Tuesday, August 06, 2019

FrankN on the latest PIE questions

It seems to me that sometimes ideas get thrown out there that are much bigger than what the data would call for; the tail wagging the dog, as it were, when it comes to some surprising aDNA sample that people then use to rewrite everything that they thought they knew about pre-history.

Now, I don't mean to say that FrankN is one such fellow; in general, I actually find him a very measured and sensible fellow, and not only one of the better contributors in the Eurogenes comments section, but also a regular actual contributor at adnaera.com too. (In fact, I wish he'd make an actual post out of the comments that I'm going to quote below, where he goes into more detail about what he's getting at.)

Anyway, this is a fascinating area of speculation for me, as long as everyone keeps their heads on straight and remembers how broad a brush we're sometimes painting with based on very little inferences of data.  Without further ado: FrankN, some of which is part of an ongoing dialogue with various other posters, most of whom I'm not posting.
Dave:
"there are also several lose ends in regards to the steppe homeland theory, such as the lack of corroboration from aDNA for the steppe origins of the Hittites and Tocharians .."
Fair enough. I'd like to add the Swat Valley samples. The Swat River appears in the Rigvedas as Suvastu, with IE etymology, so the IA samples from there should represent IE speakers. 
Now, there is pretty solid evidence of (late) Proto-Uralic borrowing from Indo-Aryan, and the most plausible zone of contact was the S. Urals during Sintashta times [Parpola argues for Sumerian as (Para-)Uralic, with Maykop as northward link, but I have a hard time to align such a scenario with the archaeological and aDNA record.] Doesn't that mean that - in spite of weak aDNA evidence - the Sintashta, ultimately Steppe origin of Indo-Aryan may be regarded as linguistically assured? 
Well, apparently Indic contains some traces of centum languages, e.g. Bangani koto, Kashmiri hata "hundred". Zoller 2016 assesses the linguistic Picture in quite some detail and concludes (emphasis is mine): 
“I have shown that at the time of Old Indo-Aryan there must have existed a linkage of lects, with Vedic just one of them. These lectal differentiations seem to suggest that the standard model of the three branches of Indo-Iranian is in need of a revision. Their existence also supports the idea of the earlier immigration of the ancestor(s) of the Outer Language which led to a strong encounter with Munda/Austro-Asiatic languages (but to a weak encounter in case of Vedic and Classical Sanskrit) which must have dominated the prehistoric linguistic area of northern India. This dominance must have extended far into prehistory because of the many parallels in the language isolate Burushaski.” 
https://www.hf.uio.no/ikos/forskning/publikasjoner/tidsskrifter/acta/volum_77/ao_2016_cpz.pdf 
IOW: Even if we accept Vedic Sanskrit as Sintashta, ultimately Steppe-derived, it apparently wasn't the first IE language arriving in India. Instead, it overformed earlier IE languages spoken there (which may, or may not, have been related to Tocharian). 
IMO, everything at the moment points to (pre-)PIE originating from the S. Caspian. This includes Matt's latest PCA (thx, Matt, btw, for the efforts you always put into these). The Khvalynsk-Progress cline there points eastwards towards Tajiks, and when extrapolated further towards "ancients", it lands somewhere between Iran_Hotu and Saraszm. [The archeological connection between those two is well established: Neolithisation of Central Asian by the Jeitun culture originated in the Alborz foothills.] 
Having said that, there are still two things that puzzle me: 
1. In terms of basic lexicon, PIE appears to have most in common with Chukotko-Kamshatkan. I also compared the PIE Swadesh 100 list with Proto-Wakashan-Nivkh-Algic (PAW) as proposed by S. Nikolaev (links under https://en.wikipedia.org/wiki/Algic_languages), with a surprisingly high number of parallels that go beyond what I term "paleo-substrate". E.g., PIE *h₁egʷʰ- “drink” and *h₂ékʷeh₂ "water" find almost perfect matches in Nikolaev's reconstruction. What is even more intriguing is that Nikolaew reconstructs *wä- "water" as generic PAW term (water body, liquid, etc.), with *hek - wä meaning "drinking water". So, the analogy isn't just restricted to the basic roots, but extends to the way the roots are combined in order to express a specific concept. 
The issue here is that Nikolaev's PAW is transpacific and ancient, so there is no point in trying to find steppe signals in modern Chukotko-Kamchatkans. We are either dealing with an early Caucasian expansion that ultimately made it to the Americas (c.f. Kennewick Man's mtDNA X2a vs. the X2 in Iran_Neo), or late Mesolithic ENA influence on PIE, e.g. connected to the introduction of "combed"/ "pseudo-corded" pottery that can ultimately be traced back to the lower Amur Epi-Paleolithic. 
