Piotr wrote
>For example, I wouldn't accept
Amerind as a genetic unit. You mention Australian, but I have doubts
even about the status of Pama-Nyungan: Robert Dixon (_The Rise and
Fall of Languages_, 1997) argues that it's not a family at all, but
the result of tens of millennia of areal diffusion in Australia's
"tangled bush".
Do you have further bibliographic details for Dixon? I don't know
much about the further reaches of Pama Nyungan, (i.e. Pama) but do
know a fair bit about Nyungar, having shared a house with the first
fluent "white" speaker of the language in about a century, currently
attempting to do what people have suggested we do with Neo PIE and
rejuvenate a language with modern non-Nyungar speaking Nyungars. It
would appear there was a cultural uniformity in the Nyungar area
going
back 39.5 thousand years, with no genetic shifts in the population
showing that there was no migration of people into the area from
outside. There was probably a movement into the SW of Western
Australia 18,000 years ago with the Ice Age maximum and the spreading
of dune fields across most of the continent. Nyungar is
clearly cognate with Yamatji and Wangkai and the languages of
the Western and Great Sandy Desert as a result. Nyungar people in
Perth have clear and accurate memories of the late Pliestocene and
early Holocene sea level rises, which argues for great continuity.
There is also evidence of continued long-distance trade across the
Pama Nyungan area, which would have had some effect on maintaining
cultural (and hence linguistic) uniformity. The "trading" of Dream
Time stories has been shown to suggest that most of these are less
than 5,000 BCE old, which corresponds nicely with the arrival of the
dingo from SE Asia, and the spread of Circumcission as an initiation
ritual (it never penetrated either SW WA (which is known to the
Western Desert people as "The Land of the Boys" as a result) or SE
Australia.
Against the vast depth of Aboriginal languages is the fact that most
Western Australian languages had the "death taboo" abaginst using the
name of those who have died, similar to that found in the Papua New
Guinean Highlands. If this had a common origin in Australoid
cultures
it is very old. This "death taboo" has been used by New Guinean
linguists to help explain the linguistic diversity of the Highlands
(eg. if you have someone called "Pandanus Nut", when that person dies
you have to invent a new way of referring to the real pandanus nut).
This is one reason why so many Aboriginal languages finish in the
word
"Jara" (eg Pitjanjatjara), as this means "The people who say
Pitjanjat
- usually for a word in another language with a sound system
something
quite different.
Piotr continues
> The African "phyla" are likewise highly controversial.
Very well, you prefer to lump and I prefer to split. But the
"weeding-out" process you describe is something I could certainly
subscribe to. The way I see it, it was mainly the recent "great
expansions" of the major modern families that obliterated the
original
diversity of some of the continents. In the Pre-Neolithic world
natural equilibrium favoured the existence of tiny families (as in
New
Guinea, which has about sixty of them), with a lot of lateral
diffusion and creolisation to make sure that the bush was really
tangled.
It was Stephen Wurm who originally showed that most could be fitted
into a Trans Papuan Phylum. More recent work suggests a single
Indo-Pacific Phylum unifying all of the smaller families - (eg
Torricelli, Sepik Ramu, Trans Papuan etc) - that as with others seems
to be due to the spread of a very deep "agricultural breakthrough"
some 28,000 years ago. Sriggs 1993 "Pliestocene Agriculture in the
Pacific" in Sahul in Review: Pliestocene Archaeology in Australia,
New Guinea and Island Melanesia" has shown that taro Caolcasia
esculenta was probably cultivated at Kilu cave 28,740 BP and
Fredricksen "Pamwacke Rockshelter on Manus Island" 1993, Research
School of Pacific Studies, ANU shows that the Canarium almond was
being cultivated from Manus, through the Bismarcks to the Solomons at
an equally early date. As farming was occurring at Kuk, 12,000 BP,
this makes Papua New Guinea the site of the world's oldest
agricultural dispursion, making the Indo Pacific Language Family a
real possibility.
Piotr continues
>But even those small-sized units were the "lucky drawers" in
the evolutionary lottery -- most of their cousins had been weeded
out.
The more recently acquired technological and cultural advantages had
the effect of increasing the size of the winners -- they enabled some
of the successful families to become middle-sized (by our standards)
or even huge.
I would strongly agree with this statement.
Thanks for your comments.
Regards
John