TDD/JofEL Winter 2013 (cid:1) Stavros Skopeteas BIELEFELD UNIVERSITY THE CAUCASIAN URUMS AND THE URUM LANGUAGE KAFKASYA URUMLARI VE URUM DİLİ Abstract Urum people identify themselves as Turkish-speaking Pontic Greeks who left Anatolia at the beginning of the 19th century. A major group immigrated to the highlands of K’vemo K’art’li, where they are still living today. They conserved the variety of Turkish that their ancestors were speaking before emigration and enriched their language by influences from the languages in their new environment, in particular from Russian. The Urum language displays substantial similarities with the Turkish dialects of Anatolia; beyond these similarities, it displays some unique developments (e.g., in vowel harmony) as well as properties that are traced back to the influences from Russian (e.g., in the use of subordinate clauses). Key Words Urum, Anatolian dialects, phonology, morphology, lexicon Özet Urum halkı kendilerini 19. yüzyılın başlarında Anadolu’dan ayrılan Türkçe konuşan Pontus Rumları olarak tanımlar. Bu halkın büyük bir bölümü, bugün halen yaşamakta oldukları K’vemo K’art’li’nin dağlık bölgesine göç etmiştir. Bu grup, atalarının göç öncesinde konuştukları Türkçenin varyasyonunu, göç ettikleri yeni çevrede bulunan dillerin (özellikle Rusçanın) etkisiyle zenginleştirilmiş bir biçimde muhafaza ettiler. Urumca Anadolu’daki Türkiye Türkçesinin diyalektleriyle önemli benzerlikler gösterir. Bu benzerliklerin ötesinde, bu dil, Rusçanın etkisine kadar geri götürebilecek (bağımlı cümlelerin kullanımında olduğu gibi) bazı özelliklerin yanı sıra kendine özgü (ünlü uyumlarında görüldüğü gibi) gelişmeler de sergilemiştir. Anahtar Kelimeler Urum, Anadolu diyalektleri, Ses bilgisi, Biçim bilgisi, Söz Varlığı TDD/JofEL Winter 2013 • Tehlikedeki Diller Dergisi/Journal of Endangered Languages Stavros Skopeteas-Caucasian Urums and Urum Language 1. Preliminaries This chapter is devoted to Caucasian Urum, a language spoken in the highlands of K’vemo K’art’li in Georgia. The basic substrate of Caucasian Urum is Anatolian Turkish; Urum people are bilingual in Russian and the currently spoken language has a great amount of borrowings from Russian, which hinders the mutual intelligibility with Turkish as spoken in Turkey. The Urum language spoken in the Caucasus has to be distinguished from the Crimean Urum spoken in Ukraine. Both linguistic communities share the same ethnonym (see Section 2.1) but this does not mean that their languages are immediately related (see Section 0). The present article is devoted to Caucasian Urum and outlines the history and current situation of the language community (Section 2) as well as the basic properties of the language of these people (Section 3). The data presented in the following were collected with native speakers in Tsalka and Tbilisi between 2009 and 2013. The cited examples are either elicited by translation (Verhoeven, Moisidi & Yordanoglou 2011) or collected from texts (Skopeteas and Moisidi 2011). Further intuition data and clarifications about data questions in corpora were elicited with Violeta Moisidi, native speaker of Urum.1 2. Caucasian Urums 2.1. Urum Ethnonyms The ethnonym Urum comes from Turkish Rum. The term originates to the Greek stem rom- ‘Roman’ (referring to citizens of the Eastern Roman empire). Rum-i-Millet in the Ottoman empire was the millet of Christian Orthodox people in the Empire – referring to a religious and not to an ethnic community. The word Urum involves a prothetic u- that generally appears in loanwords with an initial r-, e.g., u-rus ‘Russian’, u-ruset ‘Russia’ (the same words are attested in Anatolian Turkish, see urum, urus, and urusya in dialects of Erzurum; Gemalmaz 1978[III]: 318). Native speakers call their language Urum dili ‘language of Urum’ and identify themselves as Turkish- speaking Pontic Greeks (see Höfler 2011 on ethnic identity issues). Georgians use either the ethnonym Urum-eb-i (Urum-PL-NOM) or the ethnonym for Greeks berdzeni, which is used for any ethnic Greek people including the Greeks living in Greece (Sideri 2006: 26). Pontic Greek speakers of Georgia call the Urum language to turkikón ‘the Turkish (language)’, but conceive the speakers of Urum as homoethnic. 2.2. Religion Urum people are Christian Orthodox and practice their religion in Greek, Georgian and Russian churches (Karagyosov 2006; Höfler 2011:65f.). Although there are no liturgical practices in Urum, many native speakers report that they use Urum in praying (18 out of 30 Urum speakers in sociolinguistic interviews; see Sella-Mazi and Moisidi 2011). The use of Urum in religious practices is reflected in the lexicon. Urum vocabulary displays 23% loanwords in the field of religion/belief, which is less than the average proportion of loanwords in the world’s languages in religious terms – 1 I am particularly grateful to Violeta Moisidi (Tbilisi) for the Urum data collection that she compiled since 2009 and to Emrah Turan (Bielefeld) for his assistance in the comparison between Urum and Turkish. The findings about the Urum lexicon and the current language situation are the product of common research with Veronika Ries. The present article is part of the project ‘The impact of current transformational processes on language and ethnic identity: Urum and Pontic Greeks in Georgia’ funded by the VW- foundation. www.tehlikedekidiller.com 334 TDD/JofEL Winter 2013 • Tehlikedeki Diller Dergisi/Journal of Endangered Languages Stavros Skopeteas-Caucasian Urums and Urum Language estimated to 43% according to the data of the WOLD project (see Haspelmath and Tadmor 2009 for cross-linguistic facts; Ries et al. 2013 for a detailed account of the Urum vocabulary; see also 3.4 for further discussion on the lexicon). This is surprising for a Christian community but fits to the speakers’ reports that they use Urum for religious practices. The majority of the concepts about religion and belief are expressed with words of Turkish origin, e.g., allah ‘god’ or cänäm ‘hell’ (compare Turkish allah, cehennem). Russian loanwords are restricted to narrow Christian terms, e.g., gimn ‘hymn’ from Russian gimn (Гимн), episkop ‘bishop’ from Russian episkop (Епископ) (Ries et al. 2013). Note that these words are of Greek origin but are transferred in Urum in the form used in Russian; compare Russian gimn and Greek ímnos, Russian episkop and Greek epískopos. Very few terms in this field come directly from Greek, e.g., hristugin ‘Christmas’; compare Greek hristujena (Χριστούγεννα) and Russian raždjestva (Рождество). 2.3. Geography and Population The settlement areas of Urum speakers are located in the highlands of K’vemo K’art’li, in particular in the villages around the lake of Tsalka (Lat 41.6; Lng 44.1), as well as in Tetri Tsqaro (Lat 41.5 Lng 44.5) and Dmanisi (Lat 44.1 Lng 41.3). The villages around Tsalka were among the few places in the former Soviet Union in which ethnic Greek people made up more than 50% of the population (Kolossov, Galkina, and Krindatch 1998: 108). According to the 1979 census of the Georgian SSR, the ethnic Greek population in the district of Tsalka amounted to 30 811 people, whereby the vast majority of registered ethnic Greeks in this district are Urum speakers. The population shrank rapidly in the last decades as a result of the massive migration to the urban centres of Georgia (mainly Tbilisi), and from there to further destinations (Russia and Greece being the most popular targets). The majority of ethnic Greek people emigrated outside the country as documented in the counts of the National Statistic Office of Georgia, which reports a Greek population of 100 300 citizens in 1989 (1.9% of the population total) and 15 200 in 2002 (0.3% of the population total) (see National Statistics Office of Georgia 2011: 20). The ethnic Greek people of Tsalka totalled 30 811 people in 1979, 4 589 people in the 2002 census and were estimated to 1 500 people in 2005 (Wheatley 2006: 8). There are not more recent counts; the Federation of Greek communities in Tbilisi estimates that 1 000-1 500 Greeks are currently living in Tsalka and the surrounding villages. 2.4. Historical Background Caucasian Urum speakers originate in the Turkish-speaking Greek populations of Northeastern Anatolia. The settlements of these people before their emigration to the Caucasus included several cities: Kars, Giresun, Erzurum, Trabzon, Kümbet, Bayburt, and Gümüshane (see Xanthopoulou- Kyriakou 1991, Eloeva 1998, Kasapoğlu Çengel 2004: 59, Altınkaynak 2005: 39, Kalayci 2008: 144). Linguistic comparison to the dialects of Erzurum shows that the Turkish substrate of Urum displays the characteristic properties of this region (see Section 3.1). Greek populations came to the Caucasus during several waves of emigration from the beginning of the 19th century onwards (the oldest reported migration took place at the end of the Russo-Ottoman war of 1928-1929, see Fonton 1840; further migration waves are reported in association with the Crimean War, 1853-1856, and the last Russo-Ottoman war 1877-1878, see Xanthopoulou-Kyriakou 1991, Kalayci 2008: 144). In Georgia, Urum people settled in several places in K’vemo K’art’li (see Section 0). Historical sources mention 6 000 families that arrived in Tsalka and Akhaltsikhe at the end of the first Russo-Ottoman war (see Sideri 2006: 56). www.tehlikedekidiller.com 335 TDD/JofEL Winter 2013 • Tehlikedeki Diller Dergisi/Journal of Endangered Languages Stavros Skopeteas-Caucasian Urums and Urum Language Urum people live in a multilingual community and are themselves competent in different languages. They were in contact with Russian after arriving in the Russian Caucasus, which was the language of administration, education and in many cases of liturgical practices both during the Tsarist regime as well as in Soviet period (see Höfler 2011: 144f.). The impact of Russian on the language use of the Urum people is already known from early documents (see Sideri 2006: 144f.). A recent questionnaire-based sociolinguistic study (30-person sample, residents of Tsalka and Tbilisi) revealed that 93% of the Urum speakers are also competent in Russian (28 persons), 83% (25 persons) are competent in Georgian, and 33% are competent in Greek, which they either acquired in language courses in Tbilisi or during their visits to Greece (see Sella-Mazi and Moisidi 2011: 33). In the Tsalka district, Urum people were also in contact with the Armenian population, which was the second largest minority in this area (see demographic data in Wheatley 2006: 8). In the afore-mentioned sociolinguistic study, 6 out of 30 persons (20%) report that they also use Armenian in contact with friends. 