[Updated 08.28.22]

Extraordinary claims require extraordinary evidence, but the Harvard lab provides none.

David Reich & co of Harvard University’s genetics lab are now claiming that the Proto-Indo-European homeland was in the Middle East / West Asia. Expect sensationalist news stories reporting this half-baked theory as fact very soon.

Their new study claims the ancient Anatolian languages were not Indo-European, but that the ancient Anatolian and Indo-European languages are “twin daughters of a Proto-Indo-Anatolian language,” which originated around the Caucasus among Caucasian Hunter-Gatherers. Note the conspicuous removal of “European” from the new moniker.

“the link connecting the Proto-Indo-European–speaking Yamnaya with the speakers of Anatolian languages was in the highlands of West Asia, the ancestral region shared by both.”

“the homeland of the Indo-Anatolian language family was in West Asia, with only secondary dispersals of non-Anatolian Indo-Europeans from the steppe.”

Their argument

Their argument is as follows:

  • Proto-Indo-European ancestry has not yet been discovered in ancient Anatolia [untrue, by the way].
  • Proto-Indo-Europeans and ancient Anatolians both had Caucasus Hunter-Gatherer ancestry, therefore…
  • Proto-Indo-European languages originated in the Caucasus among Caucasus Hunter-Gatherers and spread to Europe and Anatolia via separate migrations.

They also bizarrely claimed that the Yamnaya had some Levantine ancestry that nobody has ever found before, including them.

Plot holes

There are some major plot holes in their argument…


  • They seem to have completely ignored all linguistic evidence. For example, PIE language shows that the Proto-Indo-Europeans were not familiar with agriculture, which was common throughout the Harvard-proposed West Asian homeland of “Indo-Anatolian.”
  • Most Indo-European Anatolian languages were found in the West of Anatolia, the opposite side to the proposed entry point of “Indo-Anatolian.”
  • For some reason, “Indo-Anatolian” (Indo-European) languages did not survive in their own homeland, but the Caucasian languages (e.g. Kartvelian) did.
  • There is a complete lack of Indo-European languages in the Middle East until Bronze Age Yamnaya-related migrants arrive, which we know because ancient Middle Eastern civilizations were literate.
  • People with the most CHG ancestry today (Georgians) speak Caucasian languages, not Indo-European languages.
  • Both the Mycenaeans and Minoans had additional Caucasian-related and supposedly “Indo-Anatolian”-speaking ancestry but only the Mycenaeans, who also had PIE steppe ancestry, spoke an Indo-European language.
  • Caucasian influence on the steppe predates the (real) Proto-Anatolian language by 2000-3000 years. We know this because Caucasian ancestry in PIE comes from a source with 100% Caucasus Hunter-Gatherer ancestry, but by the Eneolithic (5000-6000 BC), all Caucasian populations had significant additional admixture from Neolithic Anatolians and Iranians.
  • PIE culture was extremely patriarchal and the only West Asian ancestry in confirmed PIE peoples (Sredny Stog, Yamnaya, Corded Ware, etc.) came from female Caucasus Hunter-Gatherers. Did these bride-kidnapped women really impose their language on Eastern European Hunter-Gatherer men?


  • PIE expansions into the European Balkans align perfectly with the proposed date of the (real) Proto-Anatolian language, 3000-4000 BC:
    — 4500-4100 BC, Suvorovo Culture in Ukraine & Romania
    — 4000-3200 BC, Cernavoda Culture in Romania & Bulgaria
    — 3300-2700 BC, Ezero Culture in Bulgaria with affinities to Troy I (Early Bronze Age) in Northwestern Anatolia
  • There is no evidence of Indo-European culture south of the Caucasus before PIE-type genetics arose in Europe during the Eneolithic (5000-6000 BC).


  • Harvard Lab only published a handful of samples from Western Anatolia during the Bronze Age, the most obvious entry point for Indo-European peoples. This is by no means a representative sample size.
  • Their method of “proving” the lack of PIE ancestry in Anatolia was a terrible four-way admixture model that used Eastern European Hunter-Gatherers as a source instead of Yamnaya.
  • They should have used an admixed Yamnaya + Early European Farmer sample from the Balkans, which would accurately represent the genetic makeup of Proto-Anatolians.
  • It takes 7 generations (150-200 years) of ethnic/racial intermixing for autosomal DNA to get “washed out” to less than 1%. So, if Indo-European languages were spread in Anatolia via elite conquest, rather than mass migration, PIE ancestry could have been diluted pretty quickly, while the language and culture remained.
  • That being said, clear PIE ancestry has been found in a Bronze Age sample from Kaman-Kalehöyük, which was not mentioned in the study. Moreover, low amounts of EHG ancestry was detected in Anatolia in this study.
  • Haplogroup I2a1b1a2 was found in Bronze Age Western Anatolia. This haplogroup was common among Bulgarian Yamnaya and the aforementioned Ezero culture, situated on the doorstep of Anatolia. Fascinatingly, the study also neglected to mention this fact.


So, what do we think, folks? Genuine theory or more stereotypical anti-White propaganda? Maybe they’re trying to compensate for accidentally fueling the rise of White identity when they discovered that the Proto-Indo-Europeans came from Eastern Europe and were most genetically similar to modern Northern Europeans? After all, David Reich has repeatedly said the purpose of his genetics work is to encourage immigration and race-mixing. See, for example, his 2018 interview in The Atlantic:

“if you pay any attention to this world, and have any degree of seriousness, then you can’t come out feeling affirmed in the racist view of the world. You have to be more open to immigration. You have to be more open to the mixing of different peoples. That’s your own history.”

If you have anything to add to this deb00nking, post it in the comments.

UPDATE: Genetic models using data from the study show obvious Yamnaya-related ancestry in Northwest Anatolia [08.22.28]

Someone processed the data from the study, so I made some quick admixture models using Vahaduo. Here are the source coordinates, if you want to test them out: https://pastebin.com/gpPSFEqK.

These models clearly demonstrate that the samples from Northwest Anatolia (where the Proto-Indo-European derived I2a haplogroup was found) all have obvious Proto-Indo-European or “Yamnaya-related” ancestry. This is the case when using both “deep” ancestry sources (as was done in the recent Harvard Lab paper) or more contemporaneous ancestry sources.

So, why on earth did the Harvard Lab not publish an admixture model using a Yamnaya-related sample? Because doing so would prove that these Northwest Anatolians had Yamnaya-related ancestry? This is bad science verging on academic fraud. How did this get past ~100 authors and a peer review team without being corrected?

Note: The Northwest Anatolian samples are dated to 2000 BC, 2800 BC, and 3000 BC. Recent Mycenaean Greek samples, dated 1000 to 1500 BC, have been included for comparison.


A deep ancestry model without Yamnaya shows very low levels of EHG in these Northwest Anatolian samples, around 1%, which can be written off as a statistical error:

But when Yamnaya_Samara is included in the same model, it shows ~3-5% ancestry in Anatolians:

Using more proximal and contemporaneous ancestry sources, e.g. Bulgarian Yamnaya, shows slightly more Yamnaya-related ancestry in Northwest Anatolians, ~5-9%:

For reference, deep ancestry compositions of the “contemporaneous ancestry sources”: