Dwa północnokresowizmy substratowe: cybaty i kaliwo w zasobach systemu Odkrywka (badania z wykorzystaniem modelu językowego GPT-4)
Two Substrate Lexemes of Northern Borderland Polish: cybaty and kaliwo in the Resources of the Odkrywka System (A Study Using the GPT-4 Language Model)

Summary/Abstract: The authors conducted a study of two substrate lexemes of the Northern Borderland variety of Polish: cybaty and kaliwo, diving into the vast digitised collections of the Odkrywka system in the time span 1500–1939 and experimentally using AI to filter out redundant results. The study demonstrated that the use of AI enables the in- stant detection of set meanings, which is a breakthrough in finding semantic calques, common in Borderland Polish, in electronic resources. The authors extracted 43 oc- currences of the adjective cybaty and 86 occurrences of the noun kaliwo, i.e. severaltimes more than other researchers, including numerous examples from before 1939 which were previously missing. The use of the Odkrywka system made it possible to precisely identify the sources where the examined lexemes occurred, to establish the dates of their first occurrences in each meaning, to identify several previously unat- tested meanings, to provide a broader view of the geographical extent of the units, and to establish their almost zero presence in post-war collections, following the sep- aration of the North-Eastern Borderlands from Poland. The analysis of the examples found makes it possible to call into question the hypothesis of the Lithuanian origin of the lexemes cybaty and kaliwo.

• Details
• Contents