2.5. Orthography There is no writing tradition in Urum. The majority of the speakers are alphabetized in Russian (and less in Georgian) and they are not writing in Urum (Kock Kobaidze 2001: 155). Turkish translations of holy texts in the Greek alphabet (printed in Istanbul) are available in the community but they are not used in religious practices. Epitaph inscriptions in Tsalka cemetery dating back to the beginning of 20th century are written in Greek, which indicates that at least some members of the community had some knowledge of Greek and also acquired writing skills in this language; see example in (1). (1) 1859-1918 <person name> ΕΝΤΑΥΘΑ ΑΝΑΠΑΒΕΤΕ entautha anapavete here rest:3.SG Ο ΔΟΥΛΟΣ ΤΟΥ ΘΕΟΥ o doulos tou theou DEF:NOM.SG.M slave:NOM.SG.M DEF:GEN.SG.M god:GEN.SG.M ‘1859-1918; <person name>; the slave of Lord rests here’ (transcribed from the photographical collection of George Zosimidis) 3. Urum language 3.1. The Place of Urum among the Turkic Languages The historical facts indicate that the ancestors of the Urum came from several places in Anatolia (see Section 0). The Urum people in Georgia share their ethnonym and probably their origin with a community living in Ukraine (settled originally in the Crimea, and later in the neighbouring Azovian region). Some sources assume that these communities speak the same language (see Podolsky 1986: 100, Uyanık 2010, see also ethnologue report for Urum, Lewis 2009, ed.), which is not supported by the available linguistic data. Caucasian Urum is a variety of Anatolian Turkish with substantial influence of Russian. Crimean Urum, as documented in the lexicon of Garkavets (2000) and the grammatical sketch by Podolsky (1986), is a Turkic language with different substrates – especially influenced by the Turkish spoken by the Crimean Tatars – and shows lexical and grammatical properties that substantially differ from the Urum language spoken in Georgia. For instance, the www.tehlikedekidiller.com 336 TDD/JofEL Winter 2013 • Tehlikedeki Diller Dergisi/Journal of Endangered Languages Stavros Skopeteas-Caucasian Urums and Urum Language contrast between front/back non-rounded vowels is neutralized in Caucasian Urum but not in Crimean Urum (see Verhoeven 2011), Crimean Urum displays local cases (inessive and elative) that are not available in Caucasian Urum or in Turkish, etc. The Turkish substrates of both languages shows some phonological similarities which indicates that both populations originally spoke at least close-related varieties of Turkish; see (2) below. A small-size study on the lexicon reveals that Caucasian Urum is closely related to Turkish, as presented in Figure 1 (calculations were made in SplitTree4; version 4.13.1). The study is based on a sample of 137 words of the basic vocabulary (part of the Swadesh list) for which we examined cognates from five languages of interest: the aim is to estimate the relation of the two languages known as Urum (Caucasian and Crimean Urum) with Turkish (Standard variety) and with another related language that is very close to the Anatolian varieties of Turkish, namely Azerbaijanian. The comparison to Azerbaihanian is particularly relevant for the Caucasian Urum, because there are claims that Urum and Azeri people in Georgia speak one and the same language (Kock Kobaidze 2001: 154-157). Finally, Tatar was included as a control condition, in order to estimate the distances between the languages of interest in comparison to a remote Turkic language.2 The figure shows the relations between these languages in terms of a network reflecting the frequency of words of common origin. The interesting finding is that Caucasian Urum is classified next to Standard Turkish, which means that cognates are more likely between Caucasian Urum and Standard Turkish than between Caucasian Urum and Azerbaijanian. Figure 1. Urum and Turkic languages Some properties of Caucasian Urum directly relate to the Anatolian Turkish dialects, in particular to the dialects spoken in Kars and Erzurum (see Kasapoğlu Çengel 2004, Uyanık 2010). The phonological properties illustrated in (2) are reported for the dialects of Erzurum (Karahan 1996; cited from Menz 2002:199f.) and are also found in the Urum vocabulary. 2 Azerbaijanian and Tatar words were collected from Öztopçu et al. (1999), Crimean Urum words from Garkavets (2000), Caucasian Urum words from Skopeteas et al. (2011). Turkish words were provided by Emrah Turan. www.tehlikedekidiller.com 337 TDD/JofEL Winter 2013 • Tehlikedeki Diller Dergisi/Journal of Endangered Languages Stavros Skopeteas-Caucasian Urums and Urum Language (2) Common phonological properties between Caucasian Urum and Erzurum dialects a. Caucasian Urum h ~ Standard Turkish k3 halh ‘people’ (Standard Turkish halk; Erzurum halk/halh); bah-tı ‘see-PST[3]’ (Standard Turkish baktı); uzah ‘near’ (Standard Turkish uzak); gäldıh ‘we came’ (Standard Turkish geldik; Erzurum dialects:4 geldıh in Yukari Karasu, gelduh in Çoruh Boyu); harpuz ‘watermelon’ (Standard Turkish karpuz) b. Caucasian Urum ğ ~ Standard Turkish k ğıssa ‘short’ (Standard Turkish kısa; Erzurum gıssa); ğız ‘girl’ (Standard Turkish kız; Erzurum gız); ğuš ‘bird’ (Standard Turkish kuş; Erzurum guş); ğardaš ‘brother’ (Standard Turkish kardeş; Erzurum gardaş); čoğ ‘much’ (Standard Turkish çok); ğırmızı ‘red’ (Standard Turkish kırmızı); ğoyün ‘sheep’ (Standard Turkish koyun; Erzurum ğoyun); pičah/bičağ ‘knife’ (Standard Turkish bıçak; Erzurum piçah) c. metathesis yarpah ‘leaf’ (Standard Turkish yaprak; Erzurum yarpah/yarpağ; Azerbaijanian yarpaq), torpah ‘soil’ (Standard Turkish toprak; Erzurum torpah/toprah; Azerbaijanian torpak); körpi ‘bridge’ (Standard Turkish köprü; Erzurum körpi/köprü/körpü; Azerbaijanian körpü); äksi ‘old’ (Standard Turkish eski), soram ‘then’ (Standard Turkish sonra; Erzurum sona/sonam/soram, etc.); oğlu/ oğul ‘son’ (Standard Turkish oğlu) d. development of velar nasals (out of nasals) dängiz ‘sea’ (Standard Turkish deniz), donguz ‘pig’ (Standard Turkish domuz; Erzurum donguz/donuz), ingil- ‘groan’ (Standard Turkish inle-) Only a part of these properties equally applies to Crimean Urum, which supports the view that these varieties developed independently of each other. Fricativization of velar stops, (2a), is attested both in syllable-final and syllable-initial contexts (similar facts are also found in Crimean Urum halh/alh ‘people’, uzah ‘near’, etc.; Garkavets 2000). In word-initial contexts and preceding back vowels, the velar fricative is frequently voiced in Caucasian Urum, (2b), but is consistently voiceless in Crimean Urum (e.g., Caucasian Urum ğuš vs. Crimean Urum huš ‘bird’). Words with metathesis in Caucasian Urum do not necessarily involve metathesis in Crimean Urum: Caucasian Urum yarpah ‘leaf’ vs. Crimean Urum yapalah ‘leaf’ (Garkavets 2000). The velar nasals found in Caucasian Urum do not occur in Crimean Urum: deniz ‘sea’, domuz/donuz ‘pig’ (Garkavets 2000). Some further differences of Caucasian Urum to Standard Turkish do not relate to local properties of Anatolian dialects but occur in spoken Turkish varieties of several regions, e.g., voicing of voiceless consonants, e.g., barmah ‘finger’ (Standard Turkish parmak, Erzurum barmah/barmak); biš- ‘cook’ (Standard Turkish pişir-; Erzurum biş-); daš ‘stone’ (Standard Turkish taş; Erzurum daş), dut- ‘hold’ (Standard Turkish tut-), etc. 3 The replacement of velar stops with fricatives (h or ğ) distinguishes the dialects of inner Anatolia from the dialects of the Black Sea Coast, which generally preserve the velar stops (Brendemoen 1998: 237). This evidence is significant because the ethnic Greek populations of Black Sea Coast could be another potential historical source of the Urum people. 4 All cited examples from Erzurum dialects are found in Gemalmaz (1978[I]; 1978[III]). www.tehlikedekidiller.com 338 TDD/JofEL Winter 2013 • Tehlikedeki Diller Dergisi/Journal of Endangered Languages Stavros Skopeteas-Caucasian Urums and Urum Language At the morphological level, some suffixes in Caucasian Urum are similar to characteristic suffixes of the dialects of Erzurum, e.g., the second plural person in -sis/siz (see Table 5) and the imperfective suffix in -(i)er (see (16)), (see Menz 2002 referring to Karahan 1996). Syntactic properties that Urum shares in common with the Anatolian dialects is the use of converbs in -AndA (see (22)) as well as the frequent occurrence of the complementizer ki (see Section 0). 3.2. Phonology 3.2.1. Consonants The consonant inventory is identical to Turkish; see Table 1. The palatal allophone [c] of the phoneme k occurs immediately adjacent to a front vowel (palatalization), either before or after it, e.g., kök ‘thick’ is realized as [cœc]. The palatal allophone [ɟ] of the phoneme g only occurs before front vowels (palatalization), e.g., göl [ɟœl] ‘lake’. The velar allophone [ƚ] of the phoneme l occurs after back vowels (velarization); compare göl [ɟœl] ‘lake’ vs. yol [joƚ] ‘road’. The Urum transcription in this article generally follows the Turkish orthography. The main deviation is the use of haček for fricative and affricate postalveolar consonants (š for [(cid:2)], ž for [(cid:3)], č for [ʧ], ǰ for [ʤ]), which follows a common practice in orthographies of Turkic languages spoken in Slavic environment (see Azerbaijanian in Schöning 1998; Gagauz in Menz 1999). This practice is convenient due to the frequent Russian borrowings (Urum displays around 23% Russian words in narratives, following an estimation in Ries et al. 2013). Bilabial Labiod. Alveol. Postalv. Patatal Velar Glottal Plosive –voiced [p] p [t] t [c] k [k] k +voiced [b] b [d] d [ɟ] g [g] g Fricative –voiced [f] f [s] s [(cid:2)] š [x] h [h] h +voiced [v] v [z] z [(cid:3)] ž [ɣ] ğ Affricate –voiced [ʧ] č +voiced [ʤ] ǰ Nasal [m] m [n] n [ŋ] n Tap [ɾ] r Lateral [l] l [ł] l approximant [j] y Table 1. Consonant inventory (IPA values in brackets; orthography in italics) 3.2.2. Vowels A general issue in Turkic languages is the phonological distinction between mid-front vowels. In Turkish, /e/ has a mid-closed allophone [e] and a mid-open allophone [ɛ] appearing in word-final open syllables, i.e., [kel] ‘bald’ vs. [ka'lɛ] ‘castle’ (see Zimmer & Orgun 1992: 44). However, in other Turkic languages the same sounds are separate phonemes, as evinced by minimal pairs, e.g., Old Turkic älig ‘hand’ vs. elig ‘king’ (Erdal 2004: 51). A minimal pair has been identified in Urum, i.e., äl ‘hand’ vs. el ‘stranger’ (the same contrast appears in Anatolian dialects, see Brendemoen 1998: 237). This minimal pair is evidence that /e/ and /ä/ contrast in Urum. However, there is substantial phonologically conditioned variation in the realization of the mid-front vowels (depending on syllable structure and stress), which is not yet studied in detail. For many tokens, it is not yet clear, whether they are instances of the phoneme /ä/ or the phoneme /e/. Illustrative examples of the vowel inventory are given in Table 2. www.tehlikedekidiller.com 339 TDD/JofEL Winter 2013 • Tehlikedeki Diller Dergisi/Journal of Endangered Languages Stavros Skopeteas-Caucasian Urums and Urum Language Vowel Example IPA Orthography Meaning [a] [ʃar] šar ‘city’ [æ] [æl] äl ‘hand’ [œ] [ɟœl] göl ‘lake’ [e] [jel] yel ‘wind’ [o] [joƚ] yol ‘road’ [i] [it] it ‘dog’ [y] [yzyk] üzüg ‘ring’ [ɯ] [ɣɯz] ğız ‘girl’ [u] [ɣuʃ] ğuš ‘bird’ Table 2. Vowel inventory (IPA values in brackets; orthography in italics) 3.2.3. Phonological processes Assimilation processes are frequent at morpheme boundaries. The plural morpheme -lar ‘PL’ and the nominalizer -lix ‘NR’ assimilate to preceding nasals into -nar and -nix, respectively; see (6c) and Section 0. Assimilation in voicing is very frequent. The past suffix has a voiced and a voiceless allomorph, -d(I) or -t(I), assimilating to the preceding segment; see (17a-b). Stem-final voiceless consonants are voiced before vowels, e.g., ušah ‘child’ vs. ušağ-a ‘child-DAT’; arıh-sın (slim-2.SG) ‘you are slim’ vs. arığ-ım (slim-1.SG) ‘I am slim’. Vowel harmony applies in Urum with two main differences to Standard Turkish (see experimental evidence in Verhoeven 2011): (a) some suffixes that are determined by vowel harmony in Turkish are not visible for the vowel harmony in Urum; (b) for the i-suffixes, the fronting harmony does not apply to non-rounded suffixes. The first difference is illustrated in (3) by means of the 3rd person possessive suffix. The form of this suffix is invariably -i in Urum, while it is determined by vowel harmony in Turkish. (3) a. stem vowel: front and non-rounded it-i ‘dog-POSS.3.SG’, äv-i ‘house-POSS.3.SG’ b. stem vowel: front and rounded üzüg-i ‘ring-POSS.3.SG’, göl-i ‘lake-POSS.3.SG’ c. stem vowel: back and non-rounded ğız-i ‘girl-POSS.3.SG’, at-i ‘horse-POSS.3.SG’ d. stem vowel: back and rounded donguz-i ‘pig-3.SG.POSS’, yol-i ‘road-SG.POSS’ The suffixes that are determined by vowel harmony belong to two classes (similarly to Turkish, see Göksel and Kerslake 2005: 21-25): A-suffixes and I-suffixes. The A-suffixes assimilate in frontness with the last vowel of the stem, as illustrated in (4). (4) Vowel harmony: A-suffixes a. stem vowel: front it-lär ‘dog-PL’, üzüg-lär ‘ring-PL’ b. stem vowel: back ğız-lar ‘girl-PL’, yol-lar ‘road-PL’ www.tehlikedekidiller.com 340 TDD/JofEL Winter 2013 • Tehlikedeki Diller Dergisi/Journal of Endangered Languages Stavros Skopeteas-Caucasian Urums and Urum Language In Turkish, the I-suffixes assimilate in frontness and roundedness with the last vowel of the stem (Göksel and Kerslake 2005: 22). Urum differs in that the non-rounded allomorph does not assimilate in frontness: the central vowel ı appears both with back and front unrounded stem vowels (see Verhoeven 2011). This difference is illustrated by means of the genitive suffix in (5). The crucial difference to Turkish is the form of the suffix in (5a): Urum it-ın corresponds to Turkish it-in, Urum äv- ın to Turkish ev-in. The forms of the I-suffixes in the context of different stem vowels are listed for both languages in . (5) Vowel harmony: I-suffixes a. stem vowel: front and non-rounded it-ın ‘dog-GEN’, äv-ın ‘house-GEN’, äl-ın ‘hand-GEN’ b. stem vowel: back and non-rounded ğız-ın ‘girl-GEN’, at-ın ‘horse-GEN’ c. stem vowel: front and rounded üzüg-ün ‘ring-GEN’, göl-ün ‘lake-GEN’ d. stem vowel: back and rounded donguz-un ‘pig-GEN’, yol-un ‘road-GEN’ I-suffix stem vowel Urum Standard Turkish –round front i, e/ä I i back ı, a ı +round front ü, ö ü back u, o u Table 3. I-suffixes in Urum vs. Standard Turkish 3.3. Word Classes and Morphological Categories 3.3.1. Nouns Nominal morphology includes three categories: number, possession, and case. All three categories are encoded through agglutinative suffixes that attach to the stem in exactly this order, e.g., baba-lar-ım-dan (father-PL-POSS.1.SG-ABL) ‘from my fathers’. The category of number contains the plural suffix -lAr, whereby the vowel is determined by the frontness harmony; compare (6a) and (6b). With a stem-final alveolar nasal n, the plural suffix assimilates to -nar; see (6c). Nasal assimilation of the plural suffix is also attested in the Turkish dialects of Erzurum, e.g., on-nar ‘3-PL’ (Standard Turkish onlar), gan-nar ‘blood-PL’ (Standard Turkish kanlar) (Gemalmaz 1978[III]: 178; see Johanson 1998: 34 for Turkic languages in general). However, the process is productive in Urum as evinced by the fact that it is not restricted in old formations but also apply to recent borrowings from Russian, e.g., slon-nar from Russian slon (Слон). (6) a. front vowel stems göl-lär ‘lake-PL’, it-lär ‘dog-PL’, äv-lär ‘house-PL’, üzüg-lär ‘ring-PL’ b. back vowel stems at-lar ‘horse-PL’, ğız-lar ‘girl-PL’, yol-lar ‘road-PL’, donguz-lar ‘pig-PL’ www.tehlikedekidiller.com 341 TDD/JofEL Winter 2013 • Tehlikedeki Diller Dergisi/Journal of Endangered Languages Stavros Skopeteas-Caucasian Urums and Urum Language c. nasal assimilation on-nar ‘3-PL’, satan-nar ‘trader-PL’, slon-nar ’elephant-PL’, sıčan-nar ‘mouse-PL’, aslan-nar ‘lion-PL’, maimun-nar ‘monkey-PL’; oğlan-nar ‘boy-PL’ The plural suffix contrasts to the use of the stem without any number suffix, which is transnumeral in nature and obtains a singular interpretation by pragmatic inference (if no cue for plural reference is available in the context). The plural suffix is not obligatorily realized with plural referents (Bittricher et al. 2011 for corpus observations; Schüler 2013 for a study including corpus and experimental data). Two classes of factors determine the occurrence of the suffix: (a) contextual properties, i.e., discriminability of number reference through the context: the plural suffix is less likely if the plural reference can be clearly inferred from the context; (b) inherent properties of the referent: highly individuated referents (e.g., animates) are more likely to be marked for number. Noun phrases with numerals or plural quantifiers occur in texts with or without a plural suffix. When judged out of context, speakers reject the use of plural in the context of numerals; see (7a). They accept the use of plural in the context of quantifiers such as birğač/čoğ/az ‘some/many/few’, whereby they perceive a difference between animates and inanimates, such as the use of plural is less acceptable with animates. This intuition is in conflict with speech production data, which shows exactly the opposite pattern: the presence of the plural suffix is more likely with highly individuated referents, e.g., animates, and less likely with lower referents in the individuation scale, e.g., inanimates (Schüler 2013). The universal quantifier äp ‘all’ shows a different behaviour (see further discussion in 0): the plural suffix is almost always used with nouns determined by äp and speakers report that both versions (with or without suffix) are completely acceptable, see (7c) (see Schüler 2013 for quantitative results in speech production). (7) a. numeral uč ğız(*-lar) / äv(*-lär) three girl(-PL) house(-PL) ‘three girls / houses’ b. quantifier birğač/čoğ/az ğız(?-lar) / äv(-lär) some/much/few girl(-PL) house(-PL) ‘some/many/few girls / houses’ c. universal quantifier äp ğız(-lar) / äv(-lär) all girl(-PL) house(-PL) ‘all girls / houses’ The possession suffixes are cross-reference markers referring to the possessor; see (8a). In complex noun phrases these suffixes are co-referent with the dependent noun phrase (the genitive) resulting to a double morphological marking of the dependency relation, by means of possessor agreement on the head noun and a genitive case on the dependent noun phrase; see (8b). The possessive suffixes are listed in Table 4. The allomorphs with an initial vowel -ım/-ın of the 1. and 2. person occur after consonants, barmağ-ın ‘finger-POSS.2.SG’. The nasal ending (n) of the 3. person singular occurs before vowels, e.g., mama-lär-ın-i ‘udder-PL-POSS.3-ACC’; the onset (s) appears after vowels, e.g., baba-sın-a www.tehlikedekidiller.com 342
Description: