Idiostyle characteristics of lexical compatibility in the entury prose: ural stylometric project

Размещено на http://www.allbest.ru/

Ural Federal University named after the first President of Russia B.N. Yeltsin

Idiostyle characteristics of lexical compatibility in the entury prose: ural stylometric project

Mikhail Yu. Mukhin and Nikolai Yu. Mukhin

Abstract

idiostyle lexical compatibility

The article presents the project of the Ural Federal University scientists connected with the formalized study of lexical compatibility in Russian classical prose of the 19th century. The study aims to identify idiostylistic characteristics of individual-author syntagmatic. A lexical bigram - a pair of words extracted from one phrase context - is accepted as a unit of compatibility. With the help of their corpus of classical prose (works by Leo N. Tolstoy, Fyodor M. Dostoyevsky, Anton P. Chekhov, Ivan S. Turgenev and Ivan A. Goncharov), the project participants carried out a comparative statistical analysis of lexical bigrams typical for the works of each author and not found in the texts of other writers. A prerequisite for the selection of material is that one of the words constituting a bigram is often used by all authors. Thus, based on the lexical fund common to all authors, the idiostylistical peculiarities of lexical compatibility are revealed. The results of the study are presented on the example of the author's use of adverbs in the works of Fyodor M. Dostoevsky and comparison of syntagmatic characteristics of these adverbs with their textual embodiment in the works of other four authors. Conclusions are made about stylometric perspectives of formalized research of syntagmatic for idiostylistics and author's lexicography.

Keywords: Lexical statistics, stylometry, idiostyle, lexical compatibility, syntagmatic, lexical bigram.

Идиостилевые характеристики лексической сочетаемости в прозе XIX в.: уральский стилометрический проект

М. Ю. Мухин, Н. Ю. Мухин

Уральский федеральный университет им. Б. Н. Ельцина (УрФУ) Российская Федерация, Екатеринбург

Аннотация

Статья представляет проект ученых Уральского федерального университета, связанный с формализованным изучением лексической сочетаемости в русской классической прозе XIX в. Цель исследования - выявление идиостилевых характеристик индивидуально-авторской синтагматики. В качестве единицы сочетаемости принята лексическая биграмма - пара слов, извлеченных из одного фразового контекста. С помощью собственного корпуса классической прозы (произведения Л. Н. Толстого, Ф. М. Достоевского, А. П. Чехова, И. С. Тургенева и И. А. Гончарова) участники проекта провели сопоставительный статистический анализ лексических биграмм, характерных для творчества каждого автора и не встречающихся в текстах других писателей. Необходимым условием выборки материала является то, что одно из слов, составляющих биграмму, часто употребляют все авторы. Таким образом на основании общего для авторов лексического фонда выявляются идиостилевые особенности лексической сочетаемости. Результаты исследования представлены на примере авторского употребления наречий в творчестве Ф. М. Достоевского и сопоставления синтагматических характеристик этих наречий с их текстовым воплощением в произведениях других четырех авторов. Сделаны выводы о стилометрических перспективах формализованного исследования синтагматики для идиостилистики и авторской лексикографии.

Ключевые слова: лексическая статистика, стилометрия, идиостиль, лексическая сочетаемость, синтагматика, лексическая биграмма.

Introduction

The theory of idiostyle today is a multidimensional trend, striving for a systematic description of individual-author features of fictional texts. Years have passed since the works of Victor P. Grigor'ev, who proclaimed the idea of “idiostyle grammar” (Grigor'ev, 1979, 1983), and the idiostylistics has been enriched with stylometric and corpus methods. This gave serious results. It is especially noticeable when a researcher is dealing with large amounts of data that are difficult to grasp traditionally, i.e., to read and summarize or to process by hand, which is very difficult in the case of millions of words.

One of these “multiple” linguistic objects is lexical compatibility. There are attempts to create general language compatibility dictionaries, in particular (Slovar sochetaemosti slov, 2002), see also the review on the portal Gramota.ru (Dictionaries of Words Compatibility). Many works have been written on syntagmatic features of the style of specific authors, which are based on observing a system of facts - see the works (Miheev, 2010; Ruzhitsky, 2015, and many others). However, no systematic syntagmatic “idiostyle grammar” has been formed due to the volume of the material, i.e. a large number of interword semantic relations in each text.

In works on linguistic analysis of the text, there are often close terms “individual compatibility of words”, “author's syntagmatic”, etc., which until recently was also difficult to explain systematically. Statistical comparisons based on text frames allow “materializing” subjective perception of the author's style and such a voluminous component of it as lexical compatibility. From this point of view, it is possible to take a new look at the style of classical writers.

A group of researchers from Ural Federal University proposed a project to study the syntagmatic features of the 19th-century prose based on multiple corpus data More about the project: https://urgi.urfu.ru/formalization/ (the project page on the website of the Ural Federal Universi-ty); https://nauka.tass.ru/nauka/6144728 (TASS news).. The fundamental source of the project was the concept of individual compatibility proposed in the doctoral thesis (M. Mukhin, 2011) and previously approved on the material of the 20th-century prose. A lexical pair (bigram) which consists of two lexemes of significant parts of speech located in one phrase context at a distance of no more than five words is recognized as a unit of compatibility.

Parameters of the text corpus and research methodology

The corpus, based on which the study is conducted, are the works of Leo N. Tolstoy, Fyodor M. Dostoevsky, Anton P. Chekhov, Ivan S. Turgenev and Ivan A. Goncharov (about 4 million words). At present the corpus is morphologically marked, the interword homonymy within words of significant parts of speech is removed. The semantic centered approach of lexical compatibility research allows ignoring function words in general, while aware that there are also special author's possibilities of using prepositions, conjunctions or particles.

The formation of any corpus is a separate task: it must meet the criteria of representativeness and completeness. In the first stage of the corpus planning, the list of texts to be used by each author was discussed and approved. The general attitude towards the equal participation of the classical authors idiostyle determined the fact that each of them is represented in the corpus by four novels (more major works do not appear in the heritage of all writers). Besides, Anton P. Chekhov's prose was divided chronologically and conceptual-stylistically into four large fragments, recognized as independent subsections of the corpus. As a result of discussion and consultation, the following works were included in the corpus:

Ivan S. Turgenev: “Rudin”, “Home of the Gentry”, “On the Eve”, “Fathers and Sons”;

Ivan A. Goncharov: “A Common Story”, “Oblomov”, “The Precipice”, “Frigate `Pallada'”;

Fyodor M. Dostoevsky: “Crime and Punishment”, “The Idiot”, “Demons”, “The Brothers Karamazov”;

Leo N. Tolstoy: “Family Happiness”, “War and Peace”, “Anna Karenina”, “Resurrection”.

Anton P. Chekhov: chronologically

conditioned fragments representing different periods of work (prose works of 1880-1885, 1886-1887, 1888-1897, 1898-1904, conditionally understood as four large texts).

Thus, each author is represented in the corpus by four major subsections.

To introduce a unified system of coordinates in the consideration of the styles of different authors and different works of each of them, the context of frequency words used by all these authors (for instance, человек - a human, говорить - to talk, рука - a hand, лицо - a face, понимать - to understand, выходить - to go out, думать - to think, очень - very, место - a place, вдруг - suddenly, опять - again, большой - big, etc., more than a hundred) were systematized. These lexemes are considered to be one of the most used in the Russian language, which is reflected by frequency dictionaries, but for us, they must be often found in every sufficiently large work of all five discussed writers. From the sentences that include these frequency lexemes, the words that are next to them, at a distance of one to five left and right, are selected.

Thus, lexical context pairs (bigrams) are obtained, which necessarily include one of the words in the frequency list (for instance, выходить - to go out or опять - again). To understand the volume of syntagmatic sampling, it is necessary to explain that there are 1 674 948 lexical bigrams in the corpus. Therefore, it is clear why it was impossible to carry out continuous research on compatibility before - our set of bigrams is based only on the most frequent words. If we sum up all the lexical pairs used in phrase-based contexts, in any major novel the number of bigrams (i.e., in fact, contextual lexical relations) will be measured in millions.

Then, it is necessary to determine the composition of individual-author bigrams significant for the idiostyle, i.e. the material for further contextual analysis. Bigrams, which occur in more than one work of one author (especially, when they occur repeatedly) and do not appear in the texts of other authors at all, should be recognized as indicative for the study of individual-author features of lexical compatibility. This limitation is directly related to the assessment of the significance of a certain bigram for the writer's idiostyle, so that the selected material could reveal the peculiarities of lexical compatibility in the texts of a particular author and at the same time would not be strongly dependent on the volume and content of one work.

Thus, the units of special analysis in this study are lexical bigrams that are typical for one writer's texts as compared to other authors; at the same time, these bigrams are based on words frequently used by all authors. According to the research logic, the statistical comparison is followed by the stages of contextual analysis and idiostylistic generalization.

Initial research results

As a result of the lexical-statistical analysis, arrays of individual-author bigrams (1475 in total) were formed from more than 1.5 million lexical pairs, which occur in more than one work of one author and are not found in phrase contexts of works of other authors. It should be emphasized once again that such a comparative approach allows revealing specificity of author's compatibility of words and at the same time to avoid the influence of a plot of concrete work. The methodology involves comparing one author with others, as well as several works by one author. The latter is particularly important in terms of compatibility with personal names and toponyms relevant to a particular text. For example, bigrams as «сказать + Раскольников» and «Соня», «Порфирий» (in English, “say + Raskolnikov” and “Sonia”, “Porphyry”, etc.), naturally, will get to the “Crime and Punishment” frequency top, but for Fyodor M. Dostoevsky's works as a whole, they will not be indicative. For details on methodological issues see also the works of the authors (M. Mukhin, 2019; M. Mukhin & N. Mukhin, 2019).

The formed array of author's bigrams and their phrase contexts allows revealing syntagmatic priorities of the considered authors. Thus, a set of words that has the greatest number of usage with original context partners in the work of each of them - in comparison with texts of other authors is defined. In other words, each writer has their original array of vocabulary, which shows individual-author “syntagmatic activity”, i.e. has a relatively larger number of original contextual partners in the work of a particular writer than other authors.

At this stage, it has been found that in the most interesting contexts, the following words are more common among different authors than others:

Anton P. Chekhov: глаз - an eye, лицо - a face, говорить - to talk, глядеть - to look, рука - a hand, сидеть - to sit, голова - a head, нога - a leg, etc.;

Fyodor M. Dostoevsky: вдруг - suddenly, уже - already, совсем - completely, знать - to know, опять - again, теперь - now, замечать - to notice, etc.;

Ivan A. Goncharov: сказать - to say, жизнь - life, друг - a friend, дом - a house, спрашивать - to ask, давать - to give, идти - to go, надо - it is necessary, сторона - a side, etc.;

Ivan S. Turgenev: начинать - to start, человек - a human, мочь - can, очень - very, долго - long, продолжать - to continue, думать - to think, слово - a word, etc.;

Leo N. Tolstoy: время - time, отвечать - to answer, видеть - to see, делать - to do, дверь - a door, старый - old, разговор - a conversation, дело - an affair, приезжать - to come, etc.;

These lists need to be further interpreted: these words are polysemantic, and their list is semantically abstract without context. Only such general moments as, for example, the specificity of the group of nouns denoting body parts in Chekhov's works attract attention. To identify the parameters of the author's compatibility, it is necessary to consider the list of contextual partners of these words. This requires contextological analysis leading to idiostylistic conclusions.

The example of the interpretation

of the received results: Fyodor M. Dostoevsky

in comparison with other authors

For example, let us consider here a group of adverbs characterized by special compatibility in the works of Fyodor M. Dostoevsky. These words are the following: вдруг - suddenly, еще - still, сейчас - now, совсем - completely, теперь -presently, уже - already. Some words of this list have already been previously noted by Dostoevsky's researchers as significant for the idiostyle. Thus, a brilliant analysis of the contextual environment of the adverb «вдруг» (in English, “suddenly”) was carried out by Igor V. Ruzhytsky (Ruzhytsky, 2011). This adverb, along with the words «почти» (almost) and «слишком» (too much), is recognized as the so- called idioglossia in Fyodor M. Dostoevsky's idiostylistic dictionary (Slovar iazyka Dostoevskogo, 2001). However, other adverbs of our list in the dictionary of the project designed by Yuri N. Karaulov and continued today by Igor V. Ruzhytsky do not appear: they are too common and do not attract much attention. The works of Abram A. Belkin (1973), Guo Yun Yang (2011) and other authors also mention the characteristic features of Dostoevsky's semantics and text functions of adverbs. This study suggests, as mentioned earlier, selective consideration of contexts that are specific to one author and not specific to the other.

Let us take a closer look at the specifics of the original context environment of the above-mentioned adverbs. All the following contextual partners are repetitive (sometimes dozens of times) and occur in at least two of Dostoevsky's works.

We propose here first to quote the lexemes in the original in Russian with the analysis, and then give their translation into English as a separate paragraph.

[Вдруг] вскричать, завопить, отрезать, чрезвычайно и чрезвычайный (чаще для квалификации эмоций), ощущать/ ощутить, подхватить (о речи), вскинуться, раздражительно (воскликнуть и т. п.), озлиться, залепетать, в исступлении, слишком (о чувстве), стремительно (о чувстве, эмоции в движении), выказывать/ выказать, истерический (смех, ожидание, восторг, припадок), презрительно (посмотреть, сказать), ненавистен (стал, станет).

[Suddenly] to scream, to howl, to snap out, extreme and extraordinary (more often to describe emotions), to feel, to pick up (about speech), to leap, irritable (to exclaim, etc.), to get angry, to go crazy, too much (about feelings), too fast (about feelings, emotions in motion), to evince, hysterical (laughter, waiting, delight, seizure), contemptuous (look, to say), hated (to become / will become).

[Опять] рассмеяться, вскричать, прокричать, прервать (собеседника).

[Again] to laugh, to scream, to shout, interrupt (an interlocutor).

[Сейчас] вскричать.

[Now] to cry out.

[Совсем] посторонний (о чем- либо постороннем, обстоятельстве), сумасшедший, крикнуть, нечаянно, сбиваться/сбиться, (не) нуждаться (в ком-, чем-либо), обстоятельство (совсем непредвиденное, постороннее).

[Completely] external (about something external, circumstance), a madman, to yell, accidentally, to stumble, to have (no) need (in someone, something), a circumstance (quite unexpected, external).

[Теперь] припоминаться/

припомниться, разрешать/разрешить (найти решение), честь (которой теперь лишен).

[Now] to recall, to allow (to find a solution), an honour (which is now lost).

[Уже] наверно (наверняка), давеча, в исступлении (до исступления), обнаруживаться/обнаружиться (проявиться), (не) веровать.

[Already] probably (for sure), recently, in a frenzy (to go berserk), to be discovered (to reveal), (do not) believe.

It is obvious that in the context environment of frequency adverbs in Dostoevsky's texts emotive vocabulary and words that denote the degree and speed of manifestation and flow of emotions (чрезвычайно - extremely, слишком - too much, стремительно - rapidly) prevail. And the emotive vocabulary mostly indicates the extreme degree of manifestation of feelings. The special emotional richness of Fyodor M. Dostoevsky's texts is undoubtedly an idiostylistic trait and is reflected in a great number of philological studies. Nevertheless, it is important for us that Dostoevsky uses words common to all authors in special, non-emotive contexts.

We emphasize that we are talking about recurring contexts. For example, the bigram «вдруг + вскричать» (“suddenly” + “scream”, “exclaim”, “shout”, “cry” in English) is found in different novels of Dostoevsky 34 times:

На эту картину! - вскричал вдруг князь, под впечатлением внезапной мысли, - на эту картину! («Идиот»);

`That painting! ' the prince exclaimed suddenly, under the impact of a sudden thought. `That painting! (“Idiot”);

Вот это так уж ложь, Алеша! - вскричал вдруг Иван, засверкав глазами («Братья Карамазовы»);

That's such a lie, Alyosha! - Ivan suddenly shouted, shining with his eyes (“The Brothers Karamazov”);

Какое преступление? - вскричал он вдруг, в каком-то внезапном бешенстве... («Преступление и наказание»);

What crime? - He shouted suddenly, in some sudden rage... (“Crime and Punishment”);

Убир-райся, Блюм! - вскричал вдруг фон Лембке, заслышавший голос своей супруги в соседней комнате («Бесы»);

“Get away, Blum," Von Lembke cried suddenly, hearing the voice of his spouse in the next room (“Demons”) - and many other examples.

Let us concretize the statistical data on collocations. It is Dostoevsky who combined the adverb «вдруг» (“suddenly”) with the following words: завопить - to squeal (16 times), отрезать - to snap out (14), чрезвычайно - extremely (12), ощущать - to feel (10), подхватить - to pick up (about speech) (10), вскинуться - to leap (9), раздражительно - irritably (to exclaim and other verbs) (8), озлиться - to get angry (7), etc. Thus, there are not random but repeatedly occured through various works.

To explain that the above-mentioned bigrams are typical exactly for the works of Fyodor M. Dostoevsky, and not universal for the authors of the 19th century, let us consider examples of the original contextual partners of these dialects of other classic writers. The list of words here is not presented in full. Thus, the word «вдруг» (suddenly) is combined with the words «отрезвиться» (to sober up) and «живо» (vividly) at I. A. Goncharov: But suddenly he seemed to have sobered up from this frustration, woke up from a heavy meditation (“Oblomov”); The old man listened long and carefully, then suddenly waved his hand vividly... “Frigate `Pallada'”). Besides, in Tolstoy's novels the adverb “suddenly” occurs several times together with «багрово» - “crimson” (покраснеть - to flush): Pierre suddenly flushed crimson and for a long time tried not to look at Natasha (“War and Peace”). Ivan S. Turgenev's adverb “yet” is associated with the verb «потолкаться» - “to spend some time in a crowded place” (instantiations), Leo N. Tolstoy's “still from afar” can see or hear something - a still green oat grows in the field, and a face or smile can shine (the only emotive word, unlike Dostoyevsky).

The word “again” is rarely found in other writers and original context partners. Anton P. Chekhov has this “a little later” and “darkness”. Ivan A. Goncharov - «спрятаться» - “to hide”, «задумчиво» - “thoughtfully” (to look), «склонить» - “to incline” (a head), «нервы» - “nerves” and «ровный» - “straight” - for example, she has become calm again, straight, unsophisticated, sometimes even cold (“Oblomov”). Leo N. Tolstoy has the verbs «устанавливаться» - “to be established” (about a relationship or facial expression) and «вглядываться» - “to gaze”. In the works of the authors taken for the comparison, other adverbs are not included in a sufficient number of examples of the original author's compatibility. Thus, it is the texts of Fyodor M. Dostoevsky where considered words have such a rich contextual environment and where they are included in the list of the most popular lexemes. Every author uses these words, but not everyone particularly uses them. This selectivity has an idiostylistic nature, as each of these writers has its own set of lexemes, manifested in the original contexts. It is doubtful that this can be done intentionally when creating various fiction works.

Conclusion

The proposed model of syntagmatic analysis in the long-term perspective leads to the development and systematization of philological judgments about the author's lexical compatibility. What in the perception of large works can escape from the reader's view is presented here as “precious particles”, isolated from the quite ordinary, inherent in all combinations of words. An important prerequisite for such a study, as well as for lexical statistics in general, is the volume material, large corpus of texts, large works.

Besides, we can talk about the further development of the author's lexicography, for which it is critical to reflect the idiostylistic, rather than generally significant material. It seems that syntagmatic features of words, unlike the basic author's lexicon, are much more difficult to systematically present in the author's dictionary.

References

Belkin, A.A. (1973). “Vdrug” i “slishkom” v khudozhestvennoi sisteme Dostoevskogo [“Suddenly” and “Too” in Art System of Dostoevsky]. In Chitaia Dostoevskogo i Chekhova. Stat'i i razbory [Reading Dostoevsky and Chekhov. Articles and Analyses]. Moscow, Khudozhestvennaia literatura, 129-134.

Guo Yun Yang (2011). Narechiia s semantikoi neozhidannosti i iterativnosti kak markery povestvuiushchego sub“ekta v povestiakh F.M. Dostoevskogo [The Adverbs with the Semantics of Surprise and Iteration As Markers of the Narrative Subject in the Novels by F.M. Dostoevsky]. In Vestnik SPbGU [Vestnik of Saint Petersburg University], 9 (4), 130-133.

Slovari sochetaemosti slov [Dictionaries of Word Compatibility]. In Gramota.ru. Available at: http:// gramota.ru/slovari/types/17_6 (accessed 5 May 2020).

Grigor'ev, V.P. (1983). Grammatika idiostilia: V. Hlebnikov [Grammar of Idiostiles: V. Khlebnikov]. Moscow, Nauka, 225 p.

Grigor'ev, V.P. (1979). Poetika slova [Poetics of the Word]. Moscow, Nauka, 343 p.

Miheev, M.Yu. (2010). Kompiliatsiia ili... iazykovye klishe? Sravnivaia kharakternye dlia avtorskogo stilia nabory slovosochetanii [Compilation or... Language Cliches? Comparing Sets of Word Combinations Typical for the Author's Style]. In Komp 'iuternaia lingvistika i intellektualnye tekhnologii: Po materialam ezhegodnoi mezhdunarodnoi konferentsii «Dialog» [Computer Linguistics and Intellectual Technologies: On the Materials of the Annual International Conference “Dialogue”]. Available at: http://www.dialog-21. ru/dialog2010/materials/html/52.htm (accessed 5 May 2020).

Mukhin, M.Yu. (2019). Leksicheskaia statistika i stil' avtorov klassicheskoi prozy XIX v. [Lexical Statistics and Style of Classical Prose Authors of the XIX Century]. In Teoreticheskaia semantika i ideograficheskaia leksikografija: Slovar. Diskurs. Korpus [Theoretical Semantics and Ideographic Lexicography: Dictionary. Discourse. Corpus]. Yekaterinburg, 119-126.

Mukhin, M.Yu. (2011). Leksicheskaia statistika i idiosti'l avtora: korpusnoe ideograficheskoe issledovanie (na materiale proizvedenii M. Bulgakova, V. Nabokova, A. Platonova i M. Sholohova) [Lexical Statistics and Idioms of the Author: Corpus Ideographic Research (On the Material of the Works ofM. Bulgakov, V. Nabokov, A. Platonov andM. Sholokhov)]. Yekaterinburg, 383 p.

Mukhin, M.Yu., Mukhin, N.Yu. (2019). Avtorskii tezaurus i sochetaemost slov: leksiko-statisticheskie modeli individualnogo stilia [Author's Thesaurus and Compatibility of Words: Lexical-Statistical Models of Individual Style]. In Filologicheskii klass [Philological Class], 4 (58), 8-15.

Ruzhytsky, I.V. (2011). Tekstoobrazuiushhaia funktsiia slova «vdrug» v proizvedeniiah F.M. Dostoevskogo [Text-Forming Function of the Word “Suddenly” in the Works of F. Ruzhitsky. M. Dostoevsky]. In Vestnik RUDN, Teorija iazyka. Semiotika. Semantika [Vestnik of RUDN. Language Theory. Semiotics. Semantics]. 3, 18-25.

Ruzhitsky, I.V. (2015). Iazykovaia lichnost' F.M. Dostoevskogo: leksikograficheskoe predstavlenie [Linguistic Personality of F.M. Dostoevsky: Lexicographical Representation]. Moscow.

Slovar sochetaemosti slov russkogo iazyka [Dictionary of Russian Word Compatibility] (2002). Moscow, Astrel AST, 811 p.

Slovar iazyka Dostoevskogo. Leksicheskij stroi idiolekta [Dostoevsky Language Dictionary. Lexical System of Idiolect] (2001). Moscow, Azbukovnik. Available at: http://slovari.ru/default.aspx?s=0&p=227 (accessed 5 May 2020).

Размещено на Allbest.ru