ing, text translation in / other languages, etc. [1, 3]. An urgent task today is to create synonymizers - computer programs that replace words or constructions with synonyms in texts and are typically used when writing texts by authors, as well as for rewriting. Such programs can also be used as a reference system for anyone who is learning a certain language or wants to improve their knowledge. We do not know synonymizers on the material of the Ukrainian language, therefore the offered research has scientific novelty.
In modern applied linguistics an important place is occupied by research on various aspects of automatic processing of natural speech. Significant experience has been gained in creating computer programs for text conversion of various types, in particular, text editing, text annotation, text
2021 Copyright for this paper by its authors. [6: 945] (author of the afterword to the dictionary of synonyms of the Ukrainian language O. Taranenko).
According to this approach, there are several types of synonyms for different parameters
(I. Kobozeva, M. Kronhaus, V. Levitsky, O. Taranenko, etc.):
1. By completeness / incomplete coincidence of values - absolute (complete) and partial
(incomplete): родина and сім’я; іти and шкандибати.
2. By the presence / absence of semantic differences and stylistic restrictions on use - semantic,
stylistic, and semantic-stylistic: говорити, мовити, розмовляти, казати, лопотіти, базікати,
балагурити, патякати, пащекувати, перебендювати, теревенити, промовляти, ректи,
цідити, etc.
3. By morpheme features - multi-rooted and single-rooted: активний, пасивний and неактивний.
4. By the presence / absence of the influence of the context on the meaning - common and
contextual: сторіччя and століття; холодний, відсторонений and сумний (state of mind).
5. By the presence / absence of connecting restrictions: lexical and grammatical (syntactic) synonyms:
абетка and алфавіт; закритий and зачинений (about windows, doors) [
It should be noted that a broad vision of synonymy is inherent not only in Ukrainian studies but
also common in world practice [
Complete identity of the content of language units is a rather rare phenomenon. It is generally accepted that most synonyms in any language belonging to a group of partial synonyms, ie synonymous words or synonymous words and constructions have certain differences that may relate to different aspects of content. In particular, the team of researchers led by Yu. Apresyan [29, 30] emphasizes the distinction between absolute synonyms (those that have no differences) and quasisynonyms (those that may have the semantic, reference, communicative, syntactic, connective, morphological, and/or millet differences).
This aspect is extremely important when creating a synonymizer, as not all words can be replaced automatically without reservations (taking into account the context or style restrictions, etc.), so it is necessary to provide the user with the ability to view relevant notes (remarks) and independently choose from the proposed list of synonyms. Note also that different types of partial synonyms have different terminological designations in theoretical and lexicographical works, which is due to both the peculiarities of the views of researchers and terminological traditions within different linguistic schools.
Methods of analysis and synthesis, descriptive method, component analysis and distributive analysis were used to create the synonymizer. The program was developed using the Python 3.x programming language. Tkinter, CSV was used for the test version of the application. In further work, it is planned to use the open-source relational database management system MySQL, so a relational database model was created.
The process of creating a synonymizer as a computer program is a rather complex process that includes several stages: 1. Analysis of available synonymizer programs in other languages. 2. Problem statement. 3. Compiling a register of synonyms of the Ukrainian language based on authoritative lexicographical works. 4. Database design. 5. Test filling of the database. 6. Software implementation using Python. 7. Construction of the visual interface. 8. Testing of the created computer product.
It is only natural that before creating our computer program, we analysed the synonymizer
programs available on the Internet. The vast majority of available programs are based on English or
Russian [
The analysis of computer products made it possible to highlight the following features: 1. Type (version) of the program (synonymizer can be an application program that must be installed on a personal computer, online service, or mobile application). 2. Automatic, manual or mixed mode of substituting words for synonyms (i.e. some programs involve a person in the choice of synonyms, and those that do not have such an option). 3. Consideration / disregard of morphology (presence / absence of lemmatizer). 4. Presence / absence of a database of exceptions (constructions that cannot be replaced due to various reasons, first integrity of value and connecting features). 5. Taking into account / not taking into account when replacing the word the linguistic-statistical law of Zipf (George Kingsley Zipf), according to which frequency words should be replaced by frequency words, and rarely used - rarely used. 6. The presence or absence of a report on the work performed (start time, end time, the number of processed words, the register of replaced words, the list of used synonyms, the percentage of replacements). 7. The availability of additional options (for example, highlighting words that have synonyms, to attract the user's attention, the ability to choose from several proposed databases, etc.). We try to take these parameters into account when creating our program.
In the second stage, it is determined that we aim to create an experimental version of the synonymizer for Ukrainian-language texts, which requires installation on a personal computer, provides a mixed-mode of substituting words for synonyms, contains a database of exceptions and can colour words with synonyms. However, as will be shown below, in the process of testing the created program, the authors changed the task, as they concluded that it is possible in principle only to manually select synonyms by the user. It turned out that the data on synonyms available in dictionaries today is not enough for the correct automatic replacement of words in the text.
In the future, this product should be available online, should take into account the morphology and automatically generate a report on the work done.
In Ukraine, there is considerable experience in the lexicographic processing of various lexical
oppositions, including synonyms, so when compiling a register of synonyms, we seek to take into
account the achievements of modern lexicographers. The register was formed by combining all
language units presented in three authoritative dictionaries: the Dictionary of Synonyms of the
Ukrainian Language in Two Volumes (2001), the Practical Dictionary of Synonyms of the Ukrainian
Language by S. Karavansky (2014) and the Dictionary of Synonyms of the Ukrainian Language by
O. Vusyk (2013) [
Important and painstaking steps in creating a synonymizer were the design of the database and its test content. After several attempts to build a database, we settled on a database structure that includes 5 interconnected Excel spreadsheets.
Table 1 presents all the words of the Ukrainian language that have synonyms and are recorded in the dictionaries analyzed by us. Table 1 contains 5 columns: ”#”, “Word”, “Interpretation”, “Status” and “Characteristic”. Insertion column “Status” reasoned that practice showed lexeme describing the need for differentiation status word as submission database definitive words polysemantic words and homonyms are different.
The simplest case is to display unambiguous words in the database. Let's show this with the example of a pair of synonyms мовознавство and лінгвістика (see Figure 1).
In the column “Characteristic”, if available, provide information about the stylistic affiliation, normative / non-normative (dialectal, obsolete, colloquial, etc.) words. For a мовознавство and лінгвістика, these cells remained empty.
Table 2 “Synonyms” includes 4 columns: “Word”, “Synonym”, “Type” and “Characteristic”. The column “Type” indicates an absolute or partial synonym, the column “Characteristic” for partial synonyms shows the differences (semantic, stylistic, etc.), for absolute synonyms this cell remains empty. Figure 2 shows a fragment of the table for words мовознавство and лінгвістика.
Since there are different synonyms for different meanings of a word in a language, each meaning of a polysemous word should be given separately in the database and synonyms should be given for each meaning. Let us show this by the example of the word бажання which in the modern Ukrainian literary language, according to the "Dictionary of the Ukrainian language", has three meanings: 1. Aspiration, the desire to accomplish something; will. 2. An opinion expressed by someone about the desirability of doing something; wish. 3. Love attraction, passion [22] (see Figure 3).
In the created database homonyms are given as separate units. For example, four words коса with different values are shown in the table “Lexemes” as shown in Figure 5.
потребу – потребувати, мати змогу – могти and under. This phenomenon is characteristic of
many languages of the world, including Ukrainian. Verbal-noun descriptive constructions are
qualified by researchers as one of the manifestations of analytical tendencies in language and as one
of the components of the sentence paradigm. In linguistics, the term "descriptive predicate" is also
used to denote such constructions (for details, see [
It is important for our tasks that descriptive constructions can be absolute and partial synonyms for the corresponding verbs. For example, descriptive predicates are to the verb бажати equivalent (partially equivalent) are descriptive predicates мати бажання, виявити бажання, відчувати бажання, палати бажанням, палати від бажання, кипіти від бажання. The last three constructions are partial synonyms: палати бажанням, палати від бажання have a differential (additional) sema ‘intensity of state’, and кипіти від бажання have two additional seven - ‘intensity of state’ and ‘high degree’ (see Figure 7).
