The Turks in the anthroponymical database might have had Arabic Muslim, Persian, Turkic, and Mongol names. While the Arabic and Persian names are relatively easily recognisable, the identification of Turkic and Mongol ones presents difficulties caused by the obscurities of the ethnolinguistic history of the Turks of the region.
A considerable number of names in the database belong to Turks of lower- and middle-standing, which complicates the search for analogies for these names in contemporary Byzantine and Oriental textual sources that reflect primarily the life of the upper classes. From the thirteenth to the fifteenth centuries, Turkic personal names in Anatolia and the northern Black Sea region were still in the process of transition from the old pagan Turkic patterns to the standard Muslim ones. Moreover, the noted anthroponymical Islamization developed at different speeds in these two major historical areas, because the northern Cuman and Mongol regions were less Islamized, Iranicized, and Arabicized in comparison to Anatolia, where standard Muslim and Persian names had been in use since the end of the eleventh century. However, even in Anatolia the process of Islamization of personal names embraced first the upper classes. In the thirteenth century members of the Seljukid upper class used a rather complex anthroponymical pattern. The Muslim first name was often accompanied by a pagan surname, nickname, or tribal name such as Fakhr al-Dīn Doğmuş, Mubāriz al-Dīn Qara Arslan, Ṣayf al-Dīn Salur, Shams al-Dīn Oğuz, and the like. The remnants of the pagan past were more influential among the nomads and lower-class settled Turks; therefore, the devotion to pre-Islamic personal names might well have been more consistent among them. Contemporary sources reflected this rustic Turkic anthroponymical nomenclature rarely and only randomly. The process of the Islamization of Anatolian Turkic anthroponymy continued at least until the seventeenth century. Double names, numerous in early Ottoman defters, almost vanished from the registers by the seventeenth century. For instance, the Ottoman caba defter of the liva of Trabzon for 1695–1731 mentioned double Turkic-Muslim names only occasionally. Among the Turkic nomadic population of Anatolia, Islamic names had permanently displaced native Turkic ones only by the eighteenth century.
In the Turkic and Mongol areas of the northern Black Sea, in the thirteenth through the fifteenth centuries, native Turkic and Turkic-Mongol names dominated. Islam began to spread among the area’s population as late as under Khan Uzbek (1313–41) and the process of Islamization continued for centuries after his reign. This led to some conservation of native Turkic anthroponymical nomenclature. Our knowledge of the Cuman and later Mongol anthroponymics of the twelfth through the fifteenth centuries is fragmentary and unsystematic. Some Turkic names survived in Byzantine sources, especially from the “Scythian” regions, but are unique in written sources of that time and have no analogy in available Oriental sources.
Another serious difficulty of a linguistic nature is that we can only guess at the origin and local peculiarities of the Turkic dialects in the north Black Sea area and Anatolia. Several waves of Turkic nomadic expansion covered the Balkans and Byzantine Anatolia in the eleventh through the fifteenth centuries; both Oğuz and Qipchaq tribes participated in these conquests and predatory raids. In the thirteenth century, thousands of Turks of various tribes and tongues were ousted from Central Asia and northern Iran by the Mongol pressure. Nomads and sedentary Turks continued to move through Anatolia and the Balkans until the fifteenth century. Ottoman power had since the fourteenth century practiced a policy of forced resettlement of both nomadic and sedentary groups within the borders of the Ottoman sultanate. This permanent change in the ethnic pattern of the Balkans, western Anatolia, and the Pontos prevents the identification of the prevalent Turkic linguistic substratum. Because of the policy of forced resettlement, it is not always possible to rely on the retrospective analysis of the distribution of contemporary Turkic dialects. Contemporary medieval sources preserve too few traces of the living colloquial dialects of the time.
We know too little about Turkic anthroponymical nomenclature in the eleventh through the fifteenth centuries. Turkic names preserved in the Byzantine sources have no analogy in available medieval sources. It is Byzantine sources, properly elaborated, that may help in writing the ethnolinguistic history of the Balkan and Anatolian Turks. The well-known variability of spelling of proper names in the Late Byzantine texts, however, poses additional difficulties in etymologization of the Oriental personal and place-names.
The selection of names is based on general rules of transformation of Oriental lexical elements in Middle Greek, taking into account available data of the modern Turkish dialects in the Balkans and Anatolia. Despite obscurities, these Turkish dialects are often the only surviving testimony for the ethnic and linguistic past of the region. The roots of the selected Oriental names have been checked through the dictionaries of Middle and Modern Greek, as well as some Greek dialects (Pontic, Cypriot, Cappadocian), to verify their inclusion in Greek vocabulary of the time and region. That they were borrowed by Middle or Modern Greek and regional dialects in itself represents additional confirmation. Factors in the inclusion of Arabic, Persian, Turkic, and Mongol lexical elements in Middle Greek are as follows:
1. Usually, but not always, the original accent is preserved. The words of Arabic origin entered the Greek language in Iranicized or Turkicized phonetic form with the accent on the last syllable, regardless of whether it was long or short in Arabic original. If the Oriental stem is followed by a Greek root or suffix, the original accent moves in accordance with the rules of the Greek language.
2. Most Oriental words acquire the ending of the first declension, with only a small proportion belonging to the second declension. If the stem ends in a vowel, the final vowel is absorbed by the ending of the first declension -ας, -ης, and the stressed syllable of the original stem in most cases acquires a circumflex accent, merging with the ending.
3. Some names do not acquire a Greek ending and in this case do not decline, which was rather common for foreign borrowings in Middle Greek; some of these names in oblique cases have endings -α and -η.
It is often difficult to know the generation of immigrants to whom the name belonged. In the case of aristocratic patronymics, distant descendants of the founder of the family, a Scythian or a Persian, still continued to use it throughout generations. Obviously, cultural differences between the first, second, and subsequent generations were very important, especially if one takes into account the effectiveness of Byzantine assimilation mechanisms. Only in regard to first-generation immigrants can one speak about the inflow of newcomers in Byzantine society, while in the case of second and subsequent generation immigrants we are dealing with an ordinary Byzantine who for some reason had a non-Greek sobriquet or patronymic. One can ascertain with confidence to which generation a person belonged only in cases where sources provide an explicit indication. Such indications are as frequent in narrative texts as they are rare in documentary sources. Owners of Oriental names from among lower classes referred to in the sources, who unlike the aristocracy had no patronymic, as a rule belonged to immigrants of the first and, less frequently, the second generation of newcomers.
Even if individuals bearing the same byname belong to the same family, in many cases the degree of relationship between them can be established only tentatively. If the kinship degree is not explicitly recorded in the sources, there is a special scheme that may give a rough idea about the nature of kinship links between them. The known names are placed in generation scales, in which one generation equals approximately a twenty-year period, given the early marriage age of the Byzantines.
The majority of Asian newcomers registered in the database were first-generation immigrants. Often the second generation, who did not use Oriental surnames any more, were indistinguishable from the indigenous population, losing those indications of origin which their fathers’ names might give us. Only incoming noble families retained references to their Asian lineage in subsequent generations.
On adopting Christianity, immigrants had to change their Muslim or pagan names to Christian ones. If an individual of low social standing did not retain his former foreign name as a sobriquet, his Asian roots are untraceable. Consequently, some purely Christian Byzantine names might well be concealing those of Asian immigrants (as well as those of other origins) who through their names assimilated to the Greek Christian majority. Only those Asian immigrants of low social standing, whose nickname was related to their original foreign name, are traceable, as this sobriquet had become an indispensable element of their personal identification. The Asians completely changing their names to standard Christian ones (or Greek and Slavic), as well as second-generation Asians who had lost their foreign sobriquet, had become persons of “concealed identity”. Christian and Greek names, adopted by foreign immigrants, conceal from a researcher the ethnic identity of their owners.
These limitations lead us to suggest that only some Turkic immigrants preserved their Turkic nicknames, while many others, having adopted Christian, Greek, or Slavic names, became “concealed”. Consequently, the real number of Asian immigrants in Byzantine lands was most likely considerably higher than the figures given in this study. In this sense, my database reflects only the tip of the iceberg. What, at first, seems to be a modest number of Oriental names does not reflect the real extent of Asian presence in Late Byzantine society. It is plausible that unrecognizable Asians outnumbered those whose names or biographic data reveal their Asian origin.
One cannot exclude that some Oriental names in the database in fact belonged to Greeks, Slavs, Latins, or something else. However, the majority of cases registered in my anthroponymic database reflect with certainty a direct correlation between the origin of names and the ethnic affiliation of their owners.
Ancient Greek culture widely used ethnic names and, in classical times, there was usually a direct link between the ethnic origin and the origin of the name. There is no evidence of any radical change in this sense in Byzantine times.
Byzantines distinguished among Turkic nations two largest taxa: “Scythians” (Dunabian and northern Black Sea Turks and the Mongols) and “Persians” (Anatolian and Iranian Turks).
The differentiation between “Scythian” and “Persian” settlers is not usually a problem for the data gleaned from Byzantine narrative sources. Byzantine authors, as a rule, were sensitive to this basic distinction and noted it. The problem, however, arises when handling anthroponymics and microtoponymics in documentary sources. Compilers of the acts were not concerned with the “ethnic” origin of a person mentioned in an official document. If the person was a Roman citizen, that is, subject to Roman law, the law was not interested in his or his ancestors’ ethnicity. In the case of documentary sources, when defining the provenance of the bearer of an Oriental name, one may rely only on etymology of the name itself. With some reservations, I consider purely Turkic names as likely belonging to “Scythians”, while Arabic and Persian names are regarded as belonging to “Persians”.
(Source: “The Byzantine Turks, 1204-1461”, by Rustam Shukurov)
Research-Selection for NovoScriptorium: Anastasius Philoponus