LINGUIST List 2.106

Sunday, 31 Mar 1991

Disc: Language Families and Phonology

Editor for this issue: <>

Directory

1. Scott Delancey, Re: Language Families
2. , Language Families
3. , Joos and the Boas tradition
4. Pierre Martin, RE rwojcik: phonology
5. "ACAD3A::FFJAL1", Consonant cooccurrence constraints for roots

To reply to Stahlke's note: There has been some discussion of the
relevance of Greenberg's success in African classification to the
evaluation of his American work. Certainly Greenberg and Ruhlen have
not been shy about bringing it up, and at the conference on Greenberg's
work at Colorado last year Paul Newman gave a paper making essentially
your argument, i.e. Greenberg was right then, so using the same methods
he's probably right now. The conservative Americanist response to this
is to argue that American languages are much more diverse than African
languages (or at least than N-K languages), so a slipshod method that
might bring results in Africa where the languages really are related
can't automatically be imported into the Americas where there's so much
more diversity. The Greenbergian reply is that 
Americanists obviously don't know anything about African languages,
which in fact show every bit as much diversity as American languages
(that's a tricky thing to measure, needless to say), and that anyway
the method couldn't give results if applied to languages that were
really unrelated.

Scott DeLancey

One more objection to the work on Nostratic and Sino-Caucasian
has been received, in response I guess to Starostin's remark
in Sc. Am. that words for 'hand' are never borrowed. This was
taken by some to mean that perhaps the work on these two language
families is crucially based on such arbitrary dicta about possible
borrowing patterns. Such is not the case. First, the claims
about what is borrowable are based on extensive research by
Starostin and Dolgopolsky (although I have some reservations on
this whole line of their work). Second, and more to the point,
the work on Nostratic and S-C does NOT, as far as I can see,
depend on any such assumptions any more than does the research
on IE or Romance. Incidentally, I would appreciate an example,
if someone knows one of a borrowed word for 'hand'. I myself
know of one example of a borrowed word for 'heart' (which is
perhaps even better, since 'hand' words are notorious for
being unstable, whereas 'heart' words are not). The Polish
word for 'heart' was borrowed in the late Middle Ages from
Czech. Any other examples involving basic body parts (i.e.,
not things like uvula) would also be appreciated. My own
view (as also of such people as Gerard Diffloth) is that
borrowing patterns are highly culture-specific and so not
a reasonable topic for universalist speculations (or, perhaps
I should say, an even less reasonable topic for universalist
speculation than other aspects of language).

I note with interest the recent posting in praise of Greenberg's
classification of African languages. I have always wondered
what Africanists make of it. However, even if he was successful
there, it does not follow that he would be in the American case,
since the two appear to be quite different. Unless I am
mistaken, the Africanist situation at the time was that people
were making really fundamental mistakes like classifying languages
typologically (e.g., by whether they have nominal classes), taking
Bantu to be a separate family, and so on. I do not know of any
comparable mistakes in the Americanist field. So, while Greenberg
did a great job clearing up such problems in Africa, it is not
clear what he can contribute to the Americas. Of course, I may
be wrong about the scope of his Africanist achievements. I don't
think I am about the Americanist ones.

When I was studying languages (not language) at the University of Chicago, one 
of my professors, John Kunstmann, used to thunder at us: Verify your 
references!! The Joos ascription is misleading, at best. The passage is 
presumably the one on p. 228 of 

Joos, Martin, ed. 1957. *Readings in Linguistics I.* ACLS. [I'm looking at 
the 4th edition, Chicago and London: University of Chicago Press.]

J. is commenting on Hockett's Peiping Phonology: "In his [i.e. "the practicing 
analyst, in the American sense"'s EB] Boas tradition (languages can differ 
without limit as to either extent or direction),..."

Note the ascription to the Boas tradition. It is a separate question whether 
this is a reasonable characterization of that tradition as it is whether Joos 
approved of that claim or not.

Emmon Bach

There is at least one school of thought in current phonological theory which
practices a clear distinction between the purely phonic component of a language
(=phonology) and the constraints imposed on the distinctive units by morphemes
and vice-versa (moneme variants= morphology, and not morphonology!) This is the
(structuralist) functionalist approach of Andre Martinet and his Paris School.

Na-Dene languages have series cooccurrence constraints for CVC roots:
essentially, sibilants and shibilants may not cooccur. For details,
see
 Krauss, Michael E. 1964. Proto-Athabaskan-Eyak and the problem
of Na-Dene I: Phonology. IJAL 30:118-131
 Leer, Jeff. 1990. Tlingit: a portmanteau language family? In
Linguistic change and reconstruction methodology. Trends in Linguistics
Studies and Monographs 45, ed. Philip Baldi, pp. 73-98. [The information
is in note 24. Essentially, I account for the fact that Tlingit lacks
glottalized sh, whereas it has glottalized s, by proposing that the former
merged with the latter. "The evidence relates to a restriction on cooccur-
rence of different affricate series in the stem: [sibilant] and [shibilant]-
series obstruents cannot cooccur unless the [sibilant]-series obstruent is
_s'_. However, if the _s'_ in these cases represents a merger of what were
originally *_$'_ and *_s'_, these cases would not have been exceptional
before the merger took place." Note $=s-hachek.]

[End Linguist List, Vol. 2, No. 106]