Europeans, Jews and Middle Easterners

Share on FacebookShare on Google+Email this to someoneTweet about this on Twitter

Greg’s post about SNPs, Jews and evolutionary genetic parameters has been getting a lot of play around the blogs & forums. Most of it seems to be due to the persistent interest in the genetic relationship of Ashkenazi Jews to other European populations. This makes sense, since the 19th century the question of how the “Jewish race” relates to European gentiles has had some sociopolitical relevance…. But a commenter at Steve’s blog pointed out that Bauchet et al. from last year had a PC chart which included Armenians, who are I think a good proxy for northern Middle Eastern populations in general. One interesting result from surveys of Y chromosomal lineages is the finding that Jews may have more affinities with northern Levantine & Anatolian Middle Eastern populations than with southern Levantine and Arabian ones. The non-trivial female mediated input of Sub-Saharan ancestry into many Arab populations since the rise of Islam is far less evident in non-Arab Muslim populations (Kurds, Persians and Turks) as well as Middle Eastern Jews, and obviously Ashkenazi ones. But another point is that recent work suggests that the impact of historical events (e.g., the Arab conquest) might have been more demographically significant than we had previously assumed, and so Jewish affinity with northern Middle Eastern populations may reflect that these groups have been less affected by exogenous genetic inputs within the last 2,000 years.

Caution about the sample sizes of course (though I assume within the next year we’ll have much better data to go off of), but something to include into your list of priors when making phylogenetic background assumptions.

Note: I added geographic labels to the PC chart for clarification.

Update: Steve has another post up:

On the first two axes, Ashkenazi Jews are rather close to “Europeans” and “Russians.” They are similar to Yemenites (from Southern Arabian peninsula) on the first axis, but not on the second. And they are similar to Samaritans (who currently subsist on two hilltops in Israel), good, bad or indifferent, on the second axis but not on the first. They are fairly similar to the Druze (of Lebanon and Israel) on the first two axes, but not on the third.

So, Ashkenazis look pretty European on this chart compared to a few Middle Eastern groups. But, as the recent graph showed, genetics has progressed to the point where Ashkenazis (at least those with four Ashkenazi grandparents) can now be reliably distinguished from other Europeans.

The Samaritans are cousins of the Jews. But:

In the past, the Samaritans are believed to have numbered several hundred thousand, but persecution and assimilation have reduced their numbers drastically. In 1919, an illustrated National Geographic report on the community stated that their numbers were less than 150.

Like the Kalash or Sardinians the Samaritans are going to be weird outliers because of their demographic history. Inbreeding and no gene flow in for that long will do that to you (many people in the Middle East are descended from Samaritans of course, but very few Samaritans are descended from non-Samaritans).

The Yemenites are also a peculiar comparison point because they are geographic outliers in relation to other Middle Eastern populations with a long and distinct history. They have a large proportion of Sub-Saharan ancestry for an Arab group. An interesting historical note is that during the Islamic expansion Yemenite tribes were prominent in Iraq and Egypt, though I doubt they left a very strong genetic imprint in these regions.

The Druze are a better point of comparison, being a more mainstream Middle Eastern group. But that’s only relative to the Samaritans, who are at an advanced stage of pedigree collapse, or the Yemenites, who are on the geographic margins of the Near East (it is easy to argue that before Islam Yemen was more a part of the trans-Indian Ocean world than it was of the Near East). The Druze are an esoteric ethno-religious group which as been resident in the mountains of Lebanon. who have not accepted converts since 1031, so again you have a recipe for some genetic distinctiveness developing because of social norms.

All that being said…perhaps as we explore the genetics of the Middle East further we’ll find that most groups exhibit these sorts of inbred tendencies because of the prevalence of consanguinity?

Addendum: Modest levels of gene flow are very good at equilibrating and mitigating the build up of variation between groups. Islands, like Sardinia, often develop unique genetic profiles because water seems to be a powerful barrier to marriage connections. The Samaritans & Kalash have not had any gene flow in for a very long time, in both cases in part because of being embedded among Muslims who do not generally tolerate conversion to other religions, and in the case of the Kalash their geographic isolation. Some of the same issues apply to the Druze, though I suspect much more modestly (in part because Druze isolation is more recent).

Labels:

43 Comments

  1. One interesting result from surveys of Y chromosomal lineages is the finding that Jews may have more affinities with northern Levantine & Anatolian Middle Eastern populations than with southern Levantine and Arabian ones. (…) recent work suggests that the impact of historical events (e.g., the Arab conquest) might have been more demographically significant than we had previously assumed, and so Jewish affinity with northern Middle Eastern populations may reflect that these groups have been less affected by exogenous genetic inputs within the last 2,000 years. 
     
    But there’s another possible twist to this data: Jewish diaspora, particularly Askenazi one did not really begin in Palestine/Israel but in what is now Turkey speacially. During Hellenism and the Roman Empire, millions of Jews had already begun their diaspora, largely for economic reasons, and Anatolia was their main destination. It was from there that they spread into Europe, not so much from the Levant, that had been largely depleted of Jews by the Roman genocide.  
     
    Jews were not so closed to proselitism and assimilation of gentiles (converts preferably) in old times, the cases of many Jewish populations in North Africa, Yemen, Ethiopia or Russia (Khazars) shows that very clearly. We don’t have enough historical data on the Jewry of Asia Minor to judge but it’s very possible that they also assimilated many other locals, before parting to the European heartland.  
     
    Remember that Hebrews are Semitic-speakers and that their arrival to Canaan in the late Bronze Age happened inside other Semitic migrations that preceded the Arab one but had probably the same origin anyhow. The genetic contrast between Arab and non-Arab West Asians can well be product not just of the Arab conquest (after all Arab seminomads were few and the agriculturalists of the Fertile Crescent many) but of the many succesive waves of Semitic input since the 4th milennium BCE: Akkadians, Amorites, Canaanite-Phoenicians, Chaldeans, Hicsos, Assyrians, Arameans, etc. 
     
    Based on on Baucher, I would think that is the case, at least as a serious possibility: that the differences between Palestinians and Askenazi Jews are not just caused by Arab input in Palestine but also by Anatolian (and other) input in Jews. 
     
    Proabably we would need further more focused studies to discern clearly anyhow.

  2. two other areas of jewish concentration in late antiquity: alexandria & mesopotamia. 
     
    note: i’m sure there was plenty of non-trivial conversion to judaism by those who were not descended from the people of the kingdom of judah during antiquity. that being said, of late i’m a touch more skeptical of extreme claims of jewish proselytism for two reasons 
     
    1) some of the accusations of judaization among the roman elite (e.g., at the court of domitian) after the first century rebellion seem to more likely have been analogs to mccarthyism as opposed to genuine religiously motivated animus. that is, the jews were now perceived to be a congenitally anti-patriotic element and accusations of judaizing were now one of the barbs thrown at those who were out of favor. 
     
    2) some of the claims by jews about the jewishness of elite romans (emperors or their associates), or the familial relationship of famous jewish rabbis to roman emperors, etc., seem likely to be attempts by jews to associate themselves with gentile powers who were not as reflexively hostile to them. they are similar to the periodic rumors in the american muslim community that michael jackson has converted to islam (well, ok, this is before the pedophile period). or, a roman analog are the suspicions that phillip the arabian was a crytpo-christian. more likely this is retrospective projection because of the contemporary contrast with the first systematic anti-christian emperor, decius.

  3. I’m not too surprised the the Ashk Jews cluster pretty closely with the Armenians, for exactly the reasons cited.  
     
    But what *does* surprise me is that both these groups cluster so closely with the Greeks, whose historical connection is pretty tenuous. After all the ancient Greeks were supposedly Indo-Europeans from the North, not Semites, and I’d always half-suspected that modern Greeks might be mostly “Hellenized” Slavs or whatever—maybe Bulgarians or something. Either way, you’d expect the Greeks to be much, much closer to the other Europeans than to Armenians or Jews. 
     
    It would also be very interesting if our techniques become precise enough to extract Carthaginian DNA, and determine whether the Mid-Eastern ancestry of Ashk Jews is actually mostly Carthaginian rather than Judean, as I’ve often suspected.

