--- In
cybalist@egroups.com, "Guillaume JACQUES" <xiang@...> wrote
in reply to Piotr's:
> I'd like to hear an Austronesianist's opinion about the likelihood
that the dispersal of AN took place out of Sundaland ca. 8000 BC.
>
> GJ:
> Both srcheologists and linguists specialized in AN agree that AN
languages are originated from Taiwan. As I pointed out earlier on
this list, the Malayo-polynesian languages are just the last sub-
branch of one group of Formosan languages. Any reconstruction of AN
that does not take languages of Taiwan as a basis is deemed to remain
at the Malayo-polynesian level (ca 3000 BC). This has far-reaching
consequences when you try to make external comparison (for example,
PMP *mata 'eye' is now reconstructed *maCa, the segments C and t
merged in PMP and some formosan languages)/
> Agriculture seems to tell us that Austronesian arrived in Taiwan by
6000 BC, beginning of the Neolithic (that is the date given by Blust,
Starosta and Reid). Much/ data seem also to indicate that AN lang.
are originated from northern China, but this list is not the proper
place to discuss that.
> The only serious external comparisons with AN are Austro-asiatic
(mostly morphological) and Sino-tibetan (many affixes, more than 30
words in Swadesh's list of 200 basic words)
>
> In any case, Manasala doesn't seem to understand pAN reconstruction
well.
RM: Thanks to Guillaume for relieving me of some of the burden. I
suspect he is much more up to date than I. Sorry to say I haven't
read either "Eden of the East" or Manansala, but the one's theory and
the other's data (as discussed here) don't inspire much confidence--
then, too, I'm something of an old fogey and rather conservative when
it comes to historical linguistics.
Just a couple comments on some of the forms cited in Torsten
Pedersen's web-pages.
Supposedly related to Skt. math 'kill, exterminate', the several AN
forms all correctly reflect PAN *maCey (the *C correspondence is
generally Formosan affricate /ts/ vs. all non-Form. /t/; PAN *e is a
typographic convenience for schwa); there is also *paCey, and the two
forms are related by a very common morphophonemic alternation p-
~m-. Unfortunately, my Whitney's "Sanskrit roots" has math ~ manth
'shake'; perhaps some derivatives mean 'kill', but it's a stretch.
The Polynesian words for 'moon', masina/mahina et al. cannot belong
in the word-family cited. The base is sina/hina (normal PN variation
between s/h), with a fossilized ma- 'stative or adjectival' prefix.
PAN is *sinaR 'bright; light' (where *R represents the correspondence
Malay /r/, Tagalog /g/ (plus others), PN zero.
One can be led badly astray by ascribing the "d/n/l/r alternation" to
(proto) Austronesian; they are the reflexes in a variety of
individual languages, so of relatively recent development. Similarly
the "k/t" alternation-- in a mere handful of modern languages (best
known is Hawaiian) *t has shifted to /k/ (and in fact there are said
to be dialects of Hawaiian where the shift has not taken place).
I would question, frankly, whether we can even conceive of the
existence of "Austronesian" as such in 8000 BCE or so. Pre-pre-PAN
maybe-- pre-pre-PIE too, probably.
I'd be interested to discuss (off-list) with Guillaume his evidence
for a north-Chinese origin.