The Khazarian Hypothesis and the Nature of Yiddish

[Thanks to Martin W. Lewis, Dave Howard, Dmitry Pruss, Bill Poser, Lev Stesin, David Erschler, Sophia Osminkin, and Igor Solunskiy for a helpful discussion that led to this post!]

The Khazarian hypothesis, namely that Ashkenazi Jewry derives from the Khazars, has recently been revived by Eran Elhaik, a geneticist at John Hopkins University. His article “The Missing Link of Jewish European Ancestry: Contrasting the Rhineland and the Khazarian Hypotheses” appeared in December 2012 in Genome Biology and Evolution and was widely publicized before the actual publication. The two hypotheses compared in this article depict Eastern European Jews as a group that either “emerged from a small group of German Jews who migrated eastward and expanded rapidly” (Rhineland hypothesis) or “descended from the Khazars, an amalgam of Turkic clans” (Khazarian hypothesis). According to the abstract, Elhaik “applied a wide range of population genetic analyses” and found evidence to support the Khazarian hypothesis.

The Khazarian hypothesis is most strongly associated with the noted Hungarian-Jewish polymath Arthur Koestler, who advanced the hypothesis in The Thirteenth Tribe, published in 1976. The Khazars were a nomadic tribal confederacy that established a strong, trade-based state centered in the lower Volga River during the early Middle Ages. Their elite stratum definitely converted to Judaism, although it is unknown how deeply the religion penetrated into the rest of their diverse society. Although Koestler’s arguments were intriguing, solid evidence in support has been lacking, and most recent genetic studies have indicated that the Khazarian contribution to European Jewry was minimal at best. Until the publication of Elhaik’s work, the Khazarian thesis was all but defunct.

Razib Khan wrote an eminently sensible critique of the genetic, historical, and geographical errors in Elhaik’s article, focusing on his unsubstantiated use of Armenians as a genetic proxy for the Khazars. The genetic affinity between Jews—and not only those of Eastern European ancestry, but other Jewish groups as well—and Armenians has been noted at least as far back as 2000, when Nebel et al. found a connection between Jewish groups worldwide and other peoples living in the north of the Fertile Crescent, such as Kurds, Turks, and Armenians. Elhaik’s use of Armenians as the proxy for Khazarian DNA which reveals a profound lack of understanding of historical geography. Razib Khan writes:

“If you look at the modern state of Armenia this is eminently reasonable. But for most of its history Armenia was a marginal Caucasian nation, with its center of gravity further south, straddling Anatolia and western Iran, and looming over the plains of Mesopotamia. The Caucasian nature of modern Armenia is to a great extent a function of the extermination of Armenians from much of eastern Anatolia in the early 20th century.”

Thus, if one wants to be historically accurate, one should treat Armenians as a northern Fertile Crescent group, not a Caucasian one. But equally important, the Khazars were not exactly a Caucasian nation either: their capital was in the Volga delta, and most of their empire was in the steppe zone, as the map posted on the left shows. As Khan writes, and I second, the real “smoking gun” would be genetic links with East Asian populations: as the Khazars were Turkic, “they would have had substantial proportions of East Asian ancestry”, but no East Asian traces in the Jewish gene pool have been reported by Elhaik or anyone else, as far as I know.

Elhaik commits many other errors of historical geography. In the abstract alone, he claims that the Khazars “converted to Judaism… in the 8th century”, whereas it was mostly their aristocracy that converted. He takes the conjecture that “following the collapse of their empire, the Judeo-Khazars fled to Eastern Europe” as a proven fact. He talks of “Caucasus, European, and Semitic ancestries” as if these were commensurate categories, and not a mountain range, a sub-continent, and a language family. He also proposes to explain major difference among populations of the Caucasus by the early presence of Judeans in the Southern and Central Caucasus, while in actuality a huge array of genetic diversity exists in the Caucasus regardless of Jewish population.

