The Y dna distribution by haplotypes in modern Armenians : dotting the I2-s.

The first chart is based on modern commercial data from FTDNA. It shows the most important branches also known as haplotypes. As You can note, with the exception of R1b-Z2103 the rest of Y dna is quite fragmented. The number of haplotypes that score more than 3% of population is ten. In this list of top 10 the TVC lineage I2-BY423 (also known as Y16419) is in the 9th position. I think based on this chart it's obvious that any Armenian haplotype that score more than 3% can't be labeled as "very low" or "inexistent".

But this is not the whole story. In this top 10 at last 4 haplotypes have a Neolithic - Chalcolithic age ( J2- L25, G2-M406, T-L208, E-M84 ). And current paleogenetic evidence supports their relatively old expansion. Although in some cases they have also younger lineages. Like J2-L25 which has a subbranch L70. It makes the third of L25 and it expanded in Iron Age.

So if we concentrate only on Bronze Age expansions that have more relevance to modern ethnicities then the list will be arranged in a different way.

The main lineage that expanded in Early Bronze Age (Kur-Araxian culture) are the J1-Z1842 and most probably the J2-M92. It's possible that E-M84 also is partly related to KA but there is no direct proof of that theory.

The lineages that expanded in Middle Bronze Age and after are the R1b-Z2103, I2-BY423, J2-FGC15865 ( a lineage related to Van-Urmia culture ) and R1a-Z645 appears predominantly in Iron Age.

There is also a Bronze Agr lineage J1-L862. It's from Levant and Mesopotamia.

They are also lineages that expanded in Iron Age and later but their percentage points do not exceed 1.5-2%. And they are too late to be crucial for the Ethnogenesis. Even though they are important for understanding some aspects of Armenian history.

Returning to I2 in Armenia. The main question is that did it had a dramatic decrease over time given the available paleogenetic evidence?

Well to have an answer to that question one needs to compare apples to apples not to oranges. I mean that modern Armenian Y dna is gathered from a large territory stretching from Cilicia/Sebastia to Caspian Sea/Iran. If we want to understand how the level of I2 changed over time we need to compare data from equal geographic regions.

We don't have detailed Iron Age aDNA from the aforementioned region. But I think it's not hard to imagine that it's very very unlikely that I2 level exceeded 5% in Iron Age historic Armenia and it's large neighbourhoods if a systematic ancient DNA becomes available.

Even more. Inside ancient Republic of Armenia the I2 was present in high levels only in Sevan basin sites. See the second chart (Noratus, Lchashen, Sarukhan ) Karashamb in Kotayk is close to Sevan. If we look at Urartian era samples that do not have any aDNA from Sevan basin then we see a completely different picture. See the third chart. Offcourse 20 Y dna is a small sample size. But we can reproduce a similar result for LBA-EIA period if we remove all Sevan basin sites.

In any case it's obvious that in modern RoA territory there was an important change of Y dna distribution in Iron Age. I will dedicate a special thread to this subject. Obviously both I2 and R1b decreased but given current data we can say that while they decreased in RoA they must have increased in some other regions of historic Armenia. Otherwise how we would have 3% of I2 from large territory stretching from Anatolia to Iran?

In conclusion the current Armenian Y dna is dominated by haplotypes that expanded Bronze Age. Initially they had localized peaks in certain regions. But overtime they diffused out of their peak regions and modern distribution became more homogenous. I2 is not a special case in this story. The only thing that it has more limited geographic peak than R1b which in most likelihood expanded over larger regions.

PS We also should remember that native population of Sevan basin do not live anymore there. One can expect that they had a similar or higher level of I2 than Gardman Armenians

The word for horse in Hurrian

 The word for horse in Hurrian ešši / iššiya is strikingly close to Armenian word eš (genitive išoy ) meaning donkey. This resemblance becomes even more striking if we take into account the fact that Armenian word for donkey was initially meaning horse and derives from the PIE root *éḱwos. The shifts are regular and well attested for the Armenian. E.g. *ḱwón > šun meaning dog.

Many scholars noticed that the Hurrian word seems to be a loanword from a satem IE language. But few of them discussed the possibility of a loanword from the Armenian. In his paper about the word ešši 'horse' in Hurrian Vyacheslav Ivanov mentions that the closest form is the Armenian. He also criticizes Starostin's proposal that the word for horse in PIE is a loanword from North Caucasian. The phonetic shifts required from NC to PIE are far more unrealistic than the inverse. It's quite easy to demonstrate that most of the words for horse in North Caucasian languages are borrowings from Iranian languages. Starostin was also claiming that Armenians borrowed from the Hurrian and not the inverse.

Armen Petrosyan discuss this subject in his book about the Armenian Epic. He is inclined to believe that it's Hurrian who borrowed this word from the IE languages. More specifically the Armenian. He also discusses the possible etymology of Išuwa ( Ծոփք ) meaning land of horses. A land where a cult of equines persisted until medieval era.

Today we have DNA studies of ancient horse remains. And we can have better idea what theory is more likely. In the current stage of our knowledge, we know that horses were present in many parts of Eurasia. Including Anatolia and Armenian Highlands. But the currently existing domesticated horses descend from the one that was domesticated in Pontic-Caspian Steppe.

Guimaraes 2020 compares mtdna from ancient horse remains in Anatolia and South Caucasus before and after 2000BC. Their conclusion is that after 2000BC the mtdna diversity in Anatolia increases while the local mtdna lineages like P, O and X sharply drops or disappear. They conclude that numerous new horse lineages were introduced to West Asia via Caucasus at the end of third millennium BC. This fits well with the known human migrations from Catacomb culture to South Caucasus in the same period. This data strongly supports the idea that the word for horse in Hurrian is a loanword from Armenian. It must be noted that Catacomb migrants were not the only group that introduced new horse lineages. Around 1700BC a group known as Mitanni Aryans appear in West Asia and them also might have introduced both new genes and words. The word aššuwa ( aśva in Sanskrit) attested in cuneiform texts is one of them.

PS The mtdna Q in pre-2000BC South Caucasus was found in Kakhetia East Georgia dated around the 2500BC (not radiocarbon). It's remarkable because unlike P, O and X it's number increases after 2000BC. This can have two explanations. One version is that this sample Q was related to Early Transcaucasian Kurgan groups also known as Martqopi-Beden. It's age, it's Y dna (Y-HT-1) and the horse burial practice supports this version. But it is also possible that it was native to South Caucasus and became a successful because of hybridisation from a local lineage and those introduced from Steppe. Unfortunately, numerous Paleolithic equine horse remains from Armenia didn't have the necessary quality to learn their mtdna. Let's hope we will learn more about haplogroup Q in the future.

Ancient DNA charts. Before the 2500 BC.

 Ancient DNA charts. Before the 2500 BC.

The first one is from RoA (11 samples) and neighborhood. The other one is from more southern regions. Sirnak and Hajji Firuz.

Combined with Malatya region data (in total 40 samples) we can make certain conclusions.

J2 was the most frequent haplogroup in ancient Armenian Highlands and it's neighborhood. Not only it was frequent but also very diverse. M67, M319, Z6065, J2b2b, PF5197, J2a2 branches were found in those sites. J2b1 and J2b2a also are attested. This leaves little doubts about the origin of J2 haplogroup.

E1b has old presence.

G was more frequent in the south and west than in north-east of Highlands.

J1 expanded mostly in EBA.

R1b was present but only the V1636 branch.

While R1b-M269, I2a2b, R1a-M417 and many branches of Q haplogroup came later during MBA or after

All data from Antonio 2022 preprint paper examining 17 samples ( the 18th sample was not included ) from Beniamin, in one table.

 All data from Antonio 2022 preprint paper examining 17 samples ( the 18th sample was not included ) from Beniamin, in one table.

All mtdna are from theytree site. While Y dna is double checked. Both citizen's and theytree assignments are taken into account.

Genome wide the situation in Beniamin reflects the same trends as in all Republic of Armenia (Etiuni). In LBA-EIA period with the Urartian period included there is a high level of Steppe ancestry. While in post-Urartian period we see a dilution of Steppe ancestry and an apparition of autosomes similar to modern Armenians.

The Y dna distribution has the similar trend. Strong presence of R1b, I2 with some E-M84 in LBA-EIA-Urartian period. And more diverse Y dna in a post Urartian period