The last fifth table is designed to show periphrases that can be an additional resource for
synonymous substitutions in the text, for example: гарант Конституції – Президент, імперія
зла – СРСР, місто Лева – Львів, українська Прага – Львів, перша столиця – Харків, колиска
козацтва – Запоріжжя, перлина Поділля – Вінниця, перлина Карпат – Яремча and so on. The
sources of this table were the “Dictionary of media paraphrases of the XXI century” [28], own file of
paraphrases and examples given in scientific papers on this issue [
We will show a fragment of this table, covering the paraphrases of the political figure Yulia Tymoshenko, recorded in different periods of her activity in the texts of Ukrainian periodicals: газова леді, газова принцеса, Дон Кіхот у панчохах, леді Ю, українська Жанна Д’Арк, etc. (see Figure 8). Such figurative names, of course, are incomplete synonyms, as they are always evaluated, contain a negative or positive assessment of the activities or appearance of the person. In this case, the same paraphrase (for example, газова принцеса) can express different assessments depending on the communication situation, context, purpose of the author of the message. That's why it's important to be able to manually select a synonym for a computer product user.
The five steps formed the basis for the software implementation of the synonymizer using Python 3.x. The urgency of creating a separate software product is dictated by the fact that at the moment we can name a fairly small number of available software applications for professional work with the Ukrainian language. At the same time, the demand for linguistic computer processing of large volumes of Ukrainian texts is constantly growing, and user requests are becoming more professionally oriented.
The prototype of the service “Synonymizer” was implemented locally, to build and test algorithms for solving the stated problem. To do this, the basic capabilities of the Python programming language were used, in particular, the functions of the Tkinter module, libraries for working with the CSV format, and so on.
Before starting work on creating a software product “Synonymizer” it is necessary to normalize
these tables following the standards of work with relational databases [
The described program “Synonymizer” is only a prototype of the future service, so the proposed product has a minimalist interface.
After entering the text, the user clicks the Run button, the Result tab is created and the user is allowed to select a possible synonym for the words that are available in the previously described database of synonyms of the Ukrainian language. An example of the program is shown in Figure 10.
If we just enter a set of related words - synonyms (e.g., безбоязний, безстрашний, безтрепетний, відважний, відчайдушний, відчайний, дерзкий, дерзновенний, зухвалий, мужній, небоязкий, небоязливий, нелякливий, неполохливий, пискатий, сміливий, хоробрий), we get the corresponding result (see Figure 11).
The program allows you to track homonyms in the entered text as specific language units. For example, consider the sentences Кам'яна коса, пісок, красиве небо і дає світло маяк... (https://www.nikopoltoday.com/article-364/Nikopol-Akim-bulo-misto-Nikopol-10-rokiv-tomufotoekskursia). In this case, the noun "коса" is combined with the adjective кам’яна, so according to the information presented in Figure 6, a synonym for the word коса can only be the word мис, other options (волосся, волос, селезінка) are incorrect.
Figure 12 and Figure 13 show the analysis of the above case: basic (all synonyms available in the database are allowed) and extended (taking into account compatibility).
