Evolution of markemic lexis on the basis of markem research in chronological slices

This paper studies markeme composition in the texts of English writers of the latter half of the seventeenth century. The significance of the discussed problem is proved by the need to objectify and formalize semantical analysis of the text. Clear parameters that provide both the objectification of the author literary style characteristics, determination of the interlinkage between different authors and the study of the literary language formation process haven't been formulated yet. The use of mathematical methods in linguistic research allows revealing regularities all texts written in natural language are complied with. Zipfs universal laws that describe the statistic structure of a natural language text allow extracting key words from frequency word book of any text or a number of texts. Quantitative methods used to analyze literary works make it possible both to reveal regularities in any word frequency and to solve a number of other tasks critical for linguistic research. The method of markeme analysis is suggested to be used as one of the means to formalize the semantical analysis of the text. It provides a means of presenting a complete picture of literary works language markeme composition of any chronological interval or historical period. It also gives the possibility to analyze texts practically of any wordage. Besides the method of markeme analysis allows analyzing markeme specifications of individual authors or groups of authors and the influence of social and cultural processes on markeme dynamics, studying the evolution of markeme vocabulary through several chronological intervals and establishing links between different authors due to a correlation coefficient. Its value depends on the quantitative parameters of mutual markems. The links established between authors provide an opportunity to reveal the continuity in literature. The method of markeme analysis is based on the terms markeme and Index of Textual Markedness (InTeM). InTeM expresses the correspondence between the word length and its frequency. The difference between the word weight due to its frequency and the word weight due to its length represents the author's component that marks the level of significance for each individual word in the text. The texts are processed with word thematic analysis program "ProTemAl-Engl" developed in Voronezh State University. To determine the inter linkage between English writers of the latter half of the seventeenth century it is necessary to identify and analyze markeme vocabulary. Those words that are the most significant for the author are considered as markemes. They are nouns, which are the least marked part of speech, that denote the most important categories of culture, describe the inward man and his interpersonal and human relations, basic notions, etc. The potential markeme should undergo special system of filtration including grammatical, grammar and semantic, thematic and semantic, dialogical and classifying filters. This system of filters prevents including onyms and animate nouns, except words man, creature; connotative words; names of months and days of week; names of artifacts in case they do not possess connotative or symbolic meanings; words associated with a definite literary genre; addresses and classifying words. The dependence of InTeM value on the weight of word frequency restricts potential for its usein comparative markeme analysis. This is due to the fact that too often there is great difference both in number of works and their length in words written by different authors and their availability in digital form. In such cases it is appropriate to use Normalized InTeM instead of InTeM itself. Normalized InTeM is defined as the ratio of a markeme InTeM to total InTeM of all markemes of the author. The value of Normalized InTeM always belongs to the interval 0-1. The analysis of markeme composition of the studied literary works language requires the establishing of dominant and vice-dominant markemes of the chronological interval and semantic characteristic of the interval markemes. The selection of the interval markemes allows establishing general regularities that exist in the literature of the given chronological interval. Total normalized integral weight is a parameter that determines which author markemes will become the interval markeme. Its value equals the product of total normalized InTeM and the number of authors who used this markeme. Thus the interval markemes are the first 50 markemes of a general markeme list that is sorted in descending order. The computation of the correlation coefficient in each pair of authors belonging to this chronological interval is based on the value of total normalized index of textual markedness of their mutual markemes. It allows determining the similarity of authors inrelation to one another. The largest value of the correlation coefficient determines the power of maximum vector link between the authors and serves as a means to visualize basic vector links and their power for each author. This allows determining the degree of markeme similarity between the authors. Based on the number of maximum power links and their orientation two centres of attraction - large and small - were distinguished and their quantitative and semantic markeme composition was analyzed. The research results in stratification of the authors. Due to this stratification the authors get the following statuses: typical representatives of the chronological interval - nuclear writers, non-typical to the chronological interval - fringe writers, those who obtain intermediate position - medium writers.

Keywords

Authors

References