The structure of Romani
Romani dialects differ in their structure. It is therefore difficult to provide a brief structural characterisation taking into account the full dialect variation.
For a brief structural overview of Romani, download Matras’ description of the language for Encyclopedia of Languages and Linguistics, or download Matras’ description of the Parakalamos Romani dialect, as an example of a concise description of an individual Romani dialect. For a comparative overview of Romani dialects, see Matras’s (2002) book Romani: A linguistic introduction.
The present section provides a reconstructed overview of Early Romani, the forerunner of the present-day dialects of Romani. The description is based on chapter 5 of Elšík, Viktor & Matras, Yaron, Markedness and language change: The Romani sample, Berlin: Mouton de Gruyter, 2006.
The purpose of this outline is to survey the structures that constituted the point of departure for the variation which we find today among the varieties of the language. Early Romani is not attested, and one must rely on dialect comparison as well as on comparisons with other Indo-Aryan languages in order to identify archaic structures among those that are still in existence, and postulate possible forerunner forms. The comparative method has been applied in such a way for Romani since the late 19th century (cf. discussion in Matras 2002, Ch. 3), and has in recent years experienced a revival in the modern context of Romani dialectology and typology (cf. Matras, Bakker & Kyuchukov 1997, Elšík & Matras 2000, Matras 2002). We refer to this work for the detailed arguments, and limit ourselves here to a presentation of what we consider to be reasonable outcomes of reconstruction scenarios.
We define Early Romani (ER) as the latest stage in the history of the language prior to the dispersion of Romani-speaking population groups throughout Europe and the consequent split into dialects and dialect families (see also Matras 2002). Despite the absence of any documentation or even attestation, we have both linguistic and extralinguistic clues concerning the time and location where ER was spoken. The earliest documentary evidence of an expansion of the Romani population into the northern Balkans, and beyond that, into central Europe, dates from the late fourteenth century (cf Fraser 1992 for a useful overview). A strong Greek element is shared by all dialects of Romani. It includes lexical vocabulary, grammatical vocabulary (function words, including numerals, indefinite pronouns, particles and adverbs of time and phase), derivational morphology, and patterns of syntactic typology that are most likely to have emerged in contact with Greek (such as the formation of relative clauses, a preposed definite article, verb-medial word order, and a split between factual and non-factual complementisers). The most striking evidence of Greek influence is, however, the incorporation of Greek nominal inflectional class markers (for subject-case nouns and adjectives) and of Greek verb inflection class markers (including tense-aspect marking, participial inflections, and marginally also subject concord inflections for the 3SG). All this points to the powerful influence of Greek, conditioned in all likelihood through a prolonged stay in Greek-speaking territory in Byzantium (see already Miklosich 1872-1880: III). This makes it convenient to define ER as the period during which Romani was in contact with Greek, prior to the decline of Byzantium and the beginning Romani migrations out of the Balkans and the subsequent split into isolated dialects.
As for the location of ER, one should bear in mind that Byzantium of the tenth-fourteenth century did not overlap with today’s Greek-speaking area, but stretched as far as Anatolia and beyond; indeed Greek was spoken in Anatolia until the first decades of the twentieth century. This detail is sometimes overlooked in discussions of Romani history, but it could be crucial toward an understanding of the time line of Romani migrations westwards from India, and how this time line can be reconciled with linguistic evidence. Conventionally, the presence of Iranian and Armenian loans in Romani is taken as evidence of prolonged settlement periods in Iran and Armenia. For the latter, the present-day Armenian Republic north of the Ararat is often taken as the point of reference, and migration routes are pictured as having led north from Iran to Armenia, then south again along the Black Sea coast, to present-day Greece. But in tenth-century East Anatolia, Roma would have had speakers of Greek, Armenian, and Iranian languages (such as Kurdish, but possibly also Persian contacts) as their neighbours. It may therefore be quite difficult to draw a clear-cut distinction between Romani in contact with western Asian languages, and Romani in contact with Greek, and there could well have been a period of overlap, and so possibly also a rapid, rather than gradual migration from India to Byzantium.