Mokslinio ir grožinio stiliaus sakinio ilgis

LIETUVIŲ KALBOTYROS KLAUSIMAI XXXIII (1995) GRAMATIKA IR LEKSIKOLOGIJA AUDRONĖ BITINIENĖ SENTENCE LENGTH IN SCIENTIFIC AND LITERARY STYLES Functional language styles are most often studied from a qualitative perspective: attention is paid to tendencies in the selection and arrangement of certain linguistic devices. In Lithuanian stylistics, literary and scientific styles are characterized by such features as concrete imagery, emotionality, and individuality, on the one hand, and abstraction, logicality, and objectivity, on the other (Pikčilingis 1971: 287). This approach, however, can be supplemented by a quantitative aspect, since the concept of a tendency is statistical: researchers establish greater or lesser activity of linguistic devices, described by words such as “often,” “rarely,” etc., by observing their “statistical behavior.” After examining a sufficiently large sample, it is easy to calculate how often, for example, the subject is expressed by an infinitive verb in Lithuanian scientific texts, or what percentage of all complex sentences are clauses of cause. Among all the features of functional styles, sentence length holds a special place. This quantitative measure provides not only stylistic information but is also a component of sentence structure; a sentence one or two words long differs fundamentally in its structure from a 50-word sentence. On the other hand, the choice of sentences and the way they are expanded depend on the sphere of activity, the prevailing language function, the content, and other conditions governing language use. Thus, sentence length seems to imply both the qualitative and quantitative aspects of the study of functional style. For this reason, sentence length in various languages is studied quite frequently. Sentence length is studied from various perspectives. Some studies take the view that sentence length is an element of grammar, directly related to the development of the grammatical structure of language (Admoni 1966). In other cases, sentence length is examined in the context of language statistics (Meier 1964; Herdan 1960). There are also many studies that treat sentence length as a feature differentiating functional styles (Лесскис 1962; Долежел 1964; Мистрик 1967; Arader 1963; Eggers 1962; Těšitelová 1980) or individual style (Yule 1939; Вашак 1974). In Lithuanian linguistics, sentence length has been described by J. Pikčilingis. He notes that “the style of language determines one or another sentence length” (1971: 137). The purpose of this article is to compare sentence length in scientific and literary styles; this analysis continues the aforementioned tradition of stylistic research on the Lithuanian language. Every statistical study is required to provide an operational definition of the objects under examination, that is, an understanding of the scientific concept that makes it possible to define the object unambiguously. A sentence is considered to be a segment of text from one full stop to another, or to another punctuation mark indicating the formal end of a sentence (Адмони 1973: 15). The unit for measuring sentence length is the word, that is, a grapheme or sequence of graphemes separated by spaces. Such definitions specify which 169 segment of text is to be considered a sentence and which grapheme a word; this is particularly relevant in scientific texts, where abbreviated titles are fairly common and concepts are expressed by symbols, nonalphabetic signs, and the like. The research sample consisted of 34 texts representing various fields of science (humanities, natural sciences, mathematics, and technology) and substyles (theoretical, educational, and methodological), from which 10,000 sentences were randomly selected, as well as 32 prose writers’ texts (authorial narratives), from which another 10,000 sentences were analyzed. The study examined not only the length of each sentence but also certain features of sentence structure—the sentence type and the number of clauses in a complex sentence. Sentence types were determined according to the classification presented in the Lithuanian Grammar (1976). The data were processed at the Vilnius University Computing Centre using the statistical analysis program “Paula,” written in PL/1. When discussing the results obtained, two qualifications must be borne in mind: first, word length and the lexical and grammatical meanings of words are abstracted away from, that is, all the words making up a sentence are treated as equivalent; second, each sentence’s length is measured as a whole number of words, so its features, expressed in tenths or even hundredths, are to some extent an abstraction. To describe sentence length rigorously, we would have to give the percentages of sentences whose length ranges from one word to the greatest possible number of words (121 words in our sample). But this would be inconvenient and even incomprehensible to the reader. Therefore, virtually all researchers work with the principal parameters of the sentence-length frequency distribution, including the mean sentence length (x̄). This parameter, however, describes the actual distribution of sentences in texts too approximately, so other parameters are also needed—the standard deviation (s), coefficient of variation (v), and range. The asymmetry of the distribution is essential: sentence-length frequencies are generally distributed with positive skew, that is, the mean length exceeds both the modal (most frequent) and median (dividing the sample into two equal parts) sentence length (see Table 1). Table 1 Parameters of sentence-length distributions in scientific and literary styles Functional style | Mean length x̄ | Standard deviation s | Coefficient of variation v | Range | Median | Mode | Skewness coefficient | Standard error coefficient | Scientific | 18.0 | 8.8 | 48.8 | 2—57 | 16.3 | 16 | 0.79 | 0.06 | Literary | 12.3 | 9.9 | 80.5 | 1—121 | 10.4 | 5 | 0.93 | 0.04 The data presented in the table indicate that sentence length differs between scientific and literary styles. The statistical significance of this difference (taking sample size into account) is assessed as follows. Mean length t= 3.38, p<< 0.001; that is, the difference in mean length is statistically absolute. The difference in standard deviations is less pronounced (t= 2.1, p< 0.05), but it is also statistically significant. 170 The skewness coefficients differ quite significantly (t= 2.0, p< 0.05); in this case, it is rather difficult to test the statistical significance of differences in the coefficient of variation, range, mode, and median, but the absolute differences are beyond doubt. Although sentence lengths in both styles have right-skewed distributions, this feature is much more pronounced in the literary style. In the scientific style, the median and mode of the distribution differ from the mean length by two words, while in the literary style the mode differs by as many as 7 words (the median, too, differs by two words). The distribution of sentence lengths in the scientific style is more compact (a narrower range and a smaller standard deviation despite a greater mean length). In other words, sentence length varies more in literary texts than in scientific texts. For a more detailed analysis of sentence length, and especially to reveal the effects of linguistic and extralinguistic factors, it is useful to group sentences and calculate the frequencies of the resulting length groups. Sentences are grouped in a variety of ways in the literature. Some authors distinguish groups with equal and unequal interval lengths, without giving a qualitative assessment of these groups (Статистични 1967); others assign qualitative labels to grouped sentences: short sentences—up to 20 words, long sentences—from 20 to 60 words, and very long sentences—more than 60 words (Meier 1964: 186). It should be noted that the basis for grouping is generally not specified or linked to the needs of functional stylistics or syntax. In this study, sentence grouping by length intervals was based on a statistical criterion: the boundaries of the intervals were determined by the percentiles of the empirical (sample) distribution and the “jumps” in the cumulative distribution corresponding to sentence lengths. Once these boundaries had been established, five informative sentence-length groups were obtained: very short sentences—up to and including 8 words; short sentences—from 9 to 17 words; medium-length sentences—from 18 to 30 words; long