August 2011 E-mails re Origins of R1a1a Ashkenazi Levites

From: A.J. Levin

To: Mike Rothenberg

Date: August 2, 2011

[bold-faced text from Mike Rothenberg’s prior e-mail]

Dear Mr. Rothenberg,

You’re most welcome. I will attempt to answer all your questions and points, which I have left as bolded.

I assume L342.2 is one of them. Is this correct? I wasn’t aware of more than this new SNP that could be in the Ashkenazi Levite branch.

Yes, L342 is below M417, but it is several steps below. The SNP Z93, which will soon be available for order in the Advanced menu, is the earliest split in the path that leads to L342 from the path that leads to Western European M417 (for example, the Norwegian and Scottish subclades). This is what we should be testing various groups for next, other than those we already know are negative and positive for it. Between Z93 and L342 is another SNP, Z96.

I assume that Turkey, having thousands of years of various populations living there, over the years (Hittite, Greek, etc.), now has a population with a rich mosaic of diverse and intermixed ancestral backgrounds.

Yes, though one could probably say similar things about other places. Specifically on Turkey, autosomal and other studies suggest that the actual Turkic (i.e. Steppe) genetic contribution to Turkey is fairly minimal, maybe around 5-7%. Anatolia does, however, seem to be very near the point where many major yDNA haplogroups were formed and/or spread, both east and west.

Similarly, what with movement over hundreds, if not thousands, of years, I would think that people who are may consider themselves Iranian, Pakistani, etc., are referring really to nationality. They may, individually, have an ancestral patrilineal line that originated somewhere outside of their country’s boundaries and with from an ethnic group they might not be aware of.

Yes, that’s always possible. My impression of these areas is that they are highly tribal, and patrilineally so, meaning they will usually have a fair sense of their tribal identity. Once you go east of Central Europe, of course, records get poor very quickly. L342 for the moment has good correlations with the spread of nomadic Iranian speakers/early horsemen -- we are talking about something like 3,300 years ago for the subclade.

As one goes eastward from Turkey into and across Central Asia, would there be less diversity of various ancestral ethnic groups?

Seemingly so. When you get into the mountains, you generally get more endogamy. You can see this in lack of ySTR variety, and also autosomally in such groups as the Kalash.

But I wonder whether all, or most, of the Turkic groups ultimately derive from one or just a few ancestral groups, more in the eastern area, where they may have shared a more basic form of Turkic language, customs, and religion. It seems that the Tengri religion was shared over a large area by Turkic, perhaps some Mongol and other related, groups. This is why it would be valuable, in the effort to gain more R1a1a1 subbranch detailed information, to get YDNA information from more ethnic R1a1a1 Turkic people in, or from, those Central Asian areas.

Perhaps tests for the new SNPs should be run on existing, or new, samples from men with that background. I don’t know if there is a repository of DNA samples somewhere in Russia, and/or in a number of Central Asian countries, that might already have useful data. Perhaps there’s a way to gain additional participants to submit samples to FTDNA for your program, through contacts with research and other organizations in those areas.

That would be good, though I would not necessarily view followers of Tengrism as a genetic bloc. To my mind geography is usually a much bigger constant than faith; Ashkenazi Jews are really one of the few exceptions. Expulsion-era Israelites were themselves probably an amalgam of descendants of local peoples who eventually adhered to what was in effect the state religion of Jerusalem and area, with different haplogroup blends: Canaanites, Amorites, etc.

My hunch is the Khazars may themselves have been an Indo-Iranian people who were linguistically and culturally Turkicized rather than an always-Altaic people (or perhaps they were a confederacy of Iranian and Turkic people, or some blend).

We are attempting to SNP-test some Pathans from Afghanistan for L342. They seem to have extremely high rates of R1a1a1, possibly all, or largely, L342. Pathans speak an eastern Iranian language probably similar to Scythian, the main Iranian Steppe language of antiquity. Of course the more people we test from any area the better, but there are only so many resources, and several of us have already sponsored a number of kits personally. We are attempting to make what inroads we can, and that’s how we have tested samples from Siberia, Bashkir Republic, Lithuanian Tatars, Anatolian Turks, etc.

One outcome of this additional testing effort would be a better understanding of where R1a1a1 Ashkenazi-Levites fit into the overall R1a1a1 tree, in conjunction with additional data obtained from these other sources that could be used to develop more detailed branching information for the R1a1a1 tree. Perhaps here is an area where some joint planning would be helpful to both our Projects.

Yes, that would certainly be handy. For now, we are also awaiting the availability of new SNPs at a higher level such as Z93.

Maybe there is at least one source of samples for Eastern European men - perhaps in Eastern Europe itself - where R1a1 samples that are negative for M458 can be tested to see if they are positive for L342? Perhaps it is possible for such samples to exist, meaning, presumably, that men with R1a1 in Eastern Europe derived genealogically from two ‘paths’ of R1a1 ancestry.

Thanks to concerted efforts we have already confirmed that almost every branch of non-M458 Eastern European M417 is –L342. As the new SNPs identified from the 1000 Genomes Project are available, we will be able to test SNPs further up the tree rather than L342, which I expect will be scarce west of, perhaps, central Turkey and central Ukraine (except among men such as Tatars, Ashkenazim, and those with a vague history whose STRs are similar to the above groups). Of course, testing in Ukraine and Turkey themselves would be quite interesting and valuable: two historically very large and very undertested countries with lots of potential secrets in them.

Fluxus is new to me, so thank you for the referral.

Again, you’re most welcome. A cladogram updated and emailed once or twice a year (or as new project members are added) could be a great way to visualize the ‘Tree of Levites.’ Of course this depends on people testing preferably 67 or 111 markers, because at less resolution it is much harder to make conclusions.

From what I understand, as you noted, ‘civil’ surnames of Ashkenazi Levites are roughly 200 years old, ordered to be taken within both the Russian and Austro-Hungarian Empires. So they do not mean much genealogically. Yes, there are Cohens, Kahans, Kagans, etc. and Levys, Levitans, Levines etc. but also Brodys, Schneiders, Weisses, and whatever. Who could be Cohanim, Levite or ‘ordinary’ Jews. So you can’t really go by names.

Yes. I have found evidence (through a combination of autosomal and yDNA testing) of a genetic link from a common Levite ancestor before surnames to Segals. It’s not the same name, but in a sense it is. My Levin grandfather overlaps with a Segal in the same area he overlaps with his paternal cousin -- and this cousin also overlaps with the Segal in the same spot, and there are no other surname matches -- and the fact my closest 67-marker match, 65/67, is a Segal. This seems highly suggestive of a common Levite male-line ancestor about 1725-1750, whose descendants are variously Levins and Segals.

Regards,

A.J.

* * *

From: A.J. Levin

To: Mike Rothenberg

Date: August 2, 2011

Hello A.J.,

* * *

I’d appreciate your thoughts and any corrections regarding my understanding of certain aspects of R1a1 genealogy (below) that are critical to the question of determining what ethnic group is the source of Ashkenazi-Levite R1a1.

------------------------------------------

If I understand correctly from what you have written, Ashkenazi-Levite R1a1 does not derive from ‘European’ R1a1 - i.e., not from either the M458 + or M458 - subbranches from which East European R1a1 men derive. They derive, instead, from what might be called an ‘Asian’ R1a1 subbranch.

As such, Ashkenazi-Levites who have R1a1 haplogroup clearly are not derived from Israelite ancestry.

You raise the point of the Khazars deriving from an Iranian people. If so, would they have embraced the Tengri religion if Iranian, since Iranians had converted to Islam before the Khazar conversion to Judaism, I think? My understanding is that the Khazars converted to Judaism from the Tengri religion. I must admit I’m not sure of the timeline of Islam’s spread to the Central Asian people.

How would the spread of Islam across Central Asia (which, to me, includes Kazakhstan, Uzbekistan, Kyrgyzstan, Turkmenistan, Tajikistan as well as the vast Uighur territory in Western China) not have converted the Khazars if they lived that area at the time of conversion? And how is it that the Khazars were not converted to Christianity?

I don’t know the history of the Khazars before they are noted as living in the land above the Caspian and Black Seas, and between those seas. And I’ve had trouble finding anything about their early history and migration information. However, it seems to me that they might have moved there from an area where they had no exposure to either Christianity, Islam or Judaism, and so were able to take their native Tengri religion west with them. Later, after arriving in their homeland near those Seas, they ultimately had contact with Jews and apparently chose Judaism as their religion. (I’m sure you know the story, as described in Yehudah Halevi’s book The Kuzari.) Anyway, this is why I think they might have originated in a region that was more likely to the east, as in the Altai region or eastern Turkic areas, where they may have been able to hold onto their early religion.

This is why it seems so important, to me, to gain DNA information from current Turkic groups that might have evolved from stock that also were ancestors of the Khazars. I appreciate the effort to gain more information from Iranian-related ethnic R1a1 men. The total of such information from both groups, from Eastern European groups, and perhaps from others, can help build a complete R1a1a1 ‘tree’ that will, hopefully, indicate where Ashkenazi-Levite R1a1 comes from, or who it is related to among current ethnic groups.

* * *

Regards,

Mike Rothenberg

* * *

From: A.J. Levin

To: Mike Rothenberg

Date: August 3, 2011

Hello Mike,

* * *

Yes, that seems the case. There may be scattered examples of Asian L342 in Europe, especially in its poorly tested eastern extremities, but the entire L342 group at present has an estimated origin of a little over 3,000 years ago, and probably somewhere on the Steppe. This fits fairly very well with the expansion of Scythians, though more data would be useful before making any broad conclusions.

As such, Ashkenazi-Levites who have R1a1 haplogroup clearly are not derived from Israelite ancestry.

This seems likely though not 100% certain. I suppose other explanations are possible, for instance, a group of Indo-European speakers in the northern Near East may have drifted south over time (such as Hittites).

We have to be careful to distinguish between settled Iranians from Iran (Persians), and the nomadic Iranian speakers to the north -- some of whose descendants are Pathans, for example, while many others became Turkicized.

Tengriism was probably brought to the Steppe Iranians with the Turkish language: I think most scholars suggest this happened about AD 500. Why exactly there was such a cultural shift I am not sure, but even later, the entire country of Turkey also shifted from Indo-European languages to a Turkic one, and, mainly, to Islam. These things happen.

If you look at the Wikipedia page on Tengriism, there seems to be some hint that the formerly Indo-European people of the Steppe who became Turkicized had a hybrid Iranian/Turkic religion:

“In Europe, Tengrism was the religion of the Huns and of the early Bulgars who brought it to the region. It is said that the Huns of the Northern Caucasus believed in two gods. One is called Tangri han, that is Tengri Khan, who is thought to be identical to the Persian Aspandiat...”

