Vector representations of vague expressions in the Slovak language

2023 – 2026

Principal Investigator: Jana Wachtarczyková

Co-Principal Investigator: Martin Ološtiak

Researchers: Radovan Garabík, Monika Kapustová, Anna Ramšáková, Júlia Behýlová

Grant scheme and Evidence number: VEGA 2/0107/23

Annotation: The project focus is to capture various meaning-related relationships in the semantic structure of vague expressions present in the contemporary political, economic and cultural and social discourse. The goal is to identify semantic contexts in which vague words are presented and the pragmatic and commonsensical aspects of their meaning. This is a linguistic interpretation of results of word embedding: a method using the vast linguistic material in the corpus via a web interface ‘Semä’ to automatically generate a vector model quantifying the semantic similarity relationships between words in a vector-space model. The project will analyse the understanding and functioning of vague expressions in Slovak in detail through relation profiles. The overview of paradigmatic and syntagmatic relations in which a lexeme occurs will constitute the core of its profile. The vector-model approach is a new and so-far unexplored way to learn about the common-sense use conceptualization of ambiguously understood and confusedly used words.