Basic vocabulary of closely related languages in contact: case study of Turkic languages on the Crimean Peninsula

Methods

To define the genealogical affiliation of the Turkic languages of Crimea, I first apply manual subgrouping based on lexical innovation and then compare the obtained results with the inference of the computational lexicostatistical algorithms. I use three approaches which are currently most widespread in linguistic phylogeny: Maximum Parsimony Analysis, MCMC Bayesian approach, Neighbor-joining algorithm.

An important advantage of the manual subgrouping applied in the present paper is that it fits the commonly accepted requirement to build genealogical classification based on innovations (Campbell 2013: 175). This requirement is ignored by the lexicostatistical framework, where every match, whether it is an innovation or a retention, has similar value. The principle of subgrouping sufficient for our purposes is drastically trivial. Languages A and B are regarded as specifically related to each other if this pair demonstrates the highest amount of shared non-contact-induced innovations. This method was used by Leonid Kogan (2015) for the classification of the Semitic languages. When reconstructing the phylogeny of a whole family from scratch, this method leads to a vicious circle, since tree topology must be already known for most cases to distinguish between innovations and retentions. However, if the goal is merely to find the positions of newly involved taxa on a previously constructed tree, such a method is applicable.

The most important technical details on the applied computational lexicostatistical algorithms are summarized in Table 2.

Table 2. Information on the software and basic settings applied for the lexicostatistical analyses

Based on trees obtained as the result of Maximum Parsimony, a strict consensus tree was produced. The settings for Bayesian MCMC are adopted from Kassian et al. forthcoming. The full dataset and output files can be found in Supplement 2. Cognate encoding has been done within Starling software and then converted into the Nexus file with a binary matrix. The derivational drift free dataset has been used; on the principles of the derivational drift elimination see Kassian et al. forthcoming.

I have compared six wordlists of the Turkic idioms spoken until recently on the Crimean Peninsula (see Table 3) with Halich and Trakai Karaim, Turkish, Gagauz, Proto-Nogai, Proto- Kazakh-Karakalpak, Proto-Kumyk, Proto-Karachay-Balkar. Since Maximum Parsimony analysis requires an outgroup taxon, I included the Proto-Yakut list, which clearly belongs neither to the Oghuz nor to the Kipchak clade. Lists of the proto-languages are reconstructed (using methodology described in Section 3.1) by me in collaboration with Anna Dybo and Alexei Kassian within the framework of an ongoing project devoted to revision of the Turkic phylogenetic tree structure. All Karaim wordlists were collected from the sources mentioned in Section 2.1; Turkish and Gagauz ones are based on the dictionary sources as well (Parker 2008; Bogochanskaya & Torgashova 2009; Gaydarzhi et al. 1973; Sesli Sцzlьk; Rajki 2007). Collecting the lists, I was guided by the semantic specification proposed in Kassian et al. 2010.

Since some of Crimean Tatar wordlists do not meet all the modern requirements to Swadesh lists (see Section 4.3 for details), computational lexicostatistics can only play a secondary role in the present research. However, I believe that, despite somewhat faulty data, application of three different computational approaches still has some relevance.

Table 3 shows which wordlists of idioms spoken in Crimea were used in the present paper.

Table 3. 110-item Swadesh lists for the Turkic varieties of Crimea

It should be noted that three Crimean Tatar wordlists used in this research had been collected long before the semantic specification of the Swadesh list was undertaken in Kassian et al. 2010. Therefore, they are not fully compatible with my data. Three important discrepancies have been found; `all (omnes)', `to burn (intr.)', `to go' are used now instead of traditional `all (totus)', `to burn (tr.)', `to walk'. Incompatible items are marked as not attested in Polinsky's data; moreover, the original Swadesh 100-wordlist has been extended with 10 items absent from the older record. To make my lists more compatible with older ones, I deviate from Kassian et al. 2010 on three points: taking `earth (ground)' instead of `earth (soil)', `round (2D)' without `round 3D' as a synonym and not accepting medial deictic pronouns.

I now avoid labeling my data on Crimean Tatar with any terms from the traditional classification. It has been collected from two informants, a couple, about 60 years old, who were interviewed independently. They were born in Uzbekistan, now they live in the Bakhchisaray district. The wife's parents come from Duvankцy (Ukr. Verxnjosadove), the traditional territory of the Orta dialect; the husband's parents come from the nearest suburbs of Gurzuf, the traditional territory of the Coastal dialect. The two obtained wordlists correspond with each other fully as far as lexical items are concerned; at the same time, they considerably differ in phonology. The woman's idiolect lacks labial harmony and the distinction of voiced and voiceless stops in the word-onset; the voiceless uvular is a stop. So, it should be regarded as a Kipchak- based dialect. Thus, she pronounces `fat' with the uvular consonant as yax, but `forest' with a labial as taw, cf. section 3.4.1. Her husband's idiolect demonstrates consistent labial harmony, sporadic preservation of the PT initial *d, and the fricativization of the voiceless uvular; these features are typical for the Coastal (Oghuz-based) dialect. The word auz `mouth' has been found to have the typically Kipchak reflex of *agi in the both idiolects, but axir `heavy' with the Oghuz reflex. Polinsky's lists contain some inconsistencies in phonology as well; the most glaring is a sporadic *y > 3 before a in the Coastal dialect, which is expected only in Steppe Crimean Tatar, and, at the same time, the lack of this shift in some lexemes from the Steppe dialect. The distribution of reflexes of PT *g across dialects confounds expectations as well: cf. axiz in all dialects, CoCrTat. tau, yav, OrCrTat yax, dax (with d instead of expected t!). The phonology represented in Polinsky's data is completely inconsistent with existing description of the Crimean Tatar dialects. As is shown below, basic vocabulary allows us to make some clearer conclusions about the original genealogical affiliation of the dialects documented in the wordlists under consideration and about the direction of their development.

Innovations in the basic vocabulary

Turkic languages of the Crimea demonstrate discrepancies in 25 slots out of the 110-item wordlist. Cases in which the variation is caused by Persian and Arabic borrowings are excluded, i.e. only potentially autonomous innovations are taken into account. Table 4 below presents the genetically relevant features. Archaisms, borrowings, and items innovative from the Common Turkic perspective but still not informative for the current question (i.e. innovations shared by both Kipchak and Oghuz languages) are underlined. The Oghuz loans revealed above and phonological variants have been excluded from the Crimean Karaim word- list. Indexes in the superscript identify the subgroup in which cognates of the word are found, an exclamation mark labels singletons. These indexes are somewhat rough, since I ignore the fact that some of the considered words can occur sporadically in the other Turkic languages which cannot be applied to the closest relative of the studied Crimean varieties. These inaccuracies are partially clarified in the commentary immediately after the table. I use mostly Kip and Ogh meaning primarily West Kipchak and West Oghuz respectively; when possible, I refer to the low level subgroups instead of Kipchak. The full lists for the Turkic idioms of Crimea and for languages they have been compared with can be found in Supplement 2.

I. all - CrTat., Krym. epsi, cf. Tur., Gag. hepsi `all (omnes)' - the word is limited to the mentioned languages; no fully acceptable Turkic or foreign etymology (Dybo 2013: 66), the root has a probable Persian origin (cf. Rдsдnen 1969: 158), but this hypothesis faces some phonological difficulties.

4. belly - CrKar., CrTat., Krym. qursaq - assuming the stem as a basic term is innovative as a result of the elevation `paunch' > `belly'; widespread in the Kipchak subgroup (Dybo 2013: 106-108).

10. bone - CrTat., Krym. kemik, cf. Tur., Gag. kemik `bone' - result of generalization `spongy bone' > `bone'; as a basic term, limited to the mentioned languages (Dybo 2013: 173174).

II. breast - Strictly speaking, both stems are archaic: CrTat., Krym. gцkьs, kцkьs, kokus, koks reflect PT *gцkьf; CrKar. kцkrek, CrTat. kokrek reflect the same root with the fossilized diminutive suffix *gцkfek (Dybo 2013: 178). Since the derivational connection between these stems has been erased a long time ago, I believe that their distribution is informative for genealogical classification. However, in agreement with the principles of derivational drift elimination, I mark the simplex and diminutive form with one index in the lexicostatistical dataset.

Table 4. Innovations in the basic vocabulary of the Turkic varieties of Crimea

12. to burn - CrKar. kьydьr- - causative from Proto-Turkic *kцn- `to burn'; cognates have been found in other Karaim dialects, Karachay-Balkar, Kazan Tatar and in some other languages, mainly in the Middle Asian area (Dybo 2013: 189).

18. dog - CrTat., Krym. kцpek, Tur., Gag. kцpek `dog' - result of generalization `hound' > `dog'; as a basic term, limited to the mentioned languages (Dybo 2013: 189).

20. dry - SCrTat. xati - result of the semantic shift `solid' > `dry' (Rдsдnen 1969: 241).

26. fat - SCrTat. may - result of the semantic shift `butter, suet' > `fat'; as a basic term, common in Volga and Middle Asian Kipchak languages and in Nogai (Dybo 2013: 249-250).

28. feather - CrKar. yun, cf. HKar., TrKar., yun `feather / down' -result of the semantic shift `fur / down' > `feather / dawn'; with original meaning, common in the Kipchak languages (Dybo 2013: 259-260).

SCrTat. qusin - etymology is somewhat obscure; the stem should be derived from qus `bird', -in can be an old instrumental affix; exclusive isoglosses with the Nogai subgroup. The etymology proposed here is more probable than the hypothetical Persian loan mentioned in Dybo 2013: 261. Anna Dybo has as of now rejected the etymology involving Pers. kuc `fish scale' (personal communication).

CoCrTat. lelek, cf. Tur. yelek `feather' - derived from PT *ye:l `mane'; attested in all Oghuz languages, including Salar, and in Khalaj (Dybo 2013: 259).

37. hand - CrKar., CrTat. qol - result of the semantic shift `arm' > `hand'; extremely widespread across Nuclear Turkic languages, particularly in all languages belonging to the Kip- chak subgroup. Oghuz and some other languages preserve reflexes of the stem *elg `hand' (Dybo 2013: 300-307).

51. man - CrTat. aqay ~ axay - fossilized vocative form from *aqa `uncle, older relative' (Se- vortyan et al. 1974: 121). The stem looks rather like a borrowing from baby-talk (cf. Russian baby-talk words d'ad'a `man' < `uncle' and t'ot'a `woman' < `aunt'), so its relevance for the genealogical classification is questionable, cf. `woman'.

52. many - CrTat., Krym. coq, cf. Tur., Gag. coq, Az. cox `many' - reconstruction of the original semantics is somewhat difficult but its innovative nature is obvious; as a basic term, attested only in the mentioned languages (Dybo 2013: 371).

55. mountain - CrTat. bayir, cf. Gag. bayir `mountain' - result of the semantic shift `hill' > `mountain'; attested as a basic term only in the mentioned languages (Dybo 2013: 380).

76. to sleep - CrKar., CrTat., Krym. yuq-la- - derived from *uyki `sleep'; common in Kipchak languages (Dybo 2013: 473-474).

77. small - CrTat. yufaq, ifaq, cf. Tur. ufaq `small, little' - result of the generalization `small, fine (of pebble, crumb, powder etc.)' > `small, little'; attested in most languages with more specific meanings (Sevortyan et al. 1974: 560-561).

76. smoke - CrTat. duman, cf, Tur., Gag. duman `smoke' - result of the semantic shift `fog' > `smoke'; this meaning is found only in the mentioned languages (Dybo 2013: 481).

82. sun - CrTat., Krym. kunes, gьnes, kьnes - formal innovation, root extension is not quite clear, apparently constitutes an analogical rhyme with quyas `heat, blazing sun'; with this extension, this root is attested mainly in Oghuz languages but occurs beyond this subgroup as well (Dybo 2013: 488-489). Since no meaning shift has occurred, I mark reflexes *gьn and *gьnes with the same index in the lexicostatistical dataset.

83. to swim - CrTat., Krym. yalta-, yalda-, jalda- - derived from yal `horse mane' with further semantic shift `to swim or to cross a river holding the horse's mane' > `to swim'; the verb with its original meaning occurs in Kipchak languages; as the basic term for `to swim', only in Nogai (Dybo 2013: 492; Sevortyan et al. 1974: 93-94).

90. tree - CrKar., CrTat., Krym. terek, derek - result of the generalization `poplar' > `tree'; attested as the basic term for `tree' in West Kipchak and Sary Yugur (Dybo 2013: 510).

92. to go - CrTat., Krym. kit-, cf. Tur. Gag. git- - result of the semantic shift `to go away' > `to go'; attested as the basic term for `to go' in the Oghuz subgroup (Dybo 2013: 515).

93. warm - CrTat., Krym. sijax, cf. Gag. sijaq `warm' - a formal innovation, PT *isig `warm, hot (?)' extended with a diminutive suffix; found with this suffix only in the mentioned languages (Dybo 2013: 518; Rдsдnen 1969: 173-174).

99. woman - CrTat. apay - fossilized vocative form from *apa `elder sister' (Sevortyan et al. 1974: 159). The stem looks rather like a borrowing from baby-talk (cf. Russian baby-talk words d'ad'a `man' < `uncle' and t'ot'a `woman' < `aunt'), so its relevance for genealogical classification is questionable.

SCrTat. xisayaxli - the compound *qiz-ayal-ki `girl-woman-nominal suffix', metathesis of the cluster under the analogical influence of the more productive suffix -li (etymology proposed by Anna Dybo, p.c.). Due to the fact that the main component -ayal- is an Arabic loan, I mark the whole stem as a borrowing.

101. far - CrTat., Krym. uzaq ~ uzax - result of the shift `far (adj.)' > `far (adv.)'; found in numerous languages. The isogloss is not particularly informative for affiliation of Crimean Tatar and Krymchak, however, it opposes Karaim to other Turkic varieties of Crimean (Dybo 2013: 534); the same is true for `near'.

103. near - CrTat., Krym. yaqin ~ yaxin - found in numerous languages, however, the distribution points to PT *yaguk (Kar. yuwuq) as a more archaic term for `near (adv.)' from the Common Turkic perspective (Dybo 2013: 539).

Table 5 shows the amount of the innovations shared by each of the investigated wordlists with other Turkic subgroups.

Table 5. The amount of the innovations (including both inherited items and inter-Turkic borrowings) shared with other Turkic subgroups

Four innovations (`to burn', `feather', `sun', `to swim') clearly connect Crimean Karaim with the Halich and Trakai dialect and oppose it to other studied languages. Since it demonstrates only one Kipchak-looking stem, Coastal Crimean Tatar must be classified as Oghuz language. The situation with other idioms is not so transparent. Oghuz-looking innovations are predominant in all wordlists, yet at the same time they demonstrate some typically Kipchak lexemes. Two possible interpretations of this situation may be offered: 1) the Oghuz innovations can be identified as inherited ones and Kipchak as substrate loans; 2) the Oghuz lexemes can be regarded as loans from the dominant language of the region and Kipchak innovations as inherited ones. Such ambiguity indicates that there has been a language shift. The range of borrowings can be easily identified in Crimean Tatar (Steppe, Orta and my data) and in Krym- chak based on the same distributional criterion which has already been applied to the Crimean Karaim data. All lexemes with the meanings mentioned above are widespread in Oghuz languages but not in Kipchak. Taking into account the prestige status of Crimean Turkish, i.e. Coastal Crimean Tatar, extensive borrowing from it is very probable. Such interpretation is supported by the fact that the lexemes suspected to be loanwords concentrate predominantly in the less stable part of the Swadesh list: `small' - 110; `mountain' - 107; `many' - 106; `big' - 101; `far' - 100; `near' - 95; `warm' - 90; `to go' - 89; `all' - 84; `to swim' - 78; `breast' - 49; `smoke' - 40; `sun' - 39; `bone' - 38; `dog' - 16. The number after the concept indicates its position in the stability ranking (the higher the number, the less stable the concept), according to Sergei Starostin's estimations (2007b). Avoiding the discussion of every particular case, I would like to draw attention only to OrCrTat. el `hand' which is a classic example of a fake archaism.

Other lexemes are more likely to be borrowed West Oghuz innovations. It must be also emphasized that at least in some cases we should deal not with MAT-borrowings (like epsi `all'; kьnes `sun'; sijax `warm'; kopek `dog' < `hound') but rather with PAT-borrowings, i.e. with contact-induced semantic shifts (kemik `bone' < `spongy bone'; bayir `mountain' < `hill'; ifaq `small, little' < `small, fine (of pebble, crumb, powder etc.)'; kit- `to go' < `to go away').

Considering all the Oghuz-looking lexemes as borrowings, hence, irrelevant for genealogical classification, I come to the conclusion that Crimean Karaim does indeed belong to the same subgroup with other Karaim dialects; Steppe Crimean Tatar - to the Nogai subgroup; Crimean Tatar (based on my data), Middle Crimean Tatar (Polinsky's data) and Krymchak are close to West Kipchak. Any further conclusions about their proximity to a particular subgroup within Kipchak cannot be made with enough certainty.

The proximity of Steppe Crimean Tatar to the Nogai subgroup is proven by three nontrivial innovations (`fat', `feather', `to swim') which are not attested in any language from potential candidates for the closest relatives. Two of these innovations (qusin `feather' and yalda- `to swim') do not occur elsewhere in the Turkic languages. Orta Crimean Tatar, Crimean Tatar (based on my data), and Krymchak yalda- must be analyzed as a borrowing from the Steppe dialect. Despite the fact that the Steppe dialect was not a dominant idiom, some influence on its part cannot be excluded. Otherwise, such non-trivial (`to swim or to cross a river holding horse's mane' > `to swim') innovations can be regarded only as a signal of relatedness. This is less probable, since there are no other facts pointing to the specific proximity of Orta Crimean Tatar and of the dialect reflected in my data to the Nogai subgroup.

To sum up, I assume the following subgrouping based on innovations in the basic vocabulary: [Turkish, Gagauz, Coastal Crimean Tatar], [[Nogai, Steppe Crimean Tatar], [Halich Karaim, Trakai Karaim, Crimean Karaim], [Orta Crimean Tatar, Crimean Tatar (my data), Krymchak, Kumyk, Karachay-Balkar].

The alternative approach, i.e. consider Oghuz lexemes as inherited and Kipchak lexemes as borrowings, leads to difficulties. Must we mark as borrowings only lexemes looking similar to Kipchak innovations or must we regard typical Kipchak retentions as borrowings too, i.e. as fake archaisms? This question does not have a satisfactory answer. Had Kipchak innovations been considered borrowings, we would have to deal with a suspiciously archaic Oghuz idiom simultaneously overflowing with Kipchak loans. It should be noted that these fictitious Kipchak loans would have concentrated in the somewhat more stable part of the lists than the real loans considered above, cf. `belly' - 109; `fat' - 81; `to sleep' - 73; `breast' - 49; `tree' - 42; `hand' - 11. If Kipchak retentions such as sьyek `bone', it `dog', bar- `to go' etc. had been fake archaisms in the tentative Oghuz language, the mass of borrowings would have strongly contradicted the direction of influence proven by sociolinguistic factors. Thus, it seems reasonable to reject such a decision.

Results of lexicostatistical analysis

All expectations based on innovations in the basic vocabulary are confirmed by formal computational methods. Trees inferred by three applied algorithms differ only in some details. They all agree on the following points: (a) Coastal Crimean Tatar belongs to the same clade as Turkish and Gagauz; (b) other languages are included in the Kipchak subgroup; (c) Steppe Crimean Tatar is combined with Nogai and Kazakh-Karakalpak; (d) Crimean Karaim forms a clade with other Karaim dialects.

As for the internal structure of the Kipchak clade, the applied analyses are in minor disagreement with each other. Neighbor-joining and Bayesian MCMC suggests a first split into Nogai-Kazakh-Karakalpak-Steppe-Crimean-Tatar and remaining languages, followed by a split

into Karachay-Balkar-Kumyk and Karaim-Orta-Crimean-Tatar-Krymchak. Both algorithms have established that Krymchak is the closest to Crimean Tatar (according to my recently collected data) and the two idioms are related to the Orta dialect. The strict consensus tree build by Maximum Parsimony analysis shows multifurcation of the Kipchak clade into the following taxons: [Halich, Trakai, and Crimean Karaim], [Nogai, Steppe Crimean Tatar, Kazakh- Karakalpak], Orta Crimean Tatar, Crimean Tatar (my data), Krymchak, Kumyk, Karachay- Balkar. Such a structure for the tree fully fits all my assumptions made on the basis of common innovations (Section 4.4).

Table 6. Mean age, median age and 95% HPD age of the law level nodes according to Bayesian MCMC

Figure 3. Tree constructed with Bayesian MCMC algorithm in MrBayes software visualized in FigTree software. Numbers near the nodes define mean age; numbers near branches define their probability in percent

linguist diachronic phylogenetic karaite

Figure 4. Manually redrawn strict consensus tree constructed with Maximum Parsimony algorithm

The advantage of Bayesian MCMC and neighbor-joining is that they suggest a clade joining all the languages which had originated on the Crimea Peninsula. The ancestors of the Karaim and Orta Crimean Tatar speaking people came to this region when the West Kipchak languages should have already been (insignificantly) diversified. The invasion of Tatars into Crimea dates back to the 1220s (Fisher 1978: 2); around this time the adherents of Karaite Judaism migrated here from Byzantium (Jankowski 2017: 452-453) and adopted the local Kipchak language. Then the language common for Tatars and Karaims should have split due to the closeness of the Karaim community. The Krymchak speakers, groups of Rabbinic Jews heterogeneous in their origin, adopted Orta Crimean Tatar. Thus, Polinsky is right calling Krymchak an ethnolect of Crimean Tatar (see Polinsky 1992: 173-176). However, another of Polinsky's statements must be revisited. She classifies the Orta dialect together with Krymchak and even Crimean Karaim as Oghuz languages. But even if one admits it is methodologically tolerable not to exclude inter-Turkic borrowings, such an affiliation actually reflects a later language shift. Polinsky's data at least on Krymchak and Orta Crimean Tatars allows the reconstruction of their original genealogical affiliation. The identification of the borrowings plays here a crucial role. My study confirms Sevartyan's (1966) view on the Crimean Tatar dialects as three genealogically distinct items. The early separation of Nogai-Kazakh-Karakakpak-Steppe-Crimean- Tatar from remaining Kipchak languages in question corresponds to the traditional opinion that Nogai does not belong to the West Kipchak subgroup. The speakers of the Steppe dialect have massively settled in Crimea only in the beginning of the 17^th century; this was the result of Nogai migration from Lower Volga Steppes which had started a century before (see Trepavlov 2014).

A recent attempt at another revision of the Turkic classification fell victim to undetected loans as well. Martine Robbeets and Alexander Savelyev (2020) include Crimean Tatar into the Oghuz subgroup. They discuss this contradiction with previous classifications and correctly explain it by a lot of Oghuz elements in the wordlist. However, the authors do not try to exclude them despite careful elimination of all non-Turkic borrowings. Robbeets and Savelyev's wordlist of Literary Crimean Tatar (based on the Orta dialect) is compiled on the basis of Useinov 2007. 13 lexemes from this list can be treated as Oghuz loans and contact-influenced semantic innovations: el `hand', qarin `belly', kemik `bone', kokьs `breast', kцpek `dog', uzaq `far',

Figure 5. Tree constructed with neighbor-joining algorithm in Starling software

ket- `to go', eyi `good', coq `many', ufaciq `small', kьnes `sun', sijaq `warm', sivalcan `worm'. At the same time, the list contains the following lexemes typical for the Kipchak subgroup: qursaq `belly', asa- `to eat', ayt- `to say', yuqla- `to sleep', terek `tree'. My remark made on Polinsky's results is true for this study as well. If an effort is made to exclude 13 items from the dataset, we should be able to identify the original genealogical affiliation of the Orta Crimean Tatar language. That a language should contain so many borrowings from one source in its basic vocabulary is not a frequent case, but hardly a unique exception either (the abovementioned case of Riksmвl is another example of a language with a similar amount of borrowings in the basic vocabulary).

Conclusions

Detection of borrowings is a necessary procedure for the purposes of phylogenetic studies. A historical linguist should be attentive not only to external but also to intra-family loans. When closely related languages are in intimate contact with each other, the areal criterion becomes more important than the phonological one. The similar phonotactics, phonological inventories, and the minor differences in the phonological shape of words driven by historical phonological processes often make application of phonological criteria difficult or even completely impossible. Loans from a genetically related language can sometimes look like archaic items. They can be both MAT- and PAT-borrowings. These fake archaisms can be revealed based on the distributional criterion if the general direction of influence is known.

Borrowings from a closely related dominant language can strongly influence the basic vocabulary of less prestigious idioms and make the genealogical classification of the latter quite difficult. However, careful elimination of all borrowings makes it possible to identify the subgroup to which the language in question belongs. Both manual analysis of isoglosses and computational lexicostatistics give acceptable results if the dataset is free from borrowings.

References

1. Aikio, Ante. 2007. Etymological nativization of loanwords: A case study of Saami and Finnish. In: Ida Toivonen, Diane Nelson (eds.). Saami Linguistics (Current Issues in Linguistic Theory 288): 17-52. Amsterdam, Philadelphia: John Benjamins Pub. Co.

2. Aqtay, Gulayhan. 2009. Eliyahu Ben Yosef Qilci's Anthology of Crimean Karaim and Turkish Literature. Istanbul: Yildiz Dil ve Edebiyat Dizisi 8.

3. Aqtay, Gulayhan & Henryk Jankowski. 2015. A Crimean Karaim-English dictionary (Prace Karaimoznawcze 2). Poznan.

4. Bammatov, Burgan G., Nurmagomed E. Gadzhiakhmedov. 2011. Kumyksko-russkij slovar' [Kumyk-Russian dictionary]. Makhachkala.

5. Baskakov, Nikolay A. (ed.). 1956. Russko-Nogajskij slovar' [Russian-Nogai dictionary]. Moscow: Gosudarstvennoe iz- datel'stvo innostrannyx i nacyonal'nyx slovarej.

6. Baskakov, Nikolay A. 1967. Russko-karakalpakskij slovar' [Russian-karakalpak dictionary]. Moscow: Sovetskaya entsik- lopediya.

7. Baskakov, Nikolay A., Seraja Ben Mordechai Szapszal, Ananiasz Zajqczkowski. 1974. Slownik karaimsko-rosyjsko- polski [Karaim-Polish-Russian dictionary]. Moscow: Russkij Jazyk.

8. Bektaev, Kaldybay. 1996. Bol'soj kazaxsko-russkij russko-kazaxskij slovar' [Large Kazakh-Russian Russian-Kazakh dictionary]. Kazaxsij proekt razvitija gosudarstvennogo jazyka.

9. Bekturov, Shabken & Ardak Bekturova. 2001. Kazaxsko-russkij slovar' [Kazakh-Russian dictionary]. Astana: Foliant.

10. Bergsland, Knut, Hans Vogt. 1962. On the Validity of Glottochronology. Current Anthropology 3(2): 115-153.

11. Bogochanskaya, Nina N., Anna S. Torgashova. 2009. Bol'soj russko-tureckij slovar' [Large Russian-Turkish dictionary]. Moscow: Dom Slavjanskoj Knigi.

12. Bouckaert, R., P. Lemey, M. Dunn, S. J. Greenhill, A. V. Alekseyenko, A. J. Drummond, R. D. Gray, M. A. Suchard, Q. D. Atkinson. 2012. Mapping the origins and expansion of the Indo-European language family. Science 337: 957-960. doi:10.1126/science.1219669.

13. Burlak, Svetlana A., Sergei A. Starostin. 2005. Sravnitel'no-istoriceskoe jazykoznanie [Comparative-historical linguistics]. Moscow: Academia.

14. Campbell, Lyle. 2013. Historical linguistics: an introduction. 3rd edn. Edinburgh University Press.

15. Clauson, Gerard. 1972. An etymological dictionary of Pre-Thirteenth-Century Turkish. Oxford University Press.

16. Dybo, Anna V. 1996. Semanticeskaja rekonstrukcija v altajskoj йtimologii: somaticeskie terminy (plecevoj pojas) [Semantic reconstruction in the Altaic etymology: somatic terms (shoulder girdle)]. Moscow.

17. Dybo, Anna V. 2007. Lingvisticeskie kontakty rannix tjurkov. Leksiceskij fond. Pratjurkskij period [Language contacts of the early Turkic peoples: vocabulary: Proto-Turkic epoch]. Moscow: Vostochnaya Literatura.

18. Dybo, Anna V. 2013. Иtimologiceskij slovar' tjurkskix jazykov [Etymological dictionary of the Turkic languages]. Vol. 9: Иtimologiceskij slovar' bazisnoj leksiki tjurkskix jazykov [Etymological dictionary of the Turkic basic vocabulary]. Astana: Prosper Print.

19. Йtnograficeskaja karta Kryma [Ethnographic map of Crimea]. 1926. GosKrymIzdat.

20. Filonenko, W. J. 1931. Ethnografische Karte der Krimer autonomen sozialistischen Sowjetrepublik [Ethnographic map of the Crimean Autonomous Soviet Socialist Republic]. Wien.

21. Fisher, Alan W. 1978. Crimean Tatars (Studies of Nationalities in the USSR). Stanford, Calif.: Hoover Institution Press.

22. Gaydarzhi, Gavril A., Ludmila A. Pokrovskaya, Boris P. Tukan, Elena K. Koltsa. 1973. Gagauzsko-russko-moldavskij slovar' [Gagauz-Russian-Moldovan dictionary]. (Ed.) Nikolay A. Baskakov. Moscow: Sovetskaja entsiklopedija.

23. Gochiyayeva, Sofya A., Kh. I. Suyunchev. 1989. Karacaevo-balkarsko-russkij slovar' [Karachay-Balkar-Russian dictionary]. Moscow: Russkij Jazyk.

24. Goloboff, Pablo A., Santiago A. Catalano. 2016. TNT version 1.5, including a full implementation of phylogenetic morphometrics. Cladistics 32(3): 221-238. doi:10.1111/cla.12160.

25. Gray, Russell D., Quentin D. Atkinson. 2003. Language-tree divergence times support the Anatolian theory of Indo-European origin. Nature 426: 435-439. doi:10.1038/nature02029.

26. Haspelmath, Martin. 2009. Lexical borrowing: concepts and issues. In: Uri Tadmor, Martin Haspelmath (eds.). Loanwords in the World's Languages: A Comparative Handbook: 35-54. Berlin / New York: Walter de Gruyter. doi:10.1515/9783110218442.

27. Huelsenbeck, J. P., F. Ronquist. 2001. MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17: 754-755.

28. Ianbay, Iala. 2016. Krimchak Dictionary. Wiesbaden: Harrassowitz Verlag.

29. Izidinova, Seveli R. 1996. Krymskotatarskij jazyk [Crimean Tatar language]. In: Edkhyam R. Tenishev (ed.). Tjurk- skie jazyki [Turkic languages] (Jazyki Mira [World's Languages]): 298-309. Moscow.

30. Jankowski, Henryk. 2003. On the Language Varieties of Karaims in the Crimea. Studia Orientalia 95: 109-130.

31. Jankowski, Henryk. 2017. Karaim and Krymchak. In: Lily Kahn, Aaron D. Rubin (eds.). Handbook of Jewish Languages: 452-489. Leiden: Brill. doi:10.1163/9789004359543_015.

32. Johanson, Lars. 1998. The history of Turkic. In: Lars Johanson, Йva A. Csatц (eds.). The Turkic languages (Routledge Language Family Descriptions): 81-125. London: Routledge.

33. Kassian, Alexei S., George Starostin, Anna Dybo, Vasily Chernov. 2010. The Swadesh wordlist. An attempt at semantic specification. Journal of Language Relationship 4: 46-89.

34. Kassian, Alexei S., George S. Starostin, Mikhail A. Zhivlov. 2015. Proto-Indo-European-Uralic comparison from the probabilistic point of view. Journal of Indo-European Studies 43(3-4): 301-347.

35. Kassian, Alexei S., Mikhail A. Zhivlov, George S. Starostin, Artem A. Trofimov, Petr A. Kocharov, Anna Kurit- syna, Mikhail N. Saenko. Forthcoming. Rapid radiation of the Inner Indo-European languages: an advanced approach to Indo-European lexicostatistics. https://www.academia.edu/39903804/.

36. Kavitskaya, Darya. 2010. Crimean Tatar (Languages of the World/Materials 477). Mьnchen: LINCOM EUROPA.

37. Kocaoglu, Timur. 2006. Karay: the Trakai dialect (Languages of the World/Materials 458). Mьnchen: LINCOM EUROPA.

38. Kogan, Leonid E. 2015. Genealogical classification of Semitic: the lexical isoglosses. Boston, Berlin: De Gruyter.

39. Musaev, Kenesbai M. 1964. Grammatika karaimskogo jazyka: fonetika i morfologija [Grammar of the Karaim langauge: phonetics and morphology]. Moscow: Nauka.

40. Musaev, Kenesbai M. 2010. Dialekty karaimskogo jazyka [Dialects of the Karaim language]. In: Anna V. Dybo (ed.). Dialekty tjurkskix jazykov [Dialects of the Turkic languages]: 205-235. Moscow: Vostochnaya literatura.

41. Nakhleh, Luay, Donald A. Ringe, Tandy Warnow. 2005. Perfect Phylogenetic Networks: A New Methodology for Reconstructing the Evolutionary History of Natural Languages. Language 81(2): 382-420. doi:10.1353/lan.2005.0078.

42. Parker, Philip M. 2008. Webster's Turkish-English thesaurus dictionary. San Diego: ICON Classics.

43. Pekarskiy, Eduard K. 1916. Kratkij russko-jakutskij slovar' [Concise Russian-Yakut dictionary]. Saint Petersburg: Tipo- grafija imperatorskoj akademii nauk.

44. Pekarskiy, Eduard K. 1959. Slovar' jakutskogo jazyka [Dictionary of the Yakut language]. Leningrad: Izdatelstvo Akademii Nauk SSSR.

45. Pokorny, Julius. 1959. Indogermanisches etymologisches Wцrterbuch [Indo-European etymological dictionary]. 2 vols. Bern / Munich: A. Francke.

46. Polinsky, Maria. 1991. The Krymchaks: History and Texts. Ural-Altaic Yearbook 63: 123-154.

47. Polinsky, Maria. 1992. Crimean Tatar and Krymchak: classification and description. In: Howard I. Aronson (ed.). The Non-Slavic languages of the USSR: linguistic studies: 157-188. Chicago Linguistic Society, University of Chicago.

48. Poznanski, Samuel. 1913. Karдisch-tatarische Literatur. KeletiSzemle 13: 37-47.

49. Radloff, Wasilij. 1896. Der Volksliteratur der nцrdlichen tьrkischen Stдmme [The folk literature of the northern Turkish tribes]. Vol. 7: Die Mundarten der Krym. Saint Petersburg: Tipografija imperatorskoj akademii nauk.

50. Rajki, Andras. 2007. A concise Gagauz dictionary. Budapest.

51. Rдsдnen, Martti. 1957. Materialien zur Morphologie der tьrkischen Sprachen (Studia Orientalia 21). Helsinki: Societas Orientalis Fennica.

52. Rдsдnen, Martti. 1969. Versuch eines etymologisches Wцrterbuch der Tьrksprachen [Attempt at etymological dictionary of Turkic languages] (Lexica Societatis Fenno-Ugricae). Helsinki: Suomalais-Ugrilainen Seura.

53. Rebi, David. 2004. Krymcakskij jazyk: krymcaksko-russkij slovar' [Krymchak language: Krymchak-Russian dictionary]. Simferopol: Dolya.

54. Ronquist, Fredrik, Maxim Teslenko, Paul van der Mark, Daniel L. Ayres, Aaron Darling, Sebastian Hцhna, Bret Lar- get, Liang Liu, Marc A. Suchard, John P. Huelsenbeck. 2012. MrBayes 3.2: Efficient Bayesian phylogenetic inference and model choice across a large model space. Systematic Biology 61(3): 539-542. doi:10.1093/sysbio/sys029.

55. Rzymski, Christoph, Tiago Tresoldi, Johann-Mattis List, Simon Greenhill, Robert Forkel. 2019. The Database of Cross-Linguistic Colexifications, reproducible analysis of cross-linguistic polysemies. https://clics.clld.org/.

56. Sakel, Jeanette. 2007. Types of loan: matter and pattern. In: Jeanette Sakel & Yaron Matras (eds.). Grammatical borrowing in cross-linguistic perspective (Empirical Approaches to Language Typology 38): 15-29. Berlin, New York: De Gruyter Mouton.

57. Savelyev, Alexander, Martine I. Robbeets. 2020. Bayesian phylolinguistics infers the internal structure and the time-depth of the Turkic language family. Journal of Language Evolution 5(1): 39-53. doi:10.1093/jole/lzz010.

58. SesliSцzlьk. SesliSцzlьk. Online English-Turkish and Multilingual Dictionary. https://www.seslisozluk.net/ (29 July, 2020).

59. Sevortyan, Ervand V. 1966. Krymsko-tatarskij jazyk [Crimean Tatar language]. In: Viktor V. Vinogradov (ed.). Jazyki narodov SSSR, vol. 2: Tyurkskie jazyki [Turkic languages]: 234-259. Moscow: Nauka.

60. Sevortyan, Ervand V., Liya S. Levitskaya, Anna V. Dybo, Valentin I. Rassadin, Galina F. Blagova, Dmitriy M. Nasi- lov. 1974. Иtimologiceskij slovar' tjurkskix jazykov [Etymological dictionary of the Turkic languages]. 9 vols. Moscow: Nauka.

61. Stachowski, Marek. 1993. Dolganischer Wortschatz (Zeszyty naukowe Uniwersytetu Jagiellonskiego, Prace jйzyko- znawcze 1086. zesz. 114). Krakow: Nakl. Uniewersytetu Jagiellonskiego.

62. Stachowski, Marek. 1998. Dolganischer Wortschatz: Supplementband. Krakow: Ksiegarnia Akademicka.

63. Starostin, George S. 2016. From wordlists to proto-wordlists: reconstruction as `optimal selection.' Faits de langues 47(1): 177-200. doi:10.3726/432492_177.

64. Starostin, Sergei A. 2000. Comparative-historical linguistics and lexicostatistics. In: Colin Renfrew, April McMahon & Larry Trask (eds.). Time depth in historical linguistics, vol. 1: 223-265. Cambridge, England: The McDonald Institute for Archaeological Research.

65. Starostin, Sergei A. 2007a. Rabocaja sreda dlja lingvista [Linguist's workspace]. In: Sergei A. Starostin. Trudy po jazykoznaniju [Works on linguistics]: 481-496. Moscow: Jazyki slavjanskix kul'tur.

66. Starostin, Sergei A. 2007b. Opredelenie ustojcivosti bazisnoj leksiki [Defining the stability of basic lexicon]. In: Sergei A. Starostin. Trudy po jazykoznaniju [Works on linguistics]: 827-839. Moscow: Jazyki slavjanskix kul'tur.

67. Syzdykova, Rabiga G., K. Sh. Khusaiyn. 2008. Kazaxsko-russkiy slovar' [Kazakh-Russian dictionary]. Alma-Ata.

68. Tenishev, Edkhyam R., Anna V. Dybo (eds.). 2002. Sravnitel'no-istoriceskaja grammatika tjurkskix jazykov: regional'nye rekonstrukcii [Comparative grammar of Turkic languages: regional reconstructions]. Moscow: Nauka.

69. Tenishev, Edkhyam R., Anna V. Dybo (eds.). 2006. Sravnitel'no-istoriceskaja grammatika tjurkskix jazykov: pratjurkskij jazyk-osnova. Kartina mira pratjurkskogo иtnosa po dannym jazyka [Comparative grammar of Turkic languages: the Proto-Turkic language. A world-view of the Proto-Turkic people according to language data]. Moscow: Nauka.

70. Trepavlov, Vadim V. 2014. Nogajskaja orda [Nogai Horde]. In: Rafael S. Khakimov (ed.). Istorija tatar s drevnejsix vremen [History of Tatars from antiquity], vol. 4: 223-252. Kazan.

71. Useinov, Seyran M. 2007. Russko-krymskotatarskij, krymskotatarsko-russkij slovar' [Russian-Crimean Tatar, Crimean Tatar-Russian dictionary]. Simferopol: Tezis.

Размещено на Allbest.ru