LINGUIST List 5.1477

Mon 19 Dec 1994

Disc: Comparative Method

Editor for this issue: <>

Directory

1. Karl Teeter, Nostratophobia; re: 5.1448 Comparative Method
2. Karl Teeter, Re: 5.1448 Comparative Method

Message 1: Nostratophobia; re: 5.1448 Comparative Method

Let me say in response to Alexis Manaster-Ramer that nobody would be more
delighted than I should it turn out that Nostratic is a real language
family, or even "Amerind". All I want is convincing evidence. It is
certainly wrong of him to say in respect to me, and I am sure also with
respect to Victor Golla, that we are "afraid of looking at the substantive
issues surounding the Nostratic hypothesis". To the best of my knowledge,
I have not even remotely been discussing that issue, but rather my views on
the methodology of comparative grammar and linguistic history. I am
neither a "splitter" nor a "lumper",just an old-fashioned believer in
facts and proof.

 As to specific points in AMR's latest epistle:
 (1) "surely [Karl] does not mean that a comparative grammar is a
prerequisite to a reconstruction". Of course not, I can't imagine how AMR
could have gotten anything so silly out of all the things I have been
saying. Then he goes on to the undeniable statement, undeniable since
wholly circular, "we cannot demand a detailed morphological
reconstruction until the languages are accepted as related FIRST and
then ONLY if the languages in question HAVE morphology to speak of".
Totally unexceptionable, but also, it seems to me, totally irrelevant to
any position of mine. What I have been talking about in most of the
discussion to date is the crucial importance of distinguishing borrowed
material, words or structure, from that for which we must postulate
retention from a protolanguage. I have characterized the process of
comparative grammar as that of writing a grammar of the protolanguage,
and would still do so, and clearly, grammars include more than morphology.
 (2) AMR goes on to rabble-rouse a bit about the importance of
"having to look at the substantive issues"--again, entirely
unexceptionable. It is clear I and AMR agree on many things: I don't
think we can do morphological reconstruction without having morphological
data, and I consider it highly important to look at substantive issues.
 (3) One place where we do not agree I have already responded to
in my message which appeared the same time as AMR's, although he had it
earlier: the possibility of writing a grammar of proto-French-English on the
 basis of the very large
amount of common vocabulary. He says he can easily write a "comparative
grammar" of it and call it proto-Wilhelmian. What he can't do is write a
grammar of it, because it is not a language.
 (4) Finally, I have said and would repeat, that the most
important task in doing the history of a family of languages is to write
a grammar of the protolanguage. You don't need morphology, but it is
hard work, and nobody has done it yet for Nostratic or Amerind.

Message 2: Re: 5.1448 Comparative Method

Dear Helge, Sally, and "multiple recipients of list LINGUIST": Ye gods and
little fishes! I am honored and overwhelmed to receive, along
with my latest on the comparative method, no less than THREE messages in
response to mine. One of them, from Alexis Manaster-Ramer, I have
already responded to and sent to the LINGUIST list, under the title
Nostratophobia. The other two are from Helge Dyvik and Sally Thomason, and
I am grateful for their attention.
 My response. First to Helge: if it seems I am trying to "dismiss
the whole methodological discussion as meaningless" I demur; that is not
my intention. I am, nevertheless, primarily a grammarian as a scholar,
and I believe that when I do linguistic history I write a grammar of a
protolanguage, just as when I do descriptive linguistics via field work
or textual research, I write a grammar of that language based on the data I
can discover.
 Second, Sally. Thanks you for your kind words, I am glad you
have been enjoying the discussion. I regret having evidently been too
hasty in stating without elaboration, "Words may be borrowed, structures
no", so I owe Sally and the rest of you further explanation. I do make that
assertion, and in fact I believe it to be the principle that makes
historical linguistics possible. What can be borrowed must be something
that can be heard, and what one hears are utterances, words and strings:
structure, on the other hand, cannot be heard but must be deduced,
whether by linguist or language learner. So strictly speaking, structure
cannot be borrowed, in principle. This is not to deny that languages can
be learned, and when they are, as we all know, the learner proceeds
to reconstruct a mental grammar which will bear some resemblance to the
model. The question is, to what extent is such borrowing possible, and
the answer from experience seems to be, not to a very great extent,
although naturally one must pursue each individual case in detail, and
some of them will involve what appear to be borrowed structures.
The facts I have discussed, however, show why Meillet -- and many of we
his followers -- believes that "structural" correspondences are more likely
to be probative than resemblances in vocabulary. Clearly the crucial
question in comparing languages, which I have been reiterating (also
Meillet's point, and not mine) is that one must meticulously distinguish
similarities which result from borrowing from those which are retentions
from a protolanguage. I have argued that the way to do this is to set
out to write a grammar of the protolanguage.

 I hope this makes sense! It should be clear that I am not trying
to create controversy, only to understand the principles of comparative
grammar as we have inherited them. Yours, Karl