LINGUIST List 3.414

Sat 16 May 1992

Disc: Linguists & Linguistics

Editor for this issue: <>

Directory

1. John Goldsmith, Integration and disintegration in phonological theory
2. John S. Coleman, 3.405 Languages, citation
3. Michael Sikillian/Annotext, 3.405 Languages, citation

Dan Everett's comment on my dissertation (as I've pointed out to
him), and by implication on how that work fits into the work of its
antecedents, is misleading. Dan was for some reason alluding to the
fact that in the MIT version of the dissertation (though not the one
that has circulated from the IULC or published by Garland, which are
the ones usually cited) I mentioned in the acknowledgements that my
interest in tone started with reading Will Leben's 1973 dissertation:
a true fact. But the dissertation itself has a chapter, the first,
entirely devoted to the proposition that this work was a continuation
of a discussion that has been going on in American phonology since
the 1940s! Dan, I think, sees Pike as the most important theoretician
to cite in that period; in my writing, I've focused more on Bloch,
Harris, and (in my 1980 book) on Hockett, rather than Pike, but this
is more a matter of style and taste than anything else.[On the same
theme, I have a paper coming out (perhaps it has come out already) in
the Journal of Linguistics on the genealogical connections between
prosodic (firthian) phonology and autosegemmental phonology. ]

Dan has also pointed out that some of the major contributors to
phonology during this period who are still very much alive and
intellectual active have felt slighted by the lack of citation of
their work. As I tried to suggest in my paper on firthian phonology,
this is more an indictment of normal human expectations of courtesy
than it is the result of people actually forgetting about these
phonologists' good, published ideas (there is much less of that
latter sin than many people wish to believe -- a point that Geoff
Huck and I have made in a recent paper on the relation of Generative
Semantics to current syntactic theory). However -- and again from a
purely human point of view -- I wonder how many people, like myself,
who were publishing material on nonlinear phonology in, say, its
first ten years (1975 to 1985) ever received a note from one of these
contributers to the literature in the 1940s and 1950s? Speaking just
for myself, I am sure I would have been galvanized to have been
dropped a note by ... any of a number of linguists; in more recent
years, I've had the opportunity to discuss the history of the field,
in writing and in person, with a number of these linguists. But I
would have been absolutely delighted to have received such a comment,
a bit of mild reproof perhaps from an established contributor (who,
now, I can perceive as feeling left out). I never did. Anyone else?

John Goldsmith

Prompted by Margaret Winters's lament that students aren't
being given a good historical appreciation of their subject,
I picked Robins's `Short History of Linguistics' from my
shelf. Interestingly `Linguistics in the Present Century'
is the eighth and final chapter (pp. 198--233). Melville
Bell appears on 203, along with Sweet; then comes
Trubetzkoy (p. 204), Jespersen, Hjelmslev (206), Boas,
Sapir, Bloomfield (207), Harris (210), Hockett (211),
"in recent years" Pike (212), Firth and Malinowski (213),
Halliday (221), Jakobson (222), N. J. Marr (225) (remember him?),
Lamb and Chomsky (226), Katz and Postal (227), and that's it!
(The book was first published in 1967).

--- John Coleman

Let me respond in general to the "Do you speak many languages?" issue.

Let's extend the statistician analogy. Let's say we ask our statistician,
"Do you care about the individual data points?"
S: Yes, as long as they contribute to my model
I: But you see yourself as independent of the data?
S: Yes, as a statitician, my function is to discover the underlying
structure from the data points. If I collect environmental data from Bar
Harbor Maine, it is no more important than if I did from Seymour, Indiana.
What is significant is the rules, structure and theories which can be made
from the individual data points.
I: Does the domain (ie. subject area) matter to you?
S: Not except for some sentimental reason. Whether it is environmental data
or the performance of computer systems, it is all the same. The methodology
is what matters.
I: would you object to being a special branch of mathematics or computer
science?
S: No, no. Statistics is quite different.
I: Ultimately, aren't you justy applying mathematical methods under a
different guise?
S: But the environment IS different than anything else.
I: Haven't you just contradicted yourself?
S: Let me rephrase then: the results have different effects depending
on subject matter.
I: But you are not concerned with the individuals in Bar Harbor, for whom
particulates or acid rain is a large concern. You don't want to understand
the data nor make decisions based on it?
S: No, that is too remorselessly PRACTICAL. I am a theoretician. I prefer to
develop theories about statistics rather than gain any understanding of
what, say .001 m vs .005 m particulates means in a given area. I focus on
knowing ABOUT rather than knowing.

If this (fairly transparent) dialogue did not make the point, let me add
this (more topical) reference. If I know that the line:

spargens humida mella soporiferumque papaver,

and write a paper that says

sparg+ e+ ns (present participle marker)
humid+ a (neuter plural marker)

then develop a rule for the grammar used in this piece, rewite it in the
phonetic alphabet, I will know a lot, from the point of whatever linguistic
area I am coming from (transformational grammar, phonetics, etc). And I can
develop a rather *wonderful* description of the language used. But I think
it is ultimately inadequate: for this line (from Virgil's 4th Aeneid) is not the
same as a mathematical data point of .005 ppm. There is a vast difference
between a formalized description of something, of knowing about language,
and knowing a language.

* translation: sprinkling moist honey and sleep-bearing poppy .