Demic diffusion theory

I just became aware of a new(ish) theory that I quite like, although I'm still on the fence about it.  It hinges on a missed detail of the "Corded Ware Culture is 75% steppe DNA"; that it's not the same steppe DNA that later spread across Europe in a later wave; it's an R1a wave, which was followed by an R1b wave.  The Yamnaya cultural horizon seems to actually have been primarily R1b, as was the Eastern Bell Beaker culture, which populated at least most of northern Europe.  (I'm simplifying the various subtleties of different R1a and R1b lineages using them as shorthand, since they don't really change the trajectory of the theory.  In reality, there are multiple R1a and R1b lineages, but I'm talking specifically about the ones that would be markers for the eastern and western steppe respectively. To read about them in more detail, try here and here and here.) The theory is great because it attempts to be much more multidisciplinarian than most; i.e, it accounts for linguistics, archaeology, anthropology and archaeogenetics.  And it utilizes much more up-to-date archaeogenetic data than other models.  And it has some of the detail of Anthony's model, except updated.

It's really quite nice.  I'm coming around to seeing its merits.  Here's how the theory works, summarized probably a little bit too simplistically:

• Steppe is steppe is steppe, right?  Well, no, although in the past, many experts have made very little of the differences on the steppe itself and assumed that all of it lead to Proto-Indo-European in some form or another.  There are actually a number of differences on the steppe in terms of physical type, economy practiced, material culture, and even genetic signatures.  Many cultural movements and displacements/replacements/admixtures are posited to have happened on the steppes long before anyone really started migrating in large numbers beyond the steppes.  For the most part, this can be sorted into an Eastern and Western cline, with most of the traits most strongly associated with Indo-Europeanism (kurgan burials, pastoral economy, horse domestication, etc.) arising first in the Eastern steppe zone; the Volga and north Caspian region, and then spreading westward to the north Pontic region, but often in "incomplete" terms, forming a cultural cline, if you will, of cultural traits that runs east to west.
• The Don River forms a kind of loose frontier between the two zones.  The cultures of the west; Dnieper-Donets II, Sredni Stog and later Dereivka, certainly interacted with the cultures to the east, but they also interacted probably even moreso with cultures further to the west of them; like the Trypillian, the Cucuteni, etc. which were urbanized, Balkan neolithic farmers, associated with the EEF population.  The West started as riparian hunters, gatherers and fishers, and gradually adopted more settled agricultural traditions from the west, and were relatively late adopters of animal husbandry, especially horse domestication and pastoral nomadism, to the extent that they did at all, it was all influence (and even migration) from the east.  The horse-head scepters and social stratification and kurgan burials are also not associated with the earlier western cultures, and only gradually adopt it as a clear innovation/adoption from the east.  The west is also the center of the R1a-Z645 haplogroup.
• In contrast, the Samara and later Khvalynsk cultures of the east, which eventually evolved into the Repin culture, which was the immediate flashpoint of the creation of the entire Yamnaya horizon are much less settled.  They practiced a more mobile economy.  The domestication of the horse (and later the riding of the horse) happened here.  Heirarchical social stratification, evidenced in rich burials happen here first (to be fair, probably adopted/adapted from the Caucasian Maikop culture, which had clear contacts with both eastern and western steppe cultures both, in terms of trade, genetic flow and more, in the past.  Plus the Caucasus was the metallurgical hotbed for the entire region, unsurprisingly.)  When changes in economy, social structure, and material culture which make the entire region more closely resemble "typical" Proto-Indo-European, they always come from east to west.  The eastern region is also the Urheimat of the R1b-L23 haplogroup and later R1b derivations.
• But what if instead of assuming that all of this was some very early form of Proto-Indo-European that we propose that there was a linguistic frontier as well as a cultural, economic and genetic frontier?  And, in fact, why wouldn't there be?
• One may notice, as soon as one makes this piece of the puzzle click into place, that the so-called "Yamnaya" DNA in the Corded Ware actually isn't; it's not associated with the actual Yamnaya eastern horizon (which, to be fair, expanded westward to cover the territory earlier held by cultures such as Sredni Stog, Mikhailovka, Dereivka, etc.); it's specifically R1a.  R1a is often sloppily applied to the entire steppe, because later admixture between late Yamnaya and post-Yamnaya cultures like Sintashta, Andronova, etc. feature high levels of R1a, but that is not because R1a is a genetic marker associated with the Yamnaya cultural horizon, but because of clear contacts and admixture between those and adjacent eastern Corded Ware cultures.  When "steppe" invaders bring R1a to India and Pakistan, for instance, during the Chalcolithic and Early Bronze Age, they are not bringing Yamnaya R1a, they are bringing Corded Ware R1a.  The genetic history here is not hard to see in the literature, but most specialists have not made a very fine distinction between R1a and R1b and have handwaved the difference between them away, essentially.
• So, this model of the Eastern and Western steppe peoples being essentially two different peoples, both in terms of genetics, economy, culture and language (albeit certainly two populations with lots of contacts, interactions and admixture between them, and quite possibly a deeper shared ancestry) fits the archaeogenetic data quite well.  It fits the archaeological and anthropological data quite well too.  Does it fit the linguistic data, or is there a linguistic model that makes sense for it?  In fact, there is.
