Methods of automated metaphor identification (based on the concept of страх "fear")

Размещено на http://www.allbest.ru/

METHODS OF AUTOMATED METAPHOR IDENTIFICATION (BASED ON THE CONCEPT OF СТРАХ `FEAR')

Olena-Anna Hultso

Lviv Polytechnic National University Lviv

Olena Levchenko

Lviv Polytechnic National University Lviv



Introduction

In cognitive linguistics, metaphors are the most interesting subject of research due to the mechanism of their appearance, because it is reproduced in speech through experience, knowledge, human interaction with the outside world. They have an inseparable connection with the cultural paradigm of native speakers, which is so ingrained in the mind and becomes familiar to people who are not aware of these metaphors. The corpus approach in the study of metaphor helps to process metaphorical compounds in language faster.

The purpose of the article is to improve the existing method of automated identification of metaphors in GRAC with the help of semantic tags in Ukrainian texts, in particular, metaphors of the VERB+NOUN model related to the concept of СТРАХ. The experiment is based on the theory of cognitive metaphor, the theory of embodiment and conceptual analysis. The textbase is formed with the help of GRAC v.14 corpus.

Methods. Method of analysis for detailed study and use of the possibilities of applying corpus and software of corpus linguistics in research of concepts. Therefore, metaphorical modelling mapping is used to research and study the structure of metaphor. The method of contextual analysis was used to analyze texts and determine comparative associations in the process of conceptual analysis of sentences, as well as the descriptive method and the method of quantitative analysis. Cognitive-semantic and linguistic-cultural approaches help to reveal the conceptual characteristics of the concept and establish the spheres of its application.

The originality of the article lies in the improvement of methods and tools of automatic/automated identification of Ukrainian metaphorical models. Theoretically, the study is based on Conceptual Metaphor Theory. The focus of the study is on the fiction texts of the XXI century and focuses on the concept of СТРАХ and its metaphorical model VERB+NOUN. The method of automated metaphor identification was tested on the material of the GRAC corpus.

Results and conclusions. As a result of the analysis of contextual use of nominees of the researched concept, the most productive metaphorical models of the concept СТРАХ are established, with the help of semantic tags in the corpus the percentage of the most used semantic groups of verbs in the verb + noun model is revealed. In addition, the gender of semantic groups of verbs is quantified. The research proved high efficiency of the corpus-linguistic method in the metaphors studied. Although at this stage of its development, the GRAC corpus has only partial semantic tagging, in particular that of verb groups. The inaccuracies of automated metaphor identification occur due to indistinguishable homonymy, but a number of techniques have been found to help solve this problem in identifying metaphorical compounds.

Keywords: conceptual metaphor, metaphorical model, semantic corpus tagging, automated identification of metaphors.



Анотація

МЕТОДИ АВТОМАТИЧНОЇ ІДЕНТИФІКАЦІЇ МЕТАФОР (БАЗУЮТЬСЯ НА КОНЦЕПТІ СТРАХ)

Олена-Анна Гульцьо, Національний університет «Львівська Політехніка» Львів.

Олена Левченко, Національний університет «Львівська Політехніка» Львів.

Статтю присвячено вдосконаленню методів та інструментів автоматичної / автоматизованої ідентифікації українських метафоричних моделей. Дослідження ґрунтується на теорії концептуальної метафори.

Зроблено спробу вдосконалити наявну методику автоматизованої ідентифікації метафори в Генеральному регіональному анотованому корпусі української мови (ГРАК) за допомогою семантичних тегів в українськомовних текстах, зокрема метафоричної моделі «дієслово+іменник», що стосується концепту СТРАХ. З'ясовано можливості застосування корпусу та програмних засобів корпусної лінгвістики в дослідженні концептів.

Фокус дослідження зосереджено на художніх текстах XXI ст. та спрямовано на концепт СТРАХ і його структурно-семантичну модель «дієслово+іменник». Методику автоматизованої ідентифікації метафори апробовано на матеріалі корпусу ГРАК.

Проаналізовано концепт СТРАХ та укладено перелік основних моделей, за якими цей концепт метафоризується: СТРАХ ЦЕ ОБ'ЄКТ, СТРАХ ЦЕ ІСТОТА, СТРАХ ЦЕ ЛЮДИНА, СТРАХ ЦЕ СУПРОТИВНИК. Виявлено, що «метафоричний ефект» створюють дієслова, належні до низки семантичних груп, як-от: дієслова психічного простору, дієслова існування тощо.

З'ясовано, що корпусний підхід та відповідне семантичне тегування дієслів забезпечують точнішу автоматизовану індентифікацію метафоричних сполук. Засвідчено високий показник ефективності запропонованої методики. Однак на цьому етапі свого розвитку корпус ГРАК семантично розмаркований лише часково. Зокрема, ця заувага стосується нечастотних дієслів. Неточності автоматизованої ідентифікації метафори зумовлено нерозрізненою омонімією, проте запропоновано низку прийомів, які допомагають вирішити це завдання в процесі виявлення метафоричних сполук.

Ключові слова: концептуальна метафора, метафорична модель, семантичне маркування корпусу, автоматизована ідентифікація метафор.



Formulation of the problem

Three decades ago, cognitive structures and mechanisms of operation of these structures appeared in the field of research interest of the humanities. Cognitive science aims to study the processes of perception, categorization, classification and understanding of the world and the system of representation and storage of knowledge. The focus of which is language as a general cognitive mechanism and its cognitive aspect in language reflection. The problem of cognitive linguistics is the categorization of the surrounding reality, where the main thing is metaphor, as a manifestation of the semantic capabilities of the human mind. Modern interpretation of metaphor in cognitivistics characterizes it as a basic and mental operation, as a way of knowing, categorizing, conceptualizing, evaluating and explaining the world. Paradigms about its ontological and epistemological aspect became the primary preconditions for the study of metaphor due to the cognitive approach. The study of metaphor using the corpus approach demonstrates a new approach in the study of the process of metaphorization.

Today, automated identification of metaphor is an important task for a number of natural information-oriented linguistics. Research is being conducted in the following directions: (1) creation of data sets with metaphorical associations; (2) identification of new characteristics useful for identification of metaphor; (3) interlingual identification of metaphors. Also presented are various approaches to the use and creation of datasets.

Analysis of recent research and publications. However, the problem of automated / automatic identification of metaphors still remains unresolved. Attempts have been made to solve this problem on the basis of a number of languages, in particular for English. Identification of metaphors based on the metaphors of the subject verb and the object verb described in Lakoffs research [11; 13]. Using the WordNet database and an initial set of manually annotated metaphorical expressions, the system searches for similar syntax structures in the corpus [18]. There are also a number of studies based on WordNet that improve already automated processes [22].

The paper also proposes the method of the CSF (Common Semantic Features) for identifying metaphors in English [21]. Intercultural models of thought through interlingual use of metaphor were investigated. Metaphor location in English and Spanish texts is a system based on LDA modeling (Latent Dirichlet Allocation) that detects metaphors of any syntactic structure [8]. A research of the automatic identification of conceptual metaphors in the Hungarian language was conducted, which compares the effectiveness of three methods: association, dictionary and corpus. The experiment confirms the effectiveness of the corpus-linguistic method, the concept of source and target are better characterized by statistical models than by psycholinguistic factors [1]. Research using NLP for metaphors in clinical trials has become relevant. These automation techniques can help diagnose and detect the early signs of schizophrenia in patients. Statistics have shown that people diagnosed with schizophrenia use more metaphors in speech than healthy ones [7].

In the Ukrainian language, the processes of metaphor automation are described in a number of papers. Among them are metaphorical combinations Adjective+Noun based on dictionary definition based on GRAC v.3 and CYM-11 / Dictionary of Modern Ukrainian-11 [14]; algorithm for automated determination of metaphorical combinations Adjective+Noun is demonstrated by the example of the phrase white+noun [17]; advantages of the combined semantic and statistical principle of metaphorical corpus annotation and automated identification of metaphors [13]; identification of collocation metaphors of the formal model NOUN+NOUN in the semantically annotated corpus of texts [15].