My guess is the Khazars were Iranian polytheists not long before they became Tengriists, when they might have had a small infusion of Turkic genetics to go along with the language and culture. Not long after that, some of them became Jewish, though others remained pagan, Christian, or Muslim. The Khazar pre-Tengri religion was probably similar to those of settled Persians, the Greeks of old, or Hindus, or perhaps something like the Kalash religion today.

http://merachitral.blogspot.com/2010/05/kalash-religion.html

A good way to view changing of religions historically is in response to pragmatic political concerns. This may seem reductionist, but Sikhism, for example, arose as a kind of compromise religion between Hinduism and Islam, in a physical place between the two.

Khazars in effect prevented their area from being Islamicized because of war with Arabs. Within Central Asia, Islamicization started east of Khazaria, around Kazakhstan, in an area beyond their control. Khazaria had fought wars with Arabs to the south for a very long time, until there was a stalemate, with the area south of the Caucasus being under Muslim Arab control, and the area north Khazarian. They apparently tolerated a variety of religions.

Both parties probably were tired of war and I suspect Khazars wanted to appease their neighbours with a gesture of friendship: no small part of the war rhetoric was Monotheist vs. Pagan. However, they were not about to surrender a culture they had fought long and hard for to their foes.

At the same time, Khazaria was friendly with Byzantium and Armenia, which were Christian.

By choosing as a state religion Judaism while allowing other religious practices, Khazaria could in effect act as a buffer zone without offending anyone. They would not anger their existing Christian rivals, and would remove much of the religious dimension to war with the Arabs. Jews may have been dhimmi, but were still People of the Book, not ‘demon worshippers.’ But ordinary Khazars could continue to practise whatever they wanted.

I’m not a historian, of course, but this seems to fit fairly well with historians’ reconstructions of things, e.g.: http://www.khazaria.com/brook.html

* * *

A.J.

* * *

From: Michał Milewski

To: A.J. Levin

Date: August 3, 2011

Hello everyone,

Let me add one comment to your very interesting discussion.

As such, Ashkenazi-Levites who have R1a1 haplogroup clearly are not derived from Israelite ancestry.

I totally agree that we need to be very cautious when trying to draw any final conclusion in this regard. For example, there is also a possibility that R1a-L342 had been brought to Israel (or to Mesopotamia if the Babylonian Exile is considered) by Persians in the 6th-4th century BC. Are there any historical evidences of special relations between the Levites and Persians?

Regards,

Michal

* * *

From: A.J. Levin

To: Michał Milewski

Date: August 3, 2011

Michał:

That’s a good point too. We know, at minimum, that in this period of roughly 600 BC - 400 BC the Persians were key in rebuilding the Jerusalem temple. We also know that Levites had temple roles. I’m not sure how much we know beyond that.

I guess a key question is, why do we not find L342 in Sephardi or Mizrahi Levites? Possibly genetic drift, but it could also point to Khazar contribution.

In any case, the fact that L342 occurs at such great frequency among Ashkenazi Levites and not otherwise among Jews does lend something to Mike’s hypothesis that it is atypical or a late entrant, so possibly high-status individuals were accepted into the Jewish culture in a key ceremonial role after the classical biblical period.

The second key question is, who within L342 has close ySTRs to Ashkenazim? I don’t think we have a clear answer yet.

* * *

From: Mike Rothenberg

To: A.J. Levin, Michał Milewski

Date: August 3, 2011

Hello Michal,

You raise an interesting point I had never thought about before. If, as I understand, the R1a1 in Levites today is limited to Ashkenazi Levites, I’m not sure how any of the pre-Roman exile experiences of the Israelites could have produced an infusion of R1a1 that would have been limited only to Levites who much later would be the ancestors of today’s Ashkenazi Levites. I’m not a student of history, but do agree that it would be worthwhile to learn if there is any historical evidence of any special relationship between Persians and Levites.

I may be wrong, but the fact that R1a1 occurs in Levites at all is, to me, unexpected, and, as such, is a major clue. Converts to Judaism are accepted as ‘ordinary’ Jewish people (i.e., joining the great majority of Jews who have neither Cohen nor Levite male-line ancestry). It is hard to understand how, or why, regarding Ashkenazi-Levite R1a1 men, such male converts, centuries ago, were not accepted at that ‘ordinary’ level. How could they have wound up in the Levite group instead? This maybe was done as a special ‘exception’: Clearly, there would have been a prohibition of even accepting ‘high-level’ (royal?) converts as Cohens. However, it was felt that there could be an acceptable compromise in accepting such men as Levites. As such, these male converts were classified at a level that still provided a level of religious exclusivity, as in synagogue services even today. The Khazar conversion could possibly have provided such higher-class (royal?) men. Perhaps you can think of another reason for appearance of R1a1 men as Levites - and only in the Ashkenazi community. I can’t think of any other possibility. I’m unaware of any historical evidence or documentation on the nature of the Khazar conversion, operation of their synagogues, who served as Cohens and Levites, etc. These may remain unanswered questions, unless evidence is found that addresses these issues..

Again, your suggestion of seeing if there is historic evidence of a special Persian-Levite relationship is interesting and would be worthwhile to research.

Regards,

Mike Rothenberg

* * *

From: Michał Milewski

To: Mike Rothenberg

Date: August 4, 2011

Mike, A.J. and others,

Of course I didn’t want to present the putative Persian connection as being a more likely option than the Khazar hypothesis (which I consider quite convincing at the moment). All I wanted was to show that there are other possibilities worth considering.