• The Uralic languages are the other "major" European language family (other than Indo-European) that appears, especially in the northernmost part of the continent (with Hungarian being a more recent and historically attested arrival.)  For the most part, they seem clearly to have been in place long before the arrival of written history.  In addition to that, there are:
• clear linguistic clues that posit that proto-Uralic and proto-Indo-European had to have been in contact while still proto-languages.
• clear heavy linguistic interaction between later eastern Indo-Europeans of the Indo-Iranian variety that are best represented by post-Yamnaya eastern steppe cultures, such as Poltavka, Sintashta, etc.
• evidence of Uralic as a substrata in much of northern Russia and other areas that later emerged as Balto-Slavic, or other northern Indo-European linguistically, much of which corresponds geographically to the Corded Ware horizon.
• So, in the Demic diffusion model, we see the following developments:
• At some point prior to 5,000 BC, there is a common Indo-European/Uralic community.  It's possible that the two are both daughter languages of a common mother tongue, but not required that it be so, merely that they grew up in close proximity, sharing areal and cultural features.  By 5,000 BC, Indo-European and Uralic have started to go their separate ways.
• Between 5,000 and 4,000 BC, roughly, Indo-European is at the "Indo-Hittite"stage.  This is associated specifically with only a portion of the steppe; the Samara and early Khvalynsk cultures.  Starting at roughly 4,400 BC, the Suvorovo-Novodanilovka "horse chief" culture migrates from the east into the western region, probably representing the splitting off of the Anatolian group.  They continued moving westwards (Cernavoda I) and took up location in the Balkans before moving into Anatolia where they are first historically attested.
• Between 4,000 and 3,500 or so BC, we have Proto-Indo-European forming into it's "classic" form, and by the end of this era, we see a number of changes.  Late Repin/early Yamnaya (of predominately R1b lineage) go on the move.  They end up as far away as the Altai (Afanasevo culture; R1b) and cross the Don as intrusive colonizers, forming the South Bug and Lower Don groups, which make up the western half of the Yamnaya horizon.  This Yamnaya horizon displaces the R1a peoples of late Sredni Stog and other cultures already living there, who migrate northwards, becoming a principle component of the Corded Ware Horizon by about 3,300 BC.
• The next half millennium shows Indo-European essentially splitting into two major stocks; a northern stock more closely related to Tocharian, and a southern stock.  The southern stock is, by this time, in the Balkans, superimposed over what remaining non-Indo-Europeans were still there, as well as any remaining Anatolian groups.  It also remains on the steppes, while the northern group moves into northern Europe, largely at the expense of the Corded Ware horizon; eventually forming the Eastern Bell Beaker culture.
• This explains two things; 1) a number of correspondences between languages that originated in the Balkans (Phrygian, Thracian, Armenian, Greek, etc.) and the Indo-Iranian group, which maintained a southern linguistic spectrum.  2) The northern group inherited a Uralic substrate which had significant impacts on the development of very early pre-Celtic, Italic, Baltic, Slavic and Germanic languages. 3) A number of other correspondences between Balto-Slavic and Indo-Iranian, since the Indo-Iranian peoples picked up a significant admixture from eastward Corded ware peoples.  The Bell Beaker R1b peoples dominate genetically over area previously occupied by R1a Corded Ware.
• These groups continue to diffuse and differentiate into proto-language families, leaving (probably) only Balto-Slavic, Indo-Iranian, and maybe Italo-Celtic as intermediate steps in between proto-language families as we know them.  Of course, there could be anonymous language groups that spun off and were later absorbed.  Historically, we can see the loss of the Thracian and Dacian languages, for instance, to Slavic or Romance languages, the loss of Phrygian and Anatolian to Greek and Roman (and later Turkic), and shrinkage of Tocharian to Iranian, to be later replaced by Turkic, and the loss of much of the former Iranian language family territory to Turkic languages and the (almost) total loss of the Celtic languages to Romance or Germanic or other languages.  The same probably happened prehistorically as well, although we simply don't know how to tell much about it except through often somewhat cryptic linguistic clues.

Now, of course it isn't strictly speaking necessary that Sredni Stog > Corded Ware > historically attested Finnic/Uralic groups of northern Europe.  Even if it does fit the data quite well.  It could be an earlier spread of Indo-European that later was subsumed by later groups, for instance.  But it's a nice theory, a nice model, and it makes a bold attempt to tie up all lines of evidence that we currently possess into one model that works for all lines of inquiry.

https://indo-european.eu/wp-content/uploads/2018/05/indo-european-uralic-haplogroup-1.jpg

EDIT: Now, granted, this view is controversial.  See one rebuttal here, for instance.  Although, I will point out that this doesn't really rebut it so much as it is an appeal to authority.  But, it does seem to imply, at least, that there is genetic data that contradicts Carlos' Demic Diffusion theory, even though it doesn't really explain it exactly.  I think maybe we're supposed to believe that everybody who reads that blog is sufficiently well versed in the technical aspects of archaeogenetic research that its already obvious without needing to be spelled out, or something.

Although the guy there makes a serious complaint about demic diffusion model, another of his own posts seems to support it, suggesting that the steppe was the bifurcation point between R1a and R1b.  Carlos also adds a linguistic assumption to that same bifurcation.