[From: Bruce Nevin (Wed 920413 09:17:47)]
(Avery Andrews (Fri, 1 May 1992) ) --
As a memory refresher (it's been a while), you said:
John gave the student article that
gets no meaning
I replied:
It is specious to say [this] example gets no meaning because to say
so presupposes that it would occur in native English, and it would not.
And your rejoinder was:
The hypothesis I'm proposing is that it does not occur in native English
*because* it gets no meaning by the normal operations of the perceptual
system described by the grammar. This could be false, but it isn't specious.
This exemplifies the sort of problem that Bill and I also keep running
into in our discussions of language, because of the complexity of the
subject. One of us takes one perspective and one another and we talk
past each other. Both perspectives are valid. What is needed is a
perspective that subsumes both, or at least one that makes it both easy
to switch from one perspective to another, obvious when one has done so,
and obvious which one one has adopted at the moment.
Avery, your perspective here is on the perceptual control of language
within an individual living control system. What is going on inside the
black box? My perspective here is on the social context for acquiring
language. If it's not in the language as social artifact in the child's
environment--particularly those who engage the child one-on-one in the
necessary language acquisition support systems (LASS)--then the child
never has the occasion to control for it and it never gets into the
black box. And that is why it subsequently has no meaning for one who
has acquired the language. However, the example given above works just
fine in Modern Greek:
O yanis dhini ton fititi to vivlio ekino.
- John gave the student the book that.
(I don't recall the word for "article" or "paper.") In English, this
word order "gets no meaning by the normal operations of the perceptual
system" only because it does not occur in the usage of other people in
the speech community, in particular did not occur in the usage current
in the language learner's LASS. It is easy to imagine a situation where
a community of native speakers of Greek learned English vocabulary but
retained Greek syntax, and the resulting interlanguage was learned by
children in that community. This seems to be precisely the origin of
interlanguages (trade languages, koine's) like Swahili. What is needed
is a system of social arrangements for cooperative action in which the
interlanguage is more advantageous than either of the original
languages. As I recall, a trade language of the Manding Empire in West
Africa was the basis for Swahili, Black English, Black Portuguese, etc.
Keep the familiar, albeit simplified, syntax and substitute in
vocabulary known to the persons with whom you are trading. As you know,
it has been argued that the history of English sounds a lot like this
too.
Now I still have to say there is something specious about your
statement, as I quoted it. To say that the unexpected word order,
placing "that" at the end, "gets no meaning by the normal operations of
the perceptual system described by the grammar" begs the question "does
the grammar in fact describe the normal operations of the perceptual
system for control of language?" It begs this question because an
answer to the question is subordinated as a modifier of "perceptual
system" when you say "the perceptual system [which is] described by the
grammar". This is no longer a question, it is a presupposition; it
presupposes an answer to the question. Begging the question is
specious. Placing the question in the foreground is not. Then a
particular answer to the question may be false or not, as you say.
(Bill Powers (920501.1430) ) --
study the actual process of speech production and error correction while it
is going on. This will directly show what is considered an error and what
corrected form is considered OK by the person, no matter how the person
might describe or misdescribe the rules actually in effect. If you then ask
why a person corrected certain errors, the answer in many cases (with non-
linguists) will be that it just sounded wrong, and sounded better the other
way. That's probably the kind of answer you want, because it shows that the
correction was made by the actual machinery and isn't just a guess at a
verbal generalization. I suppose the subjects you want for such experiments
would know nothing about grammar or syntax, but would still speak well.
Vicki Fromkin has done a lot of work with slips of the tongue. I posted
comments on an article by Ewing and others on an experiment inducing
slips of the tongue to ascertain relationships among different levels of
control. Such work is considered too "data bound" by many in the field.
Too bad for them.
(Rick Marken (920501 17:00) )
(BTW, Rick, this is the sort of place to look for control, not in the
enforcement of rules learned in school.)
Same kind of problem I had with Avery -- what do you mean? When
Sorry, I don't mean to deny that control is involved in responding to
overt peer pressure. Control in the face of disturbance by real or
imagined social influence of others is certainly control. But this is
control for one extant variant of language usage over another. Rules
like those that say one choice of "that" vs. "which" is "proper" and
another is "improper" in a given kind of syntactic construction are
based on the relative prestige of social groups whose members learned
one way or the other at their mother's (or nanny's) knee. My purpose
was not to deny that control is involved here. My purpose rather was to
emphasize that control is involved at more basic levels in constructing
or interpreting a relative clause, whether begun with "which" or "that".
Genie could not do the latter. And of course the syntax of relative
clauses, and of modifiers in general, just scratches the surface.
It is important to realize that the mechanics of putting words together
in a construction like a relative clause (syntax) is distinct from the
semantics of which pairings of words in a given construction make sense.
Genie could put words together on a semantic basis in simple
subject-predicate constructions, but not in the more complex socially
institutionalized syntactic constructions that support subordination and
coordination of assertions.
Penni's work, so far as I have read, seems to concern itself with
higher-level discourse structure. The sentence-level syntactic
constructions are rudimentary. Her results converge with Harris's in
_The Form of Information in Science_ and _A Theory of Language and
Information_. In that work, it was shown that information structures
found in discourses of a sublanguage of a science correlated with
objects and relations in the domain of the science, and that changes in
the information structures correlated directly with changes in the
science (changes in immunologists' perception and understanding of the
domain of immunology). A (r)evolutionary development in the science
coincided in time with a change in word classes and permitted word class
sequences in the sublanguage grammar. It would be interesting to see if
the system at Xerox PARC could accomodate a domain and discourse
structures of that complexity. Like Penni's work, Harris's concentrates
on domain-driven structure, leaving structures of logical argumentation
(which appear to be superordinate over the former) for future
investigation. Unlike Penni's work, Harris's includes a comprehensive
account of sentence syntax.
(Bill Powers (920502.1200) ) --
(Avery Andrews (920501) ) --
control: but how to introduce appropriate disturbances into sentence
generation?
examples of things that are somewhat tricky to say or describe.
Most of the examples you propose, Bill, concern control for semantic
relations, that is, the correlation of words with nonverbal perceptions.
Things that are tricky to say *because* they are tricky to describe.
There are also things that are tricky to say because the conventions for
things like reducing interrupting comments to modifiers are tricky to
manage.
"The man with the green car's dog has fleas."
This is more of a problem written than in speech. "With the green
car" gets lowered intonation in speech, and "dog" pops back up to the
level of amplitude and pitch that it would have had if "with the green
car" had never interrupted between "the man" and "'s." This is the more
obvious, the longer the interruption. Within that lowered intonation,
something that differentiates can receive contrastive stress:
"The man with the GREEN car's dog has fleas." (As contrasted with the
man with the WHITE car's dog.) You can get "The GREEN-car man's dog
has fleas" with the contrastive stress, but hardly without.
Just how would you say "the Queen of England's unexpected visitor"?
It is a matter of what is called good prose style to avoid such
pitfalls, of which any language sets its users many. That famous
shibboleth, the dangling participle, provides many a hilarious example
of pernicious ambiguity. Hopefully, people will eventually get tired of
complaining about certain sentence modifiers in sentence-initial
position, because of ambiguity with their use as verb modifiers.
("Hopefully" does not modify "get tired of complaining" but rather a
zeroed performative "I say" on the sentence as a whole.) Clearly, what
applies to one sentence modifier without complaint should apply to all.
(Neither does "clearly" modify "applies," but no one complains.)
But control for unforeseen (or picked-on) ambiguity is still at a
semantic level. You have two syntactic constructions that come out as
the same sequence of words, but the constructions correlate with
different nonverbal perceptions, one of which is inappropriate, perhaps
even humorous. Underneath that, what constitutes a syntactic
construction? What perceptual control is involved in controlling the
mechanical mannerisms of the reduction system, quite apart from the
dependencies of operators and arguments? Why can't you say "This is so
much funner than doing my income tax"? (Though my kids' generation use
"funner" quite freely--the language seems to be changing there.)
Consider things like:
The man which the dog with fleas--I mean, the man with the dog with
fleas, the dog which has fleas--still thought it was his neighbor's
fault.
Any place where operator grammar sees a reduction, semantic control
applies transparently in the correlation of unreduced words to nonverbal
perceptions, but less obviously to the reduced forms of words and
phrases.
the dog--said dog has fleas--got lost
the dog which has fleas got lost
the dog with fleas got lost
Syntactic control is involved in the correlation of reduced forms to
more explicit forms. The difference may be highlighted by an ambiguity
that has perhaps escaped notice:
the dog--said dog has fleas | the dog--said dog is with fleas
the dog which has fleas | the dog which is with fleas
the dog with fleas
I imagine a Bizarro cartoon captioned "dog with fleas." There they sit,
the dog and a small crowd of fleas, side by side under their respective
beach umbrellas.
At a lower level than semantic control and syntactic control is control
of phonology and phonetics. Tips of the slung, I mean slips of the
tongue.
All of these are going on at once. Control for dialect membership
(one's own, one's companions' different dialect, a model of the dialect
associated with a more prestigious social class, a parody of a socially
inferior (or superior) dialect, and so on); control for meaning; control
for syntactic construction (within dialect); control for pronunciation
(within dialect). As Avery says, it is not always easy to tell which
kind of control has been disturbed when someone says "I mean . . ." .
Studying a single idiolect in depth (speech of a single language user)
is certainly an option. There are problems. The processes of
observation can disturb and influence what you are observing. This is
obviously so when you use yourself as an informant (though it has not
been so obvious to the recent generations who have so abundantly
demonstrated the truth of this, linguists trained to do just this and
little more). You cannot make much sense of control for conformity to or
distinction from socially marked ways of speaking without considering
the latter. You cannot make sense of variation in reference perceptions
within the same speaker from one occasion to another, and change in them
over time, except with reference to language as a social artifact.
Speech is different from the control of a pointing finger in a way that
I think is important for all the social sciences. In the usual case,
behavioral outputs are incidental byproducts of control. They are not
themselves controlled. Some other perception is controlled, and the
behavioral outputs are variable means, whatever it takes in a
disturbance-prone environment to make the controlled perception match a
reference perception in memory or imagination. With speech, however
(and with any conventionalized behavior) the form of the behavioral
outputs is itself subject to control, concurrently with the perceptions
the control of which the behavioral outputs are the variable means.
This is possible whenever there is "free" variability that is not
constrained by the contingencies of control--more than one form of
behavioral outputs can accomplish the same controlled perception. Then
choice among alternatives (or in the range of free variability) itself
is exploited as an aspect of self image, or social standing, or
relationship to others involved in the transaction, etc.
Even pointing with the finger can have a personal style, or a manner
associated with a particular community.
To accomplish this, the behavioral outputs involved in effecting control
of one perception must themselves be monitored and controlled with
respect to particular choices among their range of free variability.
I hope that this is clear, and that it contributes to a useful
discussion, especially with the sociologists and social psychologists
among us.
Bruce
bn@bbn.com