Average word length dynamics as indicator of cultural changes in society

Dynamics of average length of words in Russian and English is analysed in the article. Words belonging to the diachronic text corpus Google Books Ngram and dated back to the last two centuries are studied. It was found out that average word length slightly increased in the 19th century, and then it was growing rapidly most of the 20 century and started decreasing over the period from the end of the 20 to the beginning of the 21 century. Words which contributed mostly to increase or decrease of word average length were identified. At that, content words and functional words are analysed separately. Long content words contribute mostly to word average length of word. As it was shown, these words reflect the main tendencies of social development and thus, are used frequently. Change of frequency of personal pronouns also contributes significantly to change of average word length. The other parameters connected with average length of word were also analysed. Introduction Every language reflects a certain way of world perception, in other words presents language view of the world. All the ideas about the world around which are represented by formal word structures and set expressions form a kind of frame which is more or less shared by native speakers. Up to the present moment, there were two main approaches aimed at studying language view of the world: 1) studying of individual concepts typical for a certain language, studying linguistic and cultural stereotypes of a given language; 2) studying a language dialect on the whole and its prescientific view of the world. But creation of big diachronic text corpora and development of mathematical methods for data processing enabled absolutely new approach to studying language view of the world. The research objective is to present frequency-based approach and apply it for survey of average word length change. Some factors which influence average length of words and reflect dynamics of social development will be described in this paper. Nowadays language dynamics attracts special attention [1-11, 16]. This article gives special focus to such parameters as word length and frequency. Frequency of language units is the basis of usage-based sociolinguistic models of language evolution [12]. Word length is significantly connected with other typological parameters [21]. Word frequency and length are also fundamental parameters in studying of psychological processes of language acquisition and usage [13, 14]. The main object of sociolinguistic studies is innovation diffusion. Some models were introduced [3-7] which aim is to describe processes of linguistic structure changes. The most distinguished result in this area is S-shaped curve of innovation 1 [email protected] 2 [email protected] 3 [email protected] diffusion [10] and of word frequency effect on its change rate [16] (the less frequently a word is used, the more chances it has to be changed). The models are based on the different postulates about nature of innovations spreading. Usually postulate about independence of different words evolution is accepted [4, 12] which enables to model only one word spreading. Word form change is performed by the process of word frequency usage: frequency of its previous form decays to zero and it disappears, frequency of the new one increases from zero to a certain number. But the process of word frequency change isn`t connected only with the process of substitution of one form by another one. Word frequency can vary under the influence of sociocultural factors (and in such a way reflects these factors) no matter whether it`s used in a certain language any more or not. Processes of word frequency change under the influence of sociocultural factors are studied in this article. This refers directly to the paper devoted to quantitative analysis of cultural trends [6]. The subject for study is regularities of word average length variations and factors which cause these changes. Average word length is a cumulative parameter which reflects different processes of word frequency changes. Average word length was counted in different languages though sometimes the data don’t match. As for the English language, it makes 5.1 letters [17, 18], as for the Russian language it makes 5.28 [19]. Nevertheless, accurate quantitative analysis of dynamics of this parameter hasn`t been carried out. Obviously, average word length can not only decrease but increase in course of time, in other words a kind of wavelike process takes place. It`s known that word length of any language changes due to global processes of changing the morphological type of a language. All languages change its morphological type in the course of time (agglutinative, inflective, isolating) [15]. These change of morphological type of a language influence average word length but they happen very slowly. As a rule, it takes thousand years or so to make them visible. Typology and historical linguistics research such kind of changes [8]. As the length of words themselves hasn`t changed radically during the two last centuries, the only reason of word length change at this time can be word frequency change (including neologisms and archaisms) which in its turn is caused by pragmatic (sociocultural) and cognitive factors. Studying dynamics of average word length at relatively short periods of time (decades or several centuries) became possible after creating of big diachronic text corpora. Methods and data Impressive opportunities in this sphere of study opened after creation of digital library Google Books and means of word frequency calculation – Ngram Viewer [20]. In the survey by Michel et al. [6] it is shown how these data can be applied for analysing of cultural trends. It should be taken into account that Ngram Viewer doesn`t deal with morphological analysis that`s why data concerning not words but word forms are presented. Thus, ''word'' is further regarded as ‘word form’. Using the total set of n-grams presented on the site average word length can be counted for each year. (Using for example, MathLab.) Though the library in English comprises text dated back to 1520, great amount of texts for reliable statistic computations have appeared only since 1800. (according to the authors` recommendation [20]). Average word length can be calculated using the following formula: