Quantitative Features of the Words Representing Nonverbal Behaviour in Ian McEwan’s Fiction Oksana Melnychuka, Nataliya Bondarchukb, Ivan Bekhtab,c, Olena Levchenko b a Rivne Medical Academy, Rivne, 33017, Ukraine; b Lviv Polytechnic National University, Lviv 79013, Ukraine c Ivan Franko National University of Lviv, Lviv 79000, Ukraine Abstract The computer-assisted textual research reveals a set of quantitative features (absolute and relative frequency, quantity, rank) of the words which articulate the meaning of a fiction text. This meaning is presented as nonverbal behaviour in Ian McEwan’s novels Sweet Tooth and Solar under the research framework of quantitative analysis, in particular, computational text analysis. The set of quantitative features in each text was complemented by qualitative parameters (types of nonverbal behaviour: paralanguage, moving; groups of meaning of nonverbal behaviour: descriptive, nondescriptive) which disclose the peculiarities of contextual nonverbal behaviour through its relation to the categories of “coherence” and “character” in Ian McEwan’ fiction. Nonverbal communication possesses a great potential in conveying the meaning of the message in human interaction. Real-life nonverbal communication is transferred by a writer to a fictional text as nonverbal behaviour playing a significant role in readers’ understanding of the nonverbal behaviour of characters and thus providing the coherence of the fictional text. The two types of nonverbal behaviour components – paralanguage (voice qualities) and moving (movements of the body) – establish the core of nonverbal behaviour presentation being the most frequent types which are referred to in Ian McEwan’s novels. Words denoting paralanguage and moving are not homogenous in the researched text contexts conveying two groups of meaning of nonverbal behaviour – descriptive (interprets fiction characters’ nonverbal behaviour) and nondescriptive (explains fiction text coherence). Voyant-Tools software is applied to extract quantitative data in text corpora. Keywords 1 Nonverbal behaviour, quantitative features, paralanguage, moving, textual coherence, Voyant- Tools, textual analysis 1. Introduction Modern linguistics is closely connected with IT technologies which stimulate quantitative textual researches, in particular statistical (quantittive) text analysis. British fiction of the 21st century is a substantial source to study nonverbal behaviour. It is overflowing with nonverbal behaviour components, especially paralanguage and moving. F. Poyatos [1], B. Korte [2], J. A. Hall, M. L. Knapp [3], and A. Kendon [4] review this wealth of nonverbal components on the basis of a large and growing body of modern research in nonverbal behaviour. S. Johar [5] proves that the description of nonverbal COLINS-2022: 6th International Conference on Computational Linguistics and Intelligent Systems, May 12–13, 2022, Gliwice, Poland. EMAIL: melnychukox@gmail.com (O. Melnychuk); nataliia.i.bondarchuk@lpnu.ua (N. Bodnarchuk), ivan.bekhta@lnu.edu.ua (I. Bekhta); levchenko.olena@gmail.com (O. Levchenko). ORCID: 0000-0003-4619-363x (O. Melnychuk); 0000-0002-5772-8532 (N. Bondarchuk); 0000-0002-9848-1505 (I. Bekhta); 0000-0002- 7395-3772 (O. Levchenko). ©️ 2022 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0). CEUR Workshop Proceedings (CEUR-WS.org) behaviour contributes to both the description of characters’ nonverbal behaviour and the coherence of the fictional text. Presently, the research of nonverbal behaviour in contemporary British fiction goes far beyond simple language processing: fast and accurate systematic computer calculations, as J. Flanders and F. Yannidis argue, provide more sophisticated tools and more reliable results in quantitative analysis and interpretation [6]. In spite of the fact that a multitude of subjective convictions determines readers’ understanding of a fictional text, there are factual premises that we study in terms of text data to ground the set of quantitative features and make them complete and visible. The objective of the research is to detect a set of quantitative features of nonverbal behaviour in two novels Sweet Tooth and Solar by I. McEwan [7, 8], a British novelist, short-story writer, and screenwriter whose restrained, refined prose style highlights the horror of his dark humour and deviated subject matter. We consider the intended objective through quantitative (computational) and qualitative (interpretational) textual analysis [9-11]. Empirical data is effective while being grounded on web-based text analysis. Voyant Tools (voyant-tools.org), an open-source project, which is available through GitHub, proves to be effective in the context of quantitative and qualitative analysis of fiction texts. 2. Study reasoning Fictional text as a novel is traditionally a form of literature that has responded to social and political movements, and literary fiction in a certain period has close ties with societies and culture. The late seventies outlined a period of political, social and cultural change that reveals some of the fundamental characteristics of contemporary Britain from the end of World War II up to the present day. Literary critics regard the fiction of I. McEwan as “most continued reflection on the form of the novel, and the inherited tradition of modern (especially Anglophone) fiction and criticism” [12]. The novels under examination, Sweet Tooth and Solar by I. McEwan, are modern classic ones. Sweet Tooth explores the relationship between artistic wholeness and government propaganda to reveal diverse acceptance of literature; the boundary between real life and fiction is depicted throughout [13]. Solar is a fictional text that draws heavily on natural science and modern history references. It is a satire about a Nobel-winning physicist whose broken personal life and devious ambition make him chase a solar-energy based solution for climate change. In 2010, Solar got the Bollinger Everyman Wodehouse Prize, a British literary award for comic writing. Any description of nonverbal behaviour by a writer is ipso facto significant since “a creator of the fictional texts would not record constant background noise and redundant paralanguage of everyday life” [14]. Novels by I. McEwan are the relevant source of investigation of nonverbal behaviour. 