Genesis and structure of Romani in historical context of Indo-Aryan languages

Romski je jezik indijski jezik koji je već stoljećima raširen po Europi i mnogim dijelovima svijeta. To je novoindoarijski jezik, koji s jedne strane ima vrlo konzervativna indoarijska obilježja, a s druge strane posebna obilježja koja ga bitno izdvajaju od drugih indoarijskih jezika. Cilj je ove disertacije prikazati proces njegova razvoja iz staroindoarijskoga (vedskoga, sanskrta) preko srednjoindoarijskoga (prakrta) te njegovo izdvajanje i udaljavanje od indoarijskih jezika nakon napuštanja Indijskoga potkontinenta, i to pomoću usporedbe romskoga s drugim novoindoarijskim jezicima, posebice hindskim, uzimajući u obzir i utjecaj niza ne-indoarijskih jezika na romski tijekom njihova kontakta. Cilj je također jasno i sustavno prikazati ustroj romskoga jezika kakav je danas. Sadržaj disertacije sastoji se od tri dijela. Prvi dio čine kratak pregled dosadašnjih istraživanja o Romima i romskome jeziku, kratak opis podrijetla i migracija Roma, opis položaja romskoga kao novoindoarijskog jezika unutar indoarijske potporodice jezika te klasifikacija njegovih dijalekata. Drugi odnosno glavni dio doktorskoga rada obuhvaća opis cjelokupne romske gramatike i leksika, od fonologije preko morfologije do sintakse, s temeljitim prikazom njihova povijesnoga razvoja, te utjecaja kontakntih jezika na njegovu gramatiku i leksički inventar. Treći se dio doktorskoga rada sastoji od cjelovitog i sažetog pregleda ustroja romskoga jezika, pri čemu je uveden djelomično nov način usustavljivanja kao i nešto nove terminologije. Svrha je ovoga dijela rada u odmaku od dijakronije dati jasan uvid u romsku gramatiku sa sinkronijske perspektive, iznijeti nov, sustavniji prikaz romskoga gramatičkog ustroja, koji bi također i olakšao sâmo učenje romskoga jezika. Takav prikaz povijesnoga razvoja jezika te njegove današnje strukture nuždan je i za pothvat standardizacije romskoga jezika na globalnoj razini.1. Romani language as Indo-Aryan language The Romani people, the Roma, and their language have been the subject of scientific interest since the 16th century. The scholarly research into the Romani language began in the 18th century, while the study of the Romani linguistics and the Roma was founded in the 19th century. To date, numerous works have been written on Roma history, society, culture, literature, and language, including dictionaries and grammars. Some of them describe a particular dialect, others aim at standardizing the Romani language. Today, a number of dictionaries, lexical databases, and grammars are also available digitally. In the 18th century, due to the similarities between Romani and Hindi, it was agreed that Romani is an Indian language and that the Roma originated from the Indian subcontinent. The Romani language belongs to the Indo-Aryan branch of the Indo-Iranian subfamily of Indo-European languages. It is a New Indo-Aryan (NIA) language that evolved from Old Indo-Aryan (OIA) through Middle Indo-Aryan or Prakrit languages (Pkr), similarly as all other New Indo-Aryan languages did. However, departure from its ancient homeland of India during the Middle Indo-Aryan linguistic phase, as well as the contact with numerous other languages resulted in its present unique structure and features. Therefore, Romani today conservatively preserves some original Indo-Aryan features, which are significantly closer to Prakrit and Old Indo-Aryan (preservation of some phonemes and consonant groups, synthetic verb forms, elements of grammatical structure), than in other New Indo-Aryan languages whose speakers have remained in the Indian subcontinent. On the other hand, Romani has distinct features that clearly distinguish it from the rest of the Indo-Aryan languages (its phonological system largely corresponds to the phonological system of contact languages, including the loss of retroflex consonants; new parts of speech such as prepositions and articles; syntactic order S-V-O; vocabulary containing Middle Persian loanwords, as well as Armenian, Greek, Romance, Slavic, etc. loanwords). On the basis of the the sound changes in Romani, R. L. Turner proved that the Roma people had originated from the area of the central part of northern India. Around the 4th century BC, they migrated to the northwest of the Indian subcontinent, and then, probably around the middle of the millennium, from India to Persia, where they may have settled until the middle of the 7th century. Thereafter, the Roma stayed in Armenia until about the 10th century and in Byzantium until about the 13th century before moving to other parts of Europe and, more recently, to other parts of the world. While the Romani language in the Byzantium seems to have been a unified common form of the Early Romani, the later migrations have led to a considerable diversification of the language, both through the influence of contact languages and through internal innovations, so that today there are many Romani dialects. Because of the intense migrations of the various Romani groups, their mutual contact, and interdialectal influences, it is very difficult to determine with certainty the exact genealogy of these dialects that would satisfy all scholars and lead to the agreement among them. 2. Phonology As mentioned above, Romani phonology remained in some respects more conservative than the phonology of other New Indo-Aryan languages, where the differences between sibilants were lost, as well as single intervocalic consonants, and great simplification (assimilation, etc.) of consonant groups took place. On the other hand, the Romani language lost some older Indo-Aryan phonological features due to the influence of the contact languages. For example, there was a loss of retroflex consonants (ṭ(h), ḍ(h), ṇ), because they do not occur in non-Indo-Aryan contact languages, like Persian, Armenian or European languages. In some dialects the loss of aspirates is also in progress (ph > p, th > t, kh > k, čh > č). Moreover, some new phonemes have entered Romani phonological system – f, x, c, z, ž, depending on the dialect and the phonology of the language the dialect was in contact with. 3. Morphology The declension system of Romani also differs in part from the declension system of other New Indo-Aryan languages. On the one hand, the declension of adjectives in Romani is almost identical to that of Hindi, with the endings -o, -i, -e, and the declension of nouns and (to some extent) pronouns in Romani coincides with New Indo-Aryan, in so far as it is characterized by the addition of postpositions. On the other hand, it is important to emphasize that a significant typological change has taken place in Romani – postpositions have turned into agglutinative case endings. This can be observed, for example, in the phrase “with mother, father, sister and brother,” which is read in Hindi as mātā jī, pitā jī, behen aur bhāī ke sāth. At the very end of this phrase is a single postposition meaning “with” which refers to all nouns in the phrase. In Romani, the same phrase reads e dajasa, e dadesa, e phenjasa thaj e phralesa. Each noun in the phrase ends in -sa, which means “with”. In this case, -sa is not a postposition, but a case ending for the instrumental. In addition, the Romani language has developed a category of prepositions, which is not typical of the New Indo-Aryan languages. Since these prepositions also convey case meanings and in Romani are usually associated with the nominative form (sometimes with the locative and genitive forms) of nouns, there is a tendency to replace the case forms by prepositional constructions, which means that Romani noun inflection is currently moving from synthetic to analytic inflection in a number of dialects. The Romani partly inherited and partly borrowed affixes to derive forms of different word classes and to compare adjectives. In addition, there were similar changes in morphology by analogy, both in the formation of the intransitive perfect participle and in the comparison of adjectives. Old Indo-Aryan has a comparative suffix -tara, as in ucca-tara, which is a comparative of the Old Indo-Aryan word ucca, meaning “tall”. In Romani, the comparative suffix is -eder, as in the comparative form učeder. It arose from the Old Indo-Aryan suffix and merged with the final vowel of the Old Indo-Aryan adjectival base ending in -a, giving rise to: Old Indo-Aryan ucca-tara > Rom. uče-der > uč-eder (OIA -a-tara > Rom. -eder). By analogy, comparatives of other adjectives were formed in this way, e.g., šukareder from šukar meaning “beautiful”, giving rise to the comparative suffix -eder (OIA śukra-tara without this analogy and with assimilations would give *šukeder). Romani is the only Indo-Aryan language with a definite article. In numerous dialects, the definite article can be o, i, and e (often le). The definite article was derived from the ProtoRomani and Early Romani demonstrative pronouns (derived from the OIA so, eta-), which came in use as definite articles while the Roma lived in Byzantium, due to intense contact with the Greek language, which has a definite article in its grammatical system (Romani definite articles were vocally assimilated with Greek articles). Since the articles precede the noun, usually in combination with prepositions, articles and prepositions often merge, e.g., an-o ker, meaning “in the house”. Demonstrative pronouns in Romani are a particularly complex set of word class with opaque derivation. They show traces (in their central element) of the Old Indian demonstrative pronoun (eta-), but Romani experts emphasize that the vowel a predominates in pronouns referring to what is near, while o predominates in those referring to what is far, e.g., adava, akava, kadava, etc. versus odova, okova, kodova, etc. Matras has attempted, as far in detail as possible, to reconstruct the emergence of the present demonstrative pronouns in three phases during the Proto-Romani period. In my opinion, the initial a- can be derived from the OIA pronoun aya(m) “this” (> *ā > Rom. a), and initial o- from the OIA pronoun asau “that” (as already derived by Sampson for personal pronouns and the article; asau > *aho > Rom. o). The second element of pronominal compound stems (*lo etc.) may be derived from the OIA base eta- “that”. As the third element – as formally (but without interpretation) envisaged by Matras, in his ‘third phase’ – it seems to me that the Romani adjectives kaj “where” and daj “somewhere” (after a- or o-) have been added to the beginning of the already compound base of the pronoun, to denote (as Matras says) “definite” and “indefinite” deixis. Finally, it is through analogous generalizations and polarizations of vowels a or o in all internal syllables that from *o-da-va, *o-ka-va the forms odova, okova or adava, akava, etc. must have arisen. Although this interpretation of mine is hypothetical, it may bring a great deal of morphological and semantic order and meaning into the confusing variety of demonstrative pronouns in Romani. The analytical derivation of complex numbers is yet another feature that distinguishes Romani from other New Indo-Aryan languages. In other New Indo-Aryan languages, numbers are simply inherited from Old Indo-Aryan and became – through a series of sound changes in Prakrit – quite intransparent, e.g., Hin. caudah “14” Pkr. *-a-mano > Rom. -amno > Rom. -avno). In Romani, however, these participles have shifted meanings and often become adjectives and generate adjectival nouns, such as dikhavno “careful”, which as a noun means “supervisor” and is derived from the verb dikhel “to look”. Thus, its literal meaning is “looking”. Although these adjectives are not used in everyday speech with their literal participle meaning, and are classified as adjectives in the literature, it should be taken into consideration that they are derived exclusively from verbs. Therefore, and for the sake of systematicity of the description of the Romani language in historical perspective, they should be singled out as a separate category and defined as (originally) present (middle) participles. The grammatical structure of verb inflection in Romani is distinctly synthetic. It partly preserves the old synthetic forms (in the present tense), and partly has created new synthetic forms from the Middle Indo-Aryan analytic forms (in the perfect tense and passive). It differs from the structure of other New Indo-Aryan languages, which is almost completely analytic. Thus, in Hindi, a complete restructuring of the verb inflection system occurred in the New Indo-Aryan phase, so that the forms of verb tenses are constructed analytically, using present participles or intransitive perfect participles with the verb “to be” as copula. Also, there are distinctive complex verb compounds, which may consist of two or more verbs in different forms. These compound verbs consist of a basic verb that provides the basic semantic information, of a modifier verb that determines perfectiveness/imperfectiveness (whether the action is finished or not) and modes of action (Aktionsart), possibly followed by a modal verb in a participial form, and finally, an auxiliary verb that informs about tense or mode, person and number. The only synthetic verbal forms in Hindi are verb modes – conjunctive, imperative, and conditional to some extent. Unlike in Hindi, the present (and imperfect) in Romani is formed synthetically, using both the base and endings inherited from Old Indo-Aryan. The (present) perfect (and pluperfect) on the other hand, is formed from a perfect base derived from the intransitive perfect participle. Endings derived from the copula si are added to the base, but as stated, they may have preserved some traces of the agentive pronoun. The fusion of this perfect base and the endings gave rise (in the Early Romani period, i. e. during the stay in Byzantium) to new synthetic forms of perfect (and pluperfect) from the preceding analytic forms. The future tense can also be constructed either by the agglutinative suffix -a or analytically by means of the auxiliary verb kamel, or the particles kam or ka derived from it. Modal constructions consist of the main verb forms and modal verbs or particles (conjunctions). The new synthetic verb forms derived from the analytic ones, such as the forms of the perfect tenses, as well as the passive, arose, as mentioned, from the old periphrastic constructions of the intransitive perfect participles and auxiliary verbs, e.g., *gelo sem (gelo “gone”, sem “am”) > gelem “I have gone” (perfect); *phučlo ovel (phučlo “asked”, ovel “is”) > phučlol “(he / she) is asked” (passive). The participles and forms of the auxiliary verb si “to be”' have merged into distinctive perfect forms, which can now be analysed in the perfect verb bases and perfect endings. This is the reason why the perfect endings partially coincide with the present tense forms of the verb si “to be”. (However, they could also partially preserve the traces of agentive pronouns.) The same thing happened with the construction of the passive, except that the participle merged with the forms of the auxiliary verb ovel “to become, to be”, so that today they can also be considered (new) synthetic verb forms derived by suffixes. Denominatives are derived in Romani from adjectives and nouns following the same pattern, so that their bases merge with the auxiliary verb ovel, e.g., *baro ovel (baro “big”, ovel “becomes” > barjovel “grows”). Similarly to Hindi, Romani has no special verb meaning “to have” and therefore it uses a construction with the verb “to be” and a logical subject in the form of an oblique case/accusative, e.g., si man kher “I have a house”. The origin of this construction is not completely clear. One possibility is that this ‘accusative’ form is an Old Indo-Aryan genitive, from which most of the oblique forms emerged, so that the literal meaning of si man kher would be “(of (me) is the house)”. And the second possibility is that it is the influence of the Persian construction, which also consisted of a logical subject in the accusative and the verb “to be”: ma rā (...) ast “I have”. Unlike Hindi, the adverbial system in Romani is also distinctly synthetic. Some of the adverbs were inherited from Old Indo-Aryan, especially the adverbs ending in -e and -al, e.g., upre “up”, upral “from above”; some of them are being derived from adjectives with the suffix -e(s), e.g., lačho “good (masculine)” > lačhe(s) “well”; part consists of adverbial expressions merging prepositions with nouns, e.g., džibrše “next year”; and part consists of loan words, e.g., tehara “tomorrow” (from Greek). Some phrases were derived from adverbs of Middle Indo-Aryan origin, such as sar “how” > “as, like”; upre “up” > upro “over”; paše “near” > paša “at, next to”, etc. Romani also forms the adverbial present participle by the suffix -indo/-ando (etc.), which originates from the Old Indo-Aryan present active participle ending in -ant, cf. OIA hasant- > Rom. asando “laughing”. Most Romani conjunctions are loanwords, such as Slavic pa “so”, niti “neither”, a “and, while”, ali “but”, Romanian numa(j) “but” or Tur