  4. After all the ancient Greeks were supposedly Indo-Europeans from the North, not Semites, and I’d always half-suspected that modern Greeks might be mostly “Hellenized” Slavs or whatever—maybe Bulgarians or something. Either way, you’d expect the Greeks to be much, much closer to the other Europeans than to Armenians or Jews. 
     
     
    on the “indo-europeans from the north,” i think the genetic data for most indo-european speaking groups suggests that the spread of this language family was mostly driven by elite diffusion, or that the initial genetic signal was lost along the wave of advance. IOW, the hellenes are as indo-european genetically as the anatolian turks are mongolian. the one model where the greeks are preponderantly indo-european that i know of is renfrew’s, which posits that the indo-european ur-heimat is in anatolia. in this way the IE expansion is concomitant with that of neolithic farmers from the near east, who might have contributed most of the ancestry of the peoples of greece proper. i’ve also read that 40% of greek words seem to be of pre-greek provenance (e.g., words with nth are not indo-european). the balkans is complicated by the fact that groups like bulgarians and serbs themselves show evidence of being composites of slavs (skalveni) along with the indigenous latinized or illyrian substrate. 
     
    It would also be very interesting if our techniques become precise enough to extract Carthaginian DNA, and determine whether the Mid-Eastern ancestry of Ashk Jews is actually mostly Carthaginian rather than Judean, as I’ve often suspected. 
     
    my understanding is that carthaginians often had non-punic maternal ancestry (hannibal did from what i recall, greek i believe). so if you want to do this the best bet would be to just check mtDNA and see if was like those of berbers. the relationship of ancient hebrews and phonecians though was very close from what i recall. 
     
    anyway, spencer wells has done some stuff on the phoenicians…. 
     
    http://ngm.nationalgeographic.com/ngm/0410/feature2/index.html?fs=www3.nationalgeographic.com&fs=plasma.nationalgeographic.com

  5. re: armenians. from what i recall of byzantine history when the empire reacquired the balkans after a 3 century interlude when avars, bulgars, serbs, etc. had ruled the hinterland away from the coasts many people from anatolia were settled in these lands. many of these were of armenian origin (though hellenized). some were also levantine christians from the borderlands with the muslim world.

  6. Compliments on the delicatesse of “non-trivial female mediated input of sub-Saharan ancestry”. Shades of Jefferson, eh?

  7. Biblical accounts of Jewish history can be backed up with archaeology at the point of the return from the centuries long exile in Babylonia/Persia (told in Ezra/Nehemiah). It’s likely that the Jews had some mixing with Persians during that period which would have the result of clustering them more closely with Armenians. (armenia is right on top of Iran). same thing would explain why we cluster closer to Kurds, as well. Persia was the center of Jewish civilization even after Cyrus let the Jews return to Israel – most stayed in Persia. 
    I wonder if anyone has compared Jewish genetics to Persians? that might be interesting….. I once read something related on an anti-semitic site which I was reading just because it was bizarre (I’m Jewish) – they said that you could tell Jews were “not really from Palestine” because of our noses – Palestinians have rounded blobbier noses, and Jews tend to have thinner, bonier, beakier noses “like Persians” the website said….. Obviously it was coming from racist nonsense, due to the source, but I thought it was an interesting observation. 
     
    We Jews are such a huge mix – in my family we have Middle Eastern curly hair and big eyes, Eastern European coloring and height, Central Asian bone structure and slanting eyes, and some noses that look much more Italian or Persian than Arab – yet we’re supposedly 100% Ashkenazi – what a joke!

  8. Armenians are stereotypical for the same noses as Jews, which was actually sometimes presented as evidence that physical traits are related to cognitive traits, in this case of course the supposed racial instinct for bloodsucking, conspiracy and general deviousness (Armenians were often similarily accused). In the old classification literature they sometimes spoke of “Armenoid” noses. 
     
    Armenoid was also a wider northern Middle Eastern and Caucasian racial classification used by at least Coon which IIRC did include the noses. Unfortunately, that’s probably still the first English language source for this stuff, since I’m betting no one has funded studying the distribution of beaky nose types since about 1945…

  9. re: sojourn in babylon, note that mesopotamia had relatively few persians aside from the elite during the period of persian rule (sassanids too). the populace was mostly semitic speaking, like the jews.

  10. I would expect that Greeks are closer to, say, Italians than to Armenians or Jews, but I would guess that when one compares these 4 groups to Scots or Swedes, the Greeks and some Italians (particularly Sicilians and other southerners with significant Greek ancestry)* would cluster more with the “non-European” Armenians and Jews than with the “European” Scots and Swedes. This is not something that is surprising to anyone who has seen Greeks or knows the history of the Mediterranean basin. 
    On the other hand, phenotype and geography can be deceptive. While Spaniards often physically resemble Greeks and Italians, many Spaniards are probably more closely related to western Europeans such as the French and British than to eastern Mediterranean peoples such as Greeks. 
     
