ИСТИНА

Current stylometry methods are typically based on the analysis of the frequency of words or groups of symbols. These methods to be effective require a minimum size of the sample, which presents an important limitation. To mitigate this problem for texts in verse, nowadays multidimensional stylometric studies are proposed, which introduce specific verse indicators (Plecháс et al 2018, Shelja et al). Speaking in terms of Yuri Tynianov, in poetry the constructive principle is rhythm.. Under this dominant factor the other elements are organized (vocabulary, syntax, morphology, etc.). For this reason, the study of "rhythmical-syntactic formulas" (Osip Brik) in poetry seems to be very promising from the stylometric point of view. Following an intuition that is inspired by oral-formulaic theory and Brik's ideas, we developed a computational tool for the stylometric study of Greek hexameters, based on an indicator that links the lexis and the rhythm. Our program automatically scans hexameters and allows us to associate a word with the position it occupies within the line (we call this association topolexis). Our stylometric indicator consists of the relative frequency with which topolexis occur. In the presentation we will demonstrate its utility in the attribution of the texts and in the study of the poets' individual style, suggesting that the same principle may be extended in the future to other types of regular verse.