Research in recent years has shown increasing interest among researchers in the field of conceptual metaphor. It is widely used in the study of political discourse and in the media, sports, medicine. Anna Doquin de Saint Preux and Ocarina Masid Blanco in their article describe war and sport as metaphorical areas. In particular, how they affect the emotions, feelings and decisions of Spaniards as a result of the pandemic [5].

In the next article, the conceptual metaphors of water are explored in Javanese proverbs. Since the object of study is water, data on it are taken from the Javanese Proverbs Dictionary by Darmasoetjipta. Researchers focus on the theory of cognitive-linguistic approach and qualitative paradigm. They note that metaphor is primarily a basic embodiment of human nature, so it is an effective tool in studying the mechanism of human thinking through language. The results showed that for the Javanese the concept of water is an ideological value and a reference point in social behavior and relationships [16].

The purpose of the article is to improve the existing method of automated identification of metaphors in GRAC with the help of semantic tags in Ukrainian-language texts, in particular, metaphors of the VERB+NOUN model related to the concept of СТРАХ. The experiment is based on the theory of cognitive metaphor, the theory of embodiment and conceptual analysis. The textbase is formed with the help of GRAC v.14 corpus.

Lingiustic background. The research is based on the theory of conceptual metaphor developed by J. Lakoff and M. Johnson [12]. In cognitive linguistics, metaphor is primarily a conceptual process, so metaphorical relationships are considered to be conceptual reflections that characterize not only our use of language, but also everyday life, thought and behavior [10]. The target domain and the source domain are the basis of the conceptual metaphor. Therefore, metaphorical modeling -- mapping is used to research and study the structure of metaphor. In particular, J. Jaynes describes the role metaphor plays in the formation of the human cognitive system. He believes that metaphor helps to broaden the perception and understanding of the world and human consciousness, because the evolution of consciousness is linked to the ability to metaphoricate [9]. The assumption that the generation of abstract concepts in the mind/brain is based on specific human knowledge based on the bodily experience of the world is the main evidence for the theory of embodiment in cognitive linguistics [11; 12]. A. Deignan [2] confirms the effectiveness of corpus-linguistic results and describes that metaphorical expressions have a stable grammatical form. That is, the source domain words denote actions and properties, grammatically expressed as verbs and adjectives. Confirming the theory of blending [6], when using a metaphorical language, a person creates a so-called mixed domain, having a certain structure and relation, and accordingly linguistic specificities.



Main research presentation

At the first stage of research we create in the GRAC a subcorpus with such characteristics: style: fiction (doc.style: FIC); dating: 2000-2021 (doc.publication year); original language: Ukrainian (doc.original: UK).

For automated identification of metaphor, a method for formalizing it is structural semantic models should be developed. At the initial stage of analysis, we detect verb+noun model word combinations with страх. In the GRAC corpus with CQL query

[tag=".*verb.*"] [lemma-'страх"]

perform a search. 1164 contexts were obtained.

From the analysis of the results of the search model verb+страх it is clear that there are contexts in which the homonymy of the noun and the adverb страх is not distinguished. Examples: За містом канонада гремить, то поспішали страх як (У. Самчук); Було страх як спекотно <...> (Л. Дереш); Випить уже не було за віщо, а випить страх як хотілось (А. Кондратюк); Мій чоловік, який давно вже перетворився мені на брата, і розумів страх як багато, терпляче спостерігав за моєю депресією (О. Компанієць); <...> російської мови в місті виявилося страх як багато (Ю Андрухович). Since homonym occurs in combination страх+як we managed to extract 7 model compounds verb+страх+як from the search with queries:

[tag=".*verb.*"][lemma="crpax"][tag!= "adv:&amp;pron:int:rel"] or [tag=".*verb.*"] [lemma-'страх"] [word!-^"].

There are also cases of non-metaphorical compounds, where страх corresponds to the meaning дуже `very' and is combined with an adverb and an adjective: ...аще загримить місяця іюня, то буде страх великий... (Р. ІваничукА' <...> учням було страх цікаво, що воно буде, особливо після його перепалки (Т. Гаврилів); Коридори були страх вузькі, через деякі взагалі доводилося протискатися боком (О. Українець); А ось теперечки він став страх яким перебірливим -- вічно йому щось не до вподоби! (M. Кідрук).

There are also cases of homonymy, when the noun in the dative case of страхові `fear' [3] coincides with the adjective plural in the accusative case: associated with social insurance; who is engaged in insurance [4, p. 755]. Examples: ...і не заплатила страхові данини своєю красою (В. Барка); Тому вони й вигадали страхові компанії (А. Санченко); Мирослава закликала їх мати совість, бо вона ж за кожну сплачує і до пенсійного фонду, і платить страхові внески (С. Талан). metaphorical fear verb noun

In the verb singular, 1 person, the present tense їм `eat' we have found homonymy with the personal pronoun вони `they' in the dative case: Тільки-но склеплюють повіки, задрімають, так і увижаються їм страхи, жахи несусвітні (A. Кондратюк); Один із великих демонів-галла -- ім'я їм страх -- прохарчав <...> (O. Захарченко).

Using the semantic tags of the corpus GRAC [19, 20] we classify the verb components used in metaphorical compounds. It should be noted that semantic tagging is still going on in GRAC, due to this the search results may change later. For clarity, we note that the classification of verbs into semantic groups was based on the value of the first semantic tag (see Table 1).

Table 1. Classification of verbs into semantic groups

Semantic group	Verb component	Frequency
Verbs of psychological space відчувати `to sense', відчути `to feel', чекати `to wait'	29%
Verbs of existence	бути `to be', жити `to live',існувати `to exist', поживитися `to feather one's nest'	27%
Verbs of speech	викликати `to evoke',просити `to ask', нагадувати `to remind', заявити `to declare'	12%
Verbs of possession	мати `to have', втратити `to lose', давати `to give', надавати `to give',забрати `to take'	8%
Verbs of the beginning of existence	з'явитися `to appear', з'являтися `to appear', виникати `to arise', народитися `to born', народжуватися `to be born'	6%
Verbs of mental space	знати `to know', забути `to forgive', зрозуміти `to understand', розуміти `to understand', згадати `to mention'	5%
Verbs of movement	прийти `to come', приходити `to come', увійти `to enter', входити `to come in'	5%
Verbs change position and	повернутися `to turn', йти `to go',повертатися `to be	3%
body part movement	back'
Verbs of grasping	брати `to take', взяти `to take', тримати `to hold'	1%
Verbs of general action	додавати `to add'	1%
Verb location	залишитися `to stay'	1%
Verb body position	стояти `to stand', сидіти `to sit'	1%
Verbs of emotions	боятися `to fear'	1%
Verbs to create a physical or non-physical object	робити `to do'	1%
Verbs of non-physical influence	змінювати `to change'	0%
Verbs of modality	мусити `to must'	0%
Verbs of physiology	їсти `to eat'	0%
Verbs of wanting, allowing, willing	дозволити `to allow'	0%

The standard method of annotating a metaphor in the corpus involved manually processing expressions that are metaphorical, which is a time-consuming process. Since our classification was based on the first semantic tag, which has a direct meaning and for accuracy was compared with the dictionary meaning of the word. With this in mind, we classify the concept of СТРАХ into permanent metaphorical models. Although among the problems of automatic annotation, metaphor comes up with a concept as an image of conceptual reflections without specific linguistic forms.

The next step of the analysis is to determine the model of metaphorization of the studied concept. The analysis shows that the "metaphorical effect" is created by verbs belonging to a number of semantic/thematic groups.

The concept of СТРАХ is divided into two groups of metaphors and their subspecies. In general, we distinguish the metaphorical model FEAR IS A CREATURE, but it should be noted that in some cases we can talk about the metaphorization of invariants FEAR IS A HUMAN and FEAR IS AN ENEMY.

