Chinese and Tai (was: isoltes)

[Scott DeLancey]

On Thu, 27 Mar 1997, Miguel Carrasquer Vidal wrote:
 
> and Tocharian were discovered later].  What reconstruction *can* do,
> is to show relationships that were not quite so obvious previously,
> and, most importantly, it can show that relationships which were
> thought possible (but not obvious) previously are in fact false.  I
> would imagine, for instance, that the work done on reconstructing Old
> Chinese has been crucial in dismissing any close genetic relationship
> between Chinese and Thai or Vietnamese.
 
Well, no, as a matter of fact.  Reconstruction of Old Chinese does
a lot to strengthen the case for Sino-Tibetan, i.e. the relationship
between Chinese and Tibeto-Burman, as reconstructed OC looks a lot
more like TB than later stages do.  I think, though I'm not up-to-
date on the literature, that recent work on comparative Tai-Kadai
reconstructs a proto-language which is phonologically less similar
to Chinese than modern Tai languages are, and thus bolsters the
case against Sino-Tai.  But none of this is "crucial"; the case
against the relationship of Chinese, Tai, and Vietnamese is pretty
clear without any reconstruction of anything.
    The original basis for the notion of Sino-Tai was resemblances
in phonological structure (tones, monosyllabicity, similarities of
syllable structure) and a large body of shared vocabulary, with nice,
tolerably regular, correspondences and everything.  The original splitting
suggestion (Paul Benedict's, published in 1942) was based on the
observation that this common vocabulary is virtually all
the sorts of thing that we know are easily borrowed (technological
items, metallurgy, trade goods, conspicuous animals originally
characteristic of only one language area ('elephant', 'horse'), terms
referring to markets, etc.).  None of the basic vocabulary (pronouns, kin
terms,  natural phenomena, geographical features, body parts) is shared.
This makes a pretty good case that all this is borrowed (the borrowing
went in both directions, it turns out), and thus not evidence for
relationship.
     The splitting case is strengthened when we look at Sino-Tibetan.
It turns out that very little of this shared vocabulary turns up in
Tibeto-Burman--but the vocabulary common to Chinese and T-B contains
all kinds of items, including pronouns and other pretty basic stuff.
So Sino-Tibetan is pretty secure.  But there's no way to construct a
family tree linking Chinese to both T-B and  Tai, when the sets of
vocabulary that it shares with each are essentially disjunct.
     (So, by the way, while as others have noted, it is impossible
in principle to demonstrate that two languages are absolutely not
related at all, it *is* possible to demonstrate that they are not
related at some particular level--e.g. that, regardless of whether
English and French are related at some level, French is clearly
not Germanic, or that, even if Chinese and Tai might be remotely
related, Tai is definitely not Sino-Tibetan).
 
Scott DeLancey
Department of Linguistics
University of Oregon
Eugene, OR 97403, USA
 
delancey at darkwing.uoregon.edu
http://www.uoregon.edu/~delancey/prohp.html