The assignment of grammatical gender to (non-human) nouns is generally arbitrary: sun is masculine in Romance languages but feminine in Germanic languages and neuter in Slavic languages. However, one would assume that cognate words keep the same gender within the same language family: sun is expected to be masculine in all Romance languages and neuter in all Slavic languages. While this expectation holds on average, it is not a hard rule and there are some exceptions: apple is masculine in German but neuter in Norwegian, signal is masculine in French but feminine in Spanish.
The goal of this thesis is to investigate whether the number of such exceptions (inconsistent gender assignments across cognates of two related languages) constitutes a useful measure of linguistic distance. As part of your thesis, you will:
- find and extract cognate databases for several language families,
- compute proportions of inconsistent gender assignments across various pairs of languages,
- investigate the effect of different noun types (borrowings, derived nouns, ...) on gender assignment consistency.
The main focus of the thesis will be on several subgroups of Indo-European languages. Other grammatical features (such as prepositions/cases selected by verbs) may be studied as well.