FEAR is categorized as a discrete object, given possible metaphorical expressions, the components of which are verbs belonging to the semantic group of verbs of general action, query

[tag=".*verb.*" & semtag=".*act.*"][lemma="cTpax"]: Додавало страху й те, що від цього не втечеш (В. Кошелюк).

Non-trivial cases for analysis are examples of detailed metaphors. The search query

[tag=".*verb.*" & semtag=".*effect.*"][lemma="cTpax"]

verbs of non-physical influence occurs through the use of the following metaphor: ...і місто на мить змінює страх на цікавість (С. Талан). Analyzing the context of the query, we see that grammatical information is important. Because in the example: ... такої упевненості надає страх, так ідуть, коли знають, що не запізняться ніколи... (O. Ульяненко) the metaphor FEAR IS A CREATURE that gives a discrete object CONFIDENCE.

At the request

[tag=".*verb.*" & semtag-'AexistA''^lemma-'страх"]

of the semantic group of verbs of existence, a metaphorization occurs repeatedly, where EYES ARE A CONTAINER OF EMOTIONS: В мені не було страх (Ю. Вовк); Чи не в кожного в очах були страх і підозра до будь-кого, хто несподівано опинявся поряд (K. Циганчук). The model FEAR IS A CREATURE is used for the same request: У тобі завжди жив страх, еге страх (В. Мастєрова); У цьому світі люди вмирали і жив страх перед смертю (A. Кондратюк); Але ж ні, місто жило страхами (O. Гаврош). FEAR IS AN OBJECT: У людей споконвіку був страх смерті, наразі був страх життя (С. Талан).

According to the request

[tag=".*verb.*" & semtag=".*grasp.*"][lemma= "страх"]

all the examples correspond to the model FEAR IS A CREATURE: Учителя брав страх перед ручними гранатами <...> (П. Панч); Та знову ж почула його несподіване запитання, яке врізалося прямо в серце: «Чому ж це вас так сильно тримає страх?» (A. Войнарович); Взяв страх такий самий, як і п'ять років тому... (Н. Смик).

On the query

[tag=".*verb.*" & semtag-'ApossA''^lemma-'страх"],

which contains verbs of possession, we get metaphors that correspond to the model FEAR IS AN ENEMY: Чванився, що не має страху ні перед Богом, ні перед дідьком... (I. Керницький); Це спроба втратити страх перед тією життєсмертною невірністю і починати днювати й ночувати просто у ньому, обсівшись (O. Євтушок). In addition, there are metaphors of the model FEAR IS AN OBJECT: Додавало страху й те, що від цього не втечеш (В. Кошелюк).

Combination verbs of wanting, allowing, willing with the word fear

[tag=".*verb.*" & semtag=".*psych:want.*"][lemma="страх"]

form a metaphor FEAR IS AN ENEMY: Він мусив страх перед кіньми побороти, бо який козак без коня? (Д. Корній).

On request

[tag=".*verb.*" & semtag-'AcreateA''^lemma-'страх"],



which provides a combination of verbs create a physical or non-physical object and the word fear, get metaphors where FEAR IS A CREATURE: «Що тільки з людьми робить страх ?.. -- чудувався Лев... (A. Процайло); Ось що з нами робить страх (K. Циганчук). Similarly, the query

[tag=".*verb.*" & semtag=".*loc.*"] [lemma="страх"]

outlines the model FEAR IS A CREATURE: Голуба вже не пручалася, але в очах її залишався страх (В. Тарасов); Дякуйте держбезпеці, -- відказав лікар з особливою інтонацією -- у ній залишалися страх і острога... (В. Лис).

The semantic group of movement verbs indicates the conceptualization of fear as a creature: Так, тоді до нього приходив страх (O. Ульяненко); Разом із цією зміною напрямку прийшов страх і холо. (Н. Сняданко); З-за стіни лісу йшов страх (С. Іванюк). The query

[tag=".*verb.*" & semtag=".*move.*"][lemma= "страх"]

gives numerous examples of metaphors: А тим, хто звик боятися, загалом не дуже й важить звідки йде страх (С. Іванюк); Натомість прийшов страх: якщо гнів його так засліплює, що буде, як він не стримається біля Лесі? (A. Нікуліна); Коли стихли зойки і постріли, прийшов страх, що вбивці вкинуть його у яму разом з мертвими і закопають живцем (Н. Гуменюк).

Similarly, verbs change position and body part movement

[tag=".*verb.*" semtag-'Aloc.body^'^lemma-'страх"]

give the metaphor FEAR IS A HUMAN: У ній усе ще сидів страх (Ю. Гук); Може, мені здалося, але після моєї розмови в очах слідчого стояв страх, дикий тваринний страх, що лещатами стискав йому мозок (В. Лис).

FEAR IS A HUMAN and verb body position

[tag=".*verb.*" & semtag= ".*move:body.*"][lemma="страх"]:

Мері, певне, теж щось ніби відчула, тому що на чисте личко повернувся страх, а в очах майнула паніка (Д. Гнатко); Не приходив страх, не адреналінилася кров, спонукаючи організм до самозахисту, не шикувалися думки в контратаку (M. Меднікова).

Interesting are the metaphors obtained from queries that contain verbs of psychological space

[tag=".*verb.*" & semtag=".*psych.*"] [lemma-'страх"]:

...тверезо усвідомлюєш собі, що на шляху на тебе чекають страх, безсилля, зневіра, бажання відмовитися від боротьби (В. Строкань); verbs of emotions

[tag=".*verb.*" & semtag=".*psych:emot.*"] [lemma-'страх"]:

Тепер вони не боялися страху, бо спрага забирала розум (O. Рибчинський); Не так через те, що справді хотів кудись іти, як через те, що боявся страху та паніки (В. Івченко); verbs of speech

[tag=".*verb.*" & semtag=".*speech.*"] [lemma"страх"]:

Не проси біди та розпачу, а проси страху, щоб не спокуситися власними стражданнями і не зазнати ницої розкоші (Ю. Іздрик).

Personification is typical of the categorization of emotions. The analysis shows that such a categorization of СТРАХ within the Ukrainian language map of the world is the most frequent in the studied texts. Verbal markers of this categorization and, accordingly, metaphorization are verbs belonging to semantic groups: verbs of existence, verbs of possession, verbs of mental space, verbs of movement, verbs change position and body part movement, verbs of grasping, verbs of physiology.

In general, we distinguish the metaphorical model FEAR IS A CREATURE, but it should be noted that in some cases we can talk about the metaphorization of invariants FEAR IS A HUMAN, FEAR IS AN ENEMY.

It should be noted that the query

[tag=".*verb.*" & semtag=".*act.*"] [lemma"страх"],

in particular, we get the following context: ...Чому ж це вас так сильно тримає страх? (A. Войнарович), where fear is the subject of action. Thus, in this case, grammatical information is important for classifying the directions of metaphorization.

Another example of metaphorization of СТРАХ, which we get on the query: Я все витримаю й не дам страхові оволодіти мною (С. Талан). Metaphorization follows a model FEAR IS AN ENEMY.

Productive is the model FEAR IS A CREATURE, which is verbalized by verbs of existence

[tag=".*verb.*" & semtag-".* appearA''^lemma-'страх"]:

Тоді ще не народився страх (E. Гата); У тобі завжди жив страх, еге страх (В. Мастєрова); Але ж ні, місто жило страхами (O. Гаврош); Існує страх смерті на війні, в заметілях, у катівнях (Р. Іваничук).

Quantitative results of our experiment show percentage ratios of metaphorical compounds and non-metaphorical, semantic and non-semantic tags of the group of verbs (see Table 2).



Table 2 Quantitative results of the experiment

Name	Non-metaphorical Metaphorical	Semantically	Non-
		distinction	semantically
Results''''^			distinction
Percentage value	2% 98%	30%	68%

Problem Cases

However, the question arises as to what formal features can be distinguished homonymy. We tried to remove homonyms with queries in the following constructions:

Verb+crpax (cmpaxoe/=adjective): [tag=".*verb.*"] [lemma-'crpax" & tag!=".*adj.*"].

The query removes 12 constructions, of which only 3 constructs have homonymy.

The constructions

Verb+crpax+Adjective and Verb+crpax+Adverb

could not be clear research result from the courpus, because they remove constructions that are metaphorical.

There are also cases that cannot be automatically removed from the research because they are associated with indistinguishable homonymy: Несила жити, помирати страх (С. Пономаренко); <...>Мій керівник, професор Назаренко, досліджував страх (M. Кідрук); <...>в Таниному шлунку підняв голову удав страху (Ю. Ілюха); Не спробувавши навіть розібратися в їхньому багатстві та розмаїтті, ми ставимо на них тавро «напади страху» (Я. Мельник).



Conclusions and prospects for further research

The results of the experiment revealed (98%) metaphorical expressions of which only (30%) are semantically annotated. In our investigation, we used semantic tagging of verbs to categorically group identified verbs into verb+страх phrases. According to the conceptual analysis and the theory of conceptual metaphor, the models of metaphors found in our research were described.

As a result of the analysis of contextual use of nominees of the researched concept, it is established that for metaphorization of the concept СТРАХ. The most productive are such metaphorical models: FEAR IS AN OBJECT, FEAR IS A CREATURE, FEAR IS A HUMAN. Regarding the semantic group, they are the most quantitative verbs of psychological space and verbs of existence.

We are improving the methods of automatic identification of metaphors using semantic tags. Although we encountered indistinguishable homonymy of words in the courpus in the experiment. In particular, indistinguishable grammatical homonymy: in the context of the noun and the adverb страх `fear'. Nevertheless, due to successful search queries, we avoided a certain percentage of inaccuracies.

However, the problem of homonymy of the noun страх `fear'corresponds to the meaning of дуже `very' and is combined with an adverb and an adjective, which cannot be excluded from the study. The case of the subject їм as a verb and a pronoun is non-metaphorical, as well as the noun in the dative case of страхові `fear' and the adjective in the plural. All these search cases were eliminated by manual sorting.

In this article, we have limited the research to the construction of the verb+страх model. This method can be used to research other concepts of the human emotional sphere, as well as to verbalize their synonymous series. Whereas as a fact synonymous and phraseological series of concepts are mostly based on metaphorization. So after a complete semantic tagging of the corpus, we will resume the research of the constructions страх+verb, verb+страх with coverage of the semantic distance, as well as a comparison of such synonyms of the word fear: острах `fear', побоювання `apprehension', жах `horror', тривога `anxiety', неспокій `trouble', переляк `fright', боязнь `phobia', боязкість `timidity', ляк `scare', паніка `panic' and their models. We investigate the gender peculiarities of metaphorization of the concept of СТРАХ in the fiction texts of the 21st century, their linguistic and statistical analysis.



References

1. Babarczy А., Bencze I. M., Fekete І., Simon E. The Automatic Identification of Conceptual Metaphors in Hungarian Texts: A Corpus-Based Analysis. Methods for the automatic acquisition of Language Resources and their evaluation methods: LREC 2010 Workshop. May 23. 2010. P. 30-37.

2. Deignan A. Metaphor and corpus linguistics: Amsterdam; Philadelphia: John Benjamins, 2005. 235 p.

3. Dictionaries of Ukraine online. URL: https://lcorp.ulif.org.ua/dictua/ (accessed 02.04.2022).

4. Dictionary of Modern Ukrainian-11. Vol. 9. P. 755. URL: http://sum.in.ua/p/9/755/2 (accessed 02.04.2022).

5. Doquin de Saint Preux A., Masid Blanco O. The power of conceptual metaphors in the age of pandemic: The influence of the WAR and SPORT domains on emotions and thoughts. Elsevier.Language & Communication. November 2021. Vol. 81. P. 37-47.

6. Fauconnier. G., Turner. M. The way we think: conceptual blending and the mind's hidden complexities. New-York: Basicbooks, 2002. 440 p.

7. Gutierrez E. D., Cecchi G., Corcoran Ch., Corlett Ph. Using Automated Metaphor Identification to Aid in Detection and Prediction of First-Episode Schizophrenia: In Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing: September, 2017. Copenhagen, Denmark: Association for Computational Linguistics, 2017. P. 2923-2930.

8. Heintz I., Gabbard R., Srinivasan M., Barner D., Donald B., Freedman M., Weischedel R. Automatic Extraction of Linguistic Metaphors with LDA Topic Modeling. In Proceedings of the First Workshop on Metaphor in NLP: June, 2013. Atlanta, Georgia: Association for Computational Linguistics, 2013.P 58-66.

9. Jaynes J. The origin of consciousness in the breakdown of the bicameral mind: Boston: Mariner Books, 2000. 512 p.

10. Lakoff G. Metaphor and Thought: Cambridge University Press, 1993. 245 p.

11. Lakoff G., Espenson J., Schwartz A., The Master Metaphor List: Technical report. Berkeley: University of California, 1991.

12. Lakoff G., Johnson M. Metaphors We Live by: Chicago: University of Chicago Press, 2003. 333 p.

13. Levchenko O., Romanyshyn N. Modern approaches to automated identification of metaphor. Bulletin of Lviv University, Series Philological. 2019. Is 70. 2019. P. 288-298.

14. Levchenko O., Romanyshyn N., Dosyn D. Method of automated identification of metaphoric meaning in adjective + noun word combinations (based on the Ukrainian language), in: CEUR Workshop Proceedings, Workshop proceedings of the 8th International conference on Mathematics. Information technologies. Education, MoMLeT&DS, 2019. Vol. 2386, P. 370-380.

15. Levchenko O., Tyshchenko O., Dilai M. Automated Identification of Metaphors in Annotated Corpus (Based on Substance Terms). CEUR Workshop Proceedings. Proceedings of the 5th International Conference on Computational Linguistics and Intelligent Systems (COLINS 2021). Vol. I: Main Conference. Kharkiv, Ukraine, April 22-23, 2021. P. 16-31.

16. Noviana F., Saifudin A. Conceptual Metaphors of «water” in Javanese Proverbs from a Cognitive Linguistic Perspective. E3S Web Conf Proceedings the 6th International Conference on Energy, Environment, Epidemiology and Information System (ICENIS 2021), Section Language and Environment Vol. 317, November 05, 2021. P 7.

17. Petruk O. Y., Levchenko O. P. Identification of the metaphor model білий 'white' + noun by the method of quantitative analysis of dictionary definition: Young scientist, Series Philological Sciences. № 10 (74). 2019. P. 505-511.

18. Shutova E., Sun L., Korhonen A. Metaphor identification using verb and noun clustering, Proceedings of the 23rd International Conference on Computational Linguistics (Coling 2010). Beijing. 2010. P.1002-1010.

19. Starko V. Semantic Annotation for Ukrainian: Categorization Scheme, Principles, and Tools, CEUR Workshop Proceedings. Proceedings of the 4th International Conference on Computational Linguistics and Intelligent Systems (COLINS 2020). Vol. I: Main Conference. Lviv, Ukraine, April 23-24, 2020. P.239-248.

20. Starko V. Implementing Semantic Annotation in a Ukrainian Corpus. CEUR Workshop Proceedings. Proceedings of the 5th International Conference on Computational Linguistics and Intelligent Systems (COLINS 2021), Vol. I: Main Conference. Kharkiv, Ukraine. April 22-23, 2021. P. 435-447.

21. Tsvetkov Y., Mukomel E., Gershman A. Cross-Lingual Metaphor Detection Using Common Semantic Features. in: Proceedings of the First Workshop on Metaphor in NLP, The American Association for Computational Linguistics (NAACL-2013). Atlanta, USA, 2013. P. 45-51.

22. Wilks Y., Galescu L., Allen J., Dalton A. Automatic Metaphor Detection using Large-Scale Lexical Resources and Conventional Metaphor Extraction. in: Proceedings of the First Workshop on Metaphor in NLP: The American Association for Computational Linguistics (NAACL-2013). Atlanta, USA. 2013. P.36-44.

Размещено на Allbest.ru

...