Andreas Johansson wrote: >According to a book called "Språken och historien" (="The Languages and >History") by Tore Jansson, which happens to occupy ~2cm of my bookshelf, >there where ~10 000 languages spoken around the world at the onset of the >Neolithic. This figure is apparently based on an estimated global population >of around ten million at this time (a figure I've seen in some other sources >too), and an estimate, based on modern and recent hunter-gatherer societies, >that languages of such pre-agricultural groups tend to have something on the >order of a thousand speeches. That strikes me as rather high. Assuming (at the very best) 50% (the men) would be active hunters, it seems to me it would require quite a large territory, which could entail long absences from the home base, possible encounters with hostile others (and the home base unprotected as well). IIRC, Yanomami villages tended to be on the order of 2-300; and I have the impression other Amazonian groups (among the last untrammeled hunter-gatherers) tend to be equally small. Another largely untrammeled group, the Eskimo-- how large are their communities? The well-known S.African group, whose name I forget, no longer live in an unrestricted environment. Indonesian tribal groups that practiced both hunting-gathering and slash-burn agriculture also, according to reports (19th C), tended to be fairly small, though a bit larger than 2-300, maybe 5-600 per village. The Catch-22 of course is that as you destroy the forest for gardens, you destroy the habitat of your game, so the hunter's have to go further afield.....etc, as above. Even slow population growth eventually outstrips the nearby resources-- in that case, the village would split up and some would have to find new territory. Voilà-- instant dialect divergence, and ultimately a distinct new "language". The various groups known as Toraja in central Sulawesi apparently developed in this way.