Karaites: Their Names and Migration Routes

Abstract

The article provides an analysis of the geographic origins of Karaites in four areas where Karaite congregations were commonly found after the Middle Ages, namely, Arabic Middle East (territories of modern Iraq, Syria, Israel, and Egypt), Constantinople/Istanbul and its area, the Crimean Peninsula, and Eastern European territories belonging today to Lithuania and Ukraine. It combines available historical, onomastic, and linguistic data revealing the migrations of Karaites to and inside these regions. For the first two regions, no ambiguity exists about the roots of local Karaites. Their ancestors were Jews who adopted the Karaite version of Judaism. For the Crimean communities, various factors favor the hypothesis about the territories of the Byzantine Empire (which later became Ottoman), and more specifically, Constantinople and its area are the only major source for their development. The Karaite communities in such historical Eastern European provinces as Lithuania (properly speaking), Volhynia, and Red Ruthenia were created after migrations from Crimea to these territories. The article also discusses medieval, cultural, and potentially genetic links between Karaites and Rabbanite Jews in the areas in question. It also addresses one phonological feature, the sibilant confusion, shared by the Galician–Volhynian dialect of the Karaim language and the Lithuanian dialect of Yiddish.

1. Introduction

For many centuries, all Karaites of the world considered themselves a special group of Jews. Their major difference from other Jews, called Rabbanites by them, consists in the fact that Karaites do not recognize the authority of the Talmud and their religious rite and laws contain numerous peculiarities. Gentiles quite rarely distinguished between Karaites and Rabbanites, and if such a distinction was made, both were still seen as Jews. During the last centuries, Karaite communities existed in Middle East and Eastern Europe. Descendants of Middle Eastern Karaites usually consider themselves, even today, both ethnically and religiously, Jews who profess a special version of Judaism. The same approach also was true for Eastern European Karaites. During the 20th century, however, a major ideological shift occurred in their self-identification. Some Karaite activists took the position that, on the one hand, Karaism is not a movement inside Judaism, but rather a separate religion, and, on the other hand, Eastern European Karaites descend from ethnically Turkic population groups who started to profess this religion in the Middle Ages. Writings supporting such an approach, as well as those opposing it, are often polemical.

One can distinguish four regions where Karaite congregations were present for many centuries: (1) the Arabic Middle East (territories of modern Iraq, Syria, Israel, and Egypt); (2) Constantinople/Istanbul and its area; (3) the Crimean Peninsula; (4) Eastern European territories in modern Lithuania and Ukraine. In this article, we will discuss what is known about the migrations of Karaites in these regions, bringing together major factual elements known to us from historical documents and information that can be extracted from an onomastic and, to a lesser extent, linguistic analysis of the Karaites of these four regions. For the surnames considered, the focus is on those derived from toponyms or brought from other areas. Both these categories reveal migrations.

2. Arabic Middle East

The Karaite version of Judaism takes its roots in the 8th century CE in Babylonia.1 Over the next several centuries—and, more specifically, during the period covering the 10th and 11th centuries that is often called the “Golden Age” of Karaism—its followers spread to various parts of the Middle East, creating multiple congregations in Jerusalem, Syria, Egypt, and other centers of the Muslim world of that time. Medieval historical documents show the emergence and propagation of this religious movement among one portion of local Jews. Gradually, however, numerous Karaite communities of the area in question disappeared. Sources from the 15th to 17th centuries refer to the presence of Karaites in a very limited number of places in the Middle East. During the 1480s, Italian Jewish travelers Meshullam of Volterra and Obadiah of Bertinoro both indicated the presence of 150 Karaite families in Cairo (Mešullam da Volterra 1989, pp. 39, 53; Schwab 1866, pp. 17, 21). Ottoman tax documents for Damascus refer to about 40 Karaites in the mid-16th century (Lewis 1952, p. 24). Twenty-four Jews appearing on tax lists for the city of Hīt (now in Iraq) of 1525/1526 (Epstein 1980, p. 224) could also be Karaite: no reference to Rabbanite Jews in that place is known. The small Karaite community that existed in Jerusalem during the 17th century dwindled at the beginning of the 18th century, serving just to meet pilgrims, mainly from Damascus and Cairo (Mann 1931, pp. 742–43). In Damascus, Muslim Arabic documents regularly refer to local Karaites during the 18th century. Yet, the last reference to the Karaite quarter appears in 1811, and a sales document of 1832 refers to the unique Karaite man still living in the city (Al-‘Ulabi 2011, pp. 66–70, 86, 91, 151, 158, 159, 173, 174). In 1869, 20 Karaite families were present in Hīt.2 By 1951, only 13 families remained (Schmidinger 2019, p. 60). The Karaite congregation of Jerusalem counted nine families in 1824 and 25 persons in 1909 (Fischel 1973, p. 51; Algamil 1985, p. 18). During the 19th and 20th centuries, the only flourishing Karaite community in the Middle East was in Cairo. In 1875, a Hebrew document internal to that community listed 169 family heads, comprising 16 Cohanim and 37 Levites (Brinner 1982, p. 137). A 1946 census recorded 3834 Karaites in all of Egypt. Most lived in Cairo (El-Kodsi 1987, pp. 14, 20).

After the Arabic conquest of the Middle East in the 7th century, Arabic gradually became the vernacular language for a large portion of the local population. One of the major consequences of this linguistic shift corresponded to the use of Arabic naming patterns by people in the conquered areas. Arabic names have up to five elements, of which ism is necessarily an individual name, while the other elements—kunya (starting with abū), nasab (with ibn preceding the name of the father or some more distant ancestor), laqab (a nickname often starting with the definite article al-), and nisba (designation of the geographic provenance ending in the suffix -ī and often starting with the article al-)—can be either individual names or nicknames applied to several generations of the same family. The last element, nisba, is of particular interest for the topic of this article: it indicates the origins of the person or his family. For medieval Middle Eastern Karaites, exactly as for local Rabbanite Jews, such patterns were valid as well, along with traditional Jewish naming patterns that include an individual name, the patronymic (starting with the Hebrew ben ‘son of’ or bat ‘daughter of’), and for men who were members of specific male lineages, indications that they belong to these lineages (ha-Cohen ‘of priestly origin’ or ha-Levi ‘Levite’). These naming patterns can be illustrated by names of Karaite scholars from the 9th–11th centuries, such as Benjamin al-Nahāwandī, Daniel al-Qūmisī, Abū Yūsuf Ya‘qūb *al-Qirqisānī, *Salmon (סלמון) ben Yeroḥam, Sahl ben Maṣlīaḥ ha-Cohen, Joseph ibn Nūḥ, Abū ‘Alī Ḥasan ibn ‘Alī al-Baṣrī (called Yefet ben ‘Alī ha-Levi in Hebrew sources), and Abū al-Faraj Furqān ibn Asad (Yehoshua ben Judah in Hebrew) (Nemoy 1952, pp. 21, 30, 42; Poznanski 1908, pp. 8, 12, 20, 30, 48; Mann 1931, pp. 18, 30). Except for the first three scholars on the list, these authors are known to live in Jerusalem. In the list above, their nisba links the first two scholars to the Persian city of Nahavand (today, western Iran) and the province of Qūmis (northern Iran), respectively. The name *al-Qirqisānī (or *al-Qarqasānī) may indicate the provenance of this famous religious scholar from the town of Circesium (Arabic al-Qarqīsiyā; today, near al-Busayrah, in eastern Syria) or from Qarqasān, near Baghdad (Nemoy 1952, pp. 21, 30, 42). Al-Baṣrī is based on the name of the city of Basra, now in southern Iraq. The same patterns were also valid during the next centuries. For example, for the 15th century, we know about such Karaite scholars in Egypt as Samuel ben Moses al-Maḡribī ‘the one from Maghreb’ and David ben *Se‘aḏel ibn al-Hītī ‘the one from Hīt’ (Poznanski 1908, pp. 81–82; Nemoy 1952, pp. 196, 230). The name al-Hītī shows the presence of Karaites in Hīt in the 15th century, corroborating the idea that Jews present in that city during the next century could at least partly be Karaites. Since we have no evidence confirming the presence of Karaites in the Maghreb during that period, the nickname *al-Maḡribī could have been acquired after a travel to or via the Maghreb. Yet, for the second half of the 12th century, we know about a Karaite from Alexandria, Moses ben Abraham, who settled in the Draa area (Arabic Dar‘a) of southern Morocco and for this reason was nicknamed *Dar‘ī (דרעי).

Some Damascus documents related to Muslim court decisions for cases involving Jews (Al-‘Ulabi 2011) refer to local Karaites. During the entire period covered (1688–1811), Karaites in these sources are cited by their given names and patronymics, which start with the Arabic vernacular bin ‘son (of)’ or bint ‘daughter (of).’ Often, the name includes an indication of their religion, either as al-Yahūdī ‘the Jew’ (pp. 24, 69) or al-Yahūdī al-Qarrā ‘the Karaite Jew’ (p. 113). For some persons, their last name corresponds to a nisba: Faḍl-allāh bin Ibrāhīm al-Baḡdādī ‘Fadlallah the son of Abraham of Baghdad’ and Mūsā al-A‘jamī ‘Moses the Persian’ (pp. 91, 151). This nisba can be a nickname of the individual showing his own geographic provenance or be a hereditary family name testifying about the place of origin of the ancestors of the bearer. We cannot assert with certainty for any Karaite name in the collection of documents in question that it was already hereditary. Yet, from Karaite texts written in Hebrew, we know that Fīrūz (פירוז), not present in the Muslim collection in question, was found in Damascus from the Middle Ages until the beginning of the 18th century (Poznanski 1913, pp. 45, 50; Mann 1931, p. 1282). We can be certain that this was a real surname. In the earliest references, this name looks like a typical nasab; it is preceded by the Arabic ibn ‘son of.’ This Karaite family could have originated from Persia or any country influenced by Persian culture, where local Karaites borrowed this given name from their Muslim neighbors. In Persia, Fīroz (meaning ‘victorious’) was a popular male given name borne, among others, by several kings. The same surname, Fīrūz, appears in Cairo already in the 15th century (Poznanski 1913, p. 47). In the mid-20th century, bearers of this name still were living in the same city.

Onomastics can illustrate the existence of several other links between the Karaites of the two most important communities of the Arabic Middle East of the last centuries, those of Cairo and Damascus. The earliest available representative source of Karaite surnames from Cairo dates from 1875 (Brinner 1982, pp. 129–37). In this list, we find 36 different male given names, of which only eight are non-biblical names borrowed from Arabs; all others are biblical or post-biblical names based on Hebrew. Five of these eight Arabic names, Farag, Faḍlallāh, Manṣūr, Marzūq, and Khiḍr, appear in the only source for Karaite names of Damascus available to us that is far from being representative.3 Dimashqi ‘(one) from Damascus’, borne by 21 men, is the most common last name in the Cairo list in question. Moses (Mūsā) ha-Levi el-Dimashqi was the chief rabbi (al-ḥakhām al-akbar in Standard Arabic) of Cairo Karaites in 1856–1872. He was born in Jerusalem and, for that reason, was also known by the Arabic nickname el-Qodsi, ‘the one from Jerusalem.’ He returned to his native city in 1872, but some of his descendants in Cairo used el-Qodsi as their surname (El-Kodsi 1987, p. 267).

The Cairo list from 1875 also includes the surname Qirīmi. Most likely, this surname was brought as a ready-made form from Crimea (directly or via Istanbul), where it has been known since the Middle Ages (see Section 4 below). In any case, this name illustrates a recent link between the Karaite communities of Cairo and Crimea or Istanbul. Other names related to Crimea are those of chief rabbis invited to Cairo during the first half of the 20th century: Kefīli and Levi Babovich (both born in the Russian Empire) and Mangūbi (from Istanbul but bearing a surname betraying the provenance of his ancestors from the Crimean town of Mangup). Two other chief rabbis, both bearing the surname Cohen, came from Istanbul.4

Except for the examples cited in the previous paragraph, we find no traces of Karaite migrations to the Arabic Middle East from the three other geographic areas where Karaite communities were found during the last centuries.

3. Constantinople/Istanbul

By the 11th century, the Karaite religious tradition was already well-established in the Byzantine Empire. References to Karaites appear for such places as Constantinople and neighboring Nicomedia, Attaleia (today, Antalya, southern Turkey), Trebizond (Trabzon, northeastern Turkey), Salonica (Macedonia), and Cyprus (Ankori 1959, p. 153). The initial development of Karaism in the Byzantine territories is directly related to the geographic proximity between these areas and the Muslim territories of the Middle East, where this religious movement emerged and became widespread. Ankori (1959, pp. 85–86, 98–99, 129) suggests that the arrival of some Karaites could have occurred during the temporary occupation between the second half of the 10th century and the first half of the 11th century by Byzantium of territories of the Muslim provinces where Karaites were present: Syria, a portion of the Land of Israel, and the Edessa area (today Şanlıurfa, Turkey). Other Karaites could have migrated to neighboring Byzantium because of the relative security given to Byzantine Jews by Emperor Constantine VII (913–959), which was not disturbed by his successors. We know about a migration of some Jews from Egypt to Byzantium during the years 1016–1019, a period within the reign of the Egyptian caliph al-Ḥakīm (996–1021), a Muslim religious fanatic (Sharf 1971, pp. 108–12). In theory, at least some of these Jews could have been Karaite. Finally, we cannot exclude the possibility that some Byzantine Jews adhered to Karaite principles under the influence of Karaites who came from Muslim countries of the Middle East.5 During the second half of the 12th century, Benjamin of Tudela notes the presence of a Karaite congregation of about 500 individuals in Constantinople, covering about 20 percent of the total Jewish population of the city (Adler 1907, pp. 14, 28, 30). In the mid-15th century, Adrianople, the capital city of the Ottoman Empire, was also the most important Karaite community of the empire. After the conquest of Constantinople in 1453, Karaites from Adrianople were relocated to this new Ottoman capital city. Tax documents from 1540 show the presence in Constantinople of 130 Karaite households divided into three congregations: Edirne (the Turkish name for Adrianople) 117, Kastamonu (northern Turkey) 2, and Pravadı (today Provadia, northeastern Bulgaria) 11 (Epstein 1980, pp. 178–80). In the 18th century, the Constantinople community declined dramatically. An epistle sent in 1729 by Karaites of that city tells that the local synagogue was not restored for 193 years, only 40 Karaite families are still living in the city, and a majority migrated to Crimea over the past 20 years (Mann 1931, pp. 742, 1278–79).

The Karaite communities established in Byzantine Eastern Thrace (the area covering both Adrianople and Constantinople) used the Greek language of their Christian neighbors in their daily life.6 Religious scholars from Constantinople (12th–17th centuries) used this tongue to explain certain terms in their works written in Hebrew (Danon 1912, p. 148). It was also crucial for the selection of non-Hebrew given names by Karaites even centuries after the Ottoman conquest of the Byzantine Empire. In the Karaite cemetery of Istanbul (JCT 2016), we find about 100 tombstones from the 17th–20th centuries with given names of Greek origin. Among the examples are the female names Afedra אפ׳דרה (αφέντρα ‘mistress’), Chrisi חריסי (χρυσή ‘golden’) and its variant Chursi כורסי/חורסי, Chrisokali כרישוקלי combining Chrisi and Kali, *Despina דישפינה/דשפינה (δέσποινα ‘mistress, matron’), Fostira פ׳ושטירה/פוסטירא/פוסטירה (φωστήρα, the feminine form for ‘luminary’), Fumia פ’ומיה/פומייא (compare φούμη, a dialectal variant of φήμη ‘oracle, reputation’), Jera ג’ירה (compare Jerasopula below), Kali (καλή ‘good’), Kera כרה/כרא/כירא (κερά, a dialectal variant of κυρά ‘lady’), Manaka מאנאקה (from a diminutive form of μάνα ‘mother’), Mikri מקרי/מכרי (μικρή, an affective feminine form for ‘small’), and Panorea פנורייא (πανωραία ‘most beautiful’). Numerous female given names end in the Greek diminutive suffixes -ula (ούλα in Greek) or -opula (οπούλα): Fumula פומולה from Fumia, Chanula חנולה from biblical Hannah, Chursula כורסולה from Chursi, Archo(n)dopula ארכודופולה from Aρχοντοπούλα, ‘little princess’ (a diminutive feminine form for ‘lord, master’), and Chodula חודולה from the same base form, Kerasopula קירהצופולה/קירצופולה/כרצפול (from κεράσι ‘cherry’) and Jerasopula ג’יראסופולה which seems to be its variant, Esteropula אסתרופולה from Esther, and Saropula שרופולה from Sarah; compare also the given name Pula. Men are usually identified on their tombstone inscriptions by their Hebrew (mainly biblical) names, and for that reason, we find only a few given names of Greek origin. Among them are Afenda (from a variant of Greek αφέντης ‘lord’) and its hypocoristic form Afendopulo, as well as *Eliaupulo (אליאופלו, from Elijah), another form ending in the suffix -opulo (όπουλο), the masculine equivalent of the feminine -opula.

Many names from the above list also appear in the tombstone inscriptions of Greek-speaking (“Romaniote”) Rabbanite Jews of Constantinople: male Afe(n)da and female Afedra, Archondopula, Chanula, Chodula, Chrisokali, Chursi, Chursula, Esteropula, Fostira, Jera, Jerasopula, Kali, Manaka, Mikri, Panorea, and Pula (JCT 2016). In both Karaite and Romaniote congregations of Constantinople, we also find the biblical name Caleb, unusual for Sephardic or Ashkenazic Jews. Since only a small portion of the given names used by Greek Christians was borrowed by Jews, these examples show major cultural elements shared by these two groups of Constantinople Jews. The idea of both groups descending from the same ancestral Byzantine Jewish group provides the easiest explanation for this phenomenon.

Greek words are sources for some old Karaite surnames from Adrianople and/or Constantinople. Among the examples are *Maruli (מרולי, one of the most distinguished Karaite families during the 15th-18th centuries),7 *Mesorodi (מצורודי), *Revitsi (רביצי), Havyara, and Yeraka; compare Greek μαρούλι ‘lettuce’, μεσοροδι ‘heart of the rose’, ρεβιθι ‘chickpea’,8 χαβιάρι ‘caviar’, and γεράκι ‘hawk’, respectively. The last two surnames survived until the 20th century.

Nevertheless, the earliest known Karaite names on the territory of modern Turkey are not Greek. The last name of Judah ben Elijah *Hadassi (הדסי), a Karaite religious scholar who flourished in Constantinople in the middle of the 12th century, ends in the Hebrew suffix -i, commonly used to form demonyms. Most likely, it is a nickname revealing his provenance or that of his ancestors from the city of Edessa (today, Şanlıurfa in southeastern Turkey) (Neubauer 1866, p. 56; Ankori 1959, p. 130). The oldest identified surname, *Başyaçı (appearing in Hebrew documents as בשייצי, בשיצי, בשייאצי, or באשיאגי), is of Turkic origin.9 Present in numerous Hebrew documents, it surely was used internally within the Karaite community and not just for contacts with the Ottoman authorities. Two alternative theories can explain its Turkic roots. First, we can conjecture that the Karaites of Adrianople (where its first known bearer dwelled in mid-15th century) were bilingual, using both Greek and Turkish in their daily life. In such a linguistic context, the inception of a Turkish-based surname would not be a surprise. Secondly, we cannot exclude the possibility that at least this family originated in some area outside of Eastern Thrace, where the local Gentile majority was speaking a Turkic idiom, for example, in Anatolia or Crimea.10

The corpus of surnames used by Karaites of Constantinople during the 19th and 20th centuries (JCT 2016) included a series of names brought by Karaite migrants from other areas as ready-made forms. Only one of these names, Firuz, came from the Arabic Middle East. Several dozen other surnames belonged to recent migrants from Crimea. Since the community of the Ottoman capital city was already small,11 they constituted a significant portion of it, especially in the 20th century. The earliest references, those to Kirimi and Egiz, correspond to the 18th century. In tombstone inscriptions of the Karaite cemetery of Istanbul dating from the 19th and 20th centuries, we find references to several dozen other surnames that were clearly brought as ready-made forms by Karaite immigrants from Crimea. Among the examples are Ayvaz, Bolek, Bota, Çabak, Çıbar, Cigersiz, Cigit, Emildeş, Erinçek, Gümüş, Karakaş, Kefeli, Kokey, Koyçu, Mangubi, Ormeli, Pampu, Panpulov, Penbek, Sappak, Şişman, Tanatar, Trişkan, Tongur, and Turşu. Some start to appear already in the 1820s (most likely, they came after the Russian annexation of the Crimean Peninsula at the end of the 18th century), but the majority are known beginning with the second half of the 1850s, after the end of the Crimean War (1854–1855).

4. Crimea

A Karaite source written in Hebrew in 1278/1279 indicates the presence of both Karaite and Rabbanite Jews in Solhat (now Staryi Krym), at that time the capital city of the Crimean Tatars (Bowman 1985, p. 231). Karaite communities are known to have existed in Solhat, Kefe (Genoese Caffa, now Feodosia), Mangup (south to modern Bakhchysarai), and Qırq-Yer during the 13th to 15th centuries (Kizilov 2011, p. 114). The latter city was the Crimean Khans’ residence in the 15th century. After Tatars left it in the 17th century, it gradually became known as Çufut-Qale ‘Jewish fortress’ and was the main center for the Crimean Karaites (Kizilov 2011, p. 123). This geography shows that when former Genoese territories (including Caffa) and those of the Greek Principality of Theodoro (including Mangup) became Ottoman in 1475, Karaites were living in all the three regions: the Crimean Khanate, a vassal state of the Ottoman Empire, and the two areas directly controlled by Ottomans. The Russian censuses of Crimea (1792, 1796) show the presence of about 2300–2400 Karaites in the peninsula; 1162 lived in Çufut-Qale and environs (Prokhorov 2016, p. 72).

Several alternative explanations for the emergence of the Karaite communities in Crimea can be found in historical studies. Some authors link Crimean Tatar to the Khazars, who ruled the peninsula starting in the 7th century and disappeared from historical chronicles in the 11th century. Seraya Shapshal (Szapszał in Polish spelling), a religious and administrative leader of the Crimean and later Polish–Lithuanian Karaites during the first half of the 20th century, was one of the main proponents of this idea. The theory has neither historical nor linguistic basis. From the point of view of historiography, strong arguments against it were formulated by Ankori (1959, pp. 65–70). He notes that no link to Khazar is ever mentioned in early Karaite literature that was contemporary to the Khazars or written soon after the fall of their empire. For example, *al-Qirqisānī (who lived in the 10th century) refers to Khazars as contemporary to him but says nothing about their Karaism. In numerous Karaite writings existing for the 11th century, we find detailed accounts about the successes of Karaism in the Land of Israel and other areas, without any reference to Khazars or Crimea. To the above arguments, one can also add the fact that there is a significant chronological gap between the disappearance of the Khazars in the 11th century and the earliest references to the Crimean Karaites in the last third of the 13th century.

Linguistic data provide no corroboration for putative links between Khazars and the Crimean Karaites, either. All available data point to the fact that Karaites used in their everyday life a religiolect of the same idiom as Crimean Tatars (Musaev 1964, pp. 36–38; Shapira 2006). This idiom belongs to the Kipchak group of Turkic languages and is unrelated to the Oghur group of Turkic languages, to which the extinct Khazar language belonged.

In theory, we cannot exclude the possibility that the earliest known Crimean Karaite communities (those of the 13th and 14th centuries), like those of the local Rabbanites, included Jews of heterogeneous origins who lived in Crimea prior to the Mongol (Tatar) conquest at the end of the 1230s.12 They were not necessarily already professing Karaite Judaism at that time.13 The adoption of this version could have come later under the influence of actual Karaite migrants. Note also that the earliest references to Karaite congregations in Crimea correspond to the period when the “Golden Age” of Karaism (10th–11th centuries) was already over. As follows from the travelogue by Benjamin of Tudela and other available sources, at that time, the number of Karaite communities in the Middle East, those that could become potential sources for Crimea, was limited.

We know that during the sack of Constantinople in 1204, the Jewish quarter of Constantinople was burned (Starr 1949, p. 25). This dramatic event in the history of the Byzantine capital city could have contributed to the migration of one portion of local Karaites to other neighboring regions, such as Byzantine rump states. The Trebizond Empire was one of these states, and it included a portion of southern Crimea. As mentioned in the previous section, the presence of a Karaite congregation in Trebizond, the capital city of that empire, is also known from historical sources. The idea14 of migrations of Karaites to Crimea from Constantinople, directly or via other Byzantine possessions and, possibly also from Trebizond15 has the advantage of being based on such factual data as the existence at the end of the 12th century of a large Karaite community in Constantinople and a much smaller community in Trebizond plus the dramatic effect of the events of 1204 on the Jewish congregations of the Byzantine capital city. It also correlates well chronologically with the date of the earliest references to Karaites in Crimea. The presence of a Crimean Karaite, Aaron ben Joseph, in Constantinople at the end of the same 13th century16 shows that links between these two communities were already established during that period and indirectly corroborates the possibility of migrations in the opposite direction.17

Analysis of names shows the existence of additional Byzantine connections. The earliest known reference to a Crimean Karaite bearing a Greek given name dates from 1455/1456 and corresponds to the tombstone in Theodoro (Mangup) of a woman called אפרוסיני *Euphrosyne (Greek Ευφροσύνη).18 At that time, Theodoro was the capital of the Greek Principality of Theodoro, which split at the beginning of the 14th century from the Empire of Trebizond and existed until 1475, when it was annexed by the Ottomans. During the following centuries, the given name was commonly used by Crimean Karaites; multiple references to forms derived from it appear in Çufut-Qale.19 In the Istanbul Karaite cemetery (JCT 2016), we find only one reference to one of these forms, *Firsin פירסין (1838). Because of its absence before the 19th century, odds are high that this unique reference corresponds to a recent migrant from Crimea. These factors imply that this Greek female given name was not necessarily brought from Constantinople or any Byzantine province external to Crimea. Karaites could borrow it from Greek Christians within the Principality of Theodoro. In the Hebrew inscriptions made in the 14th-century Solhat, we find a reference to Judah the son of Elijah *ha-Adrianopoliti (האדרנופוליטי) (1363).20 The nickname of this person includes the Hebrew definite article that precedes the Greek word for an inhabitant of the city of Adrianople. The document referring to this person indicates the sale of a Torah scroll to a woman called פוסירא who likely also had Byzantine roots. Indeed, this given name could be related to the given name פוסטירה (Greek Φωστήρα ‘luminary’, Fostira in Latin characters), commonly used by both Karaites and Rabbanite Jews of Istanbul in the 17th–20th centuries (JCT 2016). In 14th-century Genoese city of Caffa, situated close to Solhat, Cacanachi appears as the second (and last) name of a local Jew (Kizilov 2011, p. 180). This name seems to be composed of Greek κόκκινος ‘red’ and the Greek diminutive suffix -άκη. Most likely, its bearer was a Karaite since in the 18th century we find its Turkicized form, *Kokinay Aqa (אקא כוכני), as a Karaite male given name.21

Kizilov (2011, pp. 108–9) emphasized the provenance of the first Karaite settlers in Crimea from the territories already taken by Mongols in the Caucasus and Central Asia.22 He suggests several indirect arguments to back his theory. First, according to Kizilov, the earliest Karaite migrants settled in the Tatar capital city of Solhat, rather than in Byzantine possessions. Secondly, Karaite settlement, according to Kizilov, is closely related to that of the Tatars, namely, initially only in the same settlements as Tatars and later, when in the middle 14th century Tatar started to expand to the western Crimea, Karaites followed them. Thirdly, Karaite legends speak about ancestors coming together with Tatars from Persia, Bukhara, and Cherkessia (in the northern Caucasus) in the 13th century. Fourthly, according to Kizilov, the first Karaite settlers present in Çufut-Qale in the middle of the 14th century commonly used Tatar given names such as (male) Tohtamış (Tokhtamysh) and Parlak, while earliest Karaite tombstones with Greek given names appeared in Mangup after 1453. Fifthly, according to Kizilov, Crimean Karaites seemed to use Crimean Tatar as their vernacular language already by the beginning of the 14th century, while their coreligionists from Constantinople used medieval Greek at that time.

None of the above arguments is logically attractive. On the one hand, the fact that the earliest reference to Crimean Karaites deals with the Tatar capital city and not the Byzantine-controlled area in southwestern Crimea can be related simply to the paucity of historical documents. In no case can it be taken as evidence about the original settlement of the first Crimean Karaite migrants being precisely in Solhat. The same factor makes doubtful any discussion of the chronological details concerning the geographic expansion of the Karaite communities in Crimea. On the other hand, since Solhat has been an important urban center since the 1270s, we cannot exclude the possibility of direct migration to it from Constantinople or its environs. These factors show the weakness of Kizilov’s first two arguments. His third argument notes that stories about the earliest Crimean Karaites coming along with Tatars were recorded only during the 19th century. They sound like typical legends related to Karaite oral tradition, the inception of which may have been rather recent. We do not find any document that would indicate that Jews were following Mongol conquerors during the expansion of the Mongol Empire in the 13th century. No available source indicates the presence of a single Karaite in Bukhara or the northern Caucasus.23

To a large extent, Kizilov’s last two arguments are irrelevant to the discussion in question.24 Indeed, these data do not provide any information about the geographic roots of local Karaites. They just indicate that the Karaite migrants could not be sufficiently numerous to maintain their original idioms in Crimea and keep given names based on these idioms. It is no surprise that descendants of these migrants adopted the local language and incorporated into the corpus of their given names some names used by the local majority.

A few references to names of Karaites in medieval Solhat that appear in Hebrew documents allow illustrating the existence of multiple origins of members of the local community. Among the examples are Hoja (I)brahim Shah (שאה כוג׳אה ברהים, איברהים כוג׳אה) (1376, 1388), with the name Shah being of Persian origin, Suleiman ha-Parsi (הפרסי ‘the Persian’) the son of David ha-Parsi (1388) (Harkavy and Strack 1875, pp. 240, 260; Akhiezer 2019a, p. 338), and Abraham ha-Ger (הגר, ‘the proselyte,’ 1321) (Harkavy and Strack 1875, p. 256). The use of such personal nicknames necessarily implies that at least during the first third of the 14th century, migrants from Persia25 and persons who recently converted to Judaism were exceptional in this Karaite community. References to Hezekiah ha-Levi (1360) and Elijah ha-Cohen (1376) (Harkavy and Strack 1875, pp. 194, 241), both with the traditional designations of belonging to the Jewish lineages, reinforce the idea about the Jewish ancestry of a majority of Karaites present in medieval Solhat. During the 19th–20th centuries, both Cohen (in the form Kogen) and Levi were commonly used by Crimean Karaites, alone or as a part (almost exclusively the first one) of double surnames. Among the examples are Kogen Avah, Kogen Ayvaz, Kogen Balıq, Kogen Baqqal, Kogen Bota, Kogen Çauş, Kogen Çiçke, Kogen Ekmekçi, Kogen Horoz, Kogen Kefeli, Kogen Qambur, Kogen Mangubi, Kogen Minaş, Kogen Nuray, Kogen Öksüz, Kogen Pembek, Kogen Rebi Oğlu, Kogen Töteş, Levi Arabacı, Levi Babovich, Levi Balaqay, Levi Bödene, Levi Botuq, Sarrı Levi, Levi Topal, and Levi Yalpaçıq.26 The high percentage of surnames revealing belonging to Jewish traditional lineages fits well the idea about the genetic links of ancestors of Crimean Karaites to other Jewish groups.

Starting at the end of the 16th century, the tombstone inscriptions in Çufut-Qale indicate the presence of newcomers from Istanbul. Their number was particularly high during the 17th century. These migrants brought a series of surnames to Crimea: Gibbor (tombstone of 1589), Fuki (1668), Sikke (1672), Begi (1672), Sinani Çelebi (1696),27 *Revitsi רביצי (1703), and Yeraka (1731) (Firkovich 1872, pp. 34, 91, 94, 101, 109; Harkavy 1876, p. 259; Fedorchuk et al., forthcoming; El’yashevich 2016, p. 50). Three of these surnames, Sikke, Begi, and *Revitsi, gradually disappeared. Other surnames were commonly used during the next period. Some of the migrants brought to Crimea some Greek given names that do not appear in earlier sources. Examples from tombstones in Çufut-Qale (Firkovich 1872) are *Kereçe כרצה (1622, p. 41),28 *Archodopula (ארכורופלא, misinterpretation of ארכודופלא) born in Istanbul, the daughter of Judah Fuki (1668, p. 91), Fumia (פומיא) born in Istanbul, the daughter of Isaac Fuki (1672, p. 94) and her son *Eliaupulo (אליהו פולו, a Greek diminutive of Eliahu) whose father was David *Sikke (סכה) from Istanbul (1679, p. 101), Fumia (פומיא) the daughter of Isaac Fuki (1698, p. 109), Fumia (פומיא) the daughter of Menahem *Revitsi (1703, Fedorchuk et al.), Panorea (פנרייא) the wife of Isaac Fuki (1737, p. 134), Panorea (פנוריא) the daughter of Eliahu Yerushalmi (1746, p. 138), and Fumula (פומולא) the daughter of Eliahu *Sikke (1664, Fedorchuk et al.). Note that we find multiple references in the tombstone inscriptions of the Karaite cemetery of Hasköy, the historical neighborhood of Istanbul (JCT 2016), for each of these given names. The only exception is Kereçe, which also appears in Crimean Russian-language sources as Kepoчa and Xepeчe (Lebedeva 2000, p. 110; Prokhorov 2016, p. 89); this form is unknown in Istanbul, where we find, however, *Kerasopula, with the same root.

The surnames and given names discussed in the previous paragraph do not cover all migrants from Istanbul. Others appear in the Crimean tombstone inscriptions without surnames. For example, the provenance from the Ottoman capital city is explicitly indicated for Rachel, who died in Mangup in 1627; Judah the son of Joseph, buried in Çufut-Qale in 1640; and Lea, a Rabbanite woman who became converted to the Karaite version of Judaism and married in Caffa in 1658 (El’yashevich 2021, p. 117; Firkovich 1872, p. 83; El’yashevich 2016, p. 51).

Certain surnames used in Crimea during the 19th century also reveal the provenance of their first Crimean bearers from Istanbul, though the exact period of migration cannot be identified from the information available to us. It is the case of Stambulu/Istanbullu and Kosdini/Kostini, both meaning ‘(one) from Istanbul/Constantinople’: the first forms are Turkic, and the last ones are Hebrew. These surnames were used alone or as second parts of double surnames such as Yefet-Kosdini and Yefetov-Stambuli, corroborating the idea that Yefet (of which Yefetov represents its Russified variant) originated in Istanbul, where it was one of the most common Karaite surnames during the 19th century. Gabbay was another surname particularly frequent in the Ottoman capital city during the same period, and for this reason, at least some of the families bearing this surname in Crimea could originate from Istanbul. The surname Haskoylü (Turkish Hasköylü) designates a person from Hasköy, a quarter in Istanbul where the Karaite synagogue and cemetery were located.29 During the 19th century, one branch of the Hacar (Hadjar) family migrated to Crimea from Istanbul.30 From the same city surely came the Başyacı family known in 19th-century Crimea. As discussed in the previous section of this article, bearers of this name were prominent in the cultural life of Istanbul for several centuries.31 Some families bearing the surnames Kogen (Cohen) and Levi could also originate from Istanbul.32

The surnames mentioned above cover only a small portion of more than 300 independent surnames used by Crimean Karaites during the 19th–20th centuries. Other surnames, a large majority, are of Turkic origin; they are derived from words taken from the Karaim/Crimean Tatar or Turkish idioms. This fact should not be misinterpreted to tell that these Turkic surnames reveal families present in Crimea for many centuries. Some of them could be just branches of families that came from Istanbul, starting with the end of the 16th century. Turkic surnames, the largest layer for Crimean Karaites, are relatively recent, implying that their consideration does not allow shedding light on the origins of these families. These surnames are based on words from the language that was vernacular for this group, and so do not reveal any element from the period preceding their inception.

Several independent factors point to a recent age of numerous Karaite surnames with Turkic roots. Available sources predating the Russian conquest of Crimea in 1783 show no use of surnames by local Karaites. Turkish taxation lists compiled in the 1680s usually provide personal names and sometimes patronymics (Efimov 2020, pp. 81–82, 132–37). In the Hebrew-language tombstone inscriptions and various documents of this and previous periods (including marriage contracts), we do not find family names either.33 People are cited almost exclusively according to the traditional Jewish pattern X ben/bat (‘son/daughter of’) Y, where X is their own given name and Y is that of their father. Sometimes the given names of men and those of their fathers are followed by indications of belonging to Jewish lineages: (ha-)Cohen or (ha-)Levi. The only actual surnames are those brought by migrants from Istanbul enumerated above in this article plus a few forms based on Crimean Tatar words, including Botuk ‘wrinkled,’ Çabaq ‘roach fish,’34 Qocaq ‘strong man,’ Şekerci ‘confectioner,’ and Turşu ‘pickled.’ The situation changes only at the end of the 18th century, as can be seen from both internal Karaite documents and official Russian sources. During the 1790s, among families listed in the Russian-language census data for Çufut-Qale (Prokhorov 2016), about 70 have names that look like surnames, while only eight (including two headed by widows) appear without such names. On the one hand, we can be sure that the majority of these 70 names became hereditary, at least from that time onward: the same surnames with but a few exceptions were still being used by Crimean Karaites during the first half of the 20th century. On the other hand, many of these names are unlikely to have become hereditary generations before the 1790s. This conclusion follows from the small frequency of various names. The list comprises five households with the name (H)acı ‘pilgrim’ that could be independent of each other; each could have had a relatively recent ancestor who made a trip to Jerusalem. Three families have the surname Kogen (Cohen). Several names appear in two households: Ap(p)aq, Arabacı, Koyle (Köylü), Mangubu (Mangubi), and Sarıban. All other names appear in only one household. This is even true for Fuki (listed as Fuka in the census data), a surname that, as discussed above, has been present since the end of the 17th century. If numerous names appearing in the list were fixed for many generations, the number of their bearers would be significantly larger and the number of different surnames used within the Crimean Karaite communities significantly smaller. In the same census data from the 1790s, certain families appear with several names, one of which is usually preceded by “po prozvaniyu” ‘nicknamed.’ Examples: Moses Mandubu and his brother Elijah Kokey nicknamed Mangub (p. 92), Abraham Arabacı and his brother Ezra Bursuk Kapucı nicknamed Rençek (p. 93), Ṣadoq Gabbay nicknamed Turşu (p. 93), Samuel Kogen Qambur and his brothers Mordecai Kogen and Aron Kogen (pp. 95–96), and Moses Qalpaqçı nicknamed Butovıç (p. 97). Except for the last form, all names different from the given names in the above list were used as Karaite surnames at the beginning of the 20th century. Indeed, in the lists compiled by Vaysenberg (1913) and Kruglevich (1916), we find Mangubi (equivalent to Mandub(u) of the census), Kokey, Arabacı, Borsuk (a variant of Bursuk), Kapucı, Erinçek (of which Rençek is a corrupted variant), Gabbay, Turşu, Qambur, and Kogen. During that period, each of these names had the status of an official surname. At the end of the 18th century, however, as follows from the census data, the status of some names was different, implying their recent inception.35

Surnames derived from Crimean toponyms show short-distance migrations. One of the earliest examples corresponds to Yehoshua Qirimi, who at the turn of the 16th century moved from Solhat to neighboring Caffa (Akhiezer 2019a, p. 348). His last name represents a Hebrew nickname based on Qırım, an alternative Tatar name of Solhat from which the modern name of the peninsula, Crimea, is derived. In sources from the 18th–20th centuries, we also find references to Mangubi (with the Hebrew demonymic suffix -i) and several surnames ending in the Turkic demonymic suffix -li/lı such as Qaleli ‘from Çufut-Qale,’ Kefeli ‘from Caffa,’ and Yaltalı ‘from Yalta.’

The surname Qavşanlı is derived from the name of the city called Qavşan in Crimean Tatar, situated several hundred kilometers from Crimea. It is quite plausible, nevertheless, that the family in question had Crimean roots. Today, the place in question belongs to the state of Moldova and is called Căușeni in Romanian. In the 17th–18th centuries, however, this place was the capital city of the so-called Budjak Horde, a vassal state of the Crimean Khanate. Qırım Giray, the ruler of the Khanate during the second third of the 18th century, built there a palace that became one of his official residences, a second in rank after the one situated in the Crimean city of Bahçısaray. The surname Qavşanlı can just indicate that its first bearer had links to that city because of his professional activities. We cannot exclude the possibility that during the period of close political links between Crimea and the area in question, some Crimean Karaite families moved to Qavşan. In this scenario, the first bearer of the surname Qavşanlı returned to the region from which he or his close ancestors originated. Varnalı represents the Turkic demonym for someone from the city of Varna. Today, it is a city in northeastern Bulgaria, but before the end of the 19th century, it was an important administrative center of the Ottoman Empire. For the Karaite family in question, the exact nature of its link to the city of Varna remains obscure.36

The surnames Çerkez ‘a Circassian’, Qalmuq ‘a Kalmyk’, Kumuk ‘a Kumyk’ (in Turkish), and Kumuklu (adjectival form based on the same root) are more likely to be nicknames related to some personal features, rather than indications of the provenance of their first bearers from the corresponding regions in the North Caucasus, where no Karaite communities are known. Common nouns çerkez and qalmuq were well known in Crimea because the Crimean Khanate had military conflicts with both these Caucasian people. Yerushalmi means ‘(one) from Jerusalem’ in Hebrew. This name appears in multiple Hebrew-language sources and is a personal nickname for someone who made a pilgrimage to Jerusalem rather than a hereditary family name.37

Onomastics allows illustrating migrations to Crimea of Karaites established in the territories that today belong to Ukraine or Lithuania but, before the 19th century, were parts of the Grand Duchy of Lithuania or the Polish Kingdom. The earliest traces correspond to the 16th–17th centuries. For that period, we find hypocoristic forms of female given names ending in the Slavic diminutive suffix -ka in Karaite sources from Crimea written in Hebrew. Examples are *Sarka (שַרַכַה), a hypocoristic form of Sarah in the pinkas of the Karaite community of Caffa in the mid-17th century (Shapira 2018, p. 63), as well as *Sarka (שרכה, 1575 and 1601) and *Anka (אנכה, 1607) from biblical Sarah and Hannah, respectively; *Menka (מנכה, 1629) and *Neska (נסכה, 1651), which could be from Hebrew Menuḥa and Neḥama, respectively, all found in the tombstone inscriptions of Çufut-Qale.38 Such Slavic forms are likely due to the Karaite migrants from the Polish–Lithuanian Commonwealth.39 In the mid-18th century the famous Karaite scholar Simḥa Isaac ben Moses from the city of Lutsk in Volhynia moved to Crimea. His descendants bore the surname spelled Łucki in Polish and Лyцкий in Russian, which means ‘(one) from Łuck’ in both Polish and Russian.40 The surname Kokizov started to be used in Crimea in the 19th century. Its first bearer migrated from the town of Kukizów, Galicia (now western Ukraine). Numerous other families present in Crimea during the 19th and 20th centuries were known as Karaites from “Poland.” Analysis of the geographic distribution of these surnames in the territory that belonged to the Polish–Lithuania Commonwealth before the last third of the 18th century allows us to identify regions of origins.41 The largest group came from Lithuania. They are Abkowicz, Dubiński, Eliaszewicz, Juchniewicz, Kapłunowski, Kobecki, Ławrecki, Ławrynowicz, Malecki, Mickiewicz, Michajłowicz, Nowicki, Pilecki, Poziemski, Rajecki, Robaczewski, Szpakowski, Zajączkowski, and Żarnowski. Bearers of the surnames Firkowicz, Gogol, Gołub, Greczny, Łukszyński, Sułtański, and Turczyn were migrants from Volhynia. Bezikowicz and Grygulewicz/Kruglewicz could have come from either Lithuania or Volhynia. Eszwowicz, Leonowicz, and Mordkowicz came from Galicia, either directly or via Volhynia or Lithuania. Two surnames based on Slavic words, Bondar (‘cooper’ in Ukrainian) and Kaczor (‘drake’ in Polish), most likely also were brought as ready-made forms by Karaite migrants from Volhynia, Galicia, or Lithuania.

Because of the large distance between Crimea and Arabic Middle East, traces of links between the Karaites communities of the two regions are significantly smaller than those discussed above between communities in Crimea and either Istanbul or Slavic Eastern Europe. Still, we know that at least two Crimean Karaite families, Massuda and one branch of Yeru, came from Egypt, both during the 19th century.42 The ancestors of the bearers of the surname Miçri ‘(one) from Egypt,’ a surname present in Crimea since at least the mid-19th century, surely also had lived in Egypt. Several Crimean Karaite Hebrew documents from the end of the 18th century (Sulimowicz-Keruth 2021, p. 701) refer to a man with the last name *Şamlı ‘one from Damascus’ in Turkish, another name indicating provenance.

In Crimea, both Rabbanite (“Krymchak”) and Karaite communities have coexisted since the Middle Ages. The fact that their daily languages were similar is not a surprise: both can be considered just as religiolects of the language spoken by the local non-Jewish (Muslim) majority: the Crimean Tatars. Yet, we also find similar developments in the Hebrew repertoire of Crimean Karaite (often called Eastern Karaim in linguistic literature) and the Krymchak vernacular idiom. Both groups pronounce the Hebrew letter tsadi (צ) as the affricate/č/. We find this sound in Karaite surnames Çoref (Karaim pronunciation of צוֹרֵף ‘jeweler’), Çaduqov (from the male given name Çaduq, the local form of biblical Ṣadoq צָדוֹק), and Miçri (from מִצְרִי ‘Egyptian’). The sounds/ḥ/and/h/exist neither in Crimean Tatar nor in Eastern Karaim or Krymchak daily idiom. This feature influenced the pronunciation of the Hebrew ḥet (ח) and he (ה). At least for Karaites, both gave rise to/g/in the initial word position; compare Gazzanov from חַזׇּן (Hebrew ‘cantor,’ which Karaites came to use to mean ‘community leader’) and Gelelovich (Slavic patronymic form from Gelel, the Karaite form of the biblical given name הִלֵּל ‘Hillel’).43 In the internal position between vowels, he yielded either/g/as in Kogen, a surname used by both Karaites and Krymchaks, from their pronunciation of כֹּהֵן ‘Cohen, priest’ (for Krymchaks, this word is pronounced kögen) or/ğ/as in the Karaite surname Beğema from בְּהֵמָה ‘animal.’44 The pronunciation of Hebrew vowels by Crimean Karaites is similar not only to that by Krymchaks, but also to that by Sephardic Jews, with/a/for qamatz (compare Yaşiş from יָשִׁישׁ),/o/for holem (compare Kogen from כֹּהֵן),/e/for segol (compare Neeman from נֶאֱמָן), and the pronounced shewa under the initial word consonant (compare Meriva from מְרִיבָה). Ribbioğlu (also spelled Rebioğlu) is composed of two Karaim elements, the Hebrew-based root ribbi ‘teacher, rabbi’ and the Crimean Tatar patronymic suffix -oğlu. The exact phonetic form of the root deviates from rabbi, the Tiberian normative pronunciation of רַבִּי. It could either be influenced by vowel harmony or be based on a non-Tiberian phonetic variant of the Hebrew form. Note the Krymchak word rebi ‘teacher.’ The sharing of peculiarities of the pronunciation of Hebrew by Crimean Karaites and Krymchaks can be due to various factors. On the one hand, it can result from mutual cultural influences between the two groups. On the other hand, at least some of these peculiarities could be due to the legacy of Jewish communities to which the ancestors of some Karaites and Krymchaks belonged, with potentially similar geographic roots in Byzantium. For example, the reflex/č/can be explained by the combination of two factors: (1) the affricate /ts/-pronunciation of tsadi by ancestors of numerous Karaites and Krymchaks and (2) /č/ being the sound closest to /ts/ in the phonetic chart of Crimean Tatar in which no /ts/ is present. Note that the /ts/-pronunciation of tsadi is valid for several Jewish groups, including Romaniotes.

5. Lithuania, Volhynia, and Galicia

At the turn of the 15th century, Jews were present in five cities of the Grand Duchy of Lithuania ruled by Vytautas: Troki (the place of the ducal residence, today, Trakai), Brest and Grodno (today, Brest and Hrodna, Belarus), and Lutsk and Vladimir in Volhynia (their current Ukrainian names are Luts’k and Volodymyr). Available official documents contain no precision about the exact group of Jews concerned, Rabbanites, or Karaites.45 The same absence of distinction characterizes all non-Jewish sources of Eastern Europe of the 15th century. Sources written in Hebrew are more helpful. From them, we learn that Karaite congregations were present in the Grand Duchy of Lithuania no later than the 15th century.46 Moreover, both Lithuanian and Crimean Karaites were in close contact with their Ottoman coreligionists at that time. A Karaite from Lutsk visited Adrianople in 1450 (Mann 1931, p. 582). During the 1480s, the Karaite dignitaries of the communities of Lutsk and Troki exchanged a series of letters with the Karaite leader of the Ottoman capital, Elijah *Başyaçı (בשייצי) (Mann 1931, pp. 1155, 1165). In one of these letters, Joseph ben Mordecai of Troki writes that the Rabbanite Jews stay in Troki only as transients. This direct testimony and certain indirect factors imply that at the turn of the 15th century, the Jews present in Troki were mainly or even exclusiveley Karaites. Moreover, we cannot exclude the possibility that among the Jews present at the same period in Lutsk, some, if not all, also were Karaites. A document of 1481 refers to a Karaite community present, along with a Rabbanite one, in the city of Kiev (Mann 1931, p. 1173). Regular references to Karaites in Lutsk are found during the next century. Censuses made during the second third of the 19th century show the presence of about 230 Karaites in the city (Sulimowicz 2017, p. 25). In Lithuania, properly speaking, during the 17th century, small Karaite congregations were found not only in Troki, but also in multiple other towns including those, particularly important during the next few centuries, of Poniewież (Polish spelling, now Panevėžys) and neighboring Nowemiasto (Naujamiestis) (Mann 1931, pp. 570–79). In the mid-17th century, the intellectual life of Lithuanian Karaites was flourishing, as can be seen, for example, from letters they exchanged with their coreligionists in Cairo, Damascus, and Jerusalem (Mann 1931, pp. 571, 1211). In 1765, 81 Karaites were recorded in Troki and 36 in neighboring villages in southern Lithuania; in northern Lithuania, the numbers were 27 in Poniewież, 102 in Nowemiasto, 9 in Poswol (now Pasvalys), and 22 in Sałaty (now Saločiai) (Gąsiorowski 2008, p. 478).47 Troki censuses record 397 Karaites in 1834 and 485 in 1858 (Sulimowicz 2017, p. 25).

In Red Ruthenia—the area corresponding to present-day western Ukraine, which had belonged to Poland since the 14th century and was renamed Galicia after its annexation by Austria after the First Partition of Poland in 1772—the earliest explicit reference to a Karaite community dates from 1475 and comes from Lwów (now L’viv), the main city of that area (Bałaban 1927, p. 15). Starting in 1578, we find regular mentions of Karaites in the neighboring city of Halicz (now Halych) (Bałaban 1927, p. 19). In that city, we find the following numbers of Karaites: 99 in 1765, 254 in 1843, and 192 in 1896 (Kizilov 2009, p. 96).

Members of the Karaite congregations in all these areas are often called “Western Karaites” in historical and linguistic studies. We have no direct evidence that would indicate their geographic provenance, but indirect factors imply that Crimea is the only plausible source area. One factor is geographic. Crimea is the only geographically close area where Karaite communities flourished in the 14th and 15th centuries; that is, the period for which we find the first direct or indirect references to the existence of Western Karaite communities.

Another group of factors relates to the general history of the area. The Grand Duchy of Lithuania was the western neighbor of the Golden Horde, of which Crimea was the southwestern province. Contacts between the two states became particularly close after 1369, with the beginning of the period of great troubles in the Golden Horde when violent conflicts opposed various descendants of Genghis Khan. Vytautas, the ruler of Lithuania, took direct part in these conflicts backing certain Tatar leaders, including Tokhtamysh (Witkowski 2013, pp. 212–17). During that period, Lithuania became a refuge for groups of Tatars led by rebellious and/or disgraced Khans. Gilbert de Lannoy, the Burgundian knight who visited Troki in 1414, writes that a great number of Muslim Tatars have lived in Troki and surrounding villages for generations (Kizilov 2003, p. 37; Witkowski 2013, p. 2018).48 Hacı I Giray (1397–1466), the man who in 1441 founded the Crimean Khanate, was born in Lithuania in one of the exiled families of Tatar nobles. Last, but not least, in 1398, Vytautas undertook a large military campaign to Crimea. Karaite legends firmly relate the inception of the Karaite congregations in Lithuania to this event. Close contacts between Karaites from Crimea and their coreligionists in Lithuania are well known for the subsequent centuries. A sales contract signed in 1394 in Solhat, appearing in a copy of a book sold in the Karaite congregation of Troki in 1598 (Akhiezer 2019a, p. 341), can serve as an illustration of the route from Crimea to Lithuania.

Historical linguistics provides additional corroboration about the Crimean roots of Western Karaites. We know that for centuries, Western Karaites used a Turkic vernacular language. For this assertion, we have direct testimony written in 1691 by Swedish Christian orientalist Gustav Peringer, who was commissioned by Charles XI of Sweden to go to Polish–Lithuanian Commonwealth and gather information about the Karaites. He indicates the use of the “Tatar” vernacular by Karaites living in several Lithuanian towns (Mann 1931, p. 571). This idiom is usually called Western Karaim in linguistic studies, with its northern dialect spoken in Lithuania and the southern one in Volhynia and Galicia. It was apparently just oral, with Hebrew, the language of religious observance, considered the only language suitable for documents internal to the Karaite communities. For this reason, the earliest known documents written in Western Karaim date only from the first half of the 18th century (Kowalski 1929, p. XX). Outside of the territories that today belong to Lithuania and Ukraine, Turkic-speaking Karaites are found only in Crimea.

More specifically, from a comparative linguistic analysis, it follows that two Turkic languages are particularly close to Western Karaim dialects used during the last centuries by Western Karaites. The first is the former speech of Crimean Muslims documented by Catholic missionaries in Codex Cumanicus (Venice 1303). The second is the Armeno-Kipchak language, an idiom developed among Armenians of medieval Crimea under the influence of local Tatars and brought by Armenian migrants from Crimea to Red Ruthenia and Podolia (today in Ukraine, but during that period within the Polish Kingdom). In other words, Western Karaim keeps multiple archaic features of the medieval Cuman language of Crimean Muslims lost in modern Crimean Tatar (Kowalski 1929, pp. LIX-LXXI; Musaev 1964, pp. 20–21, 36–38; Jankowski 2003).49 The survival of these archaisms until the 20th century in a basically non-Turkic environment, within communities that were (as discussed in the next paragraphs) at least bilingual, for a language without a developed written tradition, represents a major peculiarity of Western Karaites. This could take place only because of the strict endogamy of these communities and their voluntary cultural isolation from the surrounding population groups. It also indicates that after the initial development of these communities in the Middle Ages, they were never joined by large numbers of their coreligionists from Crimea. The presence of Muslim Tatars in the Duchy of Lithuania could also be a factor in the development of the Turkic vernacular of Western Karaites and its survival.

Certain bits of information about the past of Western Karaites can also be unraveled using onomastic tools. Direct data are meager. For the territory of modern Lithuania, the earliest Karaite, whose name is known to us from Hebrew sources, could be Samuel, nicknamed Politi (פוליטי), who seems to flourish in Troki early in the 15th century. His Greek last name is derived from a toponym ending in -πολι(ς) in Greek, most likely Adrianople or Constantinople. Several other examples correspond to names, mainly Slavic, that appear during the 18th–19th centuries. In Volhynia, we find a few surnames pointing to the Ottoman origin of their first bearers: Konstantinopolski ‘one from Constantinople’, Kostdyni (Kosdini), with the same meaning in Hebrew (maybe, the same family or even the same person as the previous one), Turczyn ‘Turk’, and Sułtański ‘of sultan.’ In neighboring Red Ruthenia (future Galicia), we find Turczynik, a diminutive of ‘Turk’ (Kizilov 2009, p. 92). Some of these families could come via Crimea; note that Kosdini was a common name there. This route is surely valid for Sinani, a surname that appeared in Volhynia after a migration from Crimea of one branch of a family that originated from Istanbul. A link between Crimean and Western Karaites is corroborated by sharing by both groups of the same Hebrew male given name Naḥamu (נחמו), which was almost non-existent in other Karaite or Rabbanite communities.50 Only one surname of Turkic origin, Qırğıy (ultimately from the identical noun meaning ‘hawk, falcon’), was shared by Crimean and Western Karaites before the middle of the 19th century (Kizilov 2011, p. 136; Sulimowicz 2015, p. 363). We do not know whether its presence in Volhynia was due to a migration of one branch of its bearers from Crimea or a local creation independent from the Crimean families but from the same common noun or, more likely, the identical male given name drawn from this common noun. In both scenarios, we deal with a migration from Crimea, directly of the surname or indirectly, via the given name from which the surname was later created.

Onomastic data provided in the previous paragraph are scant. This fact is mainly due to one dramatic difference between the names used by Western and Crimean Karaites. On the one hand, with but a few exceptions, the Crimean surnames are of Turkic (large majority) or Hebrew origin. This is no surprise because of the vernacular use of a religiolect of the Crimean Tatar and the importance of Hebrew in their religious culture. On the other hand, numerous surnames of Western Karaites, including almost all surnames from Lithuania proper, have Slavic suffixes. One can explain this phenomenon through the use of the Polish language in administrative documents in the territories in question (Yariv 2015, p. 156). Since the two suffixes in question, -ski and -owicz, are the morphological elements most commonly present in the surnames of Polish Christians, the use of these suffixes corresponds to a type of administrative standardization. However, some of these forms appear in Hebrew-language Western Karaite documents internal to the community. This factor implies that we are not dealing just with administrative labels. More important, for those surnames that are not based on Hebrew male given names, their roots also are Slavic, but for a very few exceptions.51 Moreover, the absence of Turkic influence concerns not only the surnames of Western Karaites, but their given names too. In Crimea, numerous Turkic female and some male given names were used by Karaites. Yet, we do not find a single example of a Turkic given name used by Western Karaites, even for women. All given names appearing in available sources are either of Hebrew origin or borrowed from local Slavic Christians.52 For these reasons, the names of Karaites from these regions are of no help in the analysis of Western Karaim, the vernacular Turkic idiom used internally in Karaite congregations.

This absence of names based on a vernacular language represents an unusual feature. It is particularly unusual if we note that certain surnames appeared in a natural way, well before official laws forcing the adoption of fixed family names were promulgated, while Karaite given names could in no way be affected by administrative decisions. Several factors can explain the phenomenon in question. First, it is reasonable to think that for centuries, Western Karaites were at least bilingual: in their daily lives, they used not only the Western Karaim dialects, but also the local Slavic idioms. This idea can explain the use of Slavic nicknames, some of which became surnames, and the presence of Slavic suffixes in hypocoristic forms of their given names. The relatively small number of speakers of Western Karaim, along with the non-compact character of their settlements surrounded by a non-Turkic majority population, represent other major factors (Musaev 1964, p. 6). Such a context was unfavorable for the creation of new Turkic names and for the preservation of Turkic given names that could have been brought from Crimea by the ancestors of Western Karaites. Only for biblical and some traditional, commonly used, post-biblical Hebrew given names were chances for survival relatively high.

Numerous internal migrations between communities of Western Karaites took place. This assertion is strongly supported by linguistic studies of the Western Karaite dialects, which demonstrate their close kinship.53 Onomastic evidence also offers solid support for the existence of these internal migrations. For example, the surname Stenpel is attested in Red Ruthenia, Lithuania, and Volhynia during the 17th, 18th, and 19th centuries, respectively (Mann 1931, pp. 555, 1125; Sulimowicz 2017, p. 23). Malinowski, known in Lithuania since the 16th century, appeared in Red Ruthenia and Volhynia during the following centuries (Mann 1931, p. 718, 1354; Kowalski 1929, p. XV). Nowicki was known in the 17th century in both Red Ruthenia and Lithuania (Gąsiorowski 2008, pp. 455, 471). At the end of the 17th century, a group of Karaites founded the congregation of Kukizów (now Kukeziv, Ukraine) in Red Ruthenia. Łabanos was the surname of the leader of this group (Bałaban 1927, pp. 27–29; Mann 1931, pp. 601, 614). Firko was known in Red Ruthenia at the beginning of the 18th century (Mann 1931, pp. 1270, 1272, 1354). At the end of the same century, we find its patronymic form, Firkowicz, in both Lithuania and Volhynia (Gąsiorowski 2008, p. 489; Kizilov 2009, p. 181). Magas appears in Volhynia and its patronymic form, Magasowicz, in Lithuania (Sulimowicz 2015, p. 368; Mann 1931, p. 1308). During the 18th–20th centuries, in Volhynia, we find bearers of such surnames as Abkowicz, Kapłunowski, Łopatto, Nowicki, Pilecki, Robaczewski, Rojecki, and Szpakowski, all from Lithuania, while Abrahamowicz, Eszwowicz, Leonowicz, and Mordkowicz (Mardkowicz) came there from Galicia. Kaliski (also spelled Kaлишcкий in 19th-century Russian sources from Volhynia) is surely of toponymic origin. Both the Polish and the Russian spellings coincide with the adjective meaning ‘of Kalisz.’ Kalisz, a large city in western Poland, however, is situated far from any place where a Karaite presence has ever been attested in Eastern Europe. A more plausible etymon would be Kałusz (today Kalush, Ukraine), a town in Galicia.54

A study focused on Western Karaites can also provide some clues about the history of the Rabbanite communities of Eastern Europe. In the 14th and 15th centuries, the few communities of these groups professing different variants of Judaism present in the Grand Duchy of Lithuania and Polish Red Ruthenia were seen simply as Jews by local Gentiles.55 Despite the small number of these communities, the geography of the settlements had important overlaps. Both Karaites and Rabbanites were present in the Grand Duchy of Lithuania and Polish Red Ruthenia. For that period, no data in our possession allow us to assert that the Karaites were significantly less or more numerous than the Rabbanites. Population sizes for the two groups likely were comparable.56 Sources from the 15th and 16th centuries also indicate an important overlap between the given names used by the two groups. Several Karaite given names found in Christian documents of the Grand Duchy of Lithuania (originally in Cyrillic letters) are not known to have been used by local Gentiles; neither do they correspond to East Slavic forms of biblical names, nor represent forms of Hebrew names known outside of Eastern Europe. They are Shemakh/Shamak (of uncertain origin), the Slavic-based Batko and Volchko, Hoshey and Hosva ‘Yehoshua,’ Ahron ‘Aaron,’ Levey ‘Levi,’ Yehuda ‘Judah’, and Yusko (a hypocoristic form of Judah). Nevertheless, all names in this list, except for the first two, were also used by Rabbanite Jews of the Grand Duchy vernacular at that time.57 Both groups also used Mordukhay, a form of Mordecai unusual because of its internal/u/. Bohdan, the only non-biblical given name borrowed by Karaites from local Slavic Christians, was commonly used by Lithuanian Rabbanites before the mid-17th century. For them, it was also one of the rarest examples of given names borrowed from local Gentiles (Author). Note also that one Hebrew post-biblical name, Nisan, known for both Lithuanian Rabbanites and Karaites, was rarely used in European Jewish communities (Author). Such onomastic affinity could not be fortuitous. The above factors make plausible not only the idea of close contacts between the two groups but also possible common geographic roots.58 Since the provenance of Western Karaites from Crimea is more than likely, and the Byzantine source appears particularly attractive for the Crimean Karaites, ultimate Byzantine roots with the intermediary of Crimea appear plausible for a significant portion of members of the early Rabbanite Slavic-speaking communities present in Eastern Europe as well.59

For the 16th century, we find the earliest traces of the existence of a phonological feature, a confusion between the sounds [s] and [š], that, on the one hand, is unknown in local Slavic dialects, and, on the other hand, characterized in recent centuries both the Volhynian–Galician dialect of Western Karaim and the Lithuanian dialect of Yiddish. This fact can be helpful in discovering the reasons for the Yiddish phenomenon. In a number of names appearing in the 1552 census of Karaites of Lutsk, we observe the interchangeability of [s] and [š]: Mosey/Moshey ‘Moses,’ Smoylo/Shmoylo ‘Samuel,’ Simkha/Shimkha ‘Simḥa,’ Botyus/Batyush (a hypocoristic form of Batko), and Hosva/Hoshva ‘Yehoshua.’60 The merger of these two sounds in one phoneme represents one of the main phonemic peculiarities of the Volhynian–Galician dialect of the Western Karaim language.61 The same confusion is also well known in Yiddish historical linguistics. During the last few centuries, it characterized the Yiddish dialect spoken in the territories of modern Lithuania and Belarus, and in the past, it seems to have been valid for Lutsk, too. Several illustrations can be found in the 1552 census of the Rabbanite Jews of that city (Bershadskiy 1882, vol. 2, p. 26): (1) /s/in Hosva (as for Karaites, based on Yehoshua) and (2) Senko and Shenko, two variants of a hypocoristic form based on biblical Simon. Most likely, the phonetics of the vernacular languages of both Karaite and Rabbanite Jews of Lutsk experienced the same Slavic influence.

Both groups also shared the same tragic events, the total expulsion from the Grand Duchy of Lithuania (1495–1503), the massacres and devastation during the mid-17th century related to the Cossack uprising led by Bohdan Khmelnytsky (Chmielnicki), and the anti-Jewish pogroms perpetrated on the territory of modern Ukraine by Haymadaks during the 18th century. The history of the two religious groups was to a great extent independent during the subsequent centuries. Gradually, Karaites became a small minority not only in comparison to their Christian neighbors, but also in comparison to the Rabbanites in the area. It makes sense to consider that one would expect a similar shrinking development for Rabbanites if these communities had not grown dramatically during the 15th, 16th, and first half of the 17th century as the result of a permanent flow of Ashkenazic Yiddish-speaking migrants from the West. Yet, as discussed in this section, since at least the 16th century, the immigrational flow for the Western Karaites seems to have been almost non-existent. Moreover, during the same period, some Western Karaites were migrating to Crimea.

6. Conclusions

As discussed in this article, since the end of the Middle Ages, Karaite congregations have been present in a limited number of regions in which local Karaites used different vernacular languages. The first of them is covered by Middle Eastern areas with the Arabic-speaking Muslim majority. No ambiguity exists about the roots of local Karaites who have used Arabic-based daily idioms for centuries. Their ancestors were Jews who adopted the Karaite version of Judaism. Some of them could be local. Others could come from Persia, where Karaism was founded. Karaite congregations of Cairo, Damascus, and Jerusalem were closely related. Constantinople/Istanbul represents the second Karaite center. Local Greek-speaking congregations were founded in the Byzantine period by Karaite migrants from the Arabic and Persian Middle East and/or local Jews accepting Karaite religious principles. Chronologically, the third center was created in the Crimean Peninsula. Historical and onomastic data favor the hypothesis about the Byzantine Empire (Adrianople, Constantinople) and one of the Byzantine rump states, the Trebizond Empire, being the main source for local Karaite communities that, since the Middle Ages, spoke a Turkic idiom based on Crimean Tatar, the language of the local majority. The existence of other medieval sources (Karaites from Persia or local non-Jewish converts to this version of Judaism) is factual, but no data in our possession imply that these sources were important for the genesis of Crimean Karaite communities. Multiple migrations of Karaites from Istanbul to Crimea are factual for the 16th–18th centuries. The fourth region encompasses a few areas that today belong to Lithuania and Ukraine. These communities were at least bilingual. They used on a daily basis not only local Slavic dialects (on which almost all their surnames are based) but also an internal Turkic idiom. Linguistic analysis implies that, most likely, this idiom stems from the medieval language of Crimean Tatars. This way, these West Karaite congregations were founded by migrants from Crimea.