THE STUDY OF OATH, WAR AND LOOT CONCEPTS IN SEYITBEK EPOS REGARDING COLLOCATION WITH COMPUTATIONAL LINGUISTICS METHODS


Güven M., Kubilay İ. A., Kubilay Y.

Native Languages and Cultures in the Modern Changing World, no.2, pp.45-61, 2022 (Peer-Reviewed Journal)

Abstract

Seyitbek Epos which was formed in late 17th and early 18th centuries, is a continuiation of Kurmanbek Epos, which is a co-text/second narrative of the book “Coodarbeşim-Seyitbek” published in Bişkek, 1971 by Orozbay Urmambetov’s Kyrghyz Sciences Academia, Til Cana Institute of Literature. The epos’ theme is the battles of Kyrghyz People with others, especially Kalmuks. Thanks to advances in information technologies, the analysis of languages with statistical and rule based computational linguistic methods has become possible. In this context, nowadays the subject of “Collocation” is studied with computational methods. The study of “Oath, War and Loot” concepts in Seyitbek Epos with computational modeling methods was undertaken because this text contains rich semantic conceptual-orientational collocation data fit to be used in software. The epos was digitized as a plaintext (.txt) file, the collocation vocabulary was processed in AntConc 3.5.9 (2020) software and the collocations in the text were studied using different statistical analyses such as Log-likelihood, MI score, T-score, and Dice coefficient. In this study where statistical significance was used, algorithms that measure how many times the words have appeared together were used. In this way, the subject of collocation was evaluated by comparing both collocation vocabulary findings detected by the researchers using linguistic methods, and those detected by natural language processing methods.