Piotr to Mark's point wrote
> The necessary historical and geographic conditions for two language
families to converge into one are very difficult to establish or
maintain. Japan, New Guinea, and perhaps the Caucasus would seem to
be the only possible, presently-known candidates.
They certainly have been difficult to maintain in the the Eurasian-
North and East African Oekumene (the line of plough agriculture at
1500 CE) would give you a good understanding of this area. It is
what Emmanuel Wallerstein and Andre Gundar Frank called "The
Developing World System" and what I cal the "World Development Field".
Outside this region however, I feel the conditions have been a lot
easier to establish and maintain. Australia is a classic example.
So is Latin America. Although I don't know enough about the various
groupings in North America I would suspect that they are the same.
Perhaps, Piotr, the examples you give are not the exceptions to the
Oekumeme rule but rather the areas in which the older, pre-Neolithic,
patterns of languages still prevail. It is interesting that the
areas of exception you posit are mountainous or insular. I wonder if
the Himalayas are another area with the necessary conditions (eg.
Dardic, Indic, Iranian, Hunza, Ladakh and others all in close contact
for a long time. The vocabulary of Basque also shows similar
overlays.
Regards
John