The Pontic Steppe vs. the Bactrian Homeland of the Indo-Europeans

Dienekes, the confused blogger from my previous post, continues to be confused and confusing. In the comments section of Razib’s Discover Blog, he concedes:

“As of late, I am rather more willing to give even Johanna Nichols’ 1997 model of a Bactria-Sogdiana homeland … serious thought.”

“As of late” means within the past couple of days because as recently as July 3 he believed that “Proto-Indo-Europeans lived in the Transcaucasus.”

Ironically, I was the one who apparently introduced Dienekes to Johanna Nichols’s Bactrian theory of Indo-European homeland. It was in April of 2011. I’m honored to personally know Nichols and I do admire her academic contributions. And I did use her research to back up the Out-of-America hypothesis. Dienekes immediately attacked me for my out-of-America theory of modern human origins and Johanna Nichols for her theory: “Having read the theory, I see absolutely no value to it.” Apparently, in Dienekes’s world of primitive mentality in which ideas are judged by their deictic association with other ideas, if a scholar thinks that modern humans may have expanded from the New World into the Old World, rather than vice versa, his appreciation of another scholar’s theory of Indo-European homeland makes this latter theory just as unlikely. Dienekes 2011 continued his typical scienzoid rant:

“Nichols’ hypothesis is not parsimonious because it proposes a mysterious drang nach Westen for Proto-Indo-Europeans. Parsimony dictates that Indo-European dispersals would not have a directional preference, and, barring natural obstacles, would radiate in all directions from the homeland. Plus, there is no archaeological evidence for it. On the contrary, pastoralism spread West->East across Eurasia, and agriculture spread from West Asia. There is absolutely no reasonable agent that could have affected a replacement of languages all the way to the Atlantic from an origin in Bactria.”

It looks like Dienekes 2012 is now willing to drop his “parsimonious” West Asian theory for the Bactrian theory. It’s always good to see amateurs evolve. But enough of Dienekes.

Just like out-of-America reverses the conventional logic of modern human origins but retains all of its geographic anchors, Nichols’s out-of-Bactria reverses the conventional logic of Indo-European dispersals but is just as consistent with Indo-European dialect geography. Just like out-of-America challenges the assumption that greater intragroup genetic diversity means greater age of the population on such grounds as, for example, the fact the recent colonization of the Americas post 1492 has resulted in the stupendous accumulation of world-size genetic diversity in the most recently colonized continent, Nichols showed that the diversity-equals-age assumption in linguistics is demonstrably false in the areas occupied by Indo-European and Turkic languages. (This fact also may eventually put a question mark next to the out-of-Taiwan theory of Austronesian dispersals.) She, therefore, turned her attention to the opposite pole of Indo-European dispersal zone, the easternmost one marked by Tocharian A and B attested in western China and Indo-Aryan attested in India, and suggested a Bactrian homeland for the Indo-Europeans. The apparent dialect uniformity found in the eastern extremes of Indo-European geography is explained as the product of the later sweeping expansion of Iranians across Central Asia, the Caucasus and northern Iran, which resulted in the obliteration of ancient Indo-European dialect diversity in Central Asia. (In the out-of-America model, this logic of diversity extinction manifests itself in the diversity-reducing effect of “Mongoloids,” who evolved in northern North America, in the Beringian refugium or in Northeast Asia at the end of the Ice Age, on the American Indian populations further down south.)

In Nichols’s model, Proto-Indo-European first expanded westward toward the Caspian Sea as a continuous wave and then split into two branches – one took the northern route via the steppe and spawned Slavic, Italic, Celtic and Germanic branches, the other one the southern route into Anatolia resulting in the chain of such dialects as Anatolian, Phrygian, Greek, Armenian, Thracian, Illyrian and Dacian. Both waves converged in the Balkans, which resulted in the high diversity of Indo-European languages there. Indo-Aryans went down to colonize the Indian subcontinent, while Tocharians stayed behind in close proximity to the ancestral homeland. At a later point, the expansion of Iranians sequestered the northern dialects from the southern dialects and the easternmost dialects from the more western ones. In addition to the reversal of the diversity-equals-age argument, Nichols based her theory on the observation that Mesopotamian loans were not borrowed by Indo-Europeans directly but only via a third party (contra a West Asian source of the Indo-Europeans), that there are very few loans between Indo-European and North Caucasian languages (contra the Pontic Steppe theory of Indo-European homeland).