I agree with Mike and A.J. that the fact that R1a-L342 is practically absent among non-Ashkenazi Levites makes the Khazar hypothesis more likely. On the other side, I can imagine that the alternative ‘Persian admixture’ (mentioned in my previous message) could have initially concerned just one man who left not many (surviving) male descendants during the next 300-800 years until his line expanded following their settlement in Europe. In Roman (and post-Roman) times, the most likely route to Germany was through Italy (Spain and France being the less likely alternatives). Thus, one could imagine that the R1a-L342 Levites should have left some traces there (if the whole concept is true). I haven’t seen any STR haplotypes of the R1a Jews (Levites) from Italy (likely including the famous Tuscan L342 man from the 1000 Genomes Project), but maybe this is something we should take a closer look at. If at least some of these haplotypes differ a little more than expected from the modal for the Ashkenazi-Jews clade (type A in Mayka/Gwozdz classification), it will support the thesis that L342 Ashkenazi Levites have originated from Israel rather than from the Khazar state. Otherwise the Khazar link will remain our best option.

AFAIK the closest type A-related (L342+) STR haplotype belongs to a Palestinian named Kussad/Qussad, claiming a Mamluk ancestry (and possible links to the North-Ponthic region). As this is clearly an indicative of possible Khazar/Turkic/ Sarmatian origins, his Palestinian location does not allow to easily exclude an alternative connection to ancient Near Eastern populations.

Regards,

Michal

* * *

From: A.J. Levin

To: Mike Rothenberg, Michał Milewski

Date: August 4, 2011

Michał:

Thanks for your input. I think you have done a good job of treading between the most likely explanation we have given current evidence and the need to be careful not to rush to conclusions with insufficient data.

The Ashkenazi-Levite Project does have one member who claims descent from a Spanish Jew named Levy, but I haven’t seen a GEDCOM or any corroborating evidence. Of course, it’s possible. I believe there is also evidence of Khazars going to study in the prominent rabbinical schools of Spain around a thousand years ago so a single Spanish line fitting the motif might not prove anything either way.

As to the question of whether there is L342 in Italian Levites -- an excellent point! There is a brand-new Italki Project here

http://www.familytreedna.com/public/italkim/default.aspx?section=yresults

but no R1a1a1 yet. We will have to keep an eye out.

AFAIK the closest type A-related (L342+) STR haplotype belongs to a Palestinian named Kussad/Qussad, claiming a Mamluk ancestry (and possible links to the North-Ponthic region). As this is clearly an indicative of possible Khazar/Turkic/Sarmatian origins, his Palestinian location does not allow to easily exclude an alternative connection to ancient Near Eastern populations.

Absolutely agreed. We cannot rule out two groups starting out with a common origin in the Near East, though it seems balance of probability for the moment points to the Steppe.

Another key question besides Italy would be, do we find L342 similar to Ashkenazim among non-Ashkenazi non-Tatar people in former areas Khazars and Pechenegs were, specifically, Russia just north of the Caucasus, and Ukraine, especially eastern and southern? There are very few samples from these areas.

A.J.

* * *

From: Łukasz Lubicz-Łapiński

To: Michał Milewski

Date: August 5, 2011

Hello!

Thanks. Interesting analysis!

I can only add to your points - except genetic relations and close results with Indo-Iranian world interesting is analysis of geographical spread of all Ashkenazi cluster.

Look at Igor Rozhanskii’s map (I’ve added historical maximum wide borders of both kingdoms) - there are no results from Southern and Western Europe! These few results from Germany in my opinion can’t be connected with ‘nest’ of that clade in Germany.

The main core of this branch lies just in the center of the old Polish Kingdom terrains, which are just in the neighborhood of past Charaz Kingdom. (another point – the densest results are rather in EAST part of polish kingdom that West!)

It could show us direction of migration from east - rather north from Black Sea than Byzantium-Greece way.

Regards,

Lukasz Lubicz Lapinski

P.S.

Genetic distance analysis shows that Qussad is nearest non-Ashkenazi result from R1a-’A’ clade.

* * *

From: Mike Rothenberg

To: A.J. Levin, Michał Milewski, Bennett Greenspan

Date: August 4, 2011

Hi Michal, A.J. and Bennett,

I agree with A.J., Michal. You have provided a new line of questioning I hadn’t thought of. Thanks.

There could be Ashkenazi Levites who might have moved, years ago, to areas where other Levites have lived for generations (e.g., to lands where Sephardic Jews moved after the Inquisition, such as various Mediterranean-area countries; Italy, including Jews of both Sephardic, Italic background, etc.). The possible Ashkenazi background of a paternal line may have faded from memory - and so it seems like a Sephardic Levite is L342+. Plus, the name Levy is common among both Ashkenazi and Sephardic Levites. I know of a situation that went the other way: A family of German Jewish background did genealogical work, and found, hundreds of years ago, a Sephardic ancestor.

The other question is the presence of R1a1 in Ashkenazi Levites. Really, meaning in Levites. You do present a possible reason why this may have occurred such that, ultimately, the male R1a1carrier(s) wound up in the areas of Ashkenazi settlement. But why any convert was permitted in this Levite religious classification involving temple service, below Cohens (priests) and above the remainder of Jews who don’t have either paternal-based temple service classification, seems unanswerable and completely out of the ordinary. That’s why I think there must have been something extraordinary about such convert(s). Though not an expert but with some knowledge of Jewish history, I have never heard of anything occurring, hundreds to thousands of years ago, about any ethnic or national group where its non-Jewish king, leader, etc., converted to Judaism and, in deference to his power, influence, desire to convert and have his people convert, etc., was granted the extraordinary honor of Levite status - other than the Khazar conversion. Do you have any thoughts on this? Perhaps the close SNP relationships among R1a1 Ashkenazi Levites indicate that the conversion was limited to the king and related males only. Another mystery........

* * *

Extrapolating your thought, my understanding is that [the Khazars], at least partially, originated from one of a number of Turkic groups that descended from people known as Goturks (or Gokturks) who I think originated further East in Central Asia, near China. As other Turkic offshoots developed from the Go(k)turks that, in a way, continue on today as various native ethnic groups within, e.g., the Central Asian countries, it seems worthwhile to me to try to get additional testing done on them. They should be at least partial ethnic kin to the Khazars. I would try to find where samples exist for ethnic Turkic men who have already been tested and found to be R1a1(a1). Do you have any thoughts on testing for L342 and the other new intermediate SNPs? Any suggestions on what I should say?

* * *

Regards,

Mike Rothenberg

* * *

From: Michał Milewski

To: A.J. Levin

Date: August 5, 2011

A.J.:

The Ashkenazi-Levite Project does have one member who claims descent from a Spanish Jew named Levy, but I haven’t seen a GEDCOM or any corroborating evidence.

Thank you A.J.

His STR haplotype has several less common values, but only one marker seems to show a pattern that is really unique among Ashkenazi Levites (DYS464). I don’t think this would be enough for a ‘missing link’ I was searching for, though these uncommon values include DYS444=15 that is present in all three closest relatives of type-A cluster according to Lukasz’s diagram: Kussad, Kurd and Pillai (with Al Rass family members showing even more unusual DYS444=11 or 12, which indicates a ‘shared innovation’).

Mike:

But why any convert was permitted in this Levite religious classification involving temple service, below Cohens (priests) and above the remainder of Jews who don’t have either paternal-based temple service classification, seems unanswerable and completely out of the ordinary.

Therefore maybe we should also consider a possibility of NPE, although I’m aware that such suggestions are usually met with skepticism, as it always sounds like a bad excuse (where other arguments don’t work).

Lukasz:

Look at Igor Rozhanski map ... there is no results from Southern and Western Europe!

Firstly, there are also no results showing similar haplotypes in the Eastern Europe (I mean east of the former Polish-Lithuanian Commonwealth) and such results should be expected if Khazar/Turkic origin of this clade is suggested. As was already mentioned by A.J., this may simply mean that not many people from this region have been tested so far. However, I would expect that the related haplotypes should be quite frequent among certain Turkic people (who derived from Khazar Kingdom), assuming that the royal family (in a broad sense of this term) practiced polygyny and produced many descendants (before they converted to Judaism).

Secondly, we know that L342 exists in Italy (Tuscany) and maybe in Spain (as is shown by the above example). Also, we have been notified (on DNA Forum) about the preliminary data suggesting that Z93 and Z95 are absent in Balkans and Eastern Europe while relatively frequent in Italy and Iberia.

Regards,

Michal

* * *

From: A.J. Levin

To: Michał Milewski

Date: August 5, 2011

Thank you A.J.

You’re most welcome. Yes, it’s not conclusive yet.

Of course it’s always possible.

Though if we do not find widespread L342 Ashkenazi Levite-like motifs elsewhere in this area, as we would expect if the origin were royalty, we might hypothesize that it was not royalty but a small group of Khazar priests who became Jews, and who simply never quit having a priestly function. The religious hierarchy probably would not have been completely overhauled overnight: people still needed priests, and if there was a priestly caste in Judaism, the priests were going to belong to it! Just another possibility.

A.J.

* * *

From: Michał Milewski

To: Mike Rothenberg, A.J. Levin

Date: August 12, 2011

Hi all,

In response to Mike’s remarks:

I wrote:

Therefore maybe we should also consider a possibility of NPE [...]

Now, after doing some ‘research’, I came to the conclusion that there is a series of important historical/biblical information that can explain the presence of foreign Y chromosomes among the Levites. Importantly, this alternative explanation seems to be somehow connected with my Persian hypothesis.

According to my modified hypothesis, the L342 marker could have been introduced to the Levite community by some people from the group called Nethinims. The Nethinims (or Nethenims) were foreign (non-Israelite) slaves/servants in the Second Temple serving there under the supervision of Levites. The key is that they gradually became a part of the Levitic society, as is acknowledged by many scholars.

The name Nethinim is first attested in the post-Exilic texts (Books of Ezra and Nehemiah), though Nethinims were often regarded as descendants of Gibeonites (Josh 9:1-27) or as the continuators of the old tradition, that allowed giving some lower functions in the temple (like cleaning or carrying the water and wood) to the group of captured enemies or slaves.

There were two groups of Nethinims coming to Israel with the repatriants from Babylon. The first group (392 people) came to Jerusalem in 538 BC as a part of the first wave of Jews who returned to Jerusalem under the leadership of Zerubbabel/Sheshbazzar just after the Israelites were freed by the Persian king Cyrus the Great (Ezr 1). The names of the Nethinim leaders from this group (Ziha, Hesupha, Tabbaoth, etc., listed in Ezr 2:43-58 and Neh 7:46-56) are regarded as being non-Jewish. Many of these names sound Semitic or Hurrian/Urartic to me (though I’m not an expert in this field), which would partially support the view that at least some of these people could indeed originate from the Palestine region (Gibeonites, Ammonites, Kanaanites, etc.).

