LINGUIST List 5.935

Wed 31 Aug 1994

Disc: Altaic

Editor for this issue: <>


  1. , Altaic, Pronouns, Marrism, Leninism
  2. , Re: 5.929 Altaic

Message 1: Altaic, Pronouns, Marrism, Leninism

Date: Tue, 30 Aug 94 07:09:18 EDAltaic, Pronouns, Marrism, Leninism
From: <>
Subject: Altaic, Pronouns, Marrism, Leninism

About Altaic, it seems that the debate is pretty much over. I
am tempted to declare victory (as Mary Haas once did in the
context of another language grouping) and move on.

I was very impressed with Prof. Unger's posting, especially
after I had been so very critical of his published work. This
posting does make crystal clear that there is NO NEW basis
for doubting or rejecting Altaic, and that all there is the
(I think discredited) objections of Doerfer,
Rona-Tas, Clauson, Shcherbak (and I guess I forgot to add
Sinor to the list). These have in any case been responded to
in print by various people, so those who are interested can
judge for themselves. I myself intend, in the aftermath of
this discussion, to write up (if anybody will publish it!)
a response to Nichols and Unger, so that there is something
in print that can be cited.

Prof. Vovin is quite right is pointing out that some of the
traditional anti-Altaicists, so called, have adopted the
surprising position that only binary comparison (2 languages
at a time) is permitted. This point should perhaps be
pursued because it seems to be rearing its head
in some critiques of Amerind and Nostratic as well.
I might add that some of these people
also have come up with other even more unlikely
"methodological" innovations, which they simply
assert without any argument, e.g., that you may not
ever compare two reconstructed (i.e., proto-) languages
(Doerfer), that the comparative method cannot be applied
to situations like Nostratic because the way of life and
cultures of people that far back are completely different
from anything we know of (Rona-Tas), and so on.

About pronouns, I have received some mail (incl. from
a senior colleague whose views I respect a lot and who
I did not know was following this) about my critical
attitude towards Nichols' idea that the similarities

among the pronominal systems of Turkic, Mongolic,
Tungusic, IE, Kartvelian, etc., are due to their
all coming from the same place but not to their being
genealogically related. The mail I got focused on
the question of whether a language can borrow its
pronouns from another. The answer is apparently yes,
since there are well documented cases (such as Copper
Island Aleut from Russian, or Shelta which has both
Irish and English pronouns).

However, my point is not that borrowing is impossible
(although it is rare). My point is that borrowing has
to be demonstrated, like anything else (although Eric
Hamp has argued that borrowing is a weaker hypothesis
than relatedness because it is harder to refute). It
cannot simply be asserted, as some have done. And the
situation with the Altaic pronouns (Turkic, Mongolic,
Tungusic) is parallel to that in IE. It is just as
easy to say that the similarities between Germanic and
Indo-Aryan and Slavic, say, are due to borrowing as it
is in the Altaic case. It is just that no one would
DARE to say such a thing.

Which brings me briefly to the political issues. It is
very tempting for someone who did not live in Russia
during the Soviet period to laugh off the remarks made
by Prof. Vovin, but I think it is very true that
political analogies such as the ones he drew were
unfortunately very real. And we must remember that
Soviet Russia was where much much of the fight over
Altaic took place. Clauson, a Britisher, and Rona-Tas,
a Hungarian, published their early attacks on Altaic
in a Russian journal, for example. Moreover, it seems
to me that a kind of political analysis is applicable
to the situation in the West too. We must realize that
the acceptance or rejection of an idea is not a pure
ethereal intellectual act, but a political one within
the politics of academia. Thus, it is not irrelevant
I think that in the case of Indo-European there is a very
large (numerically) body of linguists in the strict sense
of the term, whereas in the case of many other language
groups (including all the Altaic ones) there have
traditionally been a few linguists (sometimes one or maybe
two in a given country) and a whole lot of philologists
or historians. The audiences at the conferences, the reviewers
in the journals, one's colleagues in a university dept.
all make a great deal of difference, and these have been
so very different for the Altaic comparative linguist as
opposed to the Indo-Europeanist. I have no doubt that it
made a difference that people who were largely untrained
in linguistics and not very intersted in its goals were
the dominant group in Altaic studies, as indeed were some
of the leading critics of Altaic. Sinor describes himself
somewhere as a nonlinguist adding that linguists will
breathe a sigh of relief to hear him say that. Now, he
was actually more of a linguist than he admitted, but certainly
not a linguist in the same sense as Poppe or Starostin or Vovin.