2. There seems to have been significant cultural influence of late Mesolithic/ early Neolithic Ukraine on Khvalynsk and Yamnaya, far beyond its rather meager genetic contribution. Points in case are ritual trepanations (first evidence from the Dniepr Rapids area during the 6th mBC, common in Progress/Vonyushka and also Khvalynsk), use of red ochre in burials (a common UA_Mes, Baltic_HG and SHG feature), and also the "proto-Kurgans", essentially not much more than stone heaps, in Progress/Vonyushka. 
Opposite to what Anthony suggests, they hardly have anything in common with the large and richly furnished Leyla-Tepe / Maykop kurgans, but correspond well to contemporary practices in the Donetsk area. 
If the late Mesolithic Dniepr-Donetsk area was culturally so important, why not also linguistically? 
Suyindik has above cited D. Reich as follows: 
"about ten thousand years ago there were at least four major populations in West Eurasia—the farmers of the Fertile Crescent, the farmers of Iran, the hunter-gatherers of central and western Europe, and the hunter-gatherers of eastern Europe. All these populations differed from one another as much as Europeans differ from East Asians today."
So far, the linguistic implications of this observation seem to have gone widely unnoticed. Genetic isolation of course also means linguistic isolation, and a longer period of specific linguistic drift. Most of the major language families have now been reconstructed to proto-level, at a time depth of typically 5-6 ky ago (substantially longer for Afro-Asiatic). Apparently, except for mixed languages/ creoles, since then no new language family emerged. There has also been some recent work on comparing, possibly combining proto-languages into Macro-families, which bears the potential to increase the time depth by another couple of thousand years. However, any attempts to do so for PIE (Nostratic, IndoUralic etc.) have so far remained unconvincing, meaning that (pre-pre-)PIE is apparently very ancient.  
As such, I believe that when it comes to the "homeland" issue, a key question is: When and where could a family develop its specific features that set it apart from all other families. Outside the tropics, the "when" question is answered relatively easily - we are either dealing with the Younger Dryas, but more likely with the LGM, when relatively small populations concentrated in spatially confined refugia, linguistically and genetically isolated from other populations. 
The "where", of course, then needs to be answered separately for each population/ language group. As concerns (pre-pre) PIE, my best guess in this respect is the S. Caspian refugium. Theoretically, Siberia might be considered as well, but then we run out of LGM refugia out of which pre-pre-Proto Uralic and Altaic could have emerged...
"So, what if two consecutive migrations brought IE languages to India?" 
What would have been the first migration? Acc. to Adams, Andronovo already fails for Tocharian, so I can't see it having reached India. Also, the Swat_IA samples preclude significant Steppe (Andronovo) ancestry in that "outer language" area described by Zoller - even more so, if the meager Portion of Steppe ancestry in them is to be attributed not just to Sintashta-derived overforming, but actually to two Steppe migrations.  
Nah, the only plausible source of this first migration is Central Asia, and that means ultimately (directly, or with a stop-over in Zerafshan) the S. Caspian. While we still lack aDNA IVC aDNA, otherwise the NE Iranian connection is genetically obvious - be it Reich's old ANI, haplogroups of Indian goats and sheep, or the ca. 3500 BC Walnut finds in Kashmir. Modern Kashmiri Walnut trees have been established to descend from S. Caspian stock, one of the 2-3 LGM refugia for the Persian(!) Walnut (the other refugia lay around the Pamirs/ Tian Shan, that's where a/o modern Chinese trees originate from).
The Narasimhan preprint clearly distinguishes Steppe_EBA (w/o EEF/ANF admix) and Steppe_MLBA (with such admix). So these 22% Steppe_MLBA are Sintashta-related, not from Andronovo, and there is actually nothing in the pre-print suggesting Steppe_EBA (Andronovo) making it across the Hindukush. 
22% is what I call "meager", considering the impact of a similar magnitude of admixture on Basques (speculatively also Etruscans). 
In addition, look at the timescale: Steppe_MLBA ancestry appears around 2000-1500 BC in BMAC, by 1200 BC it is attested in the Swat valley. Now, the Rigvedas are commonly dated to around 1500 BC, and explicitly mention the Swat river as Suvasto (with IE etymology). Steppe_MLBA alone is not only too little, but also too late to explain this IE toponym.  
After all, the Swat is quite a major river (and was important enough for being mentionned in the Rigvedas), and such rivers tend to keep their names for a long time - c.f. Missisippi (Algonquin), Guadalquivir (Arabic), Havel (->W. Slavic Hevelli), etc. This holds even more true if an ancient resident population isn't completely replaced, but only assimilated through "elite dominance" or whatever else you want to call 22% admix.

No comments: