PRINCIPLES OF ALGORITHMIZATION OF MACHINE TRANSLATION IN THE THEORIES OF LANGUAGE EVOLUTION (BASED ON THE UKRAINIAN-ENGLISH TRANSLATION).

This paper deals with the issue of the main principles of algorithmization of machine translation in the theories of language evolution. The aim of this article is to offer the original algorithmic principles of machine translation of natural languages of the Indo-European group with the optimal mathematical dimension of the text and the accuracy acceptable for practical translation tasks based on the classic linguistic theories of A.Schleicher, I.Schmidt, etc., using the phenomenon of the weak self-similarity of Mandelbrot. One of the most disputed questions in the linguistics is the problem of machine translation in the theories of language evolution. Reference has been made to the reasons that determine the necessity of the research into the linguistic theories. Attention has been paid to the main theories of language evolution: a content-forming "family-tree" theory by A.Schleicher and J.Schmidt's theory of "waves". It is stated that the involvement of linguistic theories allows us to create algorithms that reject a large number of non-grammatical constructions and create automatic criteria for choosing the best translation variant.