sentences—from 31 to 46 words; and very long sentences—47 words or more (see Table 2). Table 2 Distribution of sentence-length groups (percentages) Sentence groups Very short, 1—8 words Short, 9—17 words Medium-length, 18—30 words Long, 31—46 words Very long, 47 words or more Style Scientific 10,4 42,8 38,3 7,8 0,7 Literary 43,0 36,2 15,8 3,8 1,2 It can be seen that the frequencies of sentence groups in scientific and literary styles differ by one “phase,” i.e., in scientific style short and medium-length sentences are most frequent, whereas in literary style very short and short sentences predominate; sentences in these groups account for about 80 percent in each style. The data show that sentences in all groups can be considered distinctive features of scientific and literary style (the differ- 171 ences are statistically significant at the level p < 0,001). The distinctive power of very short sentences is particularly pronounced. When this group is divided into subgroups, it becomes clear that sentences 1–2 and 3–5 words long are five times more frequent in literary-style texts than in scientific texts. This fact could be explained as follows. When very short sentences are used in a scientific text (it is assumed that such a text is more accessible to the addressee), the problem of redundancy arises—additional connectors are needed to link independent sentences, and in such a case they may account for up to 1/3 of the text (Schlesinger 1968: 73). Thus, the main characteristic of scientific style—its pronounced logicality—explains the scarcity of very short sentences. On the other hand, literary-style texts surpass scientific texts in the frequency of very long sentences (and in the range of variation—by a great deal): an overly complex sentence structure in a scientific text may impede the comprehension of abstract scientific information. The characteristics of literary style—imagery, vivid concreteness, artistic precision, and leaving things unsaid (Župerka 1983)—make it possible to organize a text on the basis of associations; connectors linking sentences are not always necessary, and implicit means of linking sentences predominate. The zone of overlap between the two functional styles consists of subgroups of sentences containing 12–17 words, which are almost equally frequent in both styles (14.8 percent in scientific style and 16.3 percent in literary style). Thus, such sentences may be regarded as an invariant feature of both styles, a “linguistic constant” (Долежел 1964: 24). It may be assumed that sentences in this range are optimal for comprehension and memorization: adults remember sentences 15–17 words long best, while the threshold for optimal comprehension is 20 words (Ogburn 1963: 101). Of course, these limits should not be treated as absolute, since they depend on many extralinguistic factors—the nature of the information conveyed, text type, characteristics of the intended audience, and so on. Sakinio ilgis laikomas sakinio struktūros funkcija (Лесскис 1962: 83), todėl būtina atsižvelgti į vientisinių ir sudėtinių sakinių pasiskirstymą kalbamuosiuose stiliuose bei į jų ilgį. In scientific style, complex sentences predominate (55.9 percent), while they are less frequent in literary texts (46.5 percent); the difference is statistically significant (p < 0.001). The predominance of complex sentences in scientific style is stable and depends little on extralinguistic factors (discipline, substyle, author). It is well known that scientific style is characterized by stereotypicality; the author of a scientific text must accept that their publication will resemble another work on the same subject. A distinctive syntactic manner is merely a possibility whose realization depends on the individual characteristics of the author of a scientific work as a language user (Винокур 1980: 122), not on their qualifications as a specialist. There is no such stability in literary-style texts; much depends on the author: the frequency of simple sentences ranges from 26 percent (Mušinskas) to 70 percent (Bubnys, Vilimaitė). This variation may depend on which prose syntactic type—syntagmatic (hierarchical) or fragmented (actualized)—predominates in a particular literary text. Sentence length increases in