Potencial of corpus managers in the analysis of German language teaching terms

This article explores the potential of Sketch Engine and AntConc in extracting and analyzing German terms in the standardized language testing. This field is rapidly evolving in connection with the development of language teaching knowledge, which is reflected in terms as ways of its encoding. Term extraction is one of the main stages in terminological research and, in the area under consideration, is complicated by the consubstantial character of many designations - i.e., their formal coincidence with commonly used words. Consubstantial terms acquire specialized meanings only in specific contexts, which necessitates the analysis of their collocations during term selection and raises the issue of effective tools for this process. The research is carried out within the framework of corpus linguistics, which proposes the use of both general scientific methods (description, induction, deduction, classification, modeling, comparison) and field-specific methods (the corpus method and contextual analysis). Many modern language teaching terms are absent from large German corpora like DeReKo. Consequently, there is a need to create custom corpora, for which corpus managers are used. Corpus managers, tools in computational linguistics, optimize terminological research through their diverse functions. The functions of the corpus managers considered in this study (Sketch Engine, AntConc) may include, depending on the chosen tool: automatic extraction of single-word and multi-word terms, thesaurus creation, dispersion calculation, N-gram identification, word list generation, concordance development, and collocation comparison. To assess the potential of these corpus managers in extracting German terms for standardized testing, a corpus was compiled from 16 Goethe certificate guidelines (2001-2023), totaling 379,348 tokens. Sketch Engine and AntConc show high results in analyzing German terms for standardized testing, offering structured concordances, automatic collocation identification, and term classification according to given criteria (by morphemes, parts of speech). These tools can optimize the process of term selection when working with large data volumes by supplying data based on set criteria. Despite their significant advantages, corpus managers also exhibit errors and inaccuracies due to incorrect or incomplete data processing: automatic extraction of common words instead of terms, incomplete determination of word form, and incorrect identification of collocations and semantic parameters. This indicates the need to combine computer tools with manual methods of term extraction. The authors declare no conflicts of interests.

Keywords

Authors

References