The latter observation is important in light of Matasovic’s recent paper supporting the Pontic Steppe theory on the grounds of multiple typological similarities (consonant-to-vowel ratio, tonal accent, number suppletion in personal pronouns, two-stem inflection, the presence of gender and the morphological optative and, possibly, the presence of glottalized consonants and ergativity) between Indo-European and Northeast (and less so Northwest and even less so South) Caucasian languages. Matasovic also admits that lexical loans between Indo-European and North Caucasian are surprisingly few. I counter it by highlighting three possible loans in the tight domain of affinal vocabulary suggesting marital exchange between Proto-Indo-Europeans and North Caucasians. But this may not be enough to justify the Pontic Steppe theory on linguistic grounds. Ultimately, as Matasovic admits, some of his shared typological markers depend on the kind of Proto-Indo-European (PIE) reconstruction a scholar adheres to. For instance, the presence of glottalized consonants in PIE is postulated by the Glottalic Theory but not by others. Other putative typological parallels may be accidental. For instance, how similar really is two-stem inflection in PIE (e.g., PIE *yekwr (NOM and ACC SG) vs. *yekwns (GEN SG) ‘liver’) and North Caucasian (e.g., Dargi nes ‘mother’, oblique nes-li (dative nes-li-s), plural nes-ani, oblique plural nes-an-a– (dative nes-an-a-s)? In a stark contrast to Matasovic, Nichols, who is a foremost specialist in Chechen, believes that “Northwest Caucasian is radically unlike any other Eurasian stock in most respects” (“Modeling Ancient Population Structures and Movement in Linguistics,” Annual Review of Anthropology 26 [1997], 377).

Phylogenetically speaking, in my opinion, the very fact that Tocharian competes with Anatolian for being the most divergent Indo-European language, while being attested from a much later historical period, should have generated an “eastern” theory of Indo-European homeland even prior to Nichols’s sophisticated analysis. When you have two extinct and divergent languages found in the opposite extremes of a language family’s geographic range, it means it’s just as likely that Indo-Europeans came from West China as from Anatolia or the Pontic Steppe. This means some dialects went extinct or got absorbed by a superstratum and their record may never be found. The Pontic Steppe theory was originally built on archaeological grounds and the Kurgan cultures are millenia older than the attestation of Hittite. Hence, it’s likely that the Indo-European language spoken by the inhabitants of Yamnaya and other sites would have been even more divergent than Hittite or Tocharian if we had any record of it.

A decisive test to the Pontic Steppe vs. Bactrian theories of Indo-European homeland will come from Ilija Cašule’s hypothesis of phyletic kinship between Indo-European and Burushaski. Cašule’s belief in the special relationship between Burushaski and Phrygian aside, the proposed kinship between Indo-European and Burushaski may prove Nichols absolutely right. Burushaski’s location between Tocharian in the northeast and Indo-Aryan in the south, its ostensible archaisms (laryngeals are attested, the Indo-European kinterm suffix –ter is still productive) and the layers of loans confounding the “essentially Indo-European” lexicon may serve as a remarkable confirmation of Nichols’s thesis that the original Indo-European dialect diversity in the east was obliterated by later population expansions. Ilija Cašule has kindly shared with me a draft of his upcoming paper on Indo-European and Burushaski kinship terms but asked not to discuss it on the web prior to is publication. His proposal is indeed intriguing on many levels but it’s far from proven.

Genetically, Burusho score high frequencies of Y-DNA R1b (27.8%) followed by a host of minor lineages (Firasat, S. et al. (2007) “Y-chromosomal evidence for a limited Greek contribution to the Pathan population of Pakistan,” European Journal of Human Genetics 15 (1): 121-126.). Tarim mummies, the possible proxies for Tocharians, showed up as Y-DNA R1a1a. Both R1b and R1a are found among many Indo-European-speaking populations of Europe, with the marked concentration of R1b in Western Europe vs. R1a in Eastern and Central Europe. Y-DNA hg J, with clear links to West Asia/Near East, is a standout genetic marker in southern European populations. Could it be that Proto-Indo-Europeans who resided in Bactria or the Pontic steppe were R1a/R1b-dominant, while hg J got picked up along the way from a non-Indo-European-speaking Caucasian or Central Asian source?