sounds) was to elicit specific comments
that would help us construct a CT model of the process.
It was not meant to elicit a reiteration of the reasons why a CT
model would be preferable to an open-loop "plan" model.
The first specific problems or difficulties that we (Tom and I)
encountered in thinking about how to proceed with constructing a CT
model of this phenomenon was not the distinction between
"conscious" and "automatic". We already know that "habits are
carried out without thought" and we certainly do not think that
"control theory says that control has to involve thought or even
awareness".
The specific difficulties we encountered and we spoke of had to do
with what we should consider as the reference signal.
We also laid out the facts of the case as we understand them.
Facts that the CT model we hope to construct has to take into
account. We still believe that it is a fact that to produce a
particular phoneme, a time-varying pattern of articulatory gestures
is necessary. This pattern has some invariant features. What we
mean by that is simply that to produce a sound that is heard as the
French /u/, the lips must be rounded and to produce a /b/ sound,
the vocal cords must start vibrating before the lips open to let
out a burst of air. If the cords start vibrating after the air
release no /b/ can be produced, or if the lips are not closed
completely and the air is not suddenly released again, no /b/ will
be heard.
We also believe that it is a fact - and we so stated - that
although there are invariant features to each articulatory gesture
pattern, there is a range of variation as well, which accounts for
the fact that the same syllable can be pronounced with different
accents, angrily, sadly, etc.
So, we believe that it is generally true (and, hopefully, the
linguists will concur or persuade us otherwise) that every time a
/ba/ or a /ga/ or a /tu/ is heard the corresponding patterns of
articulatory gestures contain a set of invariant features.
The linguists may be also able to tell us whether we should take it
for granted that the Japanese possess auditory images of r and l,
that is they can discriminate these phonemes even if they cannot
produce them. And we are not talking here about a statistically
averaged Japanese but about concrete individual Japanese who have
not been exposed to languages in which r and l are phonemes. Until
the linguists persuade us otherwise we will be cautious and we will
not take it for granted that once a Japanese hears /relax/ for the
first time he can experience r and l as different phonemes.
We trust it is clear that the questions regarding the Japanese'
ability of experiencing r's and l's is directly related to the
issue of what constitutes, initially, the reference signal and not
to any peculiarities of particular linguistic groups.
We also trust it is clear that we are here concerned with the
general issue of skill acquisition and not the instance of
acquisition of speech sound producing skills not because we think
that these skills are NECESSARILY of a different kind than the
skill of typing for example but because they force us to consider
issues (like the problem of reference signal) which, in the case of
the typing skill may have gone unnoticed.
Finally, we hope it is understood that our search for appropriate
reference signals does not include "higher order" references like
the "intension to communicate" etc. We wish to focus on how to
account for the acquisition of the skill in a situation where the
learner simply consents to try and produce a particular sound to
the satisfaction of a teacher.
So, to summarize, (1) in an attempt to construct a CT model that
would account for the acquiring or learning the skill or forming
the habit of pronouncing new speech sounds or "reorganizing" to
that effect, one of the initial difficulties we experience is what
to think of as reference signals.
(2) Bill's comments helped us realize that there may be no
consensus as to what constitutes the set of facts that the model
should account for. Hopefully, such consensus is possible
otherwise each of us will construct models of a variety of private
linguistic facts and worlds.
(3) We wish to reiterate our intention to construct a CT model not
a hybrid CT plus "neuromotor plan" model of this process. We hope
that the facts, including the existence of articulatory
invariances, can be accommodated within such a model because to us
also "it seems very unlikely that some behaviors would be organized
according to one fundamental principle and other behaviors - of the
same kind - according to a different one".