Details

In the given master’s dissertation, the automatic search for German loanwords in English versions of European universities’ websites is considered. Definitions of the concepts “borrowing” and “assimilation” are given; causes of this linguistic process, signs of assimilation and stages of adaptation of borrowed vocabulary are identified; main classifications and ways of borrowing of foreign words from a donor to a recipient language are considered; historical periods of the penetration of foreign lexical units of different origin into English are determined. Methods for automatic extraction of loanwords of different languages are considered as well as “Web-as-Corpus” approach. In the practical part, a corpus containing 22465 sentences of two sections: “History” and “News” was automatically built on 11 English-version websites of European universities. It was manually labeled with two tags “no German loanword” or “German loanword”. After that, the multilingual language model BERT was finetuned on it, and the method for automatic extraction of Germanisms from texts was developed. In addition, 42 found Germanisms were analyzed in detail in the aspect of orthography, pronunciation, morphology and semantics. All loanwords were divided into three groups according to their degree of assimilation: fully assimilated; partially assimilated and unassimilated words denoting concepts peculiar to the source language.