Writing about something I haven’t covered at Whitmanic yet: the intersection of population genetics and history. If you know me, you know this is one of my favorite niche passions—and what a time to be passionate about it! We’re already at the point when pretty much all cultural and linguistic mysteries with population-level implications can be resolved with the help of genomic tools.
As my friend Razib Khan has lamented, despite the boom in published research and cumulative knowledge in this exciting, interdisciplinary space, very few people have joined him in blogging its discoveries.
Sometimes this means that an amateur like me will pick up on connections that no one in either a) the research field or b) semi-popular writing has pointed out. In these occasional cases, I’m going to do my best to communicate what I notice.
It happened a few months ago in the realm of Ashkenazi Jewish population genetics—one of those topics that’s been done to death, and yet still almost never done right.
I won’t rehash discourse or facts, but one new consensus opinion among people familiar with this field is that Ashkenazi Jews do not genetically differ along geographical lines. That is, as groups, German Jews, Polish Jews, Lithuanian Jews, Ukrainian Jews, etc., are not genetically distinguishable from one another.
This somewhat defies intuition: Ashkenazi Jews have been migrating out of their urheimat in Germany for at least 700-800 years, and there’s reason to believe that Eastern European Jews have other, minor sources of ancestry—both Jewish and non-Jewish.
But in fact, there’s good evidence from both Behar et al 2013 (“No evidence from genome-wide data of a Khazar origin of Ashkenazi Jews”) and samples provided by commenters on Anthrogenica that suggests that German Jews are on average distinct from Eastern European Jews—notable for their relative lack of Eastern European ancestry, which causes them to fall somewhere between European Sephardim and Italian Jews on one hand, and Eastern Ashkenazim on the other.
More research is needed to shore this up, but no one should be surprised.
What I haven’t heard people suggest, though, is the possibility of population structure among Eastern Ashkenazim. But a few months ago, I found it.
Let’s back up for context. In March, Ancestry rolled out “Genetic Communities”, a feature that makes use of information about DNA testers’ geographical origins and assigns them to ethno-geographical clusters of distantly related people. To my surprise, unlike component-based features employed by other DNA testing companies, “Genetic Communities” wasn’t totally uninformative for Ashkenazi Jews.
In fact, it revealed a hidden population structure that matches known patterns in Eastern European Jewish cultural and linguistic geography.
People familiar with Yiddish culture have surely heard yarns about “Litvaks” and “Galitzianers”—two subgroups of Eastern European Jewry set apart by geography, cuisine, dialect, and supposedly, temperament.
“Litvaks”, you might have guessed, are Lithuanian Jews—a well-defined group with a well-attested Yiddish dialect, with roots in Greater Lithuania, a territory encompassing modern Lithuania, Latvia, Belarus, and the northeastern corner of Poland.
“Galitzianers” are Jews from Galicia, the portion of southeastern Poland and western Ukraine ruled up until World War I by the Austro-Hungarian Empire. However, there’s no such thing as “Galitzianer Yiddish”, nor was Galicia a sovereign, well-defined cultural-political realm like Greater Lithuania. Instead, Galitizianers spoke either “Polish Yiddish”, a dialect common to Jews from most parts of Poland (save for its northeastern corner) or its very close cousin “Ukrainian Yiddish”, spoken in Western Ukraine, Moldova, and Romania.
Once we’ve dispensed with the mythology of Galicia (the part of Poland-Ukraine that happened to have the highest concentration of Jews in the early modern period), we can see that the main cultural divide among Eastern European Jews actually fell along a well-known non-Jewish boundary—the internal border within the Polish-Lithuanian Commonwealth between Greater Poland (which included core Ukraine) and Greater Lithuania (which included all of Belarus).
Back to genetics. In Ancestry’s published research, it identified three very closely related, but distinct, clusters of Ashkenazi Jews. One was much smaller than the other two, and exhibited no particular geographical pattern. As for the two major clusters, take a look for yourself.
Ancestry calls this one “Jews in the Russian Empire”:
And this one’s “Jews in Central Europe”:
The fact that each is divided into 3 subgroups is beside the point for now. What’s notable is that the distribution of the first cluster corresponds almost perfectly to “Litvak territory” i.e. Greater Lithuania, whereas the second cluster comprises all Eastern Ashkenazi territories other than Greater Lithuania—including the former Austro-Hungarian territories of Czech Republic, Slovakia, Hungary, and Austria, as well as Greater Romania (Romania and Moldova, whose Jewish populations, I recently learned, were mostly supplied by 18th-19th century migrations from Galicia and the Ukrainian region of Podolia).
So population structure among Eastern Ashkenazim (perhaps among all Ashkenazim, given the assignment of German Jews to the second cluster) falls along the lines of Litvak vs. non-Litvak. For a better illustration of the correspondence, and the cohesiveness of the Litvak concept, I refer you to two maps from Dovid Katz’s exceptional Atlas of Northeastern Yiddish.
The first is of Litvak cultural territory, which, you’ll notice, includes the northeastern corner of Poland and some border regions of northern Ukraine.
The second is of Litvak linguistic territory, which, you’ll notice, adds to the picture a swath running all the way down to the Black Sea through Ukraine’s midsection, supposedly the result of late settlement patterns.
Having studied those maps, let’s revisit Ancestry’s map of Litvak genetic territory:
The match is uncanny. The only real difference is that the “Litvak cluster” seems to spill over to include most of northeastern and north-central Poland (or rather, almost the entire north of historical Poland), as well as the entirety of northern and central Ukraine.
But: if you’ll take a look once again at the map of the non-Litvak cluster, you’ll notice that it also covers these very areas—which would suggest an overlap of the two population networks in these two regions. In the case of northeastern Poland in particular, this is interesting—the “Litvak cluster” predominance extends much farther west than someone would’ve guessed based on cultural and linguistic data. Perhaps this points to an origin of the Litvak population in northern Poland and an expansion into Greater Lithuania (in addition, all three Litvak sub-clusters extend into northeastern Poland).
As for historical explanations, I’m not familiar enough with the literature on Yiddish dialectology or medieval Ashkenazi migration patterns to offer a hypothesis. All that seems clear is that Litvaks (defined broadly), among all Central/Eastern Europeans Jews, have a distinct population history. I’d be curious to know if anyone can offer support for this.
Note: I would guess that the majority of present-day Ashkenazi Jews in America, Israel, and other parts have ancestors from multiple Eastern European Jewish cultural sub-regions. This means that individual-level PCA-based support for this cluster-driven analysis might be impossible.