This bachelor's thesis examines the agreement of the semantically empty verb to be with the
subject and the subject complement in sentences following the structure noun phrase 1 – the
verb to be – noun phrase 2. A review of the relevant literature, including the theories of Jože
Toporišič, Olga Kunst-Gnamuš, and Varja Cvetko Orešnik and Janez Orešnik, is followed by
a comparison of Slovenian translations of English sentences with the same structure produced
by three artificial intelligence tools: Chat GPT, Google Gemini, and Microsoft Copilot. The
central finding of the comparison is that these tools generally translate the agreement of the
verb to be correctly, agreeing it either with the subject or with the subject complement.
|