Speakers
Description
Dictionaries have traditionally served as more than mere repositories of words; they have aimed to sketch some of the relationships between words, including semantic, collocational, or hierarchical connections. However, the physical constraints of print media often limited their scope, restricting the depiction of these relationships to cross-references, exemplifications, and, in specialized instances, etymological connections or language comparisons (in the case of bilingual dictionaries). This left users to piece together the broader semantic network on their own. The potential for lexicographical work has seen remarkable growth with the advent of digital platforms. These platforms transform our view of the lexicon from a collection of words to a dynamic network, emphasizing the significance of various word relationships. An example of such platform is Latent Dictionary project (https://latentdictionary.com) that, despite its nascent development stage, showcases the potential of digital lexicography for mapping semantic space and lexical relationships by utilizing contextual word embeddings. This method opens up new possibilities for deeply exploring the lexicon and understanding how words relate to each other. Exploring the networks of words has led to incorporating the contextual information of lexemes (collocations, embeddings). A growing body of research also highlights how a word’s grammatical behavior can uncover unique characteristics, contributing to a more precise mapping of these networks. Central to our study is the introduction of grammatical profiles into lexicographical research. The grammatical profiles, as discussed by Janda & Lyashevksaya (2011), is the relative frequency distribution of the inflected forms of a lexeme. In our study, we examine the distribution frequencies across 14 Czech noun cases, divided into 7 singular and 7 plural forms, to understand their relative frequency.
