Criteria for distinguishing homonymy from polysemous words (psycholinguistic method).

Lastochkina Elena Grigor'evna PhD in Philology Associate Professor, Department of Mari Language and Literature, Mari State University 424019, Russia, Republic of Mari El, Yoshkar-Ola, pl. Lenin str., 1 antrolea@mail.ru Other publications by this author

Polysemy (from Greek polys?mos – "polysemous") – polysemy; the ability of a word to have two or more meanings. Homonymy (from Greek. homonymia – "homonymy") is a sound and/or graphic coincidence of linguistic units that differ in meaning. Through polysemy and homonymy, the communicative and cognitive functions of language are realized: knowledge and experience are transmitted through linguistic means limited by the nature of language. ^{[1, p. 48]}.

Homonymy is the result of a historical or accidental coincidence of linguistic signs with a complete difference in their content; the result of both internal linguistic transformations (for example, a disintegrated polysemy) and changes that led to an intersection, coincidence with other languages, which allows us to speak about the manifestation of one of the aspects of translinguism. This aspect of translinguism, associated with certain coincidences in language, has always been of interest to linguistics scholars. For example, already in the linguistics of ancient times we can find the definition of homonyms, in particular in Aristotle ^{[2, p. 52]}.

Semantic differences between the uses of words may be related to social, psychological differences between people using this word. For each person, one or another word is associated with a specific system of associations, with an individual system of knowledge about the world, with individual experience ^{[3, p. 157]}. So, for example, a person who is not a specialist in the relevant field of veterinary medicine associates the word purtash with a verb in the meaning of ‘to let in, to import’, and for a specialist this word has another meaning purtash vet. ‘bloating (cow disease)'.

After analyzing the criteria for distinguishing homonymy and polysemy in the Mari language, we decided that the semantic side of the word plays the main role ^[4-6], but even in this case, we cannot say for sure whether this unit is polysemy or homonymy^{[7, p. 569]}.

As stated by S. M. Potapov: "It is necessary to know how the majority of speakers of a given language perceive the analyzed phenomena and proceed from this" ^{[8, p. 49]}.

B. N. Golovin also noted that a lexicographer should not rely only on his own linguistic flair ^{[9, p. 82]}. O. N. Seliverstova believed that hypotheses that can be obtained using component analysis "should be tested using experimental research methods" ^{[10, p. 169]}; the importance of using experimental techniques in the field of semantic research was pointed out by V. M. Pavlov ^{[11, p. 153]}; indications of linguistic consciousness and experimental methods form the basis of their dissertations by O. P. Boldin ^{[12, p. 21]}, I. V. Tubalov ^{[13, p. 249]}, applying them, in particular including in the analysis of lexical and semantic motivation.

Our experiment is based on Kachurin's experience in distinguishing polysemy from homonymy in his dissertation "The problem of distinguishing homonymy and polysemy in relation to the practice of compiling explanatory dictionaries" (2013), where we will try to offer support as a verification method on the testimony of the linguistic consciousness of native speakers using the method of psycholinguistic experiment.

The following questions were put to the interviewees: Do you think there are grounds for calling the phenomena X and Y the same, is there anything in common between them? If so, why do you think these phenomena are called the same? We also ask the respondent one more question: "How significant, unrealistic do you consider such a connection? Rate it on a scale from 1 to 5, where 1 is not obvious, far—fetched, barely detectable, fragile, 5 is obvious, very durable." This question performs two functions at once: first, it forces the respondent to think again about his own answer and thereby check himself; secondly, with the help of this question, asked to each respondent, we will be able to calculate the average result for each pair of values using the arithmetic mean formula and thus obtain a certain "the coefficient of semantic connectedness" (hereinafter referred to as KSS) of the average native speaker, thereby fulfilling the first condition (to see the average result for each case). The respondent's failure to detect a semantic connection is evaluated as zero. We propose such a distribution of pairs of values in the three named classes, depending on their KSS: homonyms — KSS<2; transitional class — 2?KSS<4; polysemy — KSS?4 ^{[14, pp. 245-246]}.

We have chosen the words taking into account such conditions: 1) at least one of the dictionaries considers these meanings as belonging to one word; 2) homonymy is recorded in all dictionaries included in the sample, but our own linguistic intuition tells us that a semantic connection is possible between the meanings [Ibid., p. 256].

Our respondents were students and graduates of the Faculty of Philology. We believe that this category of people can answer our questions objectively and give an accurate assessment of the semantic connection between words.

During the survey, we received 1000 language reactions (40 respondents were offered 26 pairs of meanings). Here are examples of each value and the total KCC:

- igeche I ‘weather’ – igeche II ‘age’: possibly age, as the passage of time and changing as the weather; CCSS = 2.22;

- II I ‘year' – II II ‘ice': no similarity found; General. = 0,91;

- iqt?rlik I is ‘the same’ – iqt?rlik II the second value is not defined by anyone (iqt?rlik II mat. ‘equality');

- Yyzhy? I ‘chapter’ – yyzhy? II ‘strength': the similarity is not specified, but it is; KSSC = 1.08;

- Yilme I ‘language as the Mari language’ – Yilme II ‘human organ': language (organ) as a means of implementing a language (a system of sound and other means); CCSS = 2.44;

- ketch I ‘the sun’ – ketch II ‘the day': the similarity is that the day is the time interval from sunrise to sunset; CCSS = 3.05;

- case I ‘partition' – case II ‘one part in performances': similarity in function, because the partition performs the function of dividing rooms into two parts, and in performances one part is divided into another conditionally; CCSS = 1.40;

- mogyr I ‘body’ – mogyr II ‘side': similarity in terms of orientation, as the right or left side of the body; CCSS = 1.45;

- need I ‘nettle' – need II ‘pike, letters. in Mari, the nettle fish’: similarity in appearance; CCSS = 0.74;

- optem I ‘yolk' – optem II ‘perga, flower pollen’, folded into honeycomb cells; CCSS = 1.14;

- sem I ‘melody' – sem II the second meaning is not defined (sem II ‘sense, mind, mind’);

- sola I ‘yal’ – sola II ‘whip‘, but in most cases the second value was considered as sola ’mows', ch., 3 l., unit h.: no similarity was found: CCSS = 0.10;

- beat I ‘rhythm’ – beat II ‘sense of proportion': similarity in function – beat as observance of certain rules, norms; CCSS = 3.2;

- volume I ‘volume of the book' – volume II ‘core': some functional similarity, volume as the basis; CCSS = 0.88;

- tylze I ‘moon’ – tylze II ‘month': functional similarity – month is a unit of time measurement associated with the Moon's orbit around the Earth; CCSS = 3.85;

- check I ‘document’ – check II ‘border': no similarity found; CCSS = 0.21;

- chyvytan I ‘leech’ – chyvytan II ‘tweezers': similarity in function; CCSS = 1.29;

- ball I ‘globe' – ball II ‘ponytail': no similarity found; CCSS = 0.28;

- shozh I ‘barley, grain crop’ – shozh II ‘eye disease': external similarity; CCSS = 0.91;

- shotlymash I ‘counting’ – shotlymash II ‘reverence': although no similarity was found, nevertheless, to read from the word count, i.e. the connection is present; KSSC = 0.80;

- shoya I ‘fairy tale’ – shoya II ‘lies’; the content of fairy tales is fiction, i.e. lies; CCSS = 2.29;

- sh?r I ‘the edge, the native place’ – sh?r II ‘milk’: no similarity was found; CCSS = 0.28;

- neck I ‘neck’ – neck II ‘coal': no similarity found; CCSS = 0.91;

- sh?m I ‘heart’ – sh?m II ‘tree bark’: a similarity was found in the fact that the heart, as the basis of a person, and the bark is of great importance to the tree; CCSS = 1.38;

- shishpyk I ‘nightingale’ – shishpyk II ‘whistle': similarity in function – both whistle; CCSS = 2.97;

- shyrpe I ‘matches’ – shyrpe II ‘splinter': similarity in appearance was found; CCSS = 2.52 [see Lastochkina 7, 15].