Finally, at present, it seems to me that politics has much to
do with attitudes towards Altaic as well. In particular,
it seems that Altaic sometimes gets knocked by people who are
really gunning after Nostratic (and after Greenberg). It is
also a political fact (or a social one, or whatever) that
there is no forum in our field (except perhaps for LINGUIST)
where such a discussion as we have just had is possible. Thus,
it is perfectly easy to disseminate harsh criticism and mis-
information about a theory like Altaic so that many linguists
who do not work in comparative ling or on these languages will
hear of and assume the worst, but there is no journal which
reaches such an audience which would be a forum for a discussion
in which the issues can be aired and both sides can be heard.
There should be (much as our friends in anthropology have one)
but there isn't. That's too bad, but at least we have LINGUIST.
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 2: Re: 5.929 Altaic

Date: Tue, 30 Aug 94 16:55:49 ESRe: 5.929 Altaic
Subject: Re: 5.929 Altaic

Reinhard Hahn is certainly right that "*everyone* who does not reject the AH
definitely considers Turkic, Mongolic and Tungusic to be Altaic" and "not *ever
yone* is totally convinced that Korean and Japanese are related to them or are
closely enough related to them to be called "Altaic". I had no intention to
misinterpret Reinhard Hahn's position, and if I produced such an impression, I
beg to accept my sincere apologies. All I wanted to say is that this *everyone*
 / *not everyone*'s position stands on a very shaky basis. Let me develop
this a little bit further. First of all, as far as I know, majority of us now,
who do reserch on comparative Altaic proper, and not on separate Altaic langua-
ges, and who actively publish during last ten years on the subject, accept the
idea of five-member Altaic family. Sergei Starostin in his book made very solid
lexicostatistical arguments in favour of inclusion of Japanese and Korean on
the same level of relationship. There are also considerations of phonetic
nature (there are no common phonetic innovations and/or archaisms in "classic"
Altaic triade as opposed to Japanese and Korean or vice versa), and morphologic
al nature (Japanese, Korean and M-Tungus share more common morphological mar-
kers than any of them with Turkic or Mongolian). On the other hand, I have
never heard any solid arguments supporting exclusion of Japanese and Korean
from Altaic; as it seems to me, it is rather based on circulating rumours,
like those which were described previously regarding the very nature of Altaic
itself. If anyone can provide such arguments,I will be very willing to hear
them and to discuss them. Otherwise, I believe, the outdated concept of "Clas-
sic" Altaic with only three members as it was held in sixties and seventies
should not be represented as majority's point of view. There is, I believe,
a scholarship problem, too. While it is enough to have reading knowledge of
English, German, and Russian (and more recently Chinese, too) in order to get
access to Turkic, Mongolian, or M-Tungus materials, unfortunately, there was
no access to reconstructed Japanese data until S. Martin's fundamental "THe
Japanese Language through Time" appeared in 1987 for a specialist in Turkic,
Mongolian, or M-Tungus, unless s/he knew the way across piles in written in
modern and Classical Japanese. Using even Kenkyuusya's Japanese-English dictio-
nary would be as fruitful for comparative Altaic purposes, as using modern
Kazakh alone to represent Proto-Turkic. The same situation exists even today
with Korean data. Though Martin's excellent "Korean Reference Grammar" intro-
duces Middle Korean data to a scholar who cannot read Korean, these data are
still quite far from Proto-Korean reconstruction. On the other hand, on the
Japanese-Korean side of Altaic, we find very similar situation. First of all,
only few scholars here are interested in Altaic as such, and many of the
latter cannot read "Sravnitel'nyi slovar' tunguso-man'zhurskikh iazykov",
which is the main source on comparative Manchu_tungus. Comparing Japanese