However, at the stage of testing the program, there were many examples in which, even taking into account the combination of words, the automatically proposed option was incorrect. In particular, in the sentence Медитувати ‒ кам'яна коса нашого пляжу ідеально підходить для цього (https://www.knyazhahora.com/news/2020/12/10/) the replacement of the word коса with a мис (see Figure 14) causes objection because it is impossible to say мис нашого пляжу, although such restrictions are not described in dictionaries. Here is another example: in the sentence Ти висока струнка, в тебе руса коса (Song “Time flows by the river”), the lexeme коса can be replaced by волосся (automatically based on compatibility with the word руса). However, such a replacement is not possible in the example: Довга дівоча коса була символом краси, естетики, дівоцтва та незайманості (https://amazing-ukraine.com/chorni-ochka-iak-teren-ta-rusa-kosa-do-poiasa-abodivocha-kosa-v-ukrainskykh-tradytsiiakh /) because the symbolic load was the girl's braid, not hair at all. Perhaps, in this case, the formal criterion for prohibiting synonymous substitution is the presence of the word символ in the sentence. Such cases convinced us of the fundamental possibility of only manual selection of synonyms by the user from the list of proposed and caused a change in the task of building a synonymizer with mixed-mode to a program with the manual mode of word substitution.
At the same time, we consider the implementation of the choice of synonyms on the material of a representative text corpus to involve statistical analysis data to be promising. Besides, synonyms can be suggested for frequent placement in combination with neighbouring words (collocations analysis).
The settings allow you to display additional information in the tables as reference materials for the correct choice of a particular synonym. This requires additional contextual analysis of the textual information entered by the user, but this is one of the immediate prospects of the software application.
After the testing phase, specialized libraries for working with text data, Python libraries for working with SQL databases, web programming technologies, etc. will be used to build the software product. Further development of the software product is planned in two directions. First of all, the formation of a full-fledged database of synonyms of the Ukrainian language will be continued by describing scheme. In parallel, the software interface will be developed, taking into account modern standards of usability ISO / IEC 25010. In the future, it is planned to build a full-fledged web application with additional useful linguistic services.
Thus, we have developed an experimental local version of the synonymizer based on the Ukrainian language. This version of the program provides the possibility for the user to choose synonyms from the list offered in the database and takes into account the database of exceptions (restrictions on synonymous replacement). The question of taking into account the immediate environment of the language unit to choose an appropriate synonym needs further elaboration. The performed testing of automatically obtained word substitutions in texts in the Ukrainian language testified to the incompleteness of the available knowledge about synonyms, the insufficiency of the information provided in the dictionaries of synonyms for the correct transformation of the text fragment and requires the solution of statistical methods (association measures) for the correct choice of synonyms.
Prospects for further research are to involve also synonymous paraphrases at the sentence level. This approach will be based on the notion of the syntactic paradigm of a sentence (see the works of M. Vsevolodova, G. Zolotova, I. Vykhovanets, A. Zahnitko, etc.). The same event of objective reality can be expressed in different sentences, the same or different structure. N. Huyvanyuk calls such sentences “co-referential syntactic units” [14]. First of all, we mean the following cases of transformation of the basic sentence [8; 15; 24]: 1. Sentence transformation (opposition “active - passive state”): Ціла бригада будує наш дім → Наш дім будується цілою бригадою (or Наш дім будовано цілою бригадою). 2. Sentence conversion - changing the direction of the relationship between the participants of the situation: Богдан – друг Євгена ↔ Євген – друг Богдана; Син вищий за батька ↔ Батько нижчий за сина; Іван купив автівку в Петра ↔ Петро продав автівку Івану; Зошит лежить на столі під книгою ↔ Книга лежить на столі на зошиті, etc. 3. Synonymous transformations of different types: Молекула складається з атомів; Молекули містять атоми; Атоми – складники молекули, etc.; Мій батько має нагороду за мужність; У мого батька є нагорода за мужність; Мого батька нагородили за мужність, etc.
However, the program implementation of the interchange of such cases involves not only the formation of appropriate databases (in particular, active-passive verbs, pairs of words- conversions) but also the formulation of clear grammatical rules of relevant transformations, determining the sequence of their application and possible limitations and exceptions.
[13] M. B. Bulakh, Komunikatyvnyi status peryfrazu v media: stylistychna aspektolohiia [Communicative status of periphrasis in the media: stylistic aspectology]: avtoref. dys. kand. filol. nauk: 10.02.01 – ukrainska mova, Kyiv, 2017. [14] M. V. Vsevolodova, Teoriya funkcional'no-kommunikativnogo sintaksisa. Fragment prikladnoj (pedagogicheskoj) modeli yazyka [Theory of functional-communicative syntax: a fragment of the applied (pedagogical) language model], URSS, Moskva, 2017. [15] N. V. Huivaniuk, Formalno-semantychni spivvidnoshennia v systemi syntaksychnykh odynyts [Formal-semantic relations of syntactic units in the system of syntactic units], Ruta, Chernivtsi, 1999. [16] Online Sinonim, 2013. URL: https://online-sinonim.ru/. [17] O. S. Vusyk, Slovnyk synonimiv ukrainskoi movy: ponad 2500 synonimichnykh hnizd [Dictionary of synonyms of the Ukrainian language: more than 2500 synonymous nests.], Navchalna knyha – Bohdan, Ternopil, 2013. [18] Paraphrase Online, 2021. URL: https://www.paraphrase-online.com/. [19] Paraphrasing Tool, 2021. URL: https://paraphrasing-tool.com/. [20] S. Karavanskyi, Praktychnyi slovnyk synonimiv ukrainskoi movy [Practical Dictionary of
Synonyms in the Ukrainian Language]. 5-te vyd., opratsovane i dopovn., BaK, Lviv, 2014. [21] SeoGenerator, 2009. URL: http://www.seogenerator.ru/. [22] Slovnyk ukrainskoi movy u 20-ty tomakh [Dictionary of the Ukrainian language in 20 volumes].
URL: https://services.ulif.org.ua/expl/Entry/index?wordid=1&page=0 [23] Synonymizer, 2006. URL: http://synonymizer.com.ar/index.php#top. [24] Ukrainska mova: Entsyklopediia [Ukrainian language: Encyclopedia]. 2-e vyd., vypr. i dop.,
Vyd-vo «Ukrainska entsyklopediia» im. M. P. Bazhana, Kyiv, 2004. [25] USyn, 2009. URL: http://usyn.ru/online.php. [26] V. A. Kuzʹmenkova, Tipologiya opisatel'nyh predikatov i ih analogov v sovremennom russkom yazyke [Typology of descriptive predicates and their analogues in modern Russian]: avtoref. diss. ... kand. filol. nauk : 10.02.01 – russkij yazyk. Moskva, 2000. [27] V. V. Levickij, Semasiologiya [Semasiology], Nova knyga, Vinnica, 2006. [28] Ya. V. Kozachok, V. M. Vasylchenko, I. A. Kostiuk ta in., Slovnyk mediinykh peryfraziv ХХІ stolittia [Dictionary of media paraphrases of the XXI century], NAU, Kyiv, 2014. [29] Yu. D. Apresyan, Izbrannyye Trudy.Tom I. Leksicheskaya semantika [Selected works. Vol. I.
Lexical semantics]: 2-e i zd., ispr. i dop. Shkola «Yazyki russkoy kul’tury», Izdatel’skaya firma «Vostochnaya literatura» RAN, Moskva, 1995. VIII s. [30] Yu. D. Apresyan, V. Yu. Apresyan, E. E. Babaeva, O. YU. Boguslavskaya, I. V. Galaktionova, M. Ya. Glovinskaya, S. A. Grigor'eva, B. L. Iomdin, T. V. Krylova, I. B. Levontina, A. V. Ptencova, A. V. Sannikov, E. V. Uryson, Novyj ob"yasnitel'nyj slovar' sinonimov russkogo yazyka [New explanatory dictionary of synonyms of the Russian language]. Vtoroe izdanie, ispravlennoe i dopolnennoе, Yazyki slavyanskoj kul'tury: Venskij slavisticheskij al'manah, Moskva; Vena, 2004. (Studia philologica)