Our words were distributed as follows:

1. Homonyms: iy I ‘year’ – iy II ‘ice’, Yyzhy I ‘chapter’ – Yyzhy II ‘power', kydezh I ‘partition’ – kydezh II ‘one part in performances', mogyr I ‘body’ – mogyr II ‘side’, nuzh I ‘nettle’ – nuzh II ‘pike’, optem I ‘yolk’ – optem II ‘parchment, pollen’, sola I ‘yal’ – sola II ‘whip’, volume I ‘volume of the book’ – volume II ‘core', check I ‘document’ – check II ‘border’, chyvytan I ‘leech' – chyvytan II ‘tweezers', ball I ‘globe' – ball II ‘horse tail’, shozh I ‘barley, grain crop’ – shozh II ‘eye disease‘, shotlymash I 'account' – shotlymash II ‘reverence’, shur I ‘edge, native place’ – sh?r II ‘milk’, sh?i I ‘neck’ – sh?i II ‘coal, sh?m I ‘heart’ – sh?m II ‘tree bark'.

2. Transitional class: igeche I ‘weather' – igeche II ‘age', yilme I ‘language like the Mari language’ – Yilme II ‘human organ', ketche I ‘sun’ – ketche II ‘day', beat I ‘rhythm’ – beat II ‘sense of proportion', tylze I ‘moon' – tilze II ‘month', shoya I ‘fairy tale‘ – shoya II ’lies', shishpyk I ‘nightingale’ – shishpyk II ‘whistle', shyrpe I ‘matches’ – shyrpe II ‘splinter'.

3. In two cases, the meaning of the second word was not determined by sem I ‘melody’ – sem II the second meaning is not determined (sem II ‘sense, mind, mind’); ikt?rlyk I ‘the same’ – ikt?rlyk II the second meaning is not determined by anyone (ikt?rlyk II mat. ‘equality').

After the survey we conducted, we were convinced that some of our hypotheses coincide with the respondents, and some differ.

So, for example, we assumed that the words yilme, keche, takt, tylze are actually not homonyms, which was confirmed by our survey. As for the words Shoya I ‘fairy tale’ and Shoya II ‘lies, lies, deception’, then we have to disagree and attribute this unit to homonyms, because although there is a semantic connection between these words, but each of them names a different phenomenon. The same can be said about the words shishpyk and shyrpe, which we will classify as homonyms. In the word igeche, the respondents proposed a different interpretation: igeche I ‘weather’ – igeche II ‘day’, so the pair ended up in an intermediate class, but if we consider the word in the dictionary form igeche I ‘weather’ – igeche II ‘age, years’, then we will classify them as homonyms.

We especially note the words ikt?rlyk and this, because the meanings of the other word were not determined: ikt?rlyk I ‘the same’ – ikt?rlyk II mate. equally; sem I ‘melody, motive' – sem II ‘sense, mind, mind’ – sem III 'method, method, means'. Most often, this word iktyrlyk is used in the first meaning of ‘the same’, and the second word is used only by specialists in the field of mathematics, and the word sem is widely known as a motif, melody: Kushtymo sem dance melody; muro sem song motif; marla sem Mari melody. Miklay yocha godsek yeratym semzhym shishkalta. G. Chemekov. Miklay is whistling a tune he has loved since childhood. The composer of the cart was the seventh, and the poet was the fifth. V. Sapaev. The composer composes gentle music, poets write beautiful poems. Bayan sem pochesh tavaltymet vele shuesh. V. Sapaev. I want to dance to the sounds of the accordion ^{[16, pp. 171-172]}.

So, after the experiment, we can say that this kind of research is effective. Although there are points that need to be corrected: we were not satisfied with the answers to the question "if there is a connection between words, then how does it manifest itself?" Very often, the respondents answered "there is a connection" or "there is no connection" and did not explain how this commonality of values could be determined. It was also not clear whether the semantic connection was being evaluated, rather than stylistic or otherwise. Therefore, as G.S. Shchur argued, "there are different types of dependencies... The relationship between the presented word and its associates is determined not by the presence of semantic commonality, but by the correlation of the corresponding components with reality" ^{[17, p. 145]}. Thus, it can be recognized that it is necessary to refer to the speech experience of as many native speakers as possible when distinguishing homonymy from polysemy.

Of course, the nature of associations is influenced by age, geographical conditions, a person's profession, as well as the level of proficiency in the Mari language, but, nevertheless, the opinions of respondents mostly coincide with lexicographic data.