    *For example, my Sicilian ancestors were from Messina, where Greek was spoken as late as the 16th and 17th centuries.

  11. “Italians” seem to have a lot of variance.

  12. also, this focus on the european-non-european difference seems a little weird. in 900 most greek speakers probably were resident outside of europe, in anatolia. just because there’s a line on the map doesn’t mean that genetic variation is going to track that well. in the 1920s there are a mass transfer of anatolian “greeks” (some of whom spoke turkish but were orthodox christians) to greece itself.

  13. Probably the closest thing genetically to the Carthaginians are the modern Maltese. Spencer Wells has pretty convincingly shown them to be distinct from the Tunisian Arab invaders of their territory and to more resemble the Lebanese and Syrians, in resemblance to the geographical origin of the ancient Phoenicians with no other possible explanation of how the genes could have got there. 
     
    They are a fascinating people in that they were always Christian (pretty much as soon as it existed), even before Europe was, and are famed for being near nutjobs in terms of fighting for their beliefs, most notably against Islam and the Nazis. It does help if you are in a highly defensible island though. 
     
    Despite their Semitic linguistic and genetic nature they are accepted as fellow Europeans by almost all members of the EU because thy are aware of the bloody fights they have engaged in which always seem to fall on the side of “western civilisation”. There is a particular affection between the British and the Maltese and a particular animosity between the French and the Maltese.

  14. Razib’s point about the majority of “Greeks” having lived in Anatolia for maybe a thousand years prior to the 1920s seems a very good one.  
     
    That might certainly allowed a great deal of intermarriage/genetic transfer over the centuries between the Anatolian Greeks and the Greece Greeks, even before the eventual population transfers after WWI.

  15. Razib -a Jewish population took root in Persia during the Babylonian captivity. Persia has always had the largest Jewish population in the Middle East (until the formation of Israel)

  16. True. Italians from different regions may not belong to a single “ethnic group”,* let alone a single genetic cluster. If there is indeed a single “Italian” ethnic group, it is of fairly recent origin; for example, few Sicilians had even heard the terms “Italy” or “Italian” prior to unification.  
    The ancestral backgrounds of Italian-Americans are highly atypical of Italy as a whole. Italian-Americans are primarily descended from Southern Italians, with a disproportionate number of those coming from Campania and western Sicily. 
     
    The debunking of the “European” vs. “non-European” distinction was in response to RKU’s comments about Greeks being related to Jews and Armenians versus “other Europeans.”

  17. The makers of the video game “Rome: Total War” chose to represent the accent of the Carthaginians by using a Maltese actor. In terms of our actual current knowledge, this is fair enough.

  18. Razib -a Jewish population took root in Persia during the Babylonian captivity. Persia has always had the largest Jewish population in the Middle East (until the formation of Israel) 
     
    me, i know. but my point is that jews tended to be resident in the aramaic speaking lands of mesopotamia, not in persia proper. to make clear my point, ctesiphon did not have any fire temples, but it did have many churches and synagogues. just because “persia” generally pushed to the euphrates does not mean that the region between the two rivers was ever populated by many persians. rather, it was a predominantly semitic speaking region.

  19. and me, i don’t really discount your general point. but i think it is important that people know that the babylonian jewry probably interacted mostly with aramaic speakers of various religions (christians and exotica such as mandaeans) as opposed to zoroastrian persians. obviously some jews would be resident in the persian heartland, but i assume there’s a reason that one would call them *babylonian* jews.

  20. True. Italians from different regions may not belong to a single “ethnic group”,* let alone a single genetic cluster. If there is indeed a single “Italian” ethnic group, it is of fairly recent origin; for example, few Sicilians had even heard the terms “Italy” or “Italian” prior to unification.  
     
    cavalli-sforza thinks that there is an arguable italian group 
    http://www.gnxp.com/blog/2006/08/10-questions-for-luigi-luca-cavalli.php 
     
    that being said, i’m just pointing out that there is a wide variance across the peninsula, and when we are scale down to this level of granularity we’re really losing a lot of information at the margins based on how we chunk the categories.