But problems with Elhaik’s paper are not limited to those of historical geography. In the discussion section, he ties his flawed findings to even worse linguistic theories, specifically to the idea that Yiddish, the language of Central and Eastern European Jews, “began as a Slavic language that was re-lexified to High German at an early date”. The preponderance of Germanic words in Yiddish is clear from the page from the Shemot Devarim, a Yiddish-Hebrew-Latin-German dictionary and thesaurus, published by Elia Levita in 1542 and reproduced on the left. The theory that Yiddish is a Slavic language was originally proposed by Paul Wexler in his 1993 book The Ashkenazic Jews: A Slavo-Turkic People in Search of a Jewish Identity, a sequel of sorts to his 1990 book in which he argues that Modern Hebrew too is a Slavic language, re-lexified to Biblical Hebrew. Considering Wexler’s book on Hebrew would take us too far afield, but what about this idea that Yiddish is really a Slavic language?

First, it seems odd that a Turkic people who moved to Europe via Romania and Hungary would somehow acquire a Slavic language on the way. No traces of Turkic, Romance, or Uralic linguistic influences have ever been identified in Yiddish. Second, no clear cases have been described in which a language changes most or all of its vocabulary drawing on that of another language. If such a transformation were to happen, the old form of the language and the new one would be mutually incomprehensible, so it is hardly likely that either its speakers or linguists would use the same label for it, as if it were the same language. Many instances of language shift have been attested, but they are always described as a shift from one language to another, not a switch from one form of a language to a mutually incomprehensible one.

But most crucially, Yiddish is not only non-Slavic in its vocabulary, but also in its grammar. This is particularly obvious if we compare Yiddish to Sorbian: the former is a Germanic language with some Slavic influences, whereas the latter is a Slavic language with Germanic influences. Let’s consider Sorbian first, which makes a fitting comparison for Yiddish because both modern Sorbian varieties—Upper and Lower Sorbian—are surrounded by German-speaking territory. Neither Sorbian-speaking group has an autonomous region of its own:

“Upper Sorbs live mostly in southern Lusatia in Saxony and Lower Sorbs in the Niederlausitz region; they have long been an ethnic minority in Germany. Sorbian monolingualism ceased before World War II; nearly all Upper Sorbian speakers are fully bilingual in German and use primarily German; Lower Sorbian is serious endangered. Both languages show significant impact of contact with German, change which has been almost entirely unidirectional.” (Lenore A. Grenoble, “Contact and the Development of the Slavic Languages”, p. 588)

One example of German grammatical influence on Sorbian is the merger of the two uses of the instrumental case. Grenoble describes “the pan-Slavic pattern” as making “a distinction between the instrumental or absolute use of the instrumental case, which does not occur with a preposition, and the instrumental of accompaniment, which is used with the preposition ‘with’” (ibid). For example, compare the Russian Ja rabotaju golovoj (literally, ‘I work head.INSTR’) with Ja rabotaju s Martinom (literally, ‘I work with Martin.INSTR’). In Upper Sorbian, however, “these two usages are collapsed and are found only with the preposition z ‘with’” (ibid). Compare the Upper Sorbian Ja dźělam z ruku ‘I work with my hand’ (instrument) and Ja rěčг z prěćelom ‘I speak with my friend’ (accompaniment).

But while this example illustrates German grammatical influence on Sorbian, it also underscores the underlying Slavic grammar of the language: like most other Slavic tongues, Sorbian maintains a rich system of seven cases (including the vocative) and marks them on the nouns themselves. In comparison, Yiddish has only four cases and marks them on the determiners such as the article ‘the’. Furthermore, Sorbian adjectives decline according to the same seven cases as nouns; when an adjective is used to modify a noun, they must agree in case. Yiddish, in contrast, do not mark cases on adjectives; instead, there are different forms are used for adjectives in attributive (i.e. modifying) and predicative positions. For example, the form of the adjective ‘good’ in Yiddish is different in ‘The good man left’ and ‘The man is good’—much like in German or Norwegian, and as it used to be in Ye Olde English period. Sorbian is also noted for its retention of the dual number, a special form used in reference to two objects, as opposed to one (singular) or three or more (plural). Common Slavic had a dual number and most modern Slavic languages retain some vestiges of it. In modern Russian, for instance, the former dual is reflected in the forms of some modern plurals in which the stem-final velars k and x palatalized into č and š (e.g. Russian oči ‘eyes’ and uši ‘ears’). The older dual is also to blame for the fact that a different form of a noun is used with ‘two’ (extended also to numerals ‘three’ and ‘four’), as opposed to ‘five’ and up: for example, one says dva mal’čika (literally ‘two boy’ in the genitive singular), but pjat’ mal’čikov (literally ‘five boys’ in the genitive plural). In modern Sorbian, two hands is ruce, but three or more hands is ruki. Yiddish, in contrast, has no dual, and neither do other Germanic languages.

