Jews and the Language of Eastern Slavs

THE BEGINNINGS OF JEWISH PRESENCE in Eastern Europe are among the most enigmatic and underexplored pages in the history of the region. The dating and localization of Jewish presence, as well as the origin and cultural characteristics of the Jewish population residing among the Slavs in the Middle Ages, are among the issues which have become a subject of tense discussion and widely diverging evaluations, often connected to extra-academic ideological agendas. The question of the spoken language of the Jews inhabiting Slavic lands during the Middle Ages is unresolved. Did all or most of these Jews speak languages of their former lands (such as German, Turkic, or Greek), did they speak local Slavic vernaculars? Or did they, perhaps, speak some Judeo-Slavic vernacular(s), which later became extinct? If the last two suggestions appear plausible, then was the Jews’ experience limited to oral usage, or can we also expect to find written evidence of Slavic literacy among medieval Jews? What impact could such literacy on the part of the Jews have had on the literary production and intellectual horizons of Slavs? Or could it even have impacted East Slavic cultural contacts with Latin Europe, in which Slavic-literate Jews may have been involved?


INTRODUCTION
THE BEGINNINGS OF JEWISH PRESENCE in Eastern Europe are among the most enigmatic and underexplored pages in the history of the region. The dating and localization of Jewish presence, as well as the origin and cultural characteristics of the Jewish population residing among the Slavs in the Middle Ages, are among the issues which have become a subject of tense discussion and widely diverging evaluations, often connected to extra-academic ideological agendas. The question of the spoken language of the Jews inhabiting Slavic lands during the Middle Ages is unresolved. Did all or most of these Jews speak languages of their former lands (such as German, Turkic, or Greek), did they speak local Slavic vernaculars? Or did they, perhaps, speak some Judeo-Slavic vernacular(s), which later became extinct? If the last two suggestions appear plausible, then was the Jews' experience limited to oral usage, or can we also expect to find written evidence of Slavic literacy among medieval Jews? What impact could such literacy on the part of the Jews have had on the literary production and intellectual horizons of Slavs? Or could it even have impacted East Slavic cultural contacts with Latin Europe, in which Slavic-literate Jews may have been involved?
The research leading to these results has received funding from the European Research Council under the European Union's Seventh Framework Programme (FP7/2007(FP7/ -2013  The significance of these questions reaches far beyond the concerns of one or two particular disciplines. As we will see, certain Slavic forms reconstructed from their Hebrew contexts may provide the earliest attestation of important linguistic data relevant for Slavic historical linguistics. An attempt to reconstruct certain isolated elements of a hypothetical Judeo-Slavic vernacular may be of interest for the study of Jewish languages. Seen in historical perspective, the existence of an East Slavicspeaking Jewry may provide an additional argument in favor of the existence of pre-Ashkenazic Jewish communities in the region. The problems I raise here are also significant for an understanding of cultural processes, where Jews and Slavs may have shared a grounding not only in terms of geographic territory, legal administration, interethnic politics and interconfessional polemics; they may have also literally had a language in common. How else can we explain the unique phenomenon of early Hebrew-Slavic translations? What was their Sitz im Leben and target audience? Should we extrapolate Jewish cultural isolationism as this is known from the late medieval and modern periods to the earlier Middle Ages, or was the situation different then? Should the early history of Eastern European Jewry be regarded as a sequel to the German Jewish story, or as an independent Slavic Jewish narrative? All these wider historical questions must be reexamined as soon as we clarify the linguistic situation of Eastern European Jewry in the Middle Ages. SLAVIC THEORY The idea that the Jewish population of early Eastern Europe 1 was Slavicspeaking before it was assimilated by German-speaking Ashkenazic migrants was first put forth in 1865 by Harkavi. 2 The concept became popular with some Jewish scholars, especially those who, like Harkavi himself, 3 were interested in proving Jewish autochthonism in Russia and Poland. 4 Dubnow, a proponent of Jewish cultural autonomy, argued against the view propounded by Harkavi,5 but the discussion dealt in fact only with the period from the sixteenth century on, for which time span Dubnow succeeded in showing that Yiddish was the main spoken language if not for all, then at least for most East European Jews. And even for this late period, Dubnow admitted a small percentage of Slavicspeakers among the Jews of Eastern Europe. His debate with Harkavi was really about relative numbers. What is more important is that it focused on the factors responsible for the use of Slavic among the Jews: arguing whether the phenomenon was a result of late assimilation or early legacy. 6 Harkavi's so-called Slavic theory 7 did not receive anything approaching proper or consistent treatment in linguistic research. Harkavi was not a Slavist and did not pinpoint the dialectal identity of diverse pieces of Slavic material in his research. Much later, the prominent linguists Jakobson and Halle promised a joint monograph on this topic, to which they attributed a great deal of significance, but in the end published only fragmentary studies. 8 An important contribution to the subject was made by students of Yiddish and other Jewish languages. Thus Weinreich was the first to propose a systematic formulation not only of the (Warsaw, 1920) 6. On this, see Weinreich, History, 88. 7. This was also known as the ''Canaanic theory.'' ''Canaan,'' associated in its original biblical context with slavery (see Gen 9.25), became a regular denomination for the Slavic lands and languages in the Hebrew usage of the early Middle Ages, at the same time when the term sclavus became a replacement for servus in Latin Europe (while in the Arab world, the word s . aqlab [pl. s . aqaliba] became widespread for ''slave''), thus reflecting early medieval realia-the quantity of Slav slaves on the Mediterranean markets; see Roman Jakobson and Morris Halle, ''The Term Canaan in Medieval Hebrew,'' in For Max Weinreich on his Seventieth Birthday (The Hague, 1964), 147-72, reprinted in Roman Jakobson, Selected Writings (The Hague, 1971, 6:858-86 (the further references are to the latter edition).
8. This monograph in preparation was announced in Jakobson and Halle, ''Term,'' 886. Cf. two more publications by Jakobson: ''Ř eč a písemnictví český ch židů v době přemyslovské,'' in Kulturní sborník ROK, ed. L. Matějka (New York, 1957), [35][36][37][38][39][40][41][42][43][44][45][46]; and, ''Iz razyskanij nad staročešskimi glossami v srednevekovyh evrejskih pamjatnikah, '' Slavica Hierosolymitana 7 (1985): 45-46. For more literature, see below, section 1.4. idea of the existence of a Slavic-speaking Jewry but also of the hypothesis of a particular Judeo-Slavic language. This idea was later developed by Wexler. 9 WEST VS. EAST SLAVIC The most basic problem hindering the development of this field, and one which I would like to address here, was a failure to differentiate among Slavic materials of different provenance. This involves, first of all, demarcating West and East Slavic data. The use of Western Slavic languages among Jews in Western Slavia and Eastern Germany is more fully documented than its counterpart for the East Slavic. 10 The differentiation between Western and Eastern Slavic materials is especially needed if we do not simply accept the Germano-centric ''Western theory'' which assumes that the Jewish population of the East Slavic lands is a result of German Jewish colonization caused by persecution or as part of the more general phenomenon of German Ostsiedlung (settlement of the East) via West Slavic lands. 11 In this framework, all East Slavic features should be dismissed as either late or nonexistent; as we will see, such an approach often contradicts the evidence. 12 However, if we do assume a Jewish presence in East Slavic territories in the pre-Lithuanian period (tenth to thirteenth centuries), before any attested migration from the West took place, indications of specific cultural characteristics of the East Slavic Jewish population become especially instructive. More on this below.
Jewish Slavic materials may range from historians' linguistic unawareness or their uncritical admission of the ''Western theory'' to a lack of sufficient linguistic grounds for such a differentiation in the extant sources. Thus, the data from Yiddish as a rule support neither chronological nor geographical identification of presumed early Slavicisms. This is due to the fact that in extant Yiddish we only encounter the result of the fusion of diverse linguistic components attested in the sources which date from centuries later than the period in question. General historical considerations usually prompted researchers to consider all early Slavic data of unclear provenance as Western, while specifically East Slavic forms were hardly ever identified. 13 Following the same pattern, Slavic glosses in medieval Hebrew literature, which incontestably belong to the early stratum of the material, were all declared by Jakobson and Halle to be Western, even specifically Czech. There was no detailed treatment of all the sources provided in support of this declaration. Jakobson's and Halle's presumed authority, as well as a dearth of Slavic linguists qualified to access medieval rabbinic texts, hindered the reexamination of this issue. In general, it should be noted that the problem was approached by historians and linguists separately; the lack of an integrative historicophilological view of the issue played its role. The topic under discussion seems the most appropriate object for a composite historico-philological approach, when linguistic data should be studied against a historical background, so that a more nuanced sense of the historical subject may be achieved by grounding it in a linguistic analysis.
Hence, an attempt at historical contextualization combined with a reassessment of early linguistic data is a long overdue, as well as the only possible way to deal with the problem. For this purpose, I suggest examining separately all the available evidence, historical and linguistic alike, which pertains to the East Slavic (''Old Russian'') language used by Jews in the pre-Lithuanian and early Lithuanian period (tenth to thirteenth centuries). 14 JEWS IN MEDIEVAL EASTERN SLAVIA Beyond linguistic data, what do we know of Jews in the East Slavic lands (Rus')? Most of the evidence indicating the existence of a Jewish population in Eastern Europe prior to the mass migration from Ashkenaz 13. See Wexler,Explorations,155, 14. Periodization of this kind seems more appropriate for our discussion than the more conventional approach focused on the Mongol conquest, considering that almost no traces remain of Jewish presence in Eastern Rus' under the Mongols. originates from territories inhabited by Eastern Slavs, specifically, from the southwestern principalities of Rus'. Here a Jewish presence is attested as early as the tenth century and must go back to the Jewish settlement in the cities of western Khazaria (such as possibly Samvatas-Kiev and Tmutorokan') that is in the territory of Rus' prior to its political formation. Beginning with the Mongol Conquest in the 1240s, evidence of the presence of Jews in Rus' becomes restricted to the territory of Galicia-Volhynia, which suffered less than other areas from the invasion. 15 The Galician-Volhynian lands were annexed to the Grand Duchy of Lithuania and Poland during the thirteenth and fourteenth centuries and subsequently became an integral part of the Polish-Lithuanian Commonwealth. By the mid-fourteenth century, the age of the separate existence of the Jews of Rus' had come to an end, at least politically, and possibly also culturally; a process of their acculturation began among the bearers of Ashkenazic culture who were arriving in Poland and Lithuania from the West. Thus, long before the divisions of Poland in the eighteenth century and the beginning of a new era in the history of the descendents of the Jews of Rus', there was in the region a Jewish population with a uniform Ashkenazic culture, which retained barely any trace of the unique tradition of its pre-Ashkenazic ancestors.
Radical evaluations are not uncommon in the research done on this pre-Ashkenazic Jewish population of early Rus'. They range from hypercritical attempts to deny the fact of such a population's existence 16 to unjustified exaggerations of its size and importance in the later formation of Eastern European Jewry. 17 None of these radical opinions is corroborated by the sources in our possession. As I have tried to show in my earlier work, the scant sources available to us, even though they do not enable us so much as to estimate the size of this community, are characterized by a high ''representativeness'' of the evidence-they pertain to nearly all aspects of Jewish life.  (Edmonton, 1988), 3-21. reflected in accounts of wide-ranging and balanced dissemination, diverse occupations that are characteristic of the period, citations of communal structure, institutions, and functions, as well as certain types of cultural activity and contacts with the local gentile environment. The continuity of the evidence of Jewish life in this territory, over which sovereignty changed, and the lack of data on persecution or economic distress, may also be indicative of Jewish settlement continuity in the region. This is the case for the entire region at least until the Mongol conquest, and, in the Western principalities annexed by Lithuania, after the conquest as well. 18 The question arises, what language or languages could these Jews master? This question is also connected to the problem of the origins of these communities and the languages of their ''old countries.'' We know nothing about German in use by Jews in these territories during this period and have only fragmental and vague indications of their knowledge of Greek and Turkic. 19 The evidence of the Jews' knowledge of East Slavic is, by contrast, incomparably richer.

East Slavic-Speaking Jews in the Middle Ages
The evidence on medieval Jews mastering, to some extent or another, East Slavic speech can be divided into (a) scant direct pieces of evidence explicitly referring to Rus' or its cities and showing cases of Slavic Jewish onomastics in Rus' (from the tenth century on), Jewish East Slavic monolingualism (eleventh century), and Jews mastering East Slavic obscene speech (thirteenth century) (sections 1.1-3 below), and (b) more abundant, but less distinct historically, the linguistic data of the Slavic glosses in Hebrew literature from the same period (eleventh-thirteenth centuries) (section 1.4).

East Slavic Jewish Personal Names
Slavic Jewish onomastics are abundantly represented in Lithuanian Rus' beginning in the late Middle Ages. For 1486, we have Lithuanian Jews named Glukhoj, Kravchik, Momotlivyj, Riabchik, Samotyka, Zubets, and so on;for 1566, Borodavka, Broda, Brova, Kislo, Kon', Koza, Kozak, Lisa, Prorok,  Shpak, Sirota, and others. 20 Some of these forms can only be East Slavic. There is also abundant evidence dating from the modern period, especially of female onomastics. 21 For earlier times, however, all we have is limited to a couple of names of the elders of the Kievan community, which appear in the ''Kievan Letter'' of the tenth century. 22 Since the names are given in Hebrew consonant writing, their vowel pronunciation remains conjectural, allowing for different interpretations. All in all, interpretatio slavica for two of these names seems more convincing than other options. I mean Gostiata (Гoст тa) for ' 'Gwst . t . ? , son of Kybr the priest'' (affswg whk rbyk rb; line 26) 23 [Paris, 1989] a waw to a yod. 27 Even so, the emendation is quite plausible, and the form itself is a well attested Jewish surname since 1566 to our days (see above).

Jewish East Slavic Monolingualism
One of the earliest attestations of Slavic-speaking Jews has been preserved in the Cairo Geniza. One of the miscellanies found there among a variety of materials, including formulae of bans and a colophon of a Bible copy owner, contains a copy of a circular letter of introduction. 28 Just as in the Kievan Letter, and following medieval Hebrew epistolary etiquette, the poetic introduction and conclusion of the letter are longer than its substantial part, 29 which reads: Ͻ. . .Ͼ Honorable, great and holy communities of the holy nation, the scattered remnant of Jeshurun. Ͻ. . .Ͼ In addition to sending you our greetings we have to apprise you of the case of Master Anon. son of Anon., who is from the community of Rus' and has been staying with us, the community of Salonica, ''the young in the flock,'' 30 and found his relative Rabbi Anon. coming from Jerusalem, the Holy City, let God establish it forever, sela. And his relative had a letter from our lord, his honorable and great holiness, our teacher 31 Anon. And he described him all the magnificence of the Land of Israel, so that his spirit impelled him also to go and to prostrate himself before the Holy Place. And he requested of us these two lines to be for him a mouth and an advocate before your honorable magnificence, so that you might give him a hand and guide him along the good and safe way from city to city and from [isle] to isle with trustworthy persons, because he knows neither the Holy Tongue nor the Greek, and not Arabic, either, Mann dates the writing to the eleventh century ''and probably even earlier.'' This paleographic assessment was challenged by Weinryb, who, citing Goitein's oral evaluation, stated that the time the letter was written ''may equally be in the 12 th -13 th century.'' 33 Although Jewish pilgrimage to the Holy Land is well attested both before the Crusades and during the period of Christian domination in Palestine, the absence of Romance languages (z[wl) in the list of languages required for the voyage to Jerusalem may indicate that the journey took place before the capture of Jerusalem by the Crusaders in 1099.
In 1047, Jewish pilgrims coming from the ''Byzantine realm'' were seen in Palestine by the Persian traveler Nā s . er-e Khosraw, as he witnesses in his Book of Travels. 34 In 1028-35, a ransomed Jewish captive from Attaleia made a pilgrimage from Alexandria to Palestine. 35 Salonica was also a major center of the messianic enthusiasm following the First Crusade in 1096 and involving attempts of mass Jewish migration to Palestine. 36 There is also a vague indication of ''Khazarian'' Jews possibly joining the movement (the text of the fragment from the letter describing the events of 1096 in Salonica is not fully clear: wklhw hyyrzkla μwqml ( 55. . 43 If our traveler was initially on a business trip, it is very probable that he visited Salonica between October 20 and the first Monday after St. Demetrius' Day (October 26), the regular dates of the fair. 44 In the eleventh century, merchants and pilgrims could still go to Byzantium by the Dnieper, and then along the eastern coast of the Black Sea through Varna and Messembria (modern Nesebar), following the route described in detail by Constantine Porphyrogenetus a century earlier in his On the Administration of the Empire, written circa the year 950: ''The monoxyla [dugouts] which come down from outer Rus' to Constantinople are from Novgorod Ͻ. . .Ͼ, and others from the city of Smolensk and from Teliutza and Chernigov and from Vyshegrad. All these come down the river Dnieper, and are collected together at the city of Kiev, also called Sambatas.'' 45 Our traveler moved, as the letter specifies, ''from island to island,'' and indeed, many islands were also regular stopping points for boats making their way from Kiev to Byzantium via the Dnieper according to Constantine: St. Gregory Island (Khortitsa) and St. Aitherios (Berezan') in the mouth of the Dnieper (ibid.). The phrase ''from island to island'' may also refer to the further section of the traveler's route-from Salonica to the Holy Land through the Greek islands (cf. Asher bar Sinai, a student from Rus' studying under the Rosh in Toledo, whose abrupt journey home can be traced through Mallorca and Sicily.) 46 Alternative routes, like the way from Halich through the Pruth river to the Danube estuary, are attested later. 47 It should be noted that a very similar circular letter was given (probably in Spain) to a French pilgrim from Rodez. This letter also refers to the traveler's language difficulties: 48 Ͻ. . .Ͼ and he came pleading and saying to us, ''I was expelled and wandered from my place to yours and . . . in a funny language, and It seems that on his way through Roman language lands, between France and Spain, the pilgrim from Rodez managed to communicate, albeit in a ''funny language' ' (ynwçl g[lb). Notice also that the expression ''be for me a mouth'' (hpl yl twyh) from Ex 4.16 is identical to the one in the Salonician letter above. This letter demonstrates that the Salonician case is not unique and may corroborate the suggestion that our document with omitted names could have served if not as a form (due to the overly specific details it contains), then at least as a model for multiple use. Both French and East Slavic Jewish pilgrims in these two documents are monolingual. They are described as the perfect opposite of the semi-legendary Jewish Radhanites who spoke ''Arabic, Persian, Roman, Frankish, Spanish, and Slavic.'' 50 Our traveler does not even know ''the Holy Tongue'' (at least as a language of oral communication). The problem of Jewish monolingualism was later addressed by the Spanish mystic Abraham Abulafia (1240 to after 1291) in his autobiography: Being dispersed among many nations and bizarre languages [twnwçlh μynwçmh], we forgot our language in its entity Ͻ. . .Ͼ and behold, the Jews dwelling among Ishmaelites speak Arabic like them, and those dwelling in the Greek land speak Greek, and in the lands of the Latins [i.e., ''Franks'' or ''Italians''-μyz[wlh twxra] they speak the Romance language [tyz [wl], and the Germans speak German, and the Turks speak Turkic, and so the rest of all of them. This testimony about the Jews speaking the language of their dwelling place need not necessarily be interpreted as testifying to their being monolingual, but in continuation Abulafia describes Jewish multilingualism in Sicily as ''the great wonder,'' an exception rather than the rule: But it is the great wonder what has happened to the Jews in all of Sicily, who speak not only Latin and Greek, like the Greeks and the Latins dwelling with them, but they have preserved the Arabic that they learned in former times, when the Ishmaelites were dwelling there. 51 Thus, our pilgrim's monolingualism was not an extraordinary private case but a predicament he shared with other members of his community. This may even follow from the plain meaning of our text: whereas Marmorstein's edition of the text has [μyr]bdm rça˜[nk tpç μa yk . . . [dwy wnya yk wtdlwm ≈ra yçna-''he knows neither . . . but only the Slavic language spoken by the people of his native land,'' Mann corrects the reading to μa yk wtdlwm ≈ra yçna [μyrb]dm˜[nk tpç-''but the people of his native land speak [only] the Slavic language.'' The discrepancy may be significant. While the first reading implies that Slavic is only one of the languages spoken (by Jews?) in Rus', the second may mean that it is the only or the main language spoken there.
Another interesting parallel to our text is an additional example of East Slavic monolingualism, which obviously was not unique to the Jews: the Annals of Burton report on archiepiscopus Ruthenius, nomine Petrus, who took part in the first Council of Lyons (1245), but neque Latinum neque Graecum neque Hebraicum novit linguam. 52 As we learn from what follows, Archbishop Peter had an interpreter with him. But, coming back to our traveler, we may well ask how could he have undertaken his voyage, especially if this was initially supposed to be a business trip, without having a knowledge of languages? It seems that in the eleventh century he could take advantage of a Slavic-speaking continuum stretching from Rus' to the Balkans with its vast Slavic-speaking population. Even in the region of Salonica (Slavic Solun' ), the native city of Cyril and Methodius, the founders of literary Slavonic, Slavs probably made up the majority of the population ever since the beginning of the seventh century, preserving their language until later than the date of our letter. 53 51. Neuman, Jews, 2:300. 52. Henry R. Luard, Annales de Burton (A.D. 1004-1263), Rolls Series 36.1 (London, 1864), 271.
53. As we can learn from the Miracles of St. Demetrius and later sources, the city virtually became a Byzantine island surrounded by large-scale Slavic settle-There were also precedents of non-Jewish merchants from Rus' who, when on a trip and, without knowing the local language, could expect to find interpreters among Slavic slaves in loco. Thus, Ibn Khordadbeh in the late ninth century writes about the Norman ''Rus'' merchants who brought goods from Northern Europe and Northwestern Rus' to Baghdad: ''[They] transport beaver hides, the pelts of the black fox and swords from the farthest reaches of the Slavic lands [saqaliba] Ͻ. . .Ͼ to Baghdad. Slavic [saqlab] slaves translate for them.'' 54 Thus, the omnipresence of Slavic slaves in medieval Europe as well as in the Muslim world may also have provided a factor favorable to the international operations undertaken by Slavic-speaking Jewish merchants.

Slavic Obscenity and ''Hebrew speech''
Isaac from Rus' appears in the work by R. Moses ben Isaac (ben ha-Nessiah) of England. R. Moses composed a grammatical study organized as an alphabetical listing of roots called Sefer ha-Shoham (The Book of the Onyx). 55 In the chapter devoted to verbal roots with an initial yod, for the entry μby we read: -''Yabam. Strong [verb]. ''Come to your brother's wife and yabem her'' [Gen 38.8]. R. Isaac of Sernegov told me that in the language of Tiras, which is Rus', sexual intercourse is called yibum. 56 [Thus], ''yabem her'' means ''have intercourse with her.'' 57 R. Isaac is known in the scholarly literature as ''Yitse of Chernigov.'' However, this form of his name does not appear in either of the two extant manuscripts of the composition. ''Yitse'' must be an erroneous reading of the common abbreviation 'jxy as this appears in the Oxford manuscript. The copy from St. Petersburg has it in the nonabbreviated form: sryt˜wçlb yk bwgnrsm qjxy 'r brh yl 'ma htwa μbyw˚yja tça la ab qzj μby htwa lw [ Evr. II A 34, fol 39v). Thus, a character by the name of ''Yitse of Chernigov,'' starring even in the titles of encyclopedic entries, 58 is one of the many erroneous inventions of medieval Judeo-Slavica.
The biblical ''Tiras'' of Gen 10.3 is also identified with Rus' in the Book of Josippon. 59 ''Sernegov'' (with occlusive g; on this see 2.2 below) is obviously Chernigov, one of the most powerful cities and principalities of Kievan Rus'. We should note that for a time the princes of Chernigov also ruled the principality of Tmutorokan, located on the coast of Taman  The gloss introduced by R. Isaac is the earliest known attestation of the most fundamental obscene item in either the medieval or the modern Slavic vocabulary. It belongs to the group of tabooed words which in the Middle Ages were closely associated with pagan practices, witchcraft, and non-Christian identities. Using such a word was considered inappropriate for a Christian; a comment to this effect appears as early as in the works of the famous Church writer Cyril of Turov (1130-82). Obscene speech was sometimes referred to as ''Jewish.'' Among the examples adduced by Uspenskij are ''Jewish speech'' (жидовское слово) and ''do not speak Jewish, do not swear in foul language'' (по жидовcкы не 58. Such as in Isaac Broydé, ''Itze (Isaac) of Chernigov,'' in Jewish Encyclopedia (New York, 1904) говорите, мaтерны не брaнитес ). 63 The term was probably used as a synonym for ''pagan'' (but never ''Hellenic,'' an incomparably more common designation for ''pagan'' in Church Slavonic literature and dependent on a similar turn of phrase in the New Testament). 64

East Slavic Glosses in Medieval Hebrew Literature 65
In addition to these pieces of ''direct'' evidence on Slavic-speaking Jews from early Eastern Europe, other important data are available, which have not been adequately introduced into the historical discussion.
Medieval Hebrew writings contain many glosses in the languages spoken by their authors, editors, and target audience. These glosses in Old French, German, Italian, Greek, and other languages were meant to clarify rare Hebrew and Aramaic forms. Very often the glosses were written by the original authors of the texts. They were considered such an integral part of the text that they continued to be copied (and later reprinted) consistently even when their readers no longer understood the languages in which they were written. Glosses in Slavic languages were for the first time identified as such by Landau, 66 assembled by Harkavi,67 and further supplemented by Grü nwald, 68  material irrelevant for our discussion. However, some of the glosses in question may rather be East Slavic, in other words, they may be based on the vernacular used in Kievan and later Lithuanian Rus'. This means that in order to solve the historical question of whom and from where were the Slavic-speaking Jews behind the glosses, we need first to undertake a philological analysis of the glosses themselves.
Only one of the glosses contains specific information on the place of origin of its author, and it turns out to be East Slavic. The author is Isaac of Chernigov, whose foray into comparative linguistics I have already noted (Sefer ha-shoham, thirteenth c.; section 1.3 above). In other cases, the identification of a gloss as made in the ''tongue of Canaan'' (˜[nk˜wçl) is not very helpful, since the Jews did not use different terms to distinguish among different Slavic languages. The same was typically true of the Slavs themselves. Thus, the Russian Primary  multas alias nationes tenet-''. . . the Slavic language, which is one of the languages that occupies many regions. It is spoken in Rus', Poland, Bohemia, and among many other nations.'' 71 Bacon's information may be based on the Itinerarium by William of Rubruck (ca. 1220-ca. 1293): Lingua Rutinorum et Polonorum et Boemorum et Sclavonorum eadem est cum lingua Wandalorum-''The language of the Russians, Poles, Bohemians, and Slavs is the same as that of the Vandals.'' 72 One more factor has to be taken into account: not only Slavs but also Jews from diverse Slavic areas could resort to Slavic as their common language of communication. This is demonstrated by the encounter of scholars and students of different Slavic origins in the rabbinic schools of Northern France and the Rhineland in the twelfth and thirteenth centuries. Thus, students of Rabbenu Tam R. (ca. 1100-ca. 1171), such as R. Moses of Kiev, R. Eliezer ben Isaac of Prague, R. Pater ben Joseph from Carinthia (Slovenia; if not from Carentan in Normandy), or R. Isaac Dorbelo, who traveled extensively through West and East Slavia, could easily sit in the same bet midrash in Ramerupt or Troyes. They could refer in their discussions to forms from different Slavic dialects without referring to the specific origin of these forms. Although Sefer ha-shoham contains a dialectal designation for the ''tongue of Tiras, which is Rus','' this does not necessarily mean that the author differentiated East Slavic from other Slavic languages. Similarly, the Cyrillic alphabet, which was shared by different Slavic languages, is described as ''Russian letters'' (see section 2.2 below). In rare cases Slavic forms were even mislabeled by later editors as ''German'' (lit. ''language of Ashkenaz'').
In most cases, however, no historical documentation is available; we are thus confined to linguistic considerations alone. The glosses pose a methodological challenge for linguistic interpretation. We have to take into account that the glosses are not transliterations but phonetic renderings which reflect the oral usage typical of their locale and historical time period (a circumstance which makes the glosses very valuable as evidence). Thus, for example, we presume that unlike g or г, which in Slavic alphabets may represent either an occlusive g or its fricative development h, Hebrew gimel or hey are not bound by a graphic tradition and will faithfully reflect their corresponding consonants. Normally, we do not find linguistically representative quantities of Slavic glosses belonging to 71 a single work or author (editor), and since the glosses are geographically and chronologically dispersed, we cannot expect much graphic consistency in this corpus (similar heterogeneity obtains in the much better documented German and French glosses of the period), although some unstable regularity was noted for the Czech materials, where the glosses are especially abundant. 73 The textual history of many of the Hebrew sources containing the glosses has not been thoroughly enough investigated. In the cases which we want to consider, the glosses normally appear in the earliest, as well as in the majority of the textual variants. It is difficult to imagine their being a late insertion into these works. It is easier to reconstruct the cultural context for adding a ''Canaanite'' gloss on Western soil in the eleventh to thirteenth centuries, during the time when the composition or very early stages of the transmission of these texts took place. This is harder to do for a later period, for which we have no evidence on Slavic-speaking Jews in German lands or on Jewish migrants from Slavia in these lands.
The crux of the problem is that in most cases we cannot distinguish a Czech, or any other West Slavic form, from an East Slavic one, because Hebrew consonant transliterations seldom reflect vocalism (except for the several sources that do provide vocalization). For the unvocalized forms, the vowels can be discerned only by means of the Hebrew matres lectionis-waw, yod, and aleph; however, this is always done in a rather ambiguous way: waw is used to convey either o or u, and yod-either i, e, ȇ or a diphthong; either may also be used to render jers (ъ and ь, respectively); an aleph without any additional mater lectionis could render a and possibly other back vowels, as well. In addition, consonant equivalency between the languages is not perfect: even though Hebrew has both š and c, the vernaculars of some transliterators did not feature these consonants, hence instead of the distribution shin-š and tsadi-c, what we often see is shin/sin without any distinction between š and s (or, sometimes, shin/sin standing also for ž, č, or c), and samekh for c and č (rendered also by tsadi and the combination fç, respectively). There may also be more complicated cases, where the Ashkenazi fricative tav may stand for s, or when the velar r and h are used interchangeably, and so on.
So what made JH so sure of their identification of the whole corpus as Czech? The implicit reasoning of JH was apparently that if they conclusively identified at least some of the forms as Czech, while the origin of the others could be interpreted either way, then all of them are most 73. See Jakobson, ''Iz razyskanij;'' Š edinová, ''Life, '' 212-13. probably Czech. This kind of extrapolating is convincing enough, but only when applied to a single text or at least to a group of texts having in common some characteristics such as date, localization, milieu, and so on. Our glosses, however, are dispersed over a variety of texts from different regions and milieus. Thus, if 'Arugat ha-bosem and Or zaru'a were authored by Bohemian Jews and contain a number of indisputable West Slavic forms, it is very probable that the other forms in the same works are also Czech. But how can this extrapolation be applicable to texts written in France, Germany, or Italy, which do not contain unequivocally West Slavic forms? This becomes especially obvious in light of the evidence presented in the first section of this essay: the considerable data on the presence and creative activity of the Jewish scholars of East Slavic origin in these countries. 74 Chronologically, the first source which we should consider is the eleventh-century Mainz Commentary on tractate H . ulin of the Babylonian Talmud, ascribed to Rabbenu Gershom of Mainz (R. Gershom ben Judah Me'or Ha-Golah, ''Light of the Exile,'' ca. 960-1028), but probably written and compiled largely by his students. 75 The text has been checked with the oldest manuscript from the Bibliotheca Angelica in Rome, dated paleographically to the twelfth or thirteenth century (Rome, Ang A.5.18;F 11669) 76 and almost identical to the Vilna edition of 1880-86 (with two minor discrepancies in vocalization; see in the footnotes below). The text is vocalized, although here, as in other cases, vocalization may be a later editorial addition and is normally ignored for purposes of linguistic reconstruction. There are four ''Canaanite'' glosses:  79. A plural form pleca would go better with the transliteration with aleph. However, the gloss refers to the Aramaic form in the singular and an aleph is attested interchanging with a waw rendering o in some Romance glosses (see, e.g., for Hebrew transliterations of ammoniaco (aqaynwma in Soncino as well as wqaynwma, wqynwma in other versions; bGit 69a; Arsène Darmesteter and David S. Blondheim, Les gloses françaises dans les commentaires talmudiques de Raschi, 2 vols. [Paris, 1929-37], 1:6), or balsamo (amçlb in MS Vatican 138 along with wmslb and wmçylb; bShab 62a; ibid., 1:11  '' bH . ul 8b). 85 The form *trebeno may be a part. pass. sg. neut. from trěbiti; cf. CS trěbiti, OCz triebiti, Pol trzebić, OR terebiti, Russ trebiti. 86 All these verbs may mean ''cleanse of unnecessary parts'' in agricultural and, occasionally, butchery contexts. The widespread use of this Slavic verb among Jews is confirmed by the Yiddish Slavicism trejb(er)n ''to cleanse meat in order to make it kosher,'' found in both Western and Eastern Yiddish. 87 The verb may have had cultic connotations even in Slavic usage because of its homonymy and its possible common etymology with the CS trȇba ''sacrifice.'' Our gloss may also be an adjective derived from trȇba. The adjective trěbъny, unlike the participle, is well attested: it was widely used in different recensions of Church Slavonic with the meaning of ''sacrificial,'' ''fit or destined for a sacrifice.'' 88 In both cases, as with mako/makъ below, the transliteration would show the jer that has not fallen in a weak position, but with aъ or its back reflex instead of a ь (in view of that the letter waw, in contrast to yod, always renders a back vowel, usually o or u). This could also indicate an assimilation of the jer' to the back vowel of the next syllable (bearing out the regularity found in the Zograph Gospel of the eleventh century). 89 84. Kupfer and Lewicki,Ź ródla,[199][200] The glosses above may all belong to any of the Slavic subgroups, with *teg/tig of unclear etymology being attested only in Czech.
R. Nathan ben Jehiel (1035-c. 1110) in his encyclopedic dictionary the 'Arukh (completed in 1101; Rome) has one Slavic gloss: • *mako/makъ (wqm; Cz má k, Pol mak, CS makъ, OR makъ) for Aram ygrp ''poppy.'' 90 Noteworthy here is that this early twelfth-century transliteration reflects the weak jer in final position as still pronounced. We find the same evidence only in East Slavic sources. 91 Furthermore, the letter waw in the transliteration tends to reflect a back vowel, making the front reflexes of jer, such as can be found in Czech and Polish (but not in Slovak), less appropriate here.  [Moscow, 1995], 51-52): forms such as the nom. gorodo, čelověko (ibid., 606, 679), and the like. The Jewish gloss is less likely to reflect the literary pronunciation of the weak jers as attested in Rus' before the fourteenth century (Boris A. Uspenskij, Istorija russkogo literaturnogo jazyka (X-XVII vv.) [Moscow, 2002], 150-55; cf. also Uspenskij, ''Russkoe knižnoe proiznošenie XI-XII vv. i ego svjaz' s južnoslavjanskoj tradiciej (čtenie erov),'' in Aktual'nye problemy slavjanskogo jazykoznanija [Moscow, 1988] cited in the commentary by Nahmanides (1194-1270) ad loc. As was shown by Moshe Altbauer, resh (for a velar r) must render h here. 94 The transliteration conforms to the late Czech form but is hardly possible for eleventh-century Czech, which still preserves its occlusive g. Even in the twelfth and thirteenth centuries, Latin transliterations almost unanimously show a g (Bog, Glupa, Gostik, Podgrad, Praga, and so on in addition to the rest of our Hebrew glosses); 95 the same holds for Slovak and Upper Sorbian. 96 Cf. also ''Bohemian'' *deg(e)t (fgyd) for the OCz dehet in the twelfth century Ra'avan's work (see below). The few cases found in Czech (like Bohuslaus in 1169) and Slovak (Behis from, possibly, *Bȇgyš' in 1138) are uncertain, and in any case they are later than our source. 97 In East Slavic the shift gϾγϽh appears much earlier; it is first attested not later than the eleventh century, when we find г and х used interchangeably. 98 For the period in question, this form may belong only to Southern dialects of East Slavic (or to literary pronunciation patterns in the East Slavic North).

R. Solomon ben
The commentary on 1 Kg 6.7 ascribed to Rashi explains Heb twbqm (hammer) as ayswr˜wçlb a''fwld ''dlwtw [doloto] in the tongue of Rus'.'' This is the second case where the East Slavic provenance of the form is explicitly mentioned, although the Polish form dłóto seems to be closer to the transliteration, which does not render a pleophonic vowel. Nevertheless, in contrast to the previous gloss, this one is not found in the earliest manuscripts, all of which date from the thirteenth to fourteenth 99 centuries, and may well be a later addition.
106. ''Like small tubs that one uses to [put] water [into] and [afterwards] drops it, and in the alien language okrin'' (˜hb˜yçmtçmç twnfq twbyr[˜ybmk ylkyçm yr''qwa z[l 'wçlbw˜ykpwçw μym; Cambridge, University Library Add. 478, 14th c.). a hypothetical Judeo-Slavic akadon going back to Gk kadion based on the printed editions. 107 Moreover, Gk κ δι ν / κ δ ς are jars for storing liquids, rather than vessels that can fit the description of a certain kind of wash-basin, as added by Rashi: açyra [ybg wl hywpk μydyhw μydgbh ta˜yxjwrw μym wb μyntwnç˚wra ylk ala [bwk˜ymk hywç[ hnyaw-''The vessel is inverted over its [the idol's] head, and it is not made as a kind of turban, but as a long vessel into which one pours water and washes one's clothes and hands.'' The Slavic word normally designates specifically flat bowls corresponding to this description. 108 There is one more gloss in the early commentary on Genesis Rabba, which is ascribed to Rashi: Beyond this, there are some glosses whose inner-Slavic affiliation cannot be ascertained. Consider the following: • The undeciphered phrase asn qmwp 126 (with the meaningless variant arkyn mqynb) for the verse hbçaw yl hçg μwqmh yl rx ''the place is tight for me, make me room, and I will settle'' (Is 49:20 128. Harkavi, Jews, 50-51, 136. 129. Sreznevskij, Materialy, 2:1726. 130. ''ynwda ylja [means] 'good,' that is˜yzálae b] in the Canaanite language'' (yljã yzálae b] 'nk 'çlb bwf ynwda).
131. ''hkymçb whsktw should be translated as abnwg, that is anAE /g in the Canaanite language'' (anAE /g 'nk˜wçlb wnyyh abnwg 'grt hkymçb whsktw). most, must; Lat mustum) appearing in many Slavic languages: Czech mošt, Pol moszcz, OR mъstъ, Russ must and most, but with a front vowel in  • *ugli (ylgwa) for Heb μjp ''coal'' (Is 54.16). 133 • *dьvrϽdьbr (rwwd) for Heb μyskr ''mountain ridge'' (Is 40.4). 134 Even though they suggest the correct Slavic equivalent, 135 Kupfer and Lewicki are still faced with two problems: (1) the correspondence is not ideal: the gloss has a waw where a b is called for; (b) how can a word meaning ''dell'' explain the Hebrew word for ''mountain ridge''? I suggest the following solutions: (1) the change bϾv must reflect a feature characteristic of West German, especially the Middle Franconian dialects spoken in the Rhineland. 136  Heb rwnth dwa ''poker.'' 146 The latter seems more probable.
I believe that in these particular cases intepretatio bohemica should be preferred as long as we do not have any additional information. These cases show, however, that JH's extrapolation cannot be accepted without reservation even for the most obviously Bohemian works; this is even more true for the earlier texts which have no evident Bohemian links. We probably need to distinguish between phonological and lexical argumentation, that is, between forms phonologically impossible for a certain region, on the one hand, and lexica accidentally unattested there, on the other. In this case, the availability of forms such as těh, attested only in Czech, or očag, attested only in Russian, will not constitute a decisive argument. The topic deserves a monograph-size study, but even this brief survey leads us to a number of preliminary conclusions: 1. Textual critical and linguistic data show that the glosses cannot be late scribal additions, as has been suggested by Weinryb, 147 not only because they are attested in the earliest manuscripts but also because they reflect early linguistic phenomena, such as the pronunciation of jers in weak positions, occlusive g in the Czech forms, and other early linguistic features. 2. The provenance of the majority of the glosses, within the Slavic realm, cannot be determined exactly, i.e., most of them may equally well be East or West Slavic in origin (and occasionally even South Slavic). 3. The extrapolation from the minority of cases (where West Slavic or specifically Czech forms can be convincingly identified) to the whole corpus, which is then in its entirety defined as Czech, is not warranted. This is because, despite the approach followed in JH, the corpus of glosses contains forms that cannot be Czech but are best attested in East Slavic, such as h/γ for the eleventh century (*sneh), pronunciation of final jers for the twelfth century (*mako), and s resulting from the second palatalization (pl. nom. dusi); see also the possibly East Slavic plečo and probable East Slavic vocables such as puskati and očagъ. 4. This kind of extrapolation is not tenable even within the limits of a single work, since forms of different provenance can coexist in the works of the same author (as in the case of the Ra'avan). This could happen for a number of reasons: 4.1. the author may have been exposed to a variety of linguistic 147. ''Beginnings, '' 482. influences. This is the situation of the Ra'avan, who probably traveled to Rus' through Bohemia; 4.2. glosses could be borrowed from one text to another, such as in the case of the gloss from the 'Arukh appearing in the same form in a commentary by Rashi; 4.3. glosses can occasionally serve comparative linguistic purposes, as in the case of jȇbo for Heb μwby, perinos for Aram swnwrp, and gunka for Aram aknwg.
There is another possible explanation for this heterogeneity, although extremely hypothetical. Consider the possibility that we are dealing here not with a conglomerate of forms belonging to different Slavic vernaculars but with consistent evidence demonstrating the existence of a ''Canaanite,'' that is, Judeo-Slavic, language which integrates both West and East Slavic forms (just as this happens in Yiddish).
Weinreich's idea of a Judeo-Slavic language was developed by Wexler mainly on the basis of the Slavicisms in Yiddish. 148 Even at the present preliminary stage we should note several indications of consistency among this heterogeneous material. For instance: duxa or the earliest attestation of the reflex tьlt for PS tl ъt (*bьlxa, *pьlt), or, alternatively, a unique rendering of the West Slavic syllabic sonant.
• All the sources except one (*snih/sneh) show an occlusive g.
• The pronunciation may have been Germanized in some cases, as in *d [ь]vr (in conformity with the attested pattern for some Slavic forms common in Western Yiddish, such as n e b e x/n e v e x, and the like).
Also noteworthy is the possibility that the glosses may testify to a certain level of literacy in Church Slavonic and even familiarity with the Church Slavonic Bible (cf. *omet? in Job 15.27). This may have far-reaching implications for the extent of Jewish involvement in early medieval Hebrew-Slavic translations (see 2.1 below). This possibility may find new and unique confirmation in the source presented in section 2.2 below.

Hebrew-Slavic Translations
The possibility of Church Slavonic literacy among Jews may be corroborated by the rich corpus of medieval Hebrew-Slavic translations of the fourteenth to sixteenth centuries, with the originals chosen for translation not only from the Judeo-Christian repository of biblical texts common to both Judaism and Christianity but also from specifically Jewish material such as medieval midrash, liturgical texts, and the like. The translations include, inter alia, Slavic versions of Hebrew translations from Latin (the Book of Josippon and Johannes de Sacrobosco's Book of the Sphere) and Arabic (such as Al-Ghazalī's Intentions of the Philosophers; Maimonides' Logical Vocabulary; on ''Physiognomy'' from Rhazes' Al-Mans . uri; and Pseudo-Aristotle's Secret of Secrets). It is also possible that the Greekspeaking Jews in Rus' took part in the translation of Judeo-Greek materials. 149 However, the form that Jewish involvement in these translations assumed, just as the question of whether or to what extent Christians took part in working on the translations has, thus far, not yet been fully elucidated. 150 We must also take into account that, for the period in question, the official language of the Great Duchy of Lithuania, including the language of juridical and financial documentation, was ''Ruthenian,'' a Western variant of the East Slavic vernacular also known as ''Old Russian,'' in its written form strongly influenced by (and, for its users, often not distinguishable from) Church Slavonic. It seems very probable that the Jews involved in economic activities in medieval Rus' could benefit from a certain level of literacy in this language, just as their compatriots in the West could by mastering Latin. This literacy finds new and unique corroboration in the abecedarium discussed below.

Cyrillic-Hebrew Abecedarium
The Cyrillic-Hebrew abecedarium of the thirteenth century has never been discussed by Slavic linguists. It is found in a Hebrew Psalter from the collection of Matthew Parker, a seventeen-century Archbishop of Canterbury, preserved in the Oriental collections of the Bodleian Library at Oxford (MS Oxford, Bodl. Or. 3). The psalter is dated paleographically to the thirteenth or the early fourteenth century and was owned or probably even copied by an English-speaking Christian. The manuscript contains several abecedaria (alphabets): the Hebrew and the Greek alphabets with the acrophony (names of the letters) and notes in Latin (ff. 75r and 74v). These are followed by the Arabic and the Cyrillic alphabets with the names of the letters spelled out above the Cyrillic characters in vocalized Hebrew writing (ff. 72v-73v).
The Cyrillic-Hebrew abecedarium belongs to the very small group of medieval Slavic (Cyrillic or Glagolitic) abecedaria in which the names of the letters are transliterated into non-Slavic (Latin and Greek) scripts. 151 In fact, only two such acrophonic alphabets are known before the end of the thirteenth century: the Paris Glagolitic-Latin alphabet, known also as the Abecedarium Bulgaricum, dated to the eleventh or early twelfth centuries, 152 and the Greek abecedarium from the ''Bandurian Legend'' created possibly in the thirteenth century but preserved only in fifteenthcentury copies. This means that our text may be the earliest Cyrillic xenographic abecedarium and the second earliest xenographic acrophonic abecedarium of any kind.
I give a detailed linguistic and historical treatment of the Cyrillic alphabet in an earlier essay in this journal, 153 where, following Olszowy-Schlanger's notice, 154 I argue that the Cyrillic alphabet must date from the same time as the rest of the text and is likely to have been produced by (or at least with the participation of) a Cyrillic-literate Jew from Rus' rather than by an English Christian Hebraist. The most likely appears to be a team made up of an English Christian Hebraist and a Slavic-literate Jew, who prepared the text using Hebrew as their common language, while their object of study was elements of the Slavic language.
The language behind the transliterations points at an East Slavic provenance of the author (or the informant) of the abecedarium. It also suggests certain Central East Slavic dialectal features, which may be pinpointed in specific areas where Jewish presence is attested in the thirteenth and fourteenth centuries. Thus, the dialect reflected in the abecedarium includes the following features: explosive g, differentiation of c and č, akanie, pronunciation of ě as ie, and possibly transformation of e to o. Taken together, these features exclude the extreme South and North, namely, Kiev and Novgorod, but go well with transitional Central East Slavic dialects characteristic in this period for Chernigov and Nogorod-Seversky, as well as for Minsk and Novogrudok. 155 The latter two areas were integrated into the Great Duchy of Lithuania by the mid-thirteenth century. It should be noted that what we know of Isaac of Chernigov dates from the same period. Besides, Jewish settlement in Lithuania, most probably going back to the Kievan and Galician-Volhynian communities of the eleventh to thirteenth centuries, is well documented only beginning in 1388, when the first charter of privileges was granted to Lithuanian Jews by Vytautas. 156 CONCLUSIONS Scant as it may seem, the evidence on the knowledge of East Slavic among early East European Jews is incomparably richer than the data on any other language they may have spoken during this period. This evidence is also very diverse and representative. It ranges from grassroots elements such as the mastering of obscene speech or the adopting of Slavic personal names, to the much more advanced level of East Slavic proficiency involving Church Slavonic literacy in its East Slavic variation. Literacy of this kind leads to participation in translation projects or even becomes an object of teaching. Our sources witness both Jewish East Slavic monolingualism, on the one hand, and Jews as possibly the first attested teachers of Slavic literacy in Latin Europe, on the other.
The emerging picture may impact different fields of knowledge and prompt a reevaluation of many historical and linguistic problems. Slavic linguistics should take into account early East Slavic forms documented in Hebrew transliterations which sometimes provide earlier attestations of these forms than the ones preserved in the Slavic written sources. The issue of an early Slavic substratum is also of crucial importance for the history of Yiddish. The very existence of an East Slavic-speaking Jewry may provide an additional argument in favor of the existence of Jewish communities in this region, who either were not of German descent or else treated their German legacy in a way very different from later Yiddish-speaking communities.
This situation contrasts strikingly with what we know of the linguistic insolubility of Yiddish-speaking Jews in their Slavic environment in the early modern and modern periods up to the beginning of the assimilation processes of the late nineteenth and early twentieth centuries. Only then do we once again encounter East European Jews who speak only Russian or Polish, or who play a leading role in the teaching of these languages in the West. The linguistic situation reflected in our early sources may indicate a peculiar type of coexistence between Jews and their Slavic neighbors, one that differs from later models of either extreme isolationism or no less extreme assimilation attested in this region. What I am suggesting is a model in which the boundaries between the two groups could take shape along confessional rather than ethno-cultural lines. 156. Stanislovas Lazutka and Edwardas Gudavichius, Privilege to Jews Granted by Vytautas the Great in 1388 (Moscow, 1993).