The case of the second group of Nethinims is a little bit different. They came to Israel much later (in 458 BC or even later) with the second wave of repatriants under the leadership of Ezra, a famous priest and writer. Before Ezra left Babylon he was troubled by the shortage of Levites and servants who were needed for maintaining the service in the temple (Ezr 8:15). Because of that shortage Ezra had to postpone (or interrupt) his journey to Israel. He then sent some respected compatriots to the place called Casiphia, where they asked the man named Iddo (a leader of the local Levite community?) to find some Levites who would be willing to go with Ezra to Israel (Ezr 8:16-17). They managed to collect 38 Levites, including Sherebiah, Hashebiah, Jeshaiah and their relatives (all with well documented Levitic ancestry, Ezr 8:17-18). They also brought with themselves 220 Nethinims (temple servants) from Casiphia - to my knowledge this is the only group of Babylonian repatriants for whom the names of their ancestors were not mentioned in the Bible (or nowhere else). They were described as servants whom David and his officials had set apart to attend the Levites (Ezr 8:20). It seems likely (if not certain) that they were non-Israelites. In my opinion, there is also a probability that there were some local people (slaves or servants) among them. The exact location of Casiphia is uncertain, though according to many biblical commentaries it shouldn’t be far from Babylon. It is, however, worth noting that according to some commentators Casiphia was located not in Mesopotamia (Babylonia) but in Persia, and some scholars/commentators place it specifically in Caspian Mountains between Media and Hyrcania (near the Caspian Sea). It has also been suggested that the Nethinims from Casiphia were employed in local silver mines (keseph means silver in Hebrew). If Casiphia was indeed located in Media (conquered by Persians in 550 BC) or east of it, it seems possible that these Nethinims were slaves/servants of Iranian origin (Medes, Parthians or Scythians) who were given under the supervision of Israelites/Levites, likely being in charge of running the silver mines (in the name of the Persian rulers), though this of course is highly speculative.

After they came to Israel, the Nethinims were living in Jerusalem in a place/district called Ophel (Neh 11:21). They were initially forbidden to marry priests, Levites or other Israelites. This, however, changed with time. It seems generally accepted that the Nethinims, along with singers and porters, soon became regarded as specific subgroups within a broad Levite community. Finally, they all merged together into a single caste/class called Levites.

I hope you’ll find all this of help.

Regards,

Michal

* * *

From: Michał Milewski

To: Mike Rothenberg, A.J. Levin

Date: August 12, 2011

Hello again,

To give you a better view of the Nethinims-Levites integration process I am citing below a long excerpt from the Encyclopaedia of Religion and Ethics (by James Hastings et al.):

-----------------------------------------

The priests of the Jerusalem Temple, as of the other sanctuaries, were Levites, and all Levites were priests performing the priestly functions. The other functions of the sanctuary were performed by the slaves given for this purpose. These functions were divided, so far as our information goes, into three main parts; singing, guarding the gates and performing the service of porters, and the more menial functions, such as cleaning. It was to those performing the latter functions that the title Nethinim was given. Originally the slaves assigned to these various functions, being presumably largely foreign captives, were heathen and sometimes uncircumcised. They themselves, however, or their descendants, continuing in the service of the Temple, became in time circumcised Israelites. At the same time, their ranks would seem to have been recruited from time to time by foreign captives; hence the protest of Ezekiel, referred to above, uncircumcised foreigners serving in the Temple.

With the overthrow of the high places and the concentration of worship at Jerusalem, in 623 BC, it became necessary to provide for all Levitical priests at the Temple. Had they become priests there, its priestly sacrificial staff would have been inordinately increased; hence the priests of that Temple resisted this reform for the sake of their own prerogatives and perquisites. The Reformation, however, marks a very distinct step in exclusivism - the separation of the Jewish people from the outside world, and of their sacred worship and its service from all outside contact.

The protest of Ezekiel, himself a priest of the Temple, and his proposition to have all Temple service performed by Levites, are evidence of the existence among the priests of the intent to preserve priestly functions and prerogatives exclusively for the descendants of the priests of the Jerusalem Temple. Hence the rejection of the claims of Levites from other sanctuaries to act as priests in their Temple. It is also evidence of a spirit of religious exclusivism, which would shut out foreigners from the nation, and particularly from the sacred religion of the nation, as represented in the Temple functions. “Let us get rid of the foreigners and utilize these Levites for the service of the Temple heretofore performed by them. The Levites deserve to be so treated as a punishment for their idolatry in the past.”

In principle this point of view was made effective in the Priestly Code, which did not, however, become the law of the Jews or of the Temple until long until long after the return from the Exile. To this extent, however, Ezekiel’s point of view seems to have become at once effective, that, with the return from the Exile, only those counted as Jews were allowed to serve in the Temple, circumcised men of the Jewish religion. Such the porters, the singers, and the Nethinim had become, and with the restoration of the Temple after the Exile they resumed their old functions. The number of so-called Levites who returned was very small in relation to the number of these three classes of Temple servants.

Precisely what the function of the Levites now was it is difficult to determine. Theoretically they seem to be distinguished from the priests on the one hand and from the porters, singers, and Nethinim on the other. The tendency was, on the one side, to push down the Levites, by making them inferior to the priests, and, on the other, to exalt the Temple servants by bringing them to an equality with the Levites. The distinction of priests, Levites, porters, singers, and Nethinim is established for the time of Nehemiah (c. 440 BC) and of Ezra (c. 380 BC). The Priestly Code, promulgated at the latter period, abolishes the distinction between those rendering any service below that of the priesthood in the Temple. They are all alike to be regarded as Nethinim, given to God for the Temple service, and, accordingly, in the records of the Chronicler (c. 300 BC) we find the old distinctions done away with. There are now only priests and Levites. This process seems to have been a gradual one, and affected first the singers. So in certain remodelled portions of Nehemiah the musicians are already regarded as Levites. This is true throughout Chronicles, and here indeed they not only no longer rank below Levites, but they are the highest class of Levites. By the time of Agrippa II. (AD 50) this attitude of exaltation of the singers at the expense of the Levites in general reached a climax, so that at that time the singers became a separate class, entitled to wear linen robes like the priests.

Door-keepers were the next to be included as Levites, as shown by 1 Ch, and Nethinim the last. With the admission of the latter, all are alike; Levites are Nethinim, and Nethinim are Levites. There is no mention in the NT of singers, porters, or Nethinim, as distinct from Levites. At that time, apparently, all Temple functions, outside of the sacrificial ones, were performed by those who were called Levites, i.e. the descendants of the ancient Levitical priests of the high places, and of the slaves attached at various times to the Temple by purchase or by capture for the performance of the various other functions, as Philo also testifies.

The Talmud, by its references, shows that Nethinim did not exist in the Talmudic period, but it also shows that the Talmudic interpreters were conscious of the foreign origin of the Nethinim, mentioned in the Biblical books. Accordingly in the Talmud Nethinim are classed with, or even placed below, bastards, and their descendants for all time forbidden to marry with Israelites.

------------------------------------------

Regards,

Michal

* * *

From: Leon Kull

To: Michał Milewski

Date: August 12, 2011

Hi Michal,

Thanks for the interesting information.

Unfortunately it doesn’t help much with R1a1 Levites case. How do you explain an absence of the Nethinim descendants between Sephardi and Mizrachi Jews?

As we all know, all R1a1 Levites are Ashkenazi; their common ancestor lived about 1400-1600 years ago. As far as I understand, every group of newcomers that joined Jews in the BC time frame should leave some traces between Sfaradim.

Regards,

-- Leon

* * *

From: Michał Milewski

To: Leon Kull

Date: August 12, 2011

Hi Leon,

You wrote:

Unfortunately it doesn’t help much with R1a1 Levites case. How do you explain an absence of the Nethinim descendants between Sephardi and Mizrachi Jews?

You are right. As I already acknowledged in one of my previous messages, the absence of R1a-L342 among non-Ashkenazi Levites is indeed something we wouldn’t expect if the mutation was present among the Levites earlier than 2000 years ago. However, although such early introduction of R1a-L342 seems not very likely (in view of the above finding), it is not completely impossible.

For example, according to ‘A History of the Jews’ by Paul Johnson, the population of Jews dropped during the first millennium AD from about 8 millions (which constituted about 10% of the total population in the Roman Empire) to only 1.5 million. This alone could substantially contribute to the ‘evaporation’ of many rare lineages in certain Jewish subpopulation (and the putative ‘Iranian/Nethinim R1a-L342’ marker would have been quite rare among Levites in the 1st century AD, not to mention its rarity within the entire Jewish population). Importantly, this was the time when Jews were leaving Israel while settling in many different locations in Africa, Asia and Europe. Therefore, it seems very likely that many rare DNA markers were enriched in some regional Jewish subpopulations while becoming totally absent in others (the bottleneck effect). Furthermore, there are some indications of the relatively less frequent presence of L342 in Italy and Spain (in fact the mutation L342.2 was first detected in Italy and in Palestine, and not among the ancestors of Northern- or Eastern-European Ashkenazis), which would be consistent with the most likely route of L342 expansion in Europe (if it indeed came from Israel).

I would also add one important remark regarding the possible date of introduction of L342 to Israel. The year 458 BC (mentioned in my previous message) seems to be the earliest possible date. However, the much later dates cannot be excluded. If Ezra left Babylon during the reign of Artaxerxes II (and not Artaxerxes I, which is still a subject of controversy), the Nethinims from Casiphia would have come to Israel as late as 398 or even 368 BC. Assuming that before they were totally integrated with the Levites in Jerusalem they were living in Israel as a separate community for at least several generations (100-200 years or even more), it seems possible that the first L342 men who were unequivocally regarded (by other Jews) as Levites would have lived not long before the Second Temple was destroyed and all the Levites from Jerusalem started living in Diaspora. This would drastically increase the chance that the small group of former Nethinims of Iranian origin would have emigrated just to one location outside the Palestine (whether it was Italy or not).

Regards,

Michal

* * *

From: Mike Rothenberg

To: Leon Kull, Michał Milewski, A.J. Levin

Date: August 12, 2011

Hello Michal and A. J.,

This has been an extremely interesting set of e-mails. I wonder, if hundreds of Nethinim were incorporated into the body of Levites in the 400-600 BCE time period, wouldn’t their descendants have numbered thousands by the time of the Roman-period exile? I think they would have been dispersed among all the dispersed Levites, who we consider as part of Mizrachim, Sephardic and Ashkenazi groups.

