Deep Dates.

junk2 Piotr writes:

I must warn you that my own view of the absolute chronology of IE splits is not quite orthodox, and that the dates I prefer are deeper than assumed by most authorities (though not as deep as Colin Renfrew's, for example). I assume that the first branchings in the IE tree should be dated to ca. 5600 BC (the split into Anatolian and Non-Anatolian), and that the ancestor of Tocharian may have been a distinct dialect already in the fifth millennium BC, if it really arose from an early branching of the Non-Anatolian subfamily. If it could be clustered with Germanic, Italo-Celtic or any other branch, we'd need a shallower, possibly Bronze Age, date. The trouble is that the demonstrations I've seen are not convincing. The oft-cited lexical equations (the Northern SALMON word allegedly corresponding to Tocharian FISH, Germanic NECK having a seeming Tocharian cognate) are doubtful, being both isolated and based on superficial similarity. Morphological affinities connecting Tocharian on the one hand with Anatolian, and on the other with Italo-Celtic represent shared archaisms, not innovations, and as such don't prove anything, though they are very important for the reconstruction of PIE. I'm not aware of any special connectons between Tocharian and Aryan, despite their geographical closeness in historical times. Presumably the Tocharians spent some times in the souther Urals or thereabouts, in the neighbourhood of Altaic-speakers, and reached the Tarym basin and China from the northeast.
My personally preferred dating for the Italo/Celtic split would be some time in the 3rd millennium BC. At the time of the Italic expansion into modern Italy the branch seems to have consisted of a northern group (Venetic), which may have extended far into central and northern Europe, and a southern one (Italic proper). I regard the Illyrian group (insofar as it is a real grouping, rather than the historical linguist's waste-paper basket) as a residual offshoot of the IE movement up the Danube and into central Europe -- the one that produced Italo-Celtic and possibly Germanic. I tend to reject other than areal (Sprachbund) connections between Germanic and Balto-Slavic, though this again is my personal opinion; the genetic unity of Germanic and Balto-Slavic is something many linguists believe in (Alexander Stolbov and I have discussed this on Cybalist about a month ago).

"I assume that the first branchings in the IE tree should be dated to ca. 5600 BC (the split into Anatolian and Non-Anatolian)"
A date this deep raises the usual objection about the Sherratts' secondary products revolution, and the words that go with them. It also raises the question of where the homeland was and what archaeological complex it is to be associated with.
While glottochronology is in bad odor, seat-of-the-pants conclusions based on what happened in Romance, Germanic and Indo-Iranian, and the time scales involved make extremely deep dates hard to take. While the example of Lithuanian is often held up (spectacularly archaic!), you hear considerably less about it's innovations.
"I regard the Illyrian group (insofar as it is a real grouping, rather than the historical linguist's waste-paper basket) as a residual offshoot of the IE movement up the Danube and into central Europe -- the one that produced Italo-Celtic and possibly Germanic."
I'd like to hear more of this. 'Up the Danube' -- from where? The Tisza, the Vardar, or right up through the Iron Gate? My own thinking has IE intrusions early on into Hungary via the Tisza (the plains of Hungary are quite suitable for grazing livestock, and are better watered to boot).
My own private theory (not provable) is that Greeks entered Greece via the Vardar, coming down from Hungary via Belgrade and Skopje. This has the advantage of separating Greeks and Thracians for a sufficient time period for the two languages to differentiate (tho', admittedly, we know less of Thracian (or Dacian) than we do of Etruscan).
As for the Anatolians, the standard model jibes well with the Ezero culture (Bulgaria, ca 3200 BCE), which seems to have been the dominate culture of that part of the world at that time; the Troad gets its 3rd millennium culture from here. The Anatolians would seem to have not gone to Greece, but instead, crossed by sea to NW Anatolia. These Ezero-ized IE-speakers would have had a higher culture than what they found in Anatolia.
Did the Trojans speak Luwian? Years ago, I read a speculation that the roots that give us Ilion, Ilium, Ilus (=the eponymous founder) and 'Luwian' may be reflexes of the same word.
As Piotr knows, I timidly keep myself within the According-to-Mallory standard model; I mostly parrot the authorities I've been reading. This model, however, best explains all the evidence we presently have. About the only amendment I would make is that the PIEs, at the final point of unity (including Anatolian) would seem to have occupied two distinct habitats, the forest-steppe boundary, and the Northern European forest itself. Geography impels us to look to Kiev as the approximate center of the PIE homeland. The proto-Balto-Slavs go up the Vistula. The pre-proto-Germanics go up the Elbe, or more speculatively, precede the B-S'ics up the Vistula or some such river and ensconce themselves in Sweden with the B-S'ics displacing those left back on the North European Plain. The Celts are right behind the Germanics, but go right up the Danube to its headwaters and down into France via the Seine.
I've become wildly speculative here. I can prove none of this, nor cite any of my authorities.
I encourage Piotr to further develop his theory. A deeper date is actually more emotionally satisfying. I would not mind if Piotr becomes one of the authors of the IE books I cite. Till then, however, I'll be an interested, friendly, and hopefully, helpful onlooker, peering over his shoulder, hoping he'll change the standard model.
Mark.