We posted (see previous post) another article, from the Times, on the IE question:
The original article in Science is behind a paywall so I am frustrated in the attempt to evaluate this material. But already I am suspicious. Here is our original commentary.
Let me say that I am not completely against this kind of hypothesis, which I find intriguing, but I suspect the experts in computer technology don’t understand the known facts here. Again, since I can’t read the original I can’t be sure what they are saying. What Anatolian languages are they referring to, if any? If they don’t have any candidate for that region, the claims are a bit ‘up in the air-ish’. They could have other statistical grounds for their claim, with no such Anatolian language (which apparently has left no trace/no longer exists in Anatolia, please don’t mention Hittite from the second millennium BCE here).
Have these researchers understood the rates of language change? It makes no sense to say that an agricultural exodus in the 10th to 9th millennium BCE plus to Europe, et al., was the source of the IE languages in use there now!
That makes no sense: as we pointed out the interval from the present with, say, English, and Homeric Greek (with Myceanean Greek its immediate antecedent) is close to three thousand years. Three thousand more back to 6 thou BCE would double the distance, back to 9 thou BCE triple that. The source language would be unrecognizable, might be something totally different. That’s not what was said: they merely indicated certain key language terms (‘mother’ nexus, etc…) as often invariant. Could be, but this method can’t be called proof. Possible, but…
It is important to ask these people if they are rightly oriented toward the data: Colin Renfrew’s claim, for example, is simply up in the air. The people who migrated from Anatolia to Europe in the Neolithic might well have done so (the genetics favors this) but could not have seeded Indo-European languages there, if by this you mean those in the forms now present: Germanic, Italic to Romance, Greek (and Slavic further east). Not possible. The rate of langauge change shows that all these had a common source (whereever that was) after 5 thou BCE (and that is stretching it, 3 thou BCE is the other limit). And this common source has to produce Vedic in parallel to Greek, and find its way into India, with Tocharian reaching China. I don’t buy it. (My dates need a tighter grip, but the basic thesis is so weird that these fuzzy dates are good enough).
So the claim would have to be they seeded some IE prototype that is no longer present, perhaps displaced by later IE overlays. That is a pretty bizarre thesis. I don’t buy it. There must be some linguistic traces, of some kind, of those Neolithic farmers who entered Europe in the Neolithic. There are, but they aren’t IE languages.
None of this contradicts what to me is a more interesting thesis: that some Anatolian group from the Neolithic exited Anatolia into Central Asia, there either seeding proto-proto-IE, or else adopting an IE prototype from some Central Asia source. But without proof there are no grounds to proceed.
Note that Homeric Greek and Vedic Sanskrit, in the mid to late 2nd Millennium BCE show clearly related languages, and clearly related invasions: and these can’t have come from Anatolia in that time frame. Their common source is clearly ca. 3000 BCE, and that has to be somewhere reasonable as a source. Again, Anatolia won’t work, because the language referenced for 8000 BCE couldn’t have had any recognizable relationship to Homeric/Vedic.
I have not read the article, but this kind of computer research, while of great interest, seems to be violating linguistic common sense, as indicated. The thesis could still be right in some different interpretation.
I will have to wait to read the original research. But computer methods here are not that trustworthy. These computational biologists have gotten evolutionary computation wrong on Darwinism, so they may be bemused by a false rigor, this suspicion must remain strong.
Nonetheless, an Anatolian exodus to central Asia (and/or Europe) with some very very early substrate proto to proto-IE is one last possibility here, and an interesting one.
I have a suspicion these researchers have no feeling for the grammar of early Greek/Sanskrit and are thinking, like Renfrew, that IE speakers entering Europe seeded the IE languages there now. NOT!
Again, it is possible that some Anatolian group, very early, migrated to Central Asia, in the same stream as the later IE. But somehow the claims for Pontic IE (with their words for chariots, etc…) have to remain as a transit stage.
We should always be wary of these questions: consider that post-Fall_of_Jerusalem Israelites in early AD, very close to Anatolia, found themselves in the same region of the IE groups, and ended up in Yiddish/Germanic zones two thousand years later. That example is unnerving: a warning to never jump to conclusions about anything here.
If archeologists have any evidence, linguistics apart, for a similar exodus of an Anatolian group to the Caucasus the original thesis here of these researchers might come alive.
There is another issue: IE languages originally had an incredible inflectional complexity: the Sanskrit (and Greek/Latin/Russian/Lithuanian) cases give some inkling of this. Sanskrit and Greek had hundreds of verb forms, and the nouns have considerable inflection. But this inflectional complexity is decreasing over time, in most cases, with some cases, such as Russian moving more slowly, and with English we see the terminal point, almost, with vestigial inflections: move/moved, thing/thing’s (genitive). What explains this inflectional decay? And its rate of change is quite considerable, from Homeric Greek to English, or Vedic to Pali to Hindi…
The decay is universal, but slower in some cases. This suggests a question: when did inflectional complexity peak? And, even more strange, when did its complexification initiate? Sanskrit suggests that this peak could not have been in the earliest purported phase of some anatolian IE. So we must posit that later IE inflectional cmplexity is relatively recent: a millennium or two before the splits into the languages we now know. This proves nothing one way or the other, but an IE as suggested would have been wihout these inflections, we suspect, since we can only posit the rapid decay of inflection once it has peaked, or come into existence.
Curioser and curioser. This is not airtight reasoning, but the issue of inflectional complexity should haunt any effort to push IE back too far. At some point inflectionality must have gone into overdrive. It must have done so some hundreds of times, since inflectional structures are very common. It is all very puzzling. A speaker of English has a hard time with this kind of thing. But in fact inflectional complexification could reoccur with a kind of baseline mongrel we have with English.
Again there is no contradiction in some proto-proto-IE without inflectional complexity. But the ‘frozen’ character of Lithuanian/Russian inflectional complexity makes this argument treacherous, although probably roughly right.