syntagmatic prose, when “the individual sentence is integrated into a single complex syntagmatic cycle” (Арутюнова 1972: 163). In fragmented prose texts, hierarchical relations are broken up, and each item of information forms the basis of an individual sentence. For example: 172 Awkwardly, they felt their way over the door handles and thresholds of house porches blackened by the rains, over ploughs and harrows left in the fields, and, having slipped through the treetops to where Kristupas and Kazimieras were digging the boat’s bottom with oakum, they licked the blackened alder leaves covering the ground; it seemed they would soon be coated with frost, and when these men went across the meadows with their nets, they would be dazzling white with hoarfrost, and the earth, locked fast by the hard frost, would thud beneath their feet (Radz PV 99). And it was night. In the house and in the yard. Night everywhere. Dark, deep, long, silent. The cover of the burning lamp beneath the ceiling crackled. Dimmed. So it would use less kerosene (Balt SJ 250). Liudas Skirmonis smiled. Bareheaded, his lips embellished by the chill. For the house. For passers-by. And finally, for that peeling brick wall (Avyž Ch S 8). The styles differ even more markedly in the length of simple and complex sentences (Table 3). Table 3 Distributions of length groups of simple and complex sentences (percentages) Sentence type | Style | Sentence groups: very short (1—8 words), short (9—17 words), medium-length (18—30 words), long (31—46 words), very long (47 or more words). Simple: Scientific 20.4; 46.9; 28.8; 3.6; 0.3. Literary 68.5; 26.9; 4.2; 0.3; 0.0. Complex: Scientific 2.4; 29.3; 50.0; 17.4; 0.9. Literary 13.7; 46.9; 29.1; 7.7; 2.6. Perhaps the most obvious feature of scientific style shown by this table is the complexity of simple sentences. Although their percentage is lower than in literary style, they are considerably longer than those in literary texts (x̄m = 14,5; x̄g = 7,4; sm = 7,6; sg = 5,1). In other words, scientific texts contain many structurally complex simple sentences, expanded by series of homogeneous sentence constituents, participial word groups, parenthetical elements, and insertions. In their capacity to convey comprehensive and factually precise information, such sentences are comparable to complex sentences, which explains their frequency in scientific texts. The simplicity of the structure of simple sentences in literary texts (of course, in comparison with scientific style) is also confirmed by the fact that sentences consisting of 1—2 words account for 5.3 percent in literary texts, but only 0.3 percent in scientific texts. The picture is different for the functioning of complex sentences. Although their mean length differs statistically (x̄m = 21,2; x̄g = 17,7; p< 0,001), this difference has no 173 qualitative significance. This becomes clear when we classify complex sentences in greater detail (Table 4). Table 4 Indicators of the distribution of complex sentence lengths Sentence type Scientific style Literary style % x̄ S % x̄ S Coordinating 11,8 20,6 7,7 15,9 14,8 7,3 Subordinating 68,2 19,7 8,1 40,5 15,5 8,0 Asyndetic 6,3 22,3 8,2 18,1 13,7 7,4 Mixed 13,7 28,5 7,8 25,5 26,3 14,5 It can be seen that complex sentences in scientific style are, on average, longer than those in literary texts; the roughly equal standard deviations indicate that this difference is invariant. The only exception is mixed sentences, whose distribution in literary texts is characterized by a larger standard deviation and, consequently, a much greater range of variation in length. Among mixed complex sentences in literary style, 1.3 percent are very short (up to 8 words) and 8.7 percent are very long (more than 47 words). For comparison, it should be noted that mixed complex sentences in scientific texts are more standardized: no very short sentences were recorded, while very long sentences account for only 0.7 percent of their total number. Skiriasi ir įvairių funkcinių stilių sudėtiniai bejungtukiai sakiniai: grožiniuose tekstuose jų daugiau negu moksliniuose, jie daug trumpesni (P < 0,001). Šią tendenciją sąlygoja svarbi grožinio stiliaus ypatybė — dinamiškumas (priešinga mokslinio stiliaus savybė — statiškumas). Palyginti trumpi