The article deals with the use of generative grammars in linguistic modelling. The description of inflection of participle words is used to model synthesis processes at the morphemic level of language.
At the present stage of development, the need to develop common and specialized linguistic systems is forcing applied and computational linguistics to go far beyond their borders - in the field of information technology [1-6]. Developing effective speech models to provide computational linguistic systems will enable a way to perform such applied linguistics tasks as analyzing and synthesizing oral and written texts, describing and indexing documents, translating texts, creating lexicographic databases, and more [9-15]. An effective tool for this type of linguistic modelling is the main part of combinatorial linguistics - the theory of generative grammars, the beginning of which started from the works of the American linguist N. Chomsky [1013, 16-24]. The advantages of generative grammar modelling are that they can equally successfully describe not only the syntactic level of speech (word formation rules), but also morpheme (word formation rules with morphemes), which can be used to automate word-translation and word-formation processes [1-2, 25-29]. For example, automatic morphological synthesis implies that, based on the set requirements for word-forms, a computational linguistic system, based on morphemes, must be able to form a similar word-form itself [6, 30-48].
have identified many different structures for describing word-forms [1, 2, 4, 5, 7, 8, 10]. With the beginning of the development of the theory of generative grammars, linguists have focused not only on the description of finished word-forms, but also on the processes of their synthesis [8]. In Ukrainian studies, a lot of useful information was gathered during morphologists researches in functional part of this case, in particular theoretical problems of morphonological description, questions of classification of morphemic and word-forming structure of derivatives of Ukrainian language, regularities of affix combinatorics, modelling of word-formation mechanism of modern Ukrainian language in vocabulary of organization distinctive verbs and suffixes with a denominative meaning, problems of word-forming motivation in the formation of derivatives, the laws implementing morphonological phenomena in Ukrainian word formation, morphonological modifications in the verb inflection, morphonological processes in inflection and word formation of adjectives modern Ukrainian literary language, etc. [4]. The dynamic approach of modern linguistics to the study of the morphological level of language, when the researcher's attention is focused on the study of morphological rules, allows effective application of the results of theoretical research in practice - for the construction of computer linguistic systems of various purposes. One of the first attempts to apply the theory of generative grammars to linguistic modeling belongs to Gladkyi and Melchuk [1-2]. 3
Linguistic support is used in all computer systems without exception, and improving human-machine communication is an important challenge that can be solved by improving the process of synthesizing texts at the morphemic level. To improve this, let review the process of linguistic modelling of the synthesis of participles in Ukrainian language with help of generative grammars. To do this, a proper description of the word systems must be finished - to determine, based on a morphological analysis, a list of corresponding morphemes, and to find out a system of rules by which any correct forms of participles can be obtained without obtaining any wrong ones. 4
The participle is a changeable, grammatically formed from verb tense forms, characterized by the features of the verb and adjective [5, 9], for example: “Тепле повітря навкруги було напоєне пахощами степових квітів” (О. Гончар); “І їхній ксьондз, небритий півзими, закутаний в тутейші килими” (Л. Костенко); “Повезли мою матір на білих волах, неоплакану матір, неоплакану матір” (Л. Костенко). The basis of the participle is the basis of the verb, so the basic properties of the verb are the properties of the participle. To illustrate, let's do a morpheme analysis (parse) and define word formation rules for several randomly selected Ukrainian verbs.
At the morphemic level, units of language are morphemes (chains of phonemes for oral language and chains of letters for written language), a strictly ordered morpheme chain forms a lexeme. It has a semantic and grammatical or only grammatical meaning [6]. A typical [6] morpheme chain for Ukrainian language lexemes can be represented as follows: [prefix] + {root + [interfix]} + [suffix] + [ending] + [postfix].
the other brackets show morphemes that may be repeated.
Let review the participle word-form as a chain of three or four morphemes: запізн + і + л + ий, змарн + і + л + ий, роздрук + ова + н + ий, запрограм + ова + н + ий, посіj + а + н + ий, розпил’ + а + н + ий). In this article, the root with the prefix is considered together as one morpheme (“word stem”). Apostrophe indicates the softness of consonants meaning that л’у = лю, л’а = ля, etc. Word forms and morphemes are recorded mainly in accepted orthography with transcription, where it is convenient to formulate rules: in some cases, j is used to indicate the softness of consonants (only paired soft ones).
(
(
included to both grammatical conjunctions (втомлений, засохлий, промоклий,
мерзлий, опалий, змарнілий, розчервонілий, намерзлий, зажурений,
усміхнений);
(
Participles endings forms (full/reduced = f / f ). Table. 3 shows a list of morphemes of all the classes listed, indicating the required characteristics. стогн-( t , d , I , a%, , ся ) усміх-( t , d , I , atem , н , ся ) люб-( t , d , II , i%, , ся ся )
(
(
consonant – or a thematic element, or a suffix for the formation of verbs of the perfect and imperfect kind of predominantly foreign origin.
IX. Morphological and phonological rules. Morphological rules are rules that relate to
phonemic sequences (in our case, phonemes for simplicity are identified with
letters), and it is necessary to take into account the morphological role of these
sequences. Phonological rules deal simply with phonemic sequences, regardless of
their morphological status. For the following set of rules the phonological rule is
only rule IV.18.
(
II
Rule Example
the 1st person singular – г-ж, з-ж, к- чешу, засвистати – засвищу,
ч, х-ш, с-ш, т-ч, ст-щ, ск-щ полоскати – полощу, for example,
пекти –печу – печений;
sound changes are only in the 1st просити – прошу, водити –
водperson singular – д-дж, т-ч, з-ж, с- жу, возити – вожу, їздити –
ш, зд-ждж, ст-щ їжджу, тремтіти – тремчу,
мостити – мощу, for example,
мостити – мощений. The only
exception is the verb бігти (and its
derivatives: бігти, перебігти,
забігти etc), in which г alternates
with the same in all personal forms,
for example:бігти – біжу, біжиш,
біжить, біжать (вибігти –
вибіжу, вибіжиш etc).
(
Rule Example
with е in the roots вистелю – вислати –висланий – вистеляючий, беру – брати –
of verbs вибраний – вибираючий, завмер – замру – завмираючий, стер
– стертий – стираючий, умерти – умру –умираючий;
with и in the roots вибирати – вибраний – вибираючий, завмирати –
завмираюof verbs чий, стирати – стертий – стираючий, умирати –
умираючий.
(
-ан(ий)
Example
завішаний, сказаний, вихований
-ен(ий)
-ян(ий)
неоцінений, нескінчений, незлічений, завішений, куплений
порівняний
(
Example бомбувати – бомбування, марширувати – марширування, вивершувати – вивершування – вивершуваний, маркувати – маркування, очікувати – очікування – очікуваний; підбілювати – підбілювання – підбілюваний; -ова- мальований, or малювати, малювання; риштований, риштовання, or риштувати, риштування; друкований, or друкувати, друкування; підпорядкований, or підпорядкувати, підпорядкування; -овува- завойовувати – завойовування – завойований; пере
мальовувати – перемальовування –перемальований.
(
(
here, and therefore such entities, formally always possible, are allowed by these rules (and are generated by the grammar constructed on the basis of these rules).
Generative grammar G V ,T , S , P is a system that contains an alphabet – set V,
and its subset T of terminal elements, starting symbol S S V and s set of products
P. Set V / T is depicted as N. Elements from N are called non-terminal. Every product
from P should contain at least one non-terminal element in the left part.
(
ple just sorted out) are essentially Group VIII rules.
(
– kind d / d / d d ; a3 – conjugation I/II ; a4 – thematic a / i / a%/ i%/ o / atem ; a5 – possibility of suffix у / у%/ н / , a6 – possibility or need of accession ся ( ся / ся / ся ся ); "Pure" participle basis (without possibility or necessity accession ся ( ся / ся / ся ся ))with the same supporting information as above; the suffix of the participle with the information to it ( x – state, y – time, a3 – conjugation, see: above);
6 и, r, n
(full / abbreviated = f / f ), r – gender category (male /female/medium = m / w / k ), n – number (singular / plural = s / s ).
O, C, Ф – morphemes without information; Cd – suffix of the verb; T – thematic element; + – the boundary between morphemes; automatically appears after those morphemes that word forms cannot end; X – soft consonant, where X - the depiction of an arbitrary consonant; j – the depiction of sound [j] (yot).