  21. I think there is probably an “Italian” group (the people of central Italy probably have the strongest membership in this group). But at the margins of Italy, distinguishing a typical Sicilian from a typical Piedmontese is likely to be easier than distinguishing the Piedmontese from a Savoyard Frenchman, or the Sicilian from a Greek.

  22. note: i’m sure there was plenty of non-trivial conversion to judaism by those who were not descended from the people of the kingdom of judah during antiquity. that being said, of late i’m a touch more skeptical of extreme claims of jewish proselytism… 
     
    We are in agreement, I think. It’s just how you understand the term “proselytism” (as something very aggreive or less so. Certainly there are enough well documented cases of non-Hebraic or mixed Judaists (“Hebraic” as ancestry and “Judaist” as religion, “Jewish” is way too ambiguous) to think of more than just accidental assimilation of very particular people. Historians seem to agree that Judaist “missions” attracted many converts in certain areas (Ethiopia specially but it seems also in North Africa, Yemen – that was once a Judaistic realm – and the really curious case of Khazars) Christianism is in fact nothing but a hyper-proselytistic form of messianic Judaism actually, but surely there were others maybe not so “aggresive” but also open to gentiles.  
     
    I would expect that Greeks are closer to, say, Italians than to Armenians or Jews… 
     
    Why? Italian archaological history is basically Western European – though with very important Greek/Anatolian/Balcanic inputs since Neolithic, affecting specially the south of the peninsula. Greek Neolithic (at the origin of nearly all European one) is probably of Asian origin (Anatolian most likely) and Greeks and Anatolians have been in narrow contact since “always”. Genetically Greeks (and some other Balcanic peoples maybe) tend to stand out in the context of Europe. 
     
    Probably the closest thing genetically to the Carthaginians are the modern Maltese. 
     
    Again why? Malta was densely populated long before the arrival of Phoenicians (and we know little about the origins of these first settlers) and their language is derived from Arabic anyhow, not from ancient Phoenician.  
     
    I think there is probably an “Italian” group (the people of central Italy probably have the strongest membership in this group). 
     
    Hmmm… Remember that Etruscans (and partly modern Tuscans) show great affinity with Anatolia. In Baucher’s study, Italians seem very mixed, but unlike Spaniards, they don’t seem to show a cluster of their own – what I admit is kind of weird, even if considering that Italy has been very open to all kind of outer influences since at least Neolithic times. Maybe with greater clustering depth there would appear such Italian core, even if mixed.

  23. Despite their Semitic linguistic and genetic nature they are accepted as fellow Europeans by almost all members of the EU because thy are aware of the bloody fights they have engaged in which always seem to fall on the side of “western civilisation”. 
     
    Really? People in Finland and Sweden (for example) seem to care fuck all about whether Maltans are Europeans or not. So those tourist islands in the Mediterranean have a few more or less swarthy natives, who cares? If Malta was the size of Turkey, people would care more about the ethnic makeup, but they would also start caring about the theocratical aspects of the island.

  24. This PC graph has the same peculiarity as the one in the previous post: SE Europeans/”Anatolians” are completely opposite to NE Europeans on one PC, but line up almost perfectly on the other. 
     
    I can’t find any reasonable explanation for this pattern. 
     
    and a particular animosity between the French and the Maltese.  
     
    If so, it is pretty one-sided. Most French people couldn’t locate Malta on a map, and the only Maltese “person” they know of is Corto Maltese.

  25. This PC graph has the same peculiarity as the one in the previous post: SE Europeans/”Anatolians” are completely opposite to NE Europeans on one PC, but line up almost perfectly on the other. 
     
    I can’t find any reasonable explanation for this pattern.
     
     
    It doesn’t really make much sense to me either: when looking at the K clusters that PC2 dimension is not visible at all (while PC1 correlates almost perfectly with K=2), but that’s what Fig 5 also shows: that Armenians and Finnish are on the right side of PC2, being low on it Basques and Spanish, so maybe it’s some element that decreases westward. Fig 3 shows that the PC1/PC2 graph is very stable, even when almost half the data is removed. 
     
    The only thing that makes sense is K=3, if we take the blue and red components as one. But in that case, I’d expect the “component” to be the yellow one, not the amalgamation of the other two (i.e. Basques and Spaniards would be high in PC2 and Easterners low, a horizontally mirrored graph). Hmmm… 
     
    Anyhow, I understand much better the K structure than the PCs, that are almost esotheric in design.

  26. Someone, by which I mean Razib, should write something about Cavalli-Sforza’s principla constituents. The different P-C’s are relics of different migrations / diffusions, so that two peoples might be similiar on one and diametrically opposed on another.  
     
    Furthermore, usually differences of this sort tell you something about history — in that case, for whatever both peoples received one influence but only one of them the other influence. 
     
    The value of genetics in history for me is to help disentangle language groups, cultural groups, religious groups, and gene-pools, which a century ago were often assumed to overlap much more substantially than they do. It also can tell you a lot about about what kind of an event something was — e.g., the Anglo-Saxon conquest of Britain.  
     
    The recent study discussed here about the British Isles was pretty startling: the biggest single component of English genes is pre-Anglo-Saxon, pre-Celtic and probably pre-Indo-European.

  27. The only thing that makes sense is K=3, if we take the blue and red components as one. But in that case, I’d expect the “component” to be the yellow one 
     
    Which “one” do you mean?

  28. Luis writes 
     
    “Remember that Hebrews are Semitic-speakers and that their arrival to Canaan in the late Bronze Age happened inside other Semitic migrations that preceded the Arab one but had probably the same origin anyhow. The genetic contrast between Arab and non-Arab West Asians can well be product not just of the Arab conquest (after all Arab seminomads were few and the agriculturalists of the Fertile Crescent many) but of the many succesive waves of Semitic input since the 4th milennium BCE: Akkadians, Amorites, Canaanite-Phoenicians, Chaldeans, Hicsos, Assyrians, Arameans, etc.” 
     
    All of this is well and good except if “northern” non-Arab middle Eastern population were descended in large measure from “southern” Akkadians, Amorites etc. they would appear “southern” and cluster with the other “southern” middle eastern populations. 
     
    Anyways linguistic affiliation doesn’t seem to mean much in this part of the world, at least in Persia (1) the Caucasus (2) or Anatolia (3). All of these areas historically had many different linguistically diverse people but they roughly cluster by geography. Basically, geographically close populations cluster together what ever their supposed linguistic backgrounds, at least in this part of the world.  
     
    1) Close genetic relationship between Semitic-speaking and Indo-European-speaking groups in Iran. Ann Hum Genet. 2008 Mar;72(Pt 2):241-52. 
     
    2)Testing hypotheses of language replacement in the Caucasus: evidence from the Y-chromosome. Hum Genet. 2003 Mar;112(3):255-61. 
     
    3)Excavating Y-chromosome haplotype strata in Anatolia. Hum Genet. 2004 Jan;114(2):127-48.

  29. Someone, by which I mean Razib, should write something about Cavalli-Sforza’s principla constituents. The different P-C’s are relics of different migrations / diffusions, so that two peoples might be similiar on one and diametrically opposed on another.  
     
    Very interesting—is this really correct? I’d been vaguely assuming that the different PC’s were simply synthetic and probably artificial, corresponding merely to the largest eigenvectors. Is it generally believed instead that they usually represent something “intuitive”, like a particular ancestry-group?

  30. Which “one” do you mean? 
     
    In Baucher’s paper (link in original post), the PC graph posted here by Razib is accompanied by a (in my opinion more interesting) Bayesian K-means structure (depth up to K=6). I am refering to that parallel clustering, where you can see at least five geographically organized clusters: K=2 contrast the “Eastern Mediterranean” (red) with the “Finnish” (blue) one, coincident with PC1 dimension, but there are other three clear clusters that can be called: “Basque”, “Iberian” and “Central-Northern European”.  
     
    If in the K=3, we make the two main “populations” (red and blue, “Eastern Med” and “Finnish”) become one, then we get PC2 (or something very very similar). Paradoxically, what we are seeing is not something that Finnish and Eastern Mediterraneans share, but something they lack: a SW component that, in deeper resolution, happens to be basically the Basque characteristic one.

  31. All of this is well and good except if “northern” non-Arab middle Eastern population were descended in large measure from “southern” Akkadians, Amorites etc. they would appear “southern” and cluster with the other “southern” middle eastern populations. 
     
    But that is not the case. For instance for Y-DNA J, you have basically two regions in West Asia: the Arab “lowland” countries and the non-Arab “highland” ones. Of course both regions and the subregions in them have been in contact since “always” and they do share a lot but there’s some of a divide, I think.  
     
    Anyways linguistic affiliation doesn’t seem to mean much in this part of the world, at least in Persia (1) the Caucasus (2) or Anatolia (3). 
     
    Don’t get me wrong: what I am saying is that, since almost 6,000 years ago, Semitic peoples have been pouring into the lowlands of West Asia (Lebanon too, ok), in succesive waves. But these waves had almost no impact in the highlands (Anatolia, Iran, Caucasus). That the descendants of Elamites became Arabized “recently” doesn’t mean that Semitic genes flowed intensely in that area, that was not previously of Semitic language or culture (first Elamite, then Iranian). It’s a case like Morocco maybe, where “Arabs” are genetically mostly Berbers genetically.  
     
    I also know that linguistic adscription in Anatolia and the Caucasus has little correlation with genetics. I do not mean to make a blanket correlation between language and genes: not at all. But rather a historical correlation between pretty well known persistent migrations of Semitic-speaking peoples and the genetic peculiarities of the Fertile Crescent, that was always the main affected area, specially in the Y-DNA aspect.

  32. I’d been vaguely assuming that the different PC’s were simply synthetic and probably artificial, corresponding merely to the largest eigenvectors. Is it generally believed instead that they usually represent something “intuitive”, like a particular ancestry-group? 
     
    You are right, I think. But the layperson can get confused about that, and surely they often do. Myself have only recently discovered the fact you mention (with some relief, btw).  
     
    Rading Cavalli-Sforza, for example, he gives the impression (or at least did in the 90s) that the PCs are something “real”. So maybe the confussion begins with geneticists themselves.

  33. What I know is that C-S’s principle constitutuents were deliberately constructed from neutral markers, in order to eliminate similarities due to selection. As I understand, they are sophisticated mathematical constructs based on gene frequencies in various populations and (since neutral) have no functional meaning. In that sense they’re artifical, but deliberately so. The goal was to define groups based on migration and intermarriage history. 
     
    C-S doesn’t use these constituents to define specific groups, but to find signs of events. For example, one constituent sets the Basques off from the rest of Europe, making it seem that the Basque country is the last refuge of the pre-Indo-European Europeans. That is, the collection of traits found mostly there is a collection of pre-Indo-European traits not found among the Indo-Europeans. (It doesn’t say that the Basques are much different than everyone else, only that the Basques are the people where these traits are concentrated.) 
     
    Other constituents vary with the Greek BC diaspora, the Kurgan cuture’s expansion, and the Saami in Finland. (Whether the Saami are a survivor population like the Basques, or an invading population I’m not sure). 
     
    There’s only a weak correlation with language group for most of the features, but a stronger one with geography. None of the genes are interpreted for phenotype — there was a deliberate effort to make it non-intuitive that way.  
     
    A lot of the stuff I’ve seen recently seems to show that genetics correlates more with history and geography than with language or culture. Neighboring peoples tend to be similar, and if they aren’t, there’s probably a historical reason.

  34. Luis, 
     
    In terms of Jews, Armenians and Indo-Europeans, I’m beginning to think that Y-DNA J2 is the signal of the initial Indo-European expansion with the advent of farming in the Northern Fertile Crescent.  
     
    Renfrew posited Western Anatolia as the homeland of the Indo-Europeans, whereas Dolgopolsky posited the Eastern Balkans – with Anatolia serving as a source of Proto-Indo-Europeans, but in either case it would be from a population probably rich in J2. Note that the Armenian language (and some of its people) is supposed to derive from the Phrygians who invaded and formed a back-migration from the Balkans. 
     
    Therefore, might it not be that the original Israelites – under Joshua etc – were a group of Northern Fertile Crescent people, at an early stage of the Indo-European expansion, who conquered the Caananites (remember the fall of the city of Jerico), but who did NOT manage to impose their language on the locals, maybe due to greater population density locally. 
     
    Might not the tale of Cain (the Caanite?) and Abel, refer to the battle between settled agriculturalists and invading Indo-European pastoralists??

  35. “I’m beginning to think that Y-DNA J2 is the signal of the initial Indo-European expansion” 
     
    Think something else.

  36. Rading Cavalli-Sforza, for example, he gives the impression (or at least did in the 90s) that the PCs are something “real”. So maybe the confussion begins with geneticists themselves. 
     
    yes. he uses his “PC-maps” to discern varous waves of advance into europe. there’s obviously some debate about this. i suspect his PC map would have shown something similar for sub-saharan african genes in the middle east, but he didn’t have the density for extra-european caucasoids. obviously we can interpret that was the recent influx of sub-saharan genes through the slave trade.

  37. “what I am saying is that, since almost 6,000 years ago, Semitic peoples have been pouring into the lowlands of West Asia (Lebanon too, ok), in succesive waves. But these waves had almost no impact in the highlands (Anatolia, Iran, Caucasus).” 
     
    How does that explain the Jews closer relations to the Anatolians and Armenians? Are you positing that despite the relatively larger sizes of the late antique Babylonian and Alexandrian diasporas, the majority of Jews today descend from the Anatolian diaspora with a heavy admixture from the locals?

  38. How does that explain the Jews closer relations to the Anatolians and Armenians? Are you positing that despite the relatively larger sizes of the late antique Babylonian and Alexandrian diasporas, the majority of Jews today descend from the Anatolian diaspora with a heavy admixture from the locals? 
     
    I don’t know for sure, of course. But wasn’t the Babylonian diaspora cause of the Samaritans distinct group, who were shunned upon by “purebreed” Jews? I guess these are not anymore among Jews, at least in significant numbers, but mutated into other ethinicities, Palestinians specially. As for the ones who remained in Mesopotamia, they are at the origin of the legend of the “lost tribes”, right? Nothing to do with real Jews we know of. 
     
    I am unaware of any Alexandrian diaspora, unless you mean the natural dispersion of Jews in the Hellenistic (and later Roman) world, for mostly economic reasons. In this case, they went specially (not only) to Anatolia, and that could explain their apparent affinity with Anatolian-Aegean peoples, if they mixed with the locals there.  
     
    Of course, another possibility to be entertained is that Hebrews would not be really Semitic in (genetic) origin but some Anatolian/Kurdish/Armenian group that became semitized prior to their migration to Canaan. But this is highly speculative and would imply that Jews would have remained hyper-pure through milennia, even before the first diaspora, even before their presumed stance in Egypt (as part of the Hicsos?), even before mythological Abraham departed from Ur, what I think is extremely unlikely, even for such an endogamic people. Their culture and mythology is very Semitic (pastoralist-nomadic, patriarchal) in any case. 
     
    Part of the problem may be in the fact that these studies focus in Askenazis (a group with a marked bottleneck – founder effect, who did part from Anatolia, via Italy, according to all I’ve read). Maybe studies on Sephardites could shed more light. Real Sephardites (not all Jews of Spanish rite, that also include many Amazigh Jews, who are largerly Berber genetically) could be closer to the historical Hebrews maybe – though I guess they are not lacking admixture anyhow.

  39. I wonder the impact of the 1st century Jewish kingdom of Adiabene (in modern-day Kurdistan) on Jewish demographics.

  40. also, this focus on the european-non-european difference seems a little weird. in 900 most greek speakers probably were resident outside of europe, in anatolia. just because there’s a line on the map doesn’t mean that genetic variation is going to track that well. in the 1920s there are a mass transfer of anatolian “greeks” (some of whom spoke turkish but were orthodox christians) to greece itself. 
     
    Are you aware of any studies on genetic differences between Greek mainlanders and descendants of Anatolian Greeks?

  41. I wonder the impact of the 1st century Jewish kingdom of Adiabene (in modern-day Kurdistan) on Jewish demographics. 
     
    I was totally ignorant of this historical state but is another example of the many different people that converted to Judaism in old times. On that light, assuming that modern-day Jews are genetically closer to old-day ones than, say, Palestinians, is assuming a lot. It’s not just the European inputs, but the likely input before they even arrived to Europe.  
     
    Btw, if my memory doesn’t fail, there was an “old” Y-DNA study on Jews that showed some apparent affinities specifically with Kurds. Probably it’s very obsolete but still maybe worth mentioning.

  42. The kingdom of Adiabene was an Assyrian client state of Parthian Persia (i.e. not Kurdish): http://en.wikipedia.org/wiki/Adiabene.

  43. Btw, if my memory doesn’t fail, there was an “old” Y-DNA study on Jews that showed some apparent affinities specifically with Kurds. Probably it’s very obsolete but still maybe worth mentioning. 
     
    i think it’s cuz they’re haplogroup J2.

a