The use of computer technologies in the study of medieval narrative sources

The process of integration of technical and humanitarian disciplines, the use of quantitative methods and mathematical modeling in historical science has opened up new opportunities for textual research and the study of narrative sources. In domestic historiography, there is a significant amount of unattributed medieval works. In the fundamental work of D.S. Lykhachev’s “Textology” often slips the thought that the establishment of the author and sources of any work depends entirely on the talent and linguistic flair of the researcher. Modern methods, the use of computer technologies make it possible not only to find the authors of anonymous texts, but also to confirm these findings with a set of convincing evidence.
The Kyiv Synopsis of the 17th century is among such anonymous works. It was created within the walls of the Kiev-Pechersk Lavra. The author of the “Synopsis” is unknown and there are several historical figures who can be recognized as its authors.
When studying the problem of authorship of the Synopsis, one should consider its text as a symbolic system consisting of the same elements. The reason for the differences between different texts is the nature of the connections between these elements. Such differences between texts determine the author’s style. Now in the laboratory of computer technologies of historical studies under the leadership of Prof. V.V. Podgaetsky created a computer program that forms mathematical models of the author’s style of each of the possible authors of the “Synopsis” and the text of the “Synopsis” itself. Unlike other attributing programs, it is focused on syntactic rather than semantic content analysis. Content analysis is applicable to any text considered as a system, as an ordered set of elements. This program treats the text as a syntactic, not a semantic system. A letter is accepted as the main element of the sign system, and not any semantic unit. The statistical relationships between these elements of the sign system and the frequency of occurrence of each symbol in the text (in percent), as well as the pairwise occurrence of these symbols, are determined. The entropy of the text is calculated on the basis of the matrix of pairwise occurrence of letter symbols, i.e. a quantitative measure of its uncertainty. The individual author’s style causes differences in the level of entropy for each author. A comparison of the entropy coefficients obtained for each of the probable authors with the coefficient of the “Synopsis” itself should indicate the author of the monument.
At the moment, experiments are being conducted with the texts of famous authors. Two excerpts from the works of L.N. Gumileva, two — D.S. Lykhacheva and one – E.V. Tarle The resulting entropy coefficient is the same for texts written by the same author, and it is not affected by the genre of the work used, the presence of quotations in it, or the size of the compared passages. In texts belonging to different authors, the entropy coefficient differs sharply. Thus, the hypothesis about the connection between the level of entropy and the peculiarities of the author’s style is confirmed.
The next stage of the work will be the determination of the author of the Kyiv Synopsis, and in the future this program can be applied to other unattributed texts, including texts of the early Middle Ages, in which it is quite difficult to distinguish the features of the author’s style in the accumulation of stamps characteristic of medieval chronicles.

I.O. Tarnapolskaya, graduate student of the Department of Historiography and Source Studies of DSU

Regional and general in history: Abstracts of the international scientific conference dedicated to the 140th anniversary of the birth of D.I. Yavornytskyi and the 90th anniversary of the XIII Archaeological Congress (November 9, 1995). Dnipropetrovsk, 1995. — 328 p.