E-mail this message to a friend

Re: LINGUIST List issue: http://linguistlist.org/issues/18/18-1697.html 
 
To answer - with some delay - Hal Schiffman's question (''...there are 
tonal languages like modern Tibetan that were originally non-tonal, i.e. 
Classical Tibetan has no tones. Would the genetic hypothesis claim that the 
genes of Tibetans have changed radically in the last two millennia?''): 
 
1. How do we know that Classical Tibetan CT, nota bene only a written 
language, had no tones? We have some indications that at least at the time 
when the second grammatical treatise, the so-called _rtags-kyi vjug-pa_, 
was written or compiled (unfortunately we can only speculate when this 
happened, probably not before the 9th century CE), the spoken language must 
have developed tonal register properties, most probably on a sub-phonemic 
level, since we can find a four-fold classification of root consonants (the 
core initial following the pre- or superscribed consonants) in terms of 
gender: 1. 'male' (CT unvoiced root consonants), 2. 'neutral' (CT unvoiced 
aspirated consonants), 3. 'female' (CT voiced consonants), and 4. 'very 
female' (CT nasals). 
 
The gender description most probably did not refer to the pitch but to the 
articulation force ('male' = forced = high; 'female' = soft = low), as the 
same or similar tonal (register or pitch) properties can be found in some 
of the modern emerging-tone dialects. While the phonemic register 
distinction is only binary (former unvoiced = high vs. former voiced = 
low), some modern emerging-tone dialects in eastern Ladakh and possibly 
elsewhere may have a sub-phonemic five way distinction: 1. high for 
originally unvoiced root consonants, 2. neutral for originally 
unvoiced-aspirated root consonants, 3. low for devoiced root consonants 
(those that originally had no initial affixes), 4. lower for still voiced 
initial consonants (originally with initial affixes), and 5. lowest for 
voiced nasals (originally without initial affixes; those nasals which had 
affixes became unvoiced and thus high). 
 
2. 'Modern Tibetan' is a quite misleading term, referring only to the 
Central Tibetan coiné and its further development in the exile or even more 
specifically only to the dialect (sometimes even only a particular 
sociolect) of Lhasa. There are many modern Tibetan varieties that are non 
tonal or emerging-tone dialects, particularly those of Amdo and Kham 
(Eastern Tibet), Ladakh (India), and Baltistan (Pakistan). 
 
3. The earliest documents of the Tibetan language date from the 7th century 
CE, the question whether genes might have changed is thus not one of 2 
millenia but hardly of 1. But why not? 
 
We have some evidence that throughout the documented history, i.e. since 
the 7th century as well as in the prehistorical era, the whole Tibetan 
speaking area suffered repeated migrations, involving not only 
Tibeto-Burman groups, but also Indo-European, Turco-Mongolic, and most 
probably other minor groups. Thus first of all, we cannot start with a 
genetically well defined Tibetan identity. And secondly, yes the genetic 
composition on the Tibetan plateau may well have changed during the last 1 
millenium. 
 
Furthermore, it is well possible that (some of) the prehistorical ancestors 
of the now Central Tibetan population were originally speaking a tonal 
language. We know from a 9th century inscription in Lhasa containing 
Chinese transcriptions of Tibetan names that the officials at the Central 
Tibetan imperial court were obviously speaking quite a different 
dialect/language than those people who were responsible for the Old Tibetan 
orthography. 
 
4. However, and this as well as the following is a comment on D.R. Ladd & 
Dan Dediu's paper, low tone (pitch or register) is apparently only one 
phonetic feature attached to voice (which seems to be the outcome of a 
bundle of accustic or articulatory features) and can also be found in the 
European languages as a specialist in phonetics once explained to me. 
According to him, many if not most phonemic pitch oppositions resulted 
diachronically from an original voice opposition. Apparently the speech 
community starts neglecting one of the features, emphasizing one of the 
other features. I really wonder whether this shift in acoustic attention 
can be triggered by a genetical disposition, particularly as this shift may 
be very gradual: As one can observe in the modern Ladakhi varieties, the 
voice distinction may partially break down without the introduction of 
pitch distinctions, or a pitch distinction may evolve while retaining the 
voice distinction to a great extent. Phonetic pitch distinctions (as 
perceived by the speakers themselves) might be 5-fold as described above or 
only binary, leading to the inability of speakers to correlate aspirated 
and voiced consonants in any meaningful way (say 'neutral') to this opposition. 
 
It would be very promising to look for possible genetic correlations to 
these quite distinct developments in a quite restricted area. It might be 
even more promising to look whether the individual abilities to perceive 
non-phonemic pitch distinctions might be correlated to the proposed genetic 
variation. Such correlation should follow from the authors reaffirmations 
that, on the one hand, tonal distinctions can be introduced into a speech 
community by language contact, but that, on the other hand, there is 
something like a genetic disposition to focus on pitch or register instead 
of on voice. 
 
5. Many tonal languages do not only have phonemic pitch distinctions, but 
also contour distinctions (level, falling, rising, etc.), which might 
likewise be connected with diachronic sound changes, e.g. in the syllable 
coda as in the Tibetan tone languages. The reasons for giving up voice 
distinctions is certainly a different one than that for reducing syllable 
codas. How does the genetic hypothesis account for this difference? 
 
6. The - at least in some cases - observable diachronic shift from voice 
distinctions to pitch distinctions seems to be at variance with the claim 
that ''the distribution of the older ... alleles ... resembles the 
distribution of tone languages'' and that the ''new alleles'' correlated to 
non-tonal languages ''seem to be spreading quickly in the human species''. 
From these statements one should have expected a diachronically observable 
shift from tonal to non-tonal languages, not the opposite way.