bejungtukiai sakiniai šiame stiliuje padeda perteikti įvykių kaitą, jų ritmą, sukurti įtampą, t.y. perima papildomą teksto semantinį krūvį. Moksliniuose tekstuose siekiama pabrėžti loginius ryšius, juos tiksliau nusakyti, todėl pirmenybė teikiama prijungiamiesiems sakiniams; ši tendencija būdinga visiems moksliniams tekstams, nepriklauso nuo mokslo šakos bei postilio. Prijungiamieji sakiniai, kurie nusako ”grynus” loginius santykius (tikslo, sąlygos, priežasties, nuolaidos ir kt.), laikytini relevantišku mokslinio stiliaus požymiu. Aprašant tikrovės reiškinius, sinchroniškai reiškiant mintį, samprotaujant vartojami pažymimieji ir aiškinamieji sakiniai (tokių prijungiamųjų sakinių moksliniame tekste irgi gausu). The possibilities for expanding a complex sentence, which depend on its “average grammatical potential” (Ярцева 1969: 174), are indicated by the number of clauses (Table 5). 174 Distribution of complex sentences by number of clauses 5 lentelė Style Number of clauses 2 3 4 5–6 7–8 9 and more. Scientific 66,0 26,0 6,0 1,8 0,2 0,0 Literary 57,6 24,2 9,7 6,4 1,3 0,8 Thus, although the average length of complex sentences in literary texts is shorter, the number of their clauses is greater. In both styles, two-clause sentences are the most common, but their frequency is higher in scientific texts (the difference is statistically significant, p < 0.001). Scientific texts contain very few sentences (none were recorded in this study) with more than 8 clauses (although even such a sentence is not easy to understand!). By contrast, literary texts include sentences with up to 15 clauses. This tendency is confirmed by the sentence complexity coefficient (the average number of clauses): in literary style it is 2.8, and in scientific style, 2.2. There is also another difference: in literary style, the length of clauses in complex sentences (x̄g = 6.3, sg = 4.3) differs little from the length of simple sentences, whereas in scientific style simple sentences are much longer than clauses in complex sentences (x̄m = 9.2, sm = 6.5). In scientific style, simple sentences are often equivalent to complex ones and assume their functions; linguists who have studied the scientific styles of other languages have drawn attention to this tendency. Literary style is different. The small difference between the length of simple sentences and that of clauses in complex sentences, together with the frequency of asyndetic and compound sentences, indicates greater independence of the clauses; the length of sentences in literary style depends above all on the number of clauses. For example, the clauses in the complex sentences of texts T1 (Aputis) and T2 (Bubnys) are almost equal in length, but the complexity coefficients of these sentences differ considerably—3.8 and 2.0. A composite measure of sentence length can be calculated by taking the length of a clause in a complex sentence as the unit (Акимова 1973: 74). According to this measure, the lengths of simple and complex sentences in literary texts have the ratio 1: 1.2: 2.8, while in scientific texts the ratio is 1: 1.6: 2.3. These measures once again confirm the phenomenon discussed above: in literary texts, independent simple sentences and clauses in complex sentences exhibit the same tendencies of expansion, whereas in scientific style simple and complex sentences share common tendencies of expansion. LITERATŪRA 1. Arader H. Sentence and Word Length in Scientific Writing. — Essays on Thinking and Writing in Science, Engineering and Business. Iowa, 1963, p. 232—238. 2. Egger's H. Zur Syntax der deutschen Sprache der Gegenwart. — Studium Generale, H. 1. 1962. p. 49—59. 175 3. Beneš E. Syntaktische Besonderheiten der deutschen wissenschaftlichen Fachsprache. — Deutsch als Fremdsprache. H. 3, 1966, S. 26—36. 4. Herdan G. Type-token mathematics. — Mouton and s’Gravenhage, 1960. 5. Yule G. U. On a Sentence Length as a Statistical Characteristic of Style in Prose. — Biometrica, 1939, N. 30. 6. Kraus J. Úvod do stylistiky pro informační pracovníky. Praha, 1977. 7. Meier H. Deutsche Sprachstatistik. Bd. 1, Hildesheim, 1964. 8. Ogburn W. F. On Scientific Writing. — In.: Essays on Thinking and Writing in Science, Engineering and Business. Iowa, 1963, p. 92—105. 9. Pikčilingis J. Lietuvių kalbos stilistika. T. 1. V., 1971. 10. Schlesinger I. M. Sentence Structure and the Reading Process. The Hague-Paris, 1968. 11. Tėšitelová M. Využití statistických metod v gramatice. Praha, 1980. 12. Župerka K. Lietuvių kalbos stilistika. V., 1983. 13. Адмони В. Г. Развитие структуры предложения в период формирования немецкого языка. Ленинград, 1966. 14. Адмони В. Г. Пути развития грамматического строя в немецком языке. Москва, 1973. 15. Акимова Г. Н. Размер предложения как фактор стилистики и грамматики. — Вопр. языкознания, 1973, № 1, с. 67—79. 16. Арутюнова И. Д. О синтаксических типах художественной прозы. Общее и романское языкознание. Москва, 1972. 17. Вашак П. Длина слова и длина предложения в текстах одного автора. — Вопросы статистической стилистики. Киев, 1974, с. 314—329. 18. Винокур Т. Г. Закономерности стилистического использования языковых единиц. Москва, 1980. 19. Долежел Л. Вероятностный подход к теории художественного стиля. — Вопр. языкознания, 1964, № 2, 19—29. 20. Лесскис Г. А. О размерах предложения в русской научной прозе 60-х годов XIX века. — Вопр. языкознания, 1962, № 2, с. 78—95. 21. Мисик Й. Математические методы в стилистике. — Вопр. языкознания, 1967, № 3, с. 42—52. 22. Статистичні параметри стилі в. Київ, 1967. 23. Ярцева В. Н. Пределы развертывания синтаксических структур в связи с объемом информации. — В кн.: Инвариантные синтаксические значения и структура предложения. Москва, 1969, с. 163—177. ŠALTINIAI MOKSLINĖ LITERATŪRA 1. Basal ykas A. Lietuvos TSR kraštovaizdis. V., 1977. 2. Bernatonis J. Mikroelementų įtaka rūgščių pieno produktų gamybai ir jų savybės. V., 1980. 3. Dagys J. Augalų ekologija. V., 1980. 4. Dav idavičius E. Galvanotechnika ir jos panaudojimas. V., 1979. 5. Dundulis B. Lietuva Napoleono agresijos metais (1807—1812). V., 1981. 6. Fiziologija. V., 1978. 7. Gečiauskas S., Klimka L. Naujieji puslaidininkiniai mikrobangų prietaisai. V., 1977. 8. Gelžbetoninės konstrukcijos. V., 1978. 9. Girdenis A. Fonologija. V., 1981. 10. Girdzijauskas J. Lietuvių eilėdara. V., 1979. 11. Gublinas A., Norvila s H., R aukcis R. Mašinų gamybos techninio paruošimo automatizavimas. V., 1978. 176 12. Jasinskas L., Adomas R., Raguotienė N. Organinė chemija. V., 1979. 13. Januškevičiūtė A. Klinikinė fonokardiologija. V., 1979. 14. Jovaiša L. Psichologinė diagnostika. K., 1975. 15. Kabailienė M. Taikomosios polinologijos pagrindai. V., 1979. 16. Kapustininkaitė T. Alksnynai. V., 1983. 17. Kareckas L. Kombinuotieji pašarai. V., 1981. 18. Kubilius J. Tikimybių teorija ir matematinė statistika. V., 1980. 19. Kumpikas L. Mašinų gamybos technologijos pagrindai. V., 1978. 20. Lozuraitis A. Tiesa ir vertybė. V., 1980. 21. Mačionis A. Stuburiniai gyvūnai. V., 1981. 22. Mickūnas S. Elektrodinamikos pagrindai. V., 1975. 23. Midvikas D. Polių darbai. V., 1978. 24. Mokslinė-techninė pažanga ir gamybos efektyvumas. V., 1983. 25. Nainys V. Virpesiai ir triukšmas pramonėje. V., 1972. 26. Pedagogika. V., 1981. 27. Psichologijos paskaitos. V., 1980. 28. Raugale A., Mitūzas H., Vaičiuvėnas V. Vaikų ligų gydymas antimikrobiniais preparatais. V., 1983. 29. Sauka D. Lietuvių tautosaka. V., 1982. 30. Svogūnų auginimas. V., 1980. 31. Šeštokas V. Keliai, transportas, miestai. V., 1976. 32. Šimakauskas S. Elektra automobilyje ir motocikle. V., 1979. 33. Urbutis V. Žodžių darybos teorija. V., 1978. 34. Vasiliauskas R. Mokinių darbštumo ugdymas. K., 1983. GROŽINĖ LITERATŪRA 1. Aputis J. Gegužė ant nulūžusio beržo. V., 1986. 2. Aputis J. Skrudėlynas Prūsi joje. V., 1989. 3. Aputis J. Keleivio novelės. V., 1985. 4. Avyžius J. Žmogus lieka žmogumi. V., 1975. 5. Avyžius J. Sodybų tuštėjimo metas. V., 1977. 6. Avyžius J. Chameleono spalvos. V., 1979. 7. Būbnys V. Alkana žemė. Po vasaros dangum. Nesėtų rugių žydėjimas. V., 1977. 8. Būbnys V. Pilnaties valandą. V., 1980. 9. Baltušis J. Pard uotos vasaros. T. II. V., 1969. 10. Baltušis J. Su kuo valgyta druska. T. I. V., 1979. 11. Baltušis J. Sakmė apie Juzą. V., 1979. 12. Bieliauskas A. Kauno romanas. V., 1973. 13. Bieliauskas A. Ramūs laikai. V., 1983. 14. Bieliauskas A. Mes dar susitiksim, Vilma! V., 1975. 15. Dirgėla P., Dirgėla P. Žaibai gęsta rudenį. V., 1971. 16. Dirgėla P., Dirgėla P. Šalavijų kalnas. V., 1977. 17. Dirgėla P., Dirgėla P. Sermenų vynas. V., 1980. 18. Granauskas R. Gyvenimas po klevu. V., 1988. 19. Granauskas R. Baltas vainikas juodam garvežiui. 1987. 20. Granauskas R. Duonos valgytojai. V., 1975. 21. Ivanauskaitė J. Mėnulio vaikai. V., 1988. 22. Ivanauskaitė J. Kaip užsiauginti baimę. V., 1979. 23. Jacinevičius L. Apysakos. V., 1981. 24. Jacinevičius L. Arbatą penktą valandą ryto. V., 1979. 25. Jacinevičius L. Atmink pradžią. V., 1986. 26. Jasukaitytė V. Stebuklinga patvorių žolė. V., 1987. 28. Jasukaitytė V. Balandė, kuri lauk s. V., 1989. 29. Kašauskas R. Mažos mūsų nuodėmės. V., 1975. 30. Kašauskas R. Suaugusiųjų žaidimai. V., 1988. 177 31. Kašauskas R. Motociklininkai. V., 1973. 32. Klimas R. Gintė ir jos žmogus. V., 1981. 33. Klimas R. Tiktai ugnis užgydys. V., 1989. 34. Klimas R. Atostogos po Aukštaitijos pinatami. V., 1976. 35. Kondrotas T. S. Įvairių laikų istorijos. V., 1982. 26. Kondrotas T. S. Pasaulis be ribų. V., 1977. 37. Kondrotas T. S. Žalčio žvilgsnis. V., 1981. 38. Lankauskas R. Džiazo vežimas. V., 1971. 39. Lankauskas R. Dar viena diena. V., 1982. 40. Lankauskas R. Nė vienas nebuvo pagailėtas. V., 1990. 41. Martinkus V. Negęsta žvaigždė paukščio pėdoje. V., 1988. 42. Martinkus V. Neb ausi prapultim amžina. V., 1987. 43. Martinkus V. Medžioklė draustinyje. V., 1983. 44. Mikelinskas J. Rugpjūčio naktį. V., 1971. 45. Mikelinskas J. Už horizonto — laisvė. V., 1986. 46. Mikelinskas J. Kur lygūs laukai. V., 1981. 47. Mykolaitis-Putinas V. Sukilėliai. K., 1979. 48. Mikulėnaitė E. Ateities lietus. V., 1967. 49. Mikulėnaitė E. Diena, naktis, diena. V., 1976. 50. Mikulėnaitė E. Ona. V., 1981. 51. Mušinskas D. Kol išaušta rytas. V., 1979. 52. Mušinskas D. Mėlynų dilgėlių naktis. V., 1982. 53. Mušinskas D. Šviesa virš Tausado. V., 1987. 54. Paukštelis J. Raštai. T. II. V., 1982. 55. Paukštelis J. Netekėk, saulele. V., 1963. 56. Paukštelis J. Čia mūsų namai. V., 1974. 57. Petkevičius V. Ko klykia gervės V., 1973. 58. Petkevičius V. Sermukšnių lietus. V., 1988. 59. Petkevičius V. Speiguoti pavasariai. V., 1965. 60. Pocius A. Ištirpę migloje. V., 1972. 61. Pocius A. Liepos šešėlyje. V., 1984. 62. Pocius A. Per laukus iki miesto. V., 1979. 63. Radzevičius B. Priešaušrio vieškeliai. V., 1979. 65. Rimkevičius V. Girėnai. V., 1989. 66. Rimkevičius V. Vilniaus novelės. V., 1967. 67. Simonait y tė I. Aukštųjų Šimonių likimas. V., 1973. 68. Simonait y tė I. Vilius Karalius. V., 1967. 69. Simonait y tė I. O buvo taip. V., 1977. 70. Sirijos Gira V. Aleksoto tiltas. V., 1979. 71. Sirijos Gira V. Bėgimas nuo šešėlio. V., 1980. 72. Sirijos Gira V. Raudonmedžio rojus. V., 1978. 73. Sluckis M. Laiptai į dangų. V., 1978. 74. Sluckis M. Saulė vakarop. V., 1976. 75. Sluckis M. Nedidelė avarija. V., 1976. 76. Šaltenis S. Riešutų duona. V., 1972. 77. Šaltenis S. Atminimų cukrus. V., 1983 78. Šaltenis S. Duokiškis. V., 1977. 79. Šavelis R. Amžinoji šviesa. V., 1987. 80. Šavelis R. Paparčio šešėly. V., 1980. 81. Šavelis R. Dievo avinėlis. V., 1974. 82. Vienuolis A. Amžinasis smuikininkas. V., 1973. 83. Vilimaitė B. Vasaros paveikslėlis. V., 1981. 84. Vilimaitė B. Tėvo vardas. V., 1987. 85. Vilimaitė B. Rojaus obuoliukai. V., 1981. 86. Zurba A. Integral as. V., 1983. 87. Zurba A. Molžemis V., 1986. 178 88. Zurba A. Inkilas už miesto. V., 1989. SENTENCE LENGTH IN SCIENTIFIC PROSE AND BELLES-LETTRES STYLE Summary The distinctive feature of scientific prose and belles-lettres styles in Lithuanian is the avarage length of sentence and its parameters. The greatest differentiation is achieved in case of very short sentences, i.e., when their length is 1—5 words. In scientific prose style such short sentences are five times less frequent than in belles-lettres style. Sentence length in both styles depends upon the frequency of simple and complex compound sentences are more frequent than simple ones. Long simple sentences of this style make it possible to convey as full and exact information as that conveyed by complex or compound sentences. Whereas in belles-lettres style complex or compound sentences are less frequent, very shot simple sentences prevail. In scientific prose style the same tendencies of development are characteristic of both simple and compound or complex sentences; in belles-lettres style – of simple sentences and their constituent parts. 179