(
nal or non-terminal symbols. In grammar G0 the rules of substitution are fol
lowing:
ing, the substitution rules are divided into groups (numbered in Roman numerals), each of which corresponds to a specific meaningful representation; this view is indicated by the group number. Group and rule numbers should not be construed as indicating the order in which they are applied: the order in which the rules of the generating grammar are applied is arbitrary.
DK DK x, y , where x act / pas ; y pres / past ; Expression DK DK x, y is used for reduction: in fact, there are not one but four rules corresponding to the valid sets of variable values x , y , for example DK DK act, pres then DK DK pas, pres and so on. In the other cases x , y , are used in the same way.
II. Implementation of grammatical values by corresponding morphemes. As
mentioned above, real Ukrainian morphemes and word forms are typed in italics.
(
tures which in this rule can acquire any values. So, for example, expression O d , a3 is a reduction for many expressions like O a1, d , a3, a4, a5 ,where a1, a3, a4, a5 acquire any valid values; accordingly expression C act, past also reductions for two expressions like C act, past, a3 . Therefore, for example, line II.1 actually contains not more than one rule, but many more different rules.
and the thematic element or suffix, if any). Here atem is used for reduction and
stands for any meaning of the sign a4 , different from atem , meaning
a / i / a%/ i%/ o . A similar meaning has the expression .
(
(
V. Implementation of a suffix for verb formation by a corresponding morpheme.
(
VI. The choice of the form of the participle ( f / f ) and the implementation of flexion by the corresponding morpheme.
1)
2)
3)
4)
5)
f
f , s ого, им,ому
f , m ий
f , w ою,ої
f , s им, ими, их
6)
7)
8)
9)
f
f , w а, y
f , k е
f , s i
10) C pas f о
VII. The implementation of the word stem of the corresponding morpheme. Group VIII
rules are not numbered because each line here is a whole set of rules whose
number is determined by the number of bases with this information.
(
(
(
(
Here X reduced string not exceeding three characters. It is about alternating о/а in the verb root of the type of гонити/ганяти, схопити/хапати, котити/катати, клонити/кланятися, кроїти/краяти, ломити/ламати, скочити/скакати, допомогти/допомагати. The consonant group, which appears to be rooted in that it alternates -o- (i.e., which separates it from the thematic element -а-/-я- before у(ю)ва- does not appear to contain more than three letters.
4. с' + Х ш + Х в' + Х вл' + Х б' + Х бл' + Х ........
д' + Х дж' + Х Тут Х = -е(є)н-, -у(ю)ва-, -ова-, -овува-.
Before suffixes -е(є)н-, -у(ю)ва-, -ова-, -овува- the ending consonant of the word stem -с'- alternates for -ш-, and the ending consonant -б'- for -бл'- (same for д'-ж, т'ч, в'-вл, etc; but in our list there is no word stems with -д'-, -т'-, -в'-).
5. д + Х д' + Х; с + Х с' + Х; … Before suffixe Х (-е(є)н-, -у(ю)ва-, -ова-, овува-) the solid ending of consonants of the foundations are softened by: д-д', с-с', etc.; принес + ти – принес' + ен+ ий.
6. нн + Ф н + о.
IX. X. Graphic-orthographic rules. Х у Х ю Х е Х є sive chain is directly derived from the previous one, then X n is output from X 0 ; the same sequence X0, X1,..., X n is called output X n from X 0 [2]. This means that X 0 does not necessarily translate into X n in one step, but through the consistent use of several substitutions. Obviously, direct inference is a separate case of inference. Here is an example of output in the presented grammar G0 (in parentheses, to the left of the output line, specifies the rule number, the application of which to the previous line gives this).
DK (I) DK pas, pres (II.3) Ot, d , I , atem, y, ся ся C pas, pres, I (III.3) O t, d , I , atem, y C pas, pres, I (VI.13) O t, d , I , atem, y oва (VII.1) O t, d , I , atem, y oва f (VII.3) O t, d , I , atem, y oва ий (VIII.10) розфарб oва ий (XI) розфарбoваий
only of terminal characters) is called complete. Of course, not every derivation beginning with an initial character is complete; in particular, such outputs beginning with an initial character that cannot be continued until full output are "deadlocked output".
0 DK (I) DK act, past (II.2) O t, d , II , atem, , ся ся C act, past, II (Ш.3) O t, d , II , atem, C act, past, II (VI.1) O t, d , II , atem, л (VII.1) O t, d , II , atem, л f (VII.3) O t, d , II , atem, л ий (XI) O t, d , II , atem, лий
not end with a chain of terminal characters (Ukrainian letters). In grammar G0 , there is no rule whose left side would be contained in the last chain of this output. This is explained by the fact that in the Ukrainian language there are no athematic verbs (the base ends in the consonant), which are cancelled by the II conjugation. Thus, the signs atem , and II , which were interpreted as independent in constructing the grammar G0 , are in fact connected. Their coherence could be taken into account, but this would make the grammar more noticeable. Formal grammar does not require at all that any derivation in it ends with a correct terminal chain: it is sufficient that any complete derivation gives the correct chain (in our case, the form of the Ukrainian participle) [2, 5]. Thus, the language generated by grammar G is the totality of all participles of the above verbs. Obviously, this lan0 guage is over. However, grammars can also produce endless languages. Thus, the use of grammar is the construction of complete inferences; the last strands of these derivations form a grammar-generated language. 5
To simulate the generation of language and texts at the morphemic level, a linguistic analysis of the set of adjectives was performed, the classes of morphemes involved in the creation of word forms were identified, and the rules for the combination of morphemes with each other were determined. Identified features of word formation can be used to construct systems of computer morphological analysis and synthesis [6]: to create a generative grammar that generates language - a set of all considered adjectives. Application of the theory of generative grammars to solve problems of applied and computational linguistics at the level of morphology allows to create systems of synthesis of language and texts, as well as to create textbooks of practical morphology, to create tables of words, to list morphemes (affixes, roots), to determine productivity , to determine the frequency of implementation in texts of different grammatical categories (categories of genus, case, number, etc.) for specific languages [6, 7].
CSIT 2018 – Proceedings 2, 34-38 (2018) 36. Lytvyn, V., Sharonova, N., Hamon, T., Cherednichenko, O., Grabar, N., Kowalska
(COLINS-2019). In: CEUR Workshop Proceedings, Vol-2362. (2019) 37. Lytvyn V., Vysotska V., Peleshchak I., Basyuk T., Kovalchuk V., Kubinska S., Chyrun L.,
of Similar Texts in Big Data. In: 2019 IEEE 14th International Scientific and Technical
(2019) 38. Vysotska V., Lytvyn V., Kovalchuk V., Kubinska S., Dilai M., Rusyn B., Pohreliuk L.,
39. Cherednichenko, O., Babkova, N., Kanishcheva, O.: Complex Term Identification for
40. Bobicev, V., Kanishcheva, O., Cherednichenko, O.: Sentiment Analysis in the Ukrainian and Russian News. In: First Ukraine Conference on Electrical and Computer Engineering (UKRCON), 1050-1055. (2017) 41. Fedushko, S., Benova, E.: Semantic analysis for information and communication threats detection of online service users. In: The 10th International Conference on Emerging
42. Antonyuk N., Chyrun L., Andrunyk V., Vasevych A., Chyrun S., Gozhyj A., Kalinina I.,
43. Chyrun, L., Chyrun, L., Kis, Y., Rybak, L.: Automated Information System for Connection to the Access Point with Encryption WPA2 Enterprise. In: Lecture Notes in Computational
44. Kis, Y., Chyrun, L., Tsymbaliak, T., Chyrun, L.: Development of System for Managers
gence and Decision Making, 1020, 405-421. (2020) 45. Chyrun, L., Burov, Y., Rusyn, B., Pohreliuk, L., Oleshek, O., Gozhyj, A., Bobyk, I.: Web
Vol-2386, 255-273. (2019) 46. Chyrun, L., Gozhyj, A., Yevseyeva, I., Dosyn, D., Tyhonov, V., Zakharchuk, M.: Web
126-142. (2019) 47. Gozhyj, A., Chyrun, L., Kowalska-Styczen, A., Lozynska, O.: Uniform Method of Operative Content Management in Web Systems. In: CEUR Workshop Proceedings, Vol-2136, 62-77. (2018) 48. Chyrun, L., Kowalska-Styczen, A., Burov, Y., Berko, A., Vasevych, A., Pelekh, I.,
ment. In: CEUR Workshop Proceedings, Vol-2386, 35-54. (2019)