Also, is it possible that some remained with those Levites who didn’t return with other Israelites to Israel? I don’t know the haplogroup variations among their current descendants, but do they include R1a1-L342 among Levites? How about those who were exiled during the Roman period and became the ancestors of today’s Sephardim, and perhaps some Mizrachim? Do they include any R1a1 L3342 men among Levites? My understanding is that R1a1-L342 is almost non-existent in those Levite groups, and occurs almost exclusively among Ashkenazi Levites. I don’t know how random dispersion of Levites, over the ages up to and including the Roman exile, could have produced such a concentration within the Ashkenazi Levite group. Also, is there any genealogical evidence that R1a1-L342 exists among current descendants (if any) of people who are known to be descendants of ethnic kin of those past Nethinim?

Regards,

Mike Rothenberg

* * *

From: A.J. Levin

To: Michał Milewski

Date: August 12, 2011

Dear Michał:

Many thanks for researching this. A Persian > Levite link as you suggest certainly seems possible genetically and historically. Though with the closest ySTR match to Ashkenazi Levites being a probable Mamluk descendant, the Nethinim link seems to me not as probable as the Khazar explanation, at least for the moment. I hope as more people are tested a clearer pattern will emerge.

Best,

A.J.

* * *

From: Michał Milewski

To: A.J. Levin

Date: August 13, 2011

A.J:

the Nethinim link seems to me not as probable as the Khazar explanation, at least for the moment.

I agree that it would be premature to dismiss the Khazar theory. Of course I like my Persian/Nethinim hypothesis, but it seems quite likely that I’m becoming a little biased because of my personal involvement in constructing this theory.

On the other side, the more I read on this subject the more doubts emerge about the suggested Khazar origin of the Levitic R1a-L342. For example, since Khazars converted (most likely) to Rabbinic/Talmudic Judaism, why would they need to have the Levites among themselves so badly? The importance of Levites was dramatically diminished with the destruction of the Second Temple, so why would the Khazars decide to appoint this function to some people of local origin (while neglecting the long tradition and written Law)? Maybe they envisioned conquering Palestine and restoring the Temple in Jerusalem, but was such idea indeed popular among the Rabbinic Jews who likely assisted in Khazar conversion in the 8th-9th century AD?

I hope as more people are tested a clearer pattern will emerge.

Yes, I also hope these results will come soon.

Regards,

Michal

* * *

From: Michał Milewski

To: Mike Rothenberg

Date: August 13, 2011

Hello Mike,

You wrote:

I wonder, if hundreds of Nethinim were incorporated into the body of Levites in the 400-600 BCE time period, wouldn’t their descendants have numbered thousands by the time of the Roman-period exile? I think they would have been dispersed among all the dispersed Levites, who we consider as part of Mizrachim, Sephardic and Ashkenazi groups.

I tried to answer this question in my response to Leon, but it seems now that I wasn’t very convincing. Let me explain it again using the example shown below. I will use some hypothetical (though quite likely) numbers.

Let’s assume that the population of all Jews at the beginning of the first Millennium was 8 million people (according to Paul Johnson). Half of them would be males, which makes 4 million men. How many of them were Levites? Let’s assume that the percentage of the Levites was similar to that reported for the returnees from Babylon, which was about 1.7% (this includes both the Levites and their assistants: singers, porters and Nethinims). This would make a total number of 68,000 Levites in 1 AD. How many of them could have been the ancestors of Nethinims of Iranian origin? Let’s try to calculate it, assuming that there have been as many as 10 Iranian (L342+) men among the 220 Nethinims (of various origin) coming from Casiphia between 458 and 369 BC. The total number of all Levites, singers, porters and Nethinims living in Jerusalem at that moment is difficult to evaluate, but it could be at least 2,000 men. The Jewish population doubled every 70-80 years, so about 400 BC the initial group of 773 Levites (L+S+P+N), who returned from Babylon in 558 BC, would expand to 1500-2000 men (and we should also include the 258 Levites and Nethinims coming from Casiphia). Ten L342 Nethinims out of 2000 Levites (L+S+P+N) makes 0.5%. Though this proportion could have been altered by the genetic drift, let’s assume that it was preserved until the end of the first Millennium BC. Continuing these calculations, 0.5% of 68,000 Levites makes 340 men of Iranian origin in 1 AD. During the next centuries the whole Jewish population was decreased by the factor of 8 according to Johnson).

So we are left with just 42 Levites carrying the L342 mutation. Is there a chance that these people were evenly distributed between the Sephardi, Ashkenazi and Mizrahi Jews? I doubt so. And what if there was only one L342+ man among the Nethinims who came from Casiphia? We would end up with just four L342+ men at the end of the first Millennium AD. It would be then very likely that they all were the members of just one family, living in a single location somewhere in Europe, so there would almost be no chance to evenly allocate these people to different Jewish subpopulations (Sephardi, Ashkenazi or Mizrahi).

(For the matter of simplicity, I did not count the Levites who did not return from Babylon, although including them would further decrease the number of L342 people in 1 AD.)

Also, is there any genealogical evidence that R1a1-L342 exists among current descendants (if any) of people who are known to be descendants of ethnic kin of those past Nethinim?

I am pretty sure that absolutely nothing is known about the descendants of people being ethnic kin of Nethinims, as we don’t really know what their ethnic identity was. They could be of various ethnic origin (Kanaanitic, Hurrian, etc.), and most of these Near Eastern ethnicities are only known from the history (they left no descendants keeping track of their ancestry). But if you mean by the above the descendants of ancient Iranians (Persians, Medes, Parthians, etc.) it is maybe too early to give you exact numbers but it seems that the majority of R1a people in the Middle East are L342+ and they all are likely to be the descendants of different Iranian tribes.

Regards,

Michal