and Korean with Manchu alone, which is best accessible language for the collea-
gues who do not read Russian, even remotely will not produce the same results
as comparison with Proto_Manchi-Tungus, with reconstruction based on all availa
ble sources. This situation does create a little correlation between the
specialists, and "Inner" and "Outer" Altaic reflect not a linguistic situation
but a major division between specialists.
 Let me now briefly stop on the point why I believe that anti-Altaicist
methodology is in contradiction to the comparative method. I, believe, that
some of what follows, was already mentioned before by N. N. Poppe and R. A.
 1) In his attack on Miller Doerfer accuses the former that he compares
Japanese not with Altaic, but with different Altaic languages. This accusation
does not make sense from standpoint of a comparativist: Indoeropeanist compa-
res , let's say, a certain Greek form, not with IE reconstruction, but with
cognate forms in various IE languages. In the same way, an Uralist compares
Finnish word with its cognates in other Uralic languages, not with PU recon-
struction. And so on.
 2) Another odd methodological cornerstone of anti-Altaistics is
a thesis that one cannot compare simultaneously more than two languages. My
question "why?" addressed to one of them was answered "because I think so".IE-
peanists, Uralists, and other specialsts in comparative linguistics usually
earn their bread by comparing many languages simultaneously, and nobody finds
it strange.
 3) Anti-Altaicists, if we scrutiny attentively their arguments, tend
to reject certain examples ad hoc, if they do not look alike. Therefore, a
principle of regular correspondences is completely replaced by a search for
look-alikes. Often one can hear something like: "how can you seriously com-
pare J isi "stone" and Turkish tash (sh stands for hushing sibilant approximate
ly as English sh in shame)?" Well, I can since there are regular correspondence
s between the two.
 4) Anti_altaicists, with their theory of omnipotent loanwords tend to
disregard the simple fact that very often their proposals do not make any
sense from the historical point of view. Such is, for example, Doerfer's
claim that Manchu-Tungus borrowed its word *moo "tree" from Mongolian
modun. Taking into consideration traditional habitat of the peoples involved,
the M-Tungus word has the same chance to be borrowed from Mongolian as it does
from Martian, not to metion all phonetical difficulties which arise with such
an interpretation.
 5) This anti-historical approach further manifests itself
 in, for example, such historically irresponsible statements as
 "when ancestors of Chuvash lived in Siberia near Mongols". We have zero
evidence for supporting this statement: but it is certainly necessary to place
Chuvash (westernmost Turkic language) near Mongols in order to justify anti-
Altaistic interpretation of zetacism and lambdaism.
 Regarding the terminology, I find Reinhard Hahn's additional arguments
for Mongolic quite acceptable. Xalxa, Chaxar, Ordos etc., however, may be
also called Central Mongolian (or East Mongolian, whatever one prefers), as
opposed to West Mongolian (Oirat & Kalmyk), North Mongolian (Buriat) and
Mongolian Outliers (Dahur, Dongxiang, Baoan, Monguor, and Moghol). I agree,
however, that his proposal is more logical. I would, however, strongly
insist on preserving Manchu-Tungus rather then Tungusic, for the followingtwo r
easons: 1) Manchu is farther from the Tungusic languages than Chuvash is from
Turkic (at least on the basis of lexicostatistical results); 2) people are less
 aware(for the time being)that Manchu is an early offshot from the rest
of the group than they are aware about the similar situation with Chuvash: ther
e are attempts to group Manchu together with Nanai and other Tungusic langua-
ges of Primor'e region, which are traditionally called South Tungusic", but
to the best of my knowledge all classifications of Turkic except one very
confusing by Baskakov classify Chuvash as a language standing quite separately.

Alexander Vovin
Miami University
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue