Multifunctional Model of Morphemes in The Turkic Group Languages (On The Example of The Kazakh and Tatar Languages) D.Sh. Suleymanov A.R. Gatiatullin A.B. Almenova Scientific Research Institute Scientific Research Institute Scientific Research Institute of Applied Semiotics, of Applied Semiotics, of Applied Semiotics, Tatarstan Academy of Sciences Tatarstan Academy of Sciences Tatarstan Academy of Sciences [email protected], [email protected] [email protected] Abstract translation systems for Turkic languages and others. Also, the presence of such models and This article describes a multifunctional computer database software, implemented as Web- model of the Turkic affixal morphemes. This model interface, will be an effective assistant to is a hierarchical system of characteristics of turkologists in conducting different morphemes belonging to different language levels: comparative studies. phonological, morphological, syntactic and semantic, and it requires a certain structure and unification in the description of characteristics of A comparative analysis among Turkic morphemes. It is a kind of “inventory” base of the languages with maximum automation of these language that can be used for different purposes; in processes requires conceptual models that particular, to perform automated comparative would appropriately and adequately describe analysis of the properties of the Turkic languages, language units both structurally and and to develop different linguoprocessors working functionally. with Turkic languages. Here, we describe the elements of the multifunctional computer model with examples on the Tatar and Kazakh languages. Another role of this model is to contribute to the process of unification of morphological Keywords: multifunctional model, affixal categories, terms and tags, which are of morpheme, Tatar language, Kazakh language. particular importance for the representation of linguistic information. The analysis of the current situation shows that in the development 1Introduction of linguoprocessors working with Turkic languages and, in particular, in the Turkic One of the problems in creating of corpus linguistics, despite the genetic, linguoprocessors for Turkic languages is a structural and typological commonness of deficit in structured data that would describe Turkic languages, there is still lack of general the properties of the Turkic language units. principles and approaches to linguistic Obviously, the presence of such databases can annotation of texts, to the system of tags for accelerate the process of development of morphemes and morphological categories. In applied computer systems working with the future this may lead to difficulties in languages of the Turkic family, such as conducting comparative researches, as well as multilingual search Turkic corpora, machine in the development of Turkic parallel corpora, Gönderme ve kabul tarihi: 09.09.2014-25.10.2014 TÜRKøYE BøLøùøM VAKFI BøLGøSAYAR BøLøMLERø VE MÜHENDøSLøöø DERGøSø (Cilt:7 - SayÕ:2) - 37 multilingual text processing systems and in Table 1. Model structure resolving of other theoretical and applied Tatar Kazakh Turkish problems. We suggest a multifunctional model that will help in the process of the comparative Functio- Properties Properties Properties study of morphemes and categories, and also nal aspect A[1].Tat A[1].Kaz A[1].Tur will serve as a unified information system on Morpho- Properties Properties Properties Turkic morphemes, available in the Internet nological A[2].Tat A[2].Kaz A[2].Tur space. Morpho- Properties Properties Properties In the proposed multifunctional model, the logical A[3].Tat A[3].Kaz A[3].Tur morphemes of each of the Turkic languages Syntactic Properties Properties Properties are described both at the morphological level aspect A[4].Tat A[4].Kaz A[4].Tur and at the level of other linguistic phenomena, Semantic Properties Properties Properties such as phonology, syntax, semantics, and at the joint of language levels, such as the aspect A[5].Tat A[5].Kaz A[5].Tur morpho-semantic and morpho-syntactic levels. The morpho-syntactic level studies the Every aspect is divided into sub-parameters, auxiliary particles and postpositions, which in and those, in turn, can be further subdivided some Turkic languages can be written as a into sub-parameters, etc. [2]. single word, while in others should be written separately. The morpho-semantic level studies In this article, we consider the fragments of the compound morphemes, such as mAgAe in the morphonological, functional and semantic Tatar language, and their counterparts in other aspects. Turkic languages. To describe the semantic aspect of the multifunctional model, the 3 Morphonological aspect authors develop a special unit in the form of situational frames [1]. The morphonological aspect is presented as a table of allomorphs. It describes all possible Currently, the authors are developing a toolkit allomorphs of a morpheme, which constitute a for filling of the proposed multifunctional superficial representation of the deep computer model. At the same time, we are description of a corresponding morpheme in filling it on the example of the Tatar and some context according to the phonological Kazakh languages units. The basic elements of rules. the model of morphemes for the Tatar language are represented in the work of the The rules are described through a set of authors [2]. This article reveals the new results production rules of the following aspect: A → concerning the description of the semantic B, where A is the context-condition of use of aspect of the model and mechanisms of this allomorph, and B is the allomorph itself. expansion of the model as a unified framework The context consists of the morphological and for the description of language units of all the phonological components: A = A1 & A2 Turkic languages. The morphological component determines 2 Model structure what morpheme is on the left in the wordform, and the phonological one regulates what The structure of this model is shown in table 1: vowels and consonants are on the left in the wordform. TÜRKøYE BøLøùøM VAKFI BøLGøSAYAR BøLøMLERø VE MÜHENDøSLøöø DERGøSø (Cilt:7 - SayÕ:2) - 38 For example, in the Tatar language the 5 Semantic aspect morpheme -nIkI has 2 allomorphs: -nıkı,- neke. The selection of the necessary allomorph To describe the meaning of the linguistic units is determined by the vowel on the left (whether in the model, the authors propose to use it is hard or soft), whereas in Kazakh there are relative-situational frames (RSF). RSF present 3 allomorphs: -nіkі,-dіkі,-tіkі, and their the implementation of a typical situation, choice is determined directly by the leftmost consisting of the name of the situation and a character. number of slots, which are the roles of the constructive elements of this situation. The 4 Functional aspect slots of the frame are filled by language constructs called syntaxemes. The structure of One of the parameters of the functional aspect syntaxemes is represented by morphemes. is that of the coherence of the morpheme. RSF has the following representation: The coherence of the morpheme indicates SituationSi whether the morpheme is free (analytical) or Role1: SintaxemI1; coherent (synthetic). This parameter is important for Turkic languages, because the Role2: Sintaxem I2; morpheme can be written as a single word in … one of the Turkic languages and separately in Role 3: Sintaxem IN; some other. End_Situation For example, the interrogative morpheme in The choice of a particular situation determines the Tatar language is written as one word, the choice of a particular type of RSF, which is whereas in the Kazakh language it is written called the base frame, with its corresponding separately. At the same time, in the Kazakh slots-roles filled with some concrete values- language it also has all the phonological syntaxemes. variants depending on the context: Tatar: -mI urmanmı ‘(is it) a forest ?’ For example, the basic RSF for the situation Kazakh: MA orman ba ‘(is it) a forest?’ that expresses the action according to the change of state has the following According to this criterion, a morpheme can be representation: classified as an affix if it is coherent, or as an auxiliary part of speech if it is free. Situation 7.2: action_statel Object: Sintaxem 5; Let us see an opposite example. The Tatar Old_state: Sintaxem121; language has a postposition belən that is New_state: Sintaxem119, 120; written separately from the main word, Time: Sintaxem78; whereas in the Kazakh language the Period: Sintaxem97; counterpart of this postposition is the Manner: Sintaxem99; morpheme -Ben with the allomorphs -ben, - End_Situation men, -pen. The authors conducted a classification of For example: semantic contexts depending on types of abıj belən ‘with the brother’ relations that participate in the formation of the aғamen ‘with the brother’ deep meaning of a given context. On the basis of this classification we obtained a system TÜRKøYE BøLøùøM VAKFI BøLGøSAYAR BøLøMLERø VE MÜHENDøSLøöø DERGøSø (Cilt:7 - SayÕ:2) - 39 consisting of 60 basic relational-situational the database for retrieval of the necessary frames. information and to run the application programmes that work with the database. Let us see an example of a syntaxeme. Syntaxeme 118 as the value of a slot. The programme for working with the model is Direction: designed as multilingual, therefore on the first 1. Number of syntaxeme: 118 page of the website there is a list of languages 2. Main word: LSG (‘physical objects’) of the Turkic group (Pic.1.). Thus, by selecting a. Syntax type: AF a language from this list, the user will be able b. Morphological type: N to work with a single language or view the c. Morphemic structure: -GA; -LAR- comparative information about all the Turkic GA; -Im-GA; -Iņ-GA; -sI-GA; - languages. IbIz-GA; -IgIz-GA; -LAr-Im-GA; -LAr- Iņ-GA; -LAr-IbIz-GA; - The user can work with a specific aspect of the LAr-IgIz-GA model. For this purpose, the main menu d. The analytical form: taba contains the option ‘Model Structure’. When 3. Main word: LSG (‘move’) chosen, it opens the list of aspects and sub- a.