3. Research method Ideas associated with contemporary British fiction expose the variety of linguistic phenomena to be studied as a set of quantitative features, e.g., absolute and relative frequency, quantity, the rank of the words. One of the productive ways of fictional text analysis is its inquiry with the help of computer software. Computational tools are privileged in extracting necessary data from a vast amount of fictional text. They allow to investigate and interpret different aspects and features of fictional texts, including the aspects of meaning. Computational approaches to understanding this phenomenon exemplify accurate lexical and semantic data. Just as researchers can implement computational linguistics in various fields and through a wide assortment of tools and procedures, the research fields include a diverse range of topics [15]. Among them, computational text analysis plays a critical role in enriching qualitative approaches with quantitative ones [16, 17]. Linguists define a style in terms of a domain of language use (e.g., what choices a particular author makes in a specific genre, or in a certain text). B. Bloch draws attention to the style of a text as to the message carried by the frequency distributions of its linguistic features [20]. Therefore, style consists of choices made from the repertoire of the language [18]. While linguistic features do not constitute a fictional text’s “meaning”, a quantitative account of linguistic features serves to ground a stylistic interpretation and to help to explain why certain groups of meaning or types of nonverbal behaviour are possible [19, 20]. The quantitative research of the words representing nonverbal behavior involves the application of the following methods: textual analysis (semantics), quantitative analysis (computational linguistic) and qualitative analysis (interpretation). Textual analysis is a holistic, systematic approach to the study of a fictional text by dividing it into parts (e. g. words) to analyze each of its elements in combination with other textual elements as a set of linguistic means that convey the semantic unity of a fictional text. In the context of a fictional text, the characteristics of different textual levels are interdependent and not isolated [6, 14, 21]. The method was used to characterize the words of a specific fictional text in connection to its content. Quantitative analysis implies handling quantitative data in linguistic study [22-24]. In our research quantitative analysis is a method of observing a fictional text in terms of its quantitative features (word frequency, word quantity, and word ranks). The quantitative analysis results obtained in the process of word/word form extraction applying computer software are presented in tables and diagrams. Being designed for a wide range of applications and users, Voyant-Tools is used in the present study as a reading and visualization environment to extract the words and indexes of their frequency; to find the context for each word among single fictional text through network analysis/keyword/feature extraction. By applying advanced analytical techniques, such as Naïve Bayes, Support Vector Machines (SVM), and other deep learning algorithms, we can explore and discover implicit relationships within structured and unstructured data. Qualitative analysis is the method of interpretation that means studying the contexts of fictional text to define “paralanguage” and “moving” types of nonverbal behaviour, to sort out descriptive and nondescriptive groups of the meaning of nonverbal behaviour. Interpretation involves making equations between linguistic forms and the meanings contracted by the function of these forms in a context of a fictional text. The method also aims to make sense of quantitative data in definite corpora in the context of a fictional text. The procedure involves the following steps of analysis as corpus processing: 1. The fictional texts Sweet Tooth (Corpus 1) and Solar (Corpus 2) by I. McEwan are selected and made available for uploading in Voyant-Tools software in pdf format as text corpora. Digitalized data are inspected for errors. 2. Data aggregation means a reduction of dimensions of the data by aggregating individual text elements into broader categories. Here, all of the data are reduced to word absolute frequency in the textual corpus. The words said, v (320), went, n (144) in Corpus 1, and the words said, v (217), came, v (108) in Corpus 2 are defined as the most frequent words in corpus concerning nonverbal behaviour presentation (see Table 1). 3. Data query is the extraction of specific data from stored items. The most frequent words, mostly verbs, and their word forms form two types of nonverbal behaviour – paralanguage (say*v; tell*, v; talk*, v; speak*, v; sound*, v; sound*, n; voice*, n) and moving (go*, v; come*, v; walk*, v; follow*, v; turn*, v), and two groups of the meaning of nonverbal behaviour – descriptive, nondescriptive. The word with the mark «*» means the word and its forms 4. We did quantitative data analysis considering the quantity and absolute frequency of the words and their word forms and applying Voyant-Tools to extract the words in two text corpora. The total quantity of “paralanguage”, “moving”, “descriptive meaning”, “nondescriptive meaning” is counted and compared in Table 2 and Table 3. 5. Qualitative data analysis considers fictional text contexts extraction by Voyant-Tools to interpret the meaning and sort out the types of nonverbal behaviour (paralanguage, moving) and groups of the meaning of nonverbal behaviour (nondescriptive, descriptive). 6. The research material consists of two corpora: Corpus 1 (Sweet Tooth) and Corpus 2 (Solar) by I. McEwan. The summary of each corpus, automatically generated by Voyant-Tools, is shown in Table 1 that exhibits that the corpora have approximately the same quantity of words and unique word forms. The vocabulary density, readability index, and average quantity of words are higher in Corpus 2. The table also shows the first fifteen more frequent words in corpora. The words said, v; went, v; came, v are the most frequent and valid in the researched texts. Table 1 Summary of research corpora Corpus 1 Sweet Tooth Corpus 2 Solar Total words 103,948 92,438 Total words exposed to 1965 1312 qualitative and quantitative analysis Unique word forms 15,093 16,420 Vocabulary density 0.145 0.178 Readability index 9.322 11.400 Average words per 14.7 18.3 sentence The most frequent said (320); didn’t (206); like beard (494); said (217); like (159); words in the corpus (206); thought (186); tom (185); time (139); thought (133); man way (180); know (171); i’d (113); way (112); just (103); years (159); time (151); knew (144); (100); away (99); came (108); went (144); just (131); told knew (97); rest (96); know (92); (124); it’s (123); room (123). room (92). 4. Theoretical linguistic background Fictional text is a complex unit, which implies the creative consciousness of an author, so each lexical component (word) in its way is chosen to represent a common single aspiration of the whole text. M. Short and G. Leech, P. Simpson, B. Bloch, and R. Barthes [25] argue that textual analysis is the “stylistic of choice” in revealing the meaning of the text, which is an objective phenomenon, subjective experience, and the intersubjective meaning. The text permanently impregnates potential meanings and a reader that perceives, identifies and actualizes them. Thus, the quantitative analysis “requires a complete description”, which is “not a list of certain elements”, but “the identification of a system of functions” [26]. A full-fledged quantitative analysis is always a holistic one: it identifies not the building material but the constructive relationships of the whole as a “complexly constructed meaning” [27]. The quantitative analysis approach divides a complex unity (fictional text) into units as products of its analysis (words), which, unlike elements, do not lose the properties of the whole entity, but present in their simplest, original form those properties of the whole entity, for the sake of which the analysis is undertaken [28]. So, we argue that quantitative analysis identifies building material to explain the whole text’s structural and semantic relationships as complexly constructed meaning. One of the most productive ways of understanding how a text works, as S. Statham notes, is to challenge it or intervene in its stylistic makeup in some way [29]. The revealing of both different types of meaning and the whole meaning of a fictional text is possible due to interpretation: it is a sense-making process or revelation of meaning. This intellectual operation cannot be reduced to an explanation, which, answering the question “why?” turns it from the present into the past. Interpretation, on the other hand, is oriented toward the future since it always explicitly or implicitly answers the question “why?” (“what” is the significance of this fact for a reader) [30]. Interpretation is a subjective process, a kind of dialogue between two realities: subjective and aesthetic ones. The interpretation is a mental turn into an object of self-interest [31]. Consequently, the interpreter’s task of a fictional text is to comprehend its semantic content “better than the author”; to understand the individual significance of the fictional text as an aesthetic experience. The purpose of interpretation in quantitative analysis is not a reconstruction of the author’s intention but the construction of meaning [32, 33]. It is a creative process of interaction of the reader with the text and its internal dialogue with its personality. Hence, there are not and cannot be two identical readings of the same fictional text – the same reader doesn’t make an identical sense every time he rereads a fictional text, i. e., a reader may ignore or not actualize the components of nonverbal behaviour that are objectively present in the text. Quantitative analysis is precisely that field, which establishes a particular sector of the adequacy of reader co-creation; it identifies the boundaries beyond which there is an area of the reader’s destructive reception. 4.1. Representation of nonverbal communication and nonverbal behaviour in a fictional text Nonverbal communication, in general, is associated with human nonverbal behaviour, and it ranges from aspects of voice to gestures, movements, and interpersonal spatial positioning, accessed by the vision and other senses. M. Danesi defined nonverbal communication as a group of human attributes or actions in which words are not involved but which have a shared social meaning. The criteria on the components of nonverbal communication also varied; it encompassed all kinds of nonverbal elements ranging from bodily signals to architecture [34]. The meanings of nonverbal communication (nonverbal behaviour) usually permeate the fictional text perceived, identified and understood by a reader. For the purpose of quantitative analysis, the term “nonverbal communication” refers to the forms of nonverbal behaviour exhibited by characters. A writer addresses a fictional text to a reader not directly but through a kind of “inner vision”, “inner hearing”, and “empathy for the characters”. This kind of impact is organized by the semiotic activity of the author (a narrator, to be exact), using a definite set of lexical units or words. “A fictional text is a purely intentional object and is a product of the author’s conscious” [35]. X. Jiang proves the importance of nonverbal communication in fiction, especially paralanguage (voice qualities) and kinetics (body movements) and how it contributes to an effective relationship between the text, the writer and the reader [36]. Nonverbal communication in a fictional text is transferred from real-life nonverbal communication presented by the author through the words which describe the characters’ nonverbal behaviour and are the way to contribute to textual coherence. Such transferring accounts for how the fictional text interlocks with the semantic process, notably those of “moving” and “saying”, and how these processes influence characters. As part of fictional text, the words denoting nonverbal behaviour are under two types of meaning: descriptive (denotes nonverbal behaviour and contributes to the characters’ nonverbal behaviour in fictional text) and nondescriptive (denotes nonverbal behaviour and contributes to fictional text coherence). Thus, representation of nonverbal behaviour is an indication of certain stylistic characteristics more accentuated in some writers than in others, and therefore an important touchstone for the quantitative analysis of fictional texts supported by processing techniques of computational linguistics. 5. Results and discussion This section manifests a computer-assisted case study of the words representing nonverbal behaviour in a fictional text through some quantitative data which expose a set of quantitative features in text corpora. Thus, we identified 1965 words in proper contexts to show the difference between two groups of nonverbal behaviour (paralanguage, moving) and two types of meaning (descriptive, nondescriptive). We delivered and visualized the results of the research as follow: Figure 1, Figure 2 (the most frequent words in corpora and their relative frequencies); Table 2, Table 3, Table 4, Table 5 (absolute frequencies, contexts, quantity of the words); Figure 3, Figure 4, Figure 5, Figure 6 (absolute frequencies of the words); Table 6 (rank and absolute frequencies of the words). The significance of nonverbal behaviour presentation is connected to the frequency of words denoting paralanguage and moving in text corpora: Corpus 1 (Figure 1) and Corpus 2 (Figure 2). Figure 1: Diagram of relative frequencies of the most frequent words in text segments of Corpus 1 The diagram indicates relative frequencies of the most significant words of nonverbal behaviour presentation in Corpus 1 (said, v (320); say, v (100); told, v (132); went, v (144)) and their distribution among 10 segments of the corpus text. The most unstable word said, v is the most frequent one and has a higher frequency index in the 9th text segment. The rest of the words do not have such sharp fluctuations, and they are approximately the same relative frequencies throughout the text. Figure 2: Diagram of relative frequencies of the most frequent words in text segments of Corpus 2 The diagram conveys relative frequencies of the most significant words for nonverbal behaviour presentation in Corpus 2 (come, v (220); said, v (217); went, v (84); going, v (64)) and their distribution among 10 segments of the corpus text. The most unstable word said, v is the most frequent and has higher frequency usage in the 10th text segment. Compared to Corpus 1 the relative frequency of the word said, v falls in the 5th segment and rises sharply in the 10th segment of the textual corpus. The relative frequencies of the rest of the words, except going, v which increases in the 2nd and 8th segments, do not have such sharp fluctuations; they have approximately the same relative frequencies throughout the text corpus. The word said, v has a higher relative frequency in Corpus 1 – 0.0005 than in Corpus 2 – 0.0004. The analysis of the representation of nonverbal behaviour in fiction was extended by considering word forms of the most frequent words, which are the same in each of two corpora. They are the following words and word forms: say*, v (said, v; say, v; saying, v; says, v); tell*, v (told, v; tell, v; telling, v; tells, v); talk*, v (talked, v; talk, v; talking, v; talks, v); speak*, v (spoke, n, spoken, v; speaking, v; speaks, v); sound*, v (sounded, v; sound, v; sounding, v; sounds, v); sound*, n (sounds, n); voice*, n (voices, n) for “paralanguage”; and go*, v (went, v; go, v; going, v; goes, v); come*, v (came, v; come, v; coming, v; comes, v); walk*, v (walked, v; walk, v; walking, v; walks, v); follow*, v (followed, v; follow, v; following, v; follows); turn*, v (turned, v; turn, v; turning, v; turns, v) for “moving”. The following tables (Table 2, Table 3, Table 4, Table 5) contain quantitative and qualitative results of the research. Quantity and absolute frequencies of the words denote two types of nonverbal behaviour (paralanguage and moving) and two groups of meaning (descriptive, nondescriptive). Qualitative results are presented as a textual locale for each word form according to descriptive or nondescriptive meaning. Descriptive meaning of nonverbal behaviour is defined as the meaning which contributes to characters’ nonverbal behaviour, e.g., the words denoting nonverbal behaviour have descriptive meaning if they are followed by some description of voice qualities or the way the movement is done; for example: to say coolly; to tell in flat voice; to go slowly; to turn minimally. The descriptive meaning of “moving” is direct and may not be followed by the adjective or the adverb which describes it, but it may indicate the direction: to come into the room; to go through a door; to turn onto the side. Nondescriptive meaning of nonverbal behaviour is defined as meaning which contributes to textual coherence, e.g., the words denoting nonverbal behaviour have nondescriptive meaning if they are not followed by any description of voice qualities or the way the movement is done; for example, he said that …; he told us something …; I was going to see …; It would come soon … Nondescriptive meaning of “moving” is often indirect, metaphorical: to go mad; her drinking came; to follow the logic. Table 2, Table 3, Table 4 and Table 5 contain a total quantity of words denoting two types of representation of nonverbal behaviour and two groups of meaning in corpora. Table 2 Quantity and absolute frequency of the words denoting paralanguage presentation in Corpus 1 Word / Descriptive meaning of Nondescriptive meaning of absolute nonverbal behaviour/quantity nonverbal behaviour/quantity frequency/ quantity Paralanguage Say, v* 462 43 419 Said, v 320 ‘This isn’t exactly chatty,’ she Tom had said he didn't want to see said coolly … (p. 10). - 38 the reviews … (p. 159) - 282 Say, v 100 He said, ‘I like it when you say I don’t find this easy to say, but I’m brilliant.’ (p. 88) - 3 deeply disappointed.’ (p. 23) - 97 Saying, v 31 ‘Our cover,’ she kept saying in Tony was saying, ‘You know where a loud whisper (p. 48). - 2 this all has to lead, don’t you?’ (p. 23) - 29 Says, v 11 Though Monica never says so, it is clear she doesn’t believe him (p. 91). - 11 Tell, v* 288 7 281 Told 132 A man came out … and told me … she ignored questions and told us in a nervous, pleasant way that I nothing … (p. 67) - 128 should wait (p. 54). - 4 Tell, v 128 He was going to tell me in his He wouldn’t tell me what it was (p. own way … (p. 22). - 2 12). - 118 Telling, v 24 … while I stood there watching him, wondering whether he was telling the truth (p. 40). - 24 Tells, v 12 She pauses, and then she tells She is sick of her life, she tells him, him, in that same flat voice, that sick of being financially dependent … all his climbing gear has been (p. 92) - 11 taken too (p. 93). - 1 Talk, v* 122 - 122 Talked, v 30 So we talked of other … (p. 119) - 30 Talk, v 49 I need to talk to Serena (p. 146). - 49 Talking, v 42 ‘But she’s very kind really and she’ll like talking to you (p. 145). - 42 Talks, v 1 He watches her closely as she talks, and knows that every word is a lie (p. 95). - 1 Speak, v* 82 6 76 Spoke, v 27 We spoke in identical tones, we She spoke of her Syrian doctor, I were socially confident … (p. 30) - spoke of Jeremy Mott, but not of Tony 6 Canning (p. 31). - 29 Spoken, v 9 Within a week my mother had spoken to my headmaster (p. 8). - 9 Speak, v21 He went to speak but was stuck for words (p. 86). - 21 Speaking, v 15 Everyone was speaking of ‘the crisis’ (p. 112) - 15 Speaks, v 2 As he speaks to the desk sergeant, he feels a bit of a cad or a snitch (p. 92). - 2 Sound, v* 37 37 - Sounded, v 21 I mumbled something modest but it sounded dismissive… (p. 13) - 21 Sound, v 10 I said, hoping I didn’t sound like I was pleading (p. 88). - 10 Sounding, v 3 …he couldn’t stop himself from sounding abject one moment and over-emphatic … (p. 149) - 3 Sounds, v 3 It sounds very promising (p. 166). - 3 Sound, n* 18 18 Sound, n 16 …I thought I heard the sound of a voice (p. 173). - 16 Sounds, n 2 What then of my politely muted sounds? (p. 111) - 2 Voice, n* 38 36 2 Voice, n 31 Yes, she agreed in her I thought I heard the sound of a affectless voice … (p. 93) - 29 voice. 173 - 2 Voices, n 7 In low voices we talked office gossip for the first quarter of an hour (p. 68). - 7 Total (words Total (words and word forms Total (words and word forms of and word forms of descriptive meaning of nondescriptive meaning of of paralanguage): paralanguage): 147 paralanguage): 900 1047 Table 3. Quantity and absolute frequency of the words denoting moving presentation in Corpus 1 Word / Descriptive meaning of Nondescriptive meaning of absolute nonverbal behaviour/quantity nonverbal behaviour/quantity frequency/ quantity Moving Go, v* 394 263 131 Went, v 144 Desolate, I went slowly along My light-headed alliterative prose Great Marlborough Street (p. 27) - went down well (p. 10). - 26 118 Gone, v 35 ... and a half later, the men had Everyone had gone mad, so everyone gone (p.17). - 8 said (p. 19). - 27 Go, v 109 … Tony asked me if I would go It was a matter of letting my eyes for a longish walk in the woods (p. and thoughts go soft, like wax …(p. 9). - 17). - 64 45 Going, v 85 … she was in the crowd going I wasn’t going to see the New Year in out through the door (p. 47). - 67 with you (p. 176). - 18 Goes, v 21 He goes upstairs and lies on the As far as hand-holding goes, it won’t bed … (p. 94). - 6 be much of an imposition (p. 103). - 15 Come, v* 249 116 133 Came, v 116 I came out of the cubicle, … drinking came later (p. 108). - 69 splashed my face … (p. 102). - 47 Come, v 83 If he comes looking for you, His Foundation money hadn’t you're to turn … (p. 168). - 49 arrived, but he was sure it would come soon (p. 111). - 34 Coming, v 38 I heard Shirley coming and I knew I had an interview coming up quickly put the pile in order … (p. … (p. 20) - 23 52) - 15 Comes, v 12 The person who comes in with … so the money that comes to me a holdall and sets down … (p. 94). - each month is not simply an impersonal 5 … (p. 103). - 7 Walk, v* 84 84 - Walked, v 41 We walked in silence. He didn’t know what to say (p.40). - 41 Walk, v 20 It takes him an hour to walk the mile to his house (p. 94). - 20 Walking, v We knew what we were 21 walking towards (p. 108). - 21 Walks, v 2 … Sebastian walks back towards the Street … (p. 91) - 2 Follow, v* 56 22 34 Followed, v With the music fading behind There followed a session of small talk 27 me, I followed their directions … (p. 54). - 9 (p. 83) - 18 I didn’t follow the logic of this, but I Follow, v 10 said nothing (p. 8). - 10 Following, v I started out following her The following day I arrived home in 18 path, the one she describes in her the cathedral close with all my memoir… (p. 28) - 3 belongings (p. 28). - 15 Follows, v 5 She establishes that there is a brother and follows him to London … (p. 64) - 1 Turn, v* 135 54 81 Turned, v 91 Then he turned and glanced Lately the weather had turned mild… round the room, looking for me… (p. 144) - 51 (p. 140) - 40 Turn, v 28 … he managed to turn onto his This in turn delays the repayment to side (p. 73) - 7 Monica’s brother … (p. 93) - 21 Turning, v 12 …and I saw in front of me a … nothing resolved, thoughts still restrained movement of heads turning, when I heard footsteps on the tilting or turning minimally (p. 46) stairs (p. 117). - 8 -4 Turns, v 4 He turns and standing before The DI switches off the projector and him … (p. 91) - 3 turns up the lights (p. 94). - 1 Total (words Total (words and word forms Total (words and word forms of and word forms of descriptive meaning of nondescriptive meaning of moving): of moving): 918 moving): 539 379 Total (words Total (words and word forms Total (words and word forms of and ward forms of descriptive meaning of nondescriptive meaning of of paralanguage paralanguage and moving): 686 paralanguage and moving): 1279 and moving): 1965 Table 4. Quantity and absolute frequency of the words denoting paralanguage presentation in Corpus 2 Word/absolute Descriptive meaning of Nondescriptive meaning of frequency / nonverbal behaviour/quantity nonverbal behaviour/quantity quantity Paralanguage Say, v* 302 60 242 Said, v 217 “Solar energy?” Beard said She said she did not mind what he mildly (p. 27) - 52 did (p. 9) - 165 Say, v 54 He was starting to say … he had no idea what he wanted conversationally (p. 108). - 4 to say (p. 42). - 50 Saying, v 27 … deep female voice behind … he was saying, but it seemed him saying kindly … (p. 61) - 4 too abstract (p. 162). - 23 Says, v 4 ... she says she’s wrong (p. 195) - 4 Tell,v* 126 1 125 Told, v 66 … she told him plainly to go He told himself that things are away (p.38). - 1 often not as bad as you think (p. 131). - 65 Tell, v 42 But he would tell no one (p. 58). - 42 Tells, v 1 His doctor tells him that not thinking about that thing make it go away (p. 260). - 1 Telling, v 17 She seemed on the point of telling him something else … (p. 207) - 17 Talk, v* 67 3 64 Talked, v 12 … he talked about his work and travels … (p. 164) - 12 Talk, v 33 He had to talk fast (p. 140). - 1 She had come to talk … (p. 47). - 32 Talking, v 22 … laughing and talking at a I’m talking to a lawyer in Oregon relaxed, normal pitch (p. 90). - 2 (p. 252). - 20 Talks, v - Speak, v* 73 13 60 Spoke, v 34 … Tom Aldous spoke with the When Beard’s turn came, he spoke lilting confidence of a prize pupil… to the point (p. 96). - 22 (p. 28) - 12 Spoken, v 5 These half-truths were the best words he had ever spoken (p. 177). - 5 Speak, v 22 Beard thought it important to speak first (p. 43). - 22 Speaking, v 11 … and speaking in a measured, She was nervous speaking in husky tone… (p. 33) - 1 public… (p. 128) - 10 Speaks, v 1 He still speaks at conferences… (p. 259) - 1 Sound, v* 12 7 1 Sounded, v 10 It always sounded like a lie (p. 63). - 10 Sound, v 1 … and then, determined to sound grave rather than querulous, he said… (p. 163). - 1 Sounding, v 1 When he heard himself sounding off, he was not at all convinced… (p.183). - 1 Sounds - Voice, n* 51 38 13 Voice, n 40 … though her voice was as … she would hear his voice but not bright as ever (p. 13). - 32 his words (p.17). - 8 Voices, n 11 … the sound of children’s His turn to listen to voices through voices approaching … (p. 45). - 6 the wall? (p. 24) - 5 Sound, n* 36 28 8 Sound, n 35 … from inside came a muffled At that thought he heard a sound sound of bare feet … (p. 24). - 27 above him … (p. 42) - 8 Sounds, n 1 … from the galley … came the smell of frying meat and garlic and the sounds of spoons … (p. 61) - 1 Total (words Total (words and word forms Total (words and word forms of and word forms of of descriptive meaning of nondescriptive meaning of paralanguage): 667 paralanguage): 154 paralanguage): 513 Table 5 Quantity and absolute frequency of the words denoting moving presentation in Corpus 2 Word/absolute Descriptive meaning of Nondescriptive meaning of frequency/quantity nonverbal behaviour/quantity nonverbal behaviour/quantity Moving Go, v* 267 179 88 Went, v 84 … Beard wondered … as he left After two or three glasses of the one office and went glumly toward white, the red went down the next … (p. 31) - 66 painlessly, like water, at least at first (p. 72). 18 Gone, v 29 His groin was so tender that he Even the hangers were gone (p. waited until the others had gone 9). - 17 inside… (p. 69) - 12 Go, v 84 Beard preferred to go around … the demand for energy will go alone … (p.26). - 60 on rising as the world's population expands … (p. 173) - 24 Going, v 68 … before going in he found a I’m going to talk to Aldous, then litter bin and disposed of the plastic I’m going to take him with me to bag (p. 81). - 41 Design (p. 33). - 27 Goes, v 2 Honestly, it goes deeper … (p. 82). - 2 Come, v* 220 96 124 Came, v 108 And here he came, a gaunt The machine came to life at first parchment-faced fellow … (p. 105). touch (p. 68). - 72 - 36 Come, v She had come to talk, not to … long-running sinecures had (participle 2) 23 listen (p. 47). - 13 recently come to an end … (p. 19) - 10 Come, v 54 I thought I’d come and have a “Come on, man. Let’s go!” (p. 54) look round (p. 43). - 25 - 29 Coming, v 31 Tarpin … was coming toward But he, Beard, had had many him with a firm stride (p. 45). - 20 affairs himself … and, probably it was coming to an end (p.92). - 11 Comes, v 4 … he'll be arrested if he phones … though at times he comes or writes or comes within 500 yards closer to being pathetic … (p. 260). - of our house (p. 212). - 2 2 Walk, v* 39 39 Walked, v 24 … he walked unhurriedly down the garden path … (p. 88). - 24 Walk, v 4 He asked if he could at least walk with her across the parks (p. 180). - 4 Walking, v 11 … and it seemed he was walking directly toward it now (p. 109). - 11 Walks, v - - Follow, v* 24 6 18 Followed, v 9 … so Beard followed a narrow … the sound of a hiss followed by concrete path … (p. 41) - 3 a whiplike crack… (p. 154) - 6 Follow, v 2 “I follow you,” Jan said (p. 54). 2 Following, v 10 Within seconds he was bouncing During the following week, some across the plain, following through commentators agreed with her (p. the sight holes (p. 55). - 1 97). - 9 Follows, v 1 A confrontation follows (p. 86). 1 Turn, v* 95 32 63 Turned, v 62 … he groaned and turned angrily … spoiled generation turned its on his side (p. 65). - 25 backs on the fathers who fought the war (p. 60). - 37 Turn, v 21 Rather than turn and have his It would be difficult to turn her face ripped away, he hunched his from this calm, seductive mode (p. shoulders (p. 68). - 5 161). - 16 Turning 12 Hammer, turning to Beard, By turning his shoulder into the looked like he was about to go room, Beard was able to prompt his down on one knee (p. 247). - 2 host … (p. 135) - 10 Turns - - Total (words Total (words and word forms of Total (words and word forms of and word forms descriptive meaning of moving): nondescriptive meaning of moving): 645 352 moving): 293 Total (words Total (words and word forms of Total (words and word forms of and word forms of descriptive meaning of moving and nondescriptive meaning of moving moving and paralanguage): 506 and paralanguage): 806 paralanguage): 1312 The absolute frequencies of words according to types of representation of nonverbal behaviour and groups of meaning are manifested in Figure 3 (Corpus 1) and Figure 4 (Corpus 2). 900 1000 800 539 600 379 400 147 200 0 Paralanguage Moving Nondescriptive meaning Descriptive meaning Figure 3: The diagram of absolute frequencies of words as to types of representation of nonverbal behaviour and groups of meaning in Corpus 1 600 513 500 352 400 293 300 154 200 100 0 Paralanguage Moving Nondescriptive meaning Descriptive meaning Figure 4: The diagram of absolute frequencies of words as to types of representation nonverbal behaviour and groups of meaning in Corpus 2 Paralanguage is a type of nonverbal behaviour; it prevails in the group of nondescriptive meaning in both fictional texts. It means that a writer uses words denoting paralanguage to indicate textual coherence. The descriptive meaning of moving is significant for the characters’ description. The quantity of words belonging to nondescriptive meaning is almost the same in both Corpora. At the same time, the quantity of words used to describe paralanguage in nondescriptive meaning is two times larger in Corpus 1. The quantity of words denoting moving is approximately the same in two groups of meaning. The absolute frequencies of words and word forms representing nonverbal behaviour in the texts are shown in Figure 5 (Corpus 1) and Figure 6 (Corpus 2). Walk, v* 84 Sound, n* 18 Sound,v* 37 Voice,n* 2 36 Follow, v* 22 34 Speak, v* 6 76 Turn, v* 54 81 Talk, v* 122 Go, v* 131 263 Come, v* 116 133 Tell, v* 7 281 Say, v* 43 419 0 50 100 150 200 250 300 350 400 450 Nondescriptive meaning Descriptive meaning Figure 5: The diagram of absolute frequencies of words and their word forms, which represent nonverbal behaviour in Corpus 1 Walk,v* 39 Sound,v* 1 11 Sound,n* 8 28 Voice,n* 13 38 Follow,v* 6 18 Speak, v* 13 60 Turn,v * 32 63 Talk,v* 3 64 Go, v* 88 179 Come, v* 96 124 Tell, v* 1 125 Say, v* 60 242 0 50 100 150 200 250 300 Nondescriptive meaning Descriptive meaning Figure 6: The diagram of absolute frequencies of words and their word forms, which represent nonverbal behaviour in Corpus 2 The words and the word forms of walk, v*, sound*, n, sound, v*, voice, v*, go*, v, come*, v are productive in creating descriptive meaning in fictional texts. The words and word forms of tell*, v, say*, v, talk*, v, speak, * v are not usually used for this purpose in both corpora. Ranks and absolute frequencies of the words denoting representation of nonverbal behaviour in Corpus 1 and Corpus 2 are shown in Table 6. Table 6. Ranks and absolute frequency of the words denoting representation nonverbal behaviour Rank Corpus 1 Corpus 2 Nondescriptive Descriptive Nondescriptive Descriptive meaning meaning meaning meaning 1 Say, v* 419 Go, v* 263 Say, v* 242 Go, v* 179 2 Tell, v* 218 Come, v* 116 Tell, v* 125 Come, v* 96 3 Come, v* 133 Walk, v* 84 Come, v* 124 Say, 60 4 Go, v* 131 Turn, v* 54 Go, v* 88 Walk, v* 39 5 Talk, v* 122 Say, v* 43 Talk, v* 64 Voice, n* 38 6 Turn, v* 81 Sound, v* 37 Turn, v* 63 Turn, v* 32 7 Speak, v* 76 Voice, n* 36 Speak, v* 60 Sound, n* 28 8 Follow, v* 34 Follow, v* 22 Follow, v* 18 Speak, v* 13 9 Voice, n* 2 Sound, n*18 Voice, n* 13 Sound, v* 11 10 Sound, v* - Tell, 7* Sound, n* 8 Follow, v* 6 11 Sound, n* - Speak, v* 6 Sound, v* 1 Talk, v* 3 12 Walk, v* - Talk, v* - Walk, v* - Tell, v* 1 The rank of the words denoting paralanguage and moving as nondescriptive meaning of paralanguage coincides in two text corpora. The words sound*, v, sound*, n, walk*, v denote only descriptive meaning, and the word talk*, v denotes only nondescriptive meaning in Corpus 1. The word walk*, v denotes only descriptive meaning in Corpus 2. The rank of the words denoting paralanguage and moving as descriptive meaning is different: only the word rank of goes*, v; come,* v coincides. There are the same word ranks of follow*, v (Corpus 1) and turn,* v (Corpus 2). 6. Conclusion Computer-assisted textual research has a considerable effect in studies disclosing the meaning of fiction texts. The words, denoting nonverbal behaviour in I. McEwan’s novels Sweet Tooth and Solar are important in textual interpretation. Computation text analysis has served to exhibit two types of nonverbal behaviour and two groups of meaning which were defined by a set of quantitative features: absolute, relative frequency, and the rank of the words and word forms a writer uses to create the coherence or to describe nonverbal characters’ behaviour in fictional texts. Qualitative data was received due to fictional contexts interpretation. Representation of nonverbal behaviour in case studies of I. McEwan’s contemporary fiction comprises the following most frequent words and word forms: say*, v; tell*, v; talk*, v; speak*, v; sound*, v; sound*, n; voice*, n to describe paralanguage; and go*, v; come*, v; walk*, v; follow*, v; turn*, v to describe moving. The words and word forms of walk, v*, sound*, n, sound, v*, voice, v*, go*, v, come*, v are usually used by the writers to describe characters’ nonverbal behaviour while providing fiction text coherence is mostly bounded to the words tell*, v, say*, v, talk*, v, speak*, v. The importance of paralanguage is determined by its absolute frequency of the word say*, v – 462 and 320 in Corpus 1 and Corpus 2 respectively. The total quantity of words and word forms denoting paralanguage and moving is 1965 in both corpora. The same rank index of the words which denote the nondescriptive meaning of paralanguage and moving in two text corpora manifests and proves the idea that coherence is a text-forming and stable category. The words that describe characters’ nonverbal behaviour contribute to creating dynamism in fictional texts. The set of quantitative data of the present research is the ground for further development of software for linguistic tasks to study the author’s writing style. Supplemented by the data about collocations, obtained practical results might contribute to the study of corpus linguistics, in particular to compile a corpus digital dictionary of words and word forms of nonverbal behaviour represented in contemporary British fiction. 7. References [1] M. Danesi, Nonverbal communication, in: M. Danesi (Ed.), Understanding Nonverbal Communication: A Semiotic Guide, Bloomsbury Academic, London, 2022, pp. 1–26. doi.10.5040/9781350152670.ch-001. [2] M.G. Frank, A. Solbu, Nonverbal communication: evolution and today, in: R. Sternberg, A. Kostić, (Eds.), Social Intelligence and Nonverbal Communication, Palgrave Macmillan, London, 2020, pp,119-162. doi. 10.1007/978-3-030-34964-6_5. [3] J. A. Hall, M. L. Knapp (Eds.), Nonverbal Communication: Handbooks of Communication Science, Walter de Gruyter, Berlin, Boston, 2013. [4] S. Trenholm, Thinking Through Communication: An Introduction to the Study of Human Communication, 9th ed, Routledge, New York, 2021. doi:10.4324/9781003016366. [5] S. Johar, Emotion, Affect and Personality in Speech. The Bias of Language and Paralanguage, Springer, New Delhi, 2016. [6] J. Flanders, F. Yannidis (Eds.), The Shape of Data in Digital Humanities: Modeling Texts and Text-based Resources, Routledge, New York, London, 2019. [7] I. McEwan, Sweet Tooth, Alfred A. Knopf, Canada, 2012. [8] I. McEwan, Solar, Doubleday, New York, 2010. [9] I. Bekhta, N. Hrytsiv, Computational linguistics tools in mapping emotional dislocation of translated fiction, in: CEUR Workshop Proceedings, Vol., 2870, Proceedings of the 5th International Conference on Computational Linguistics and Intelligent Systems, COLINS 2021, Volume I: Main Conference, Lviv, Ukraine, 2021, pp. 685-699. [10] N. Bondarchuk, I. Bekhta, Quantitative characteristics of lexical-semantic groups representing weather in weather news stories (based on British online press), in: CEUR Workshop Proceedings, Vol., 2870, Proceedings of the 5th International Conference on Computational Linguistics and Intelligent Systems, COLINS 2021, Volume I: Main Conference, Lviv, Ukraine, 2021, pp. 799– 810 [11] O. Levchenko, N. Romanyshyn, D. Dosyn, Method of automated identification of metaphoric meaning in adjective + noun word combination (based on the Ukrainian language), in: CEUR Workshop Proceedings, Vol., 3286, Workshop Proceedings of the 8th International Conference on Mathematics. Information technologies. Education, MoMLeT&DC, 2019, pp. 3370-380. [12] D. Head, Ian McEwan: Contemporary British Novelists. Manchester University Press, Manchester, 2007. [13] B. Tonkin, Ian McEwan: Why I’m revisiting the seventies. The Independent, 2012. URL: https://www.independent.co.uk/arts-entertainment/books/features/ian-mcewan-why-i-m- revisiting-the-seventies-8076683.html [14] M. H. Eaves, D. Leathers, Successful Nonverbal Communication: Principles and Applications, 5th ed, Routledge, New York, 2018. doi. 10.4324/9781315542317. [15] C. Bowern, Computational phylogenetics, Annual Review of Linguistics 4 (2018) 281-296. [16] D. McIntyre, B. Walker, Corpus Stylistics: Theory and Practice, Edinburgh University Press, Edinburgh, 2019. [17] E. Purssell, N. McCrae, Reviewing Quantitative Studies: Meta-analysis and Narrative Approaches, in E. Purssell, N. McCrae (Eds.), How to Perform a Systematic Literature Review, Springer, Cham, 2020, pp. 69-101. doi: 10.1007/978-3-030-49672-2_7. [18] M. H. Short, G. N. Leech (Eds.), Style in Fiction: A Linguistic Introduction to English Fictional Prose, 2nd ed., Pearson ESL, London, 2007. [19] A. Gibbons, S. Whiteley, Contemporary Stylistics: Language, Cognition, Interpretation, Edinburgh University Press, Edinburgh, 2022. doi.10.1515/9780748682782. [20] D. McIntyre, B. Walker, Corpus Stylistics: Theory and Practice, Edinburgh University Press, Edinburgh, 2022. doi: 10.1515/9781474413220. [21] G. Ignatow, R. Mihalcea, An Introduction to Text Mining, SAGE Publications, Thousand Oaks, 2018. doi:10.4135/9781506336985. [22] J. Kuhn, Computational text analysis within the Humanities: how to combine working practices from the contributing fields? Lang Resources & Evaluation 53 (2019) 565–602. doi:10.1007/s10579-019-09459-3. [23] S. M. Rasinger, Quantitative Research in Linguistics: an Introduction, 2nd ed., Bloomsbury, London, 2013. [24] N. Nørgaard, Multimodal Stylistics of the Novel: More Than Words, 1st ed, Routledge, New York, 2018. doi:10.4324/9781315145556. [25] M. L. Jockers, T. Underwood, Text-mining the humanities, in: S. Schreibman, R. Siemens, J .Unsworth (Eds.), A New Companion to Digital Humanities, Wiley-Blackwell, West Sussex, 2016, pp. 291-306. [26] J. C. Ho, Using computational text analysis with social media text, SAGE Research Methods Cases, 2022. doi:10.4135/9781529603590. [27] E. Alsop, Making Conversation in a Modernist Fiction: Theory Interpretation Narrative, Ohio State University Press, Ohio, 2019. [28] G. Liveley, Narratology, Oxford University Press, Oxford, 2021. [29] S. Statham, The year’s work in stylistics 2020, Language and Literature 30/4 (2021) 407-433. doi:10.1177/09639470211056687. [30] T. Ogata, T. Akimoto (Eds.). Post-Narratology Through Computational and Cognitive Approaches, IGI Global, 2019. doi: 10.4018/978-1-5225-7979-3. [31] J. Attard, The Connection between literature and aesthetics: is it problematic? Symposia Melitensia 14 (2018) 79-90. [32] B. Stopel, Aesthetic appreciation and the dependence between deep and surface interpretation, Journal of Literary Theory, 14 (2020) 94-119. doi: 10.1515/jlt-2020-0006. [33] W. Bisschop, Interpretation and aesthetic appreciation, Journal of Literary Theory, 14 (2020) 1-9. doi: 10.1515/jlt-2020-0001. [34] M. Danesi. Understanding Nonverbal Communication: A Semiotic Guide, Bloomsbury Publishing, London, 2021. [35] O. Hargie, Skilled Interpersonal Communication: Research, Theory and Practice, 7th ed, Routledge, London, 2021. doi: 10.4324/9781003182269. [36] X. Jiang, (Ed.), Types of Nonverbal Communication, IntechOpen, London, 2021. doi: 10.5772/intechopen.83004.