When it comes to syntax, Sorbian and Yiddish differ as to the order of major sentence constituents: Yiddish exhibits the Verb-Second phenomenon: the finite verb—auxiliary if one is present, and the lexical verb otherwise—must appear in the second position in a main clause. For example, in the sentence Oyfn veg vet dos yingl zen a kats (literally ‘On the way will the boy see a cat’), the auxiliary vet ‘will’ appears right after the prepositional phrase oyfn veg ‘on the way’ rather than after the subject dos yingl ‘the boy’, unlike in the corresponding English sentence On the way, the boy will see a cat, where both ‘on the way’ and the subject must precede the verb. In fact, English is the only Germanic language that lacks the Verb-Second (except in some marginal sentence types). As illustrated below, Dutch, German, Swedish all exhibit the Verb-Second pattern (as does Icelandic):

a. Dutch:

Gisteren          las       ik         dit        boek.

yesterday         read     I           the       book

b. German:

Gestern           las       ich       dieses Buch.

yesterday         read     I           the       book

c. Swedish:

Igår                 läste    jag       denna bok.

yesterday         read     I           the       book

In contrast to Yiddish and other Germanic languages, Sorbian does not have the Verb-Second pattern. In sum, the grammatical profiles of Yiddish and Sorbian are completely different, with Sorbian being a bona fide Slavic language and Yiddish a well-behaved Germanic one.

This is not to say that Yiddish has not had any grammatical influences from Slavic languages. Along with numerous borrowed nouns, especially terms of a culturally or geographically specific nature such as plant and animal names, foods, etc., Yiddish picked up a few morphological and syntactic patterns from its Slavic neighbors. For instance, Yiddish has borrowed a number of derivational morphemes, such as the agentive -nik (as in nudnik ‘bore’ from nudne ‘boring’) and the diminutives -tshik and -ke. An example of Slavic syntactic influence on Yiddish comes from the domain of multiple interrogatives. Note that in English—as in other Germanic languages—only one question word can appear in the beginning: Who bought what? but not *Who what bought? In Slavic languages all question words must appear sentence-initially, as in the Russian Kto čto kupil? (literally ‘Who what bought?’). Interestingly, multiple questions in Yiddish reveal both its Germanic roots and its contact with Slavic languages. In accordance with the Germanic pattern, one can place only one question word sentence-initially (in which case the other question word must appear before the non-finite lexical verb, if there is one), as in Ver hot vos gekoyft? (literally, ‘who has what bought?’). Alternatively, one can employ the Slavic pattern and place all question words sentence-initially, as in Ver vos hot gekoyft? (literally, ‘Who what has bought?’). Yet, despite those Slavic influences, Yiddish retains the overall grammatical profile of a Germanic language, while Sorbian—despite centuries of German influence—retains its Slavic grammar. In this respect, the two languages could not be more different.

Sources:

Grenoble, Lenore A. (2010) “Contact and the Development of the Slavic Languages”. In: Raymond Hickey (ed.) The Handbook of Language Contact. Wiley Blackwell. Pp. 581-597.

Koestler, Arthur (1976) The Thirteenth Tribe: The Khazar Empire and its Heritage. NY: Random House.

Nebel A, Filon D, Weiss D, Weale M, Faerman M, Oppenheim A, Thomas M (2000) “High-resolution Y chromosome haplotypes of Israeli and Palestinian Arabs reveal geographic substructure and substantial overlap with haplotypes of Jews”. Hum Genet 107:630–641.

Wexler, Paul (1990) The schizoid nature of modern Hebrew: A Slavic language in search of a Semitic past (Mediterranean language and culture monograph series). Wiesbaden: Otto Harrassowitz.

Wexler, Paul (1993) The Ashkenazic Jews: A Slavo-Turkic People in Search of a Jewish Identity. Colombus, OH: Slavica.

GeoCurrents.info provides map-illustrated analyses of current events, delving into the often overlooked geographic and cultural contexts that shape these world developments. Led by Stanford University Senior Lecturer Martin W. Lewis and linguist Asya Pereltsvaig.