Speakers
Description
In the introductory part of the presentation, the authors will present the Croatian Web Dictionary – Mrežnik project (Hudeček & Mihaljević, 2020; Hudeček, Mihaljević & Jozić, 2024). Mrežnik consists of three modules – the module for adult native speakers of Croatian, the module for students and the module for non-native speakers learning Croatian. These modules have different approaches to semantic relations. Entries, entry words, and subentries are connected by links according to various criteria (feminine/masculine pairs, word formation, verbal aspect, semantic relations; Hudeček & Mihaljević, 2019).
This presentation will focus on recording these semantic relations in Mrežnik: synonyms, antonyms, hypernyms, hyponyms, co-hyponyms, meronyms, and holonyms (Cruse, 2006; Murphy, 2016). These relations and different approaches to them in different modules are illustrated by a number of examples from Mrežnik. Complex synonymic and antonymic relations are shown using the example of two Croatian near-synonyms, voljeti (‘to love’, but also ‘to like’), ljubiti (‘to love, to like’, but also ‘to kiss’), and mrziti (‘to hate’). Although native Croatian speakers would mostly say that voljeti and ljubiti are synonyms and mrziti is their antonym, after corpus analysis with word sketches and word sketch differences in SketchEngine, a complex network of synonymic and antonymic relations emerges, as all these words have several closely related meanings and establish the relations of synonymy and antonymy only in some meanings. In some meanings, they are also connected to other verbs. The approach to hyperonymy, hyponymy, and co-hyponymy in Mrežnik is illustrated by entries of terms that denote case (Nominative, Genitive, Dative,
Accusative, Vocative, Locative, Instrumental). The hyperonym is the term case, which occurs in all seven case definitions. The definitions are similar in structure and differ only in the element of number (e.g., first case, second case, etc.) and the questions asked for each case. Each case has six other cases as co-hyponyms. These co-hyponyms are connected by links. The approach to entries denoting case is compared with the approach in the terminological database Struna (Nahod, 2023, pp. 3–10) and the Dictionary of Croatian Linguistic Terminology (Mihaljević & Hudeček, 2024).
Although meronyms and holonyms are regularly mentioned in works on semantics, lexicology, and lexicography, most dictionaries do not record meronymy-holonymy relations. This is not surprising, as “meronymy is generally not as central as other -onymic relations, since it is not a logical relation” (Murphy, 2016, p. 446). At the beginning of the work on Mrežnik, the linking of meronyms and holonyms was considered. This will be illustrated using the example of days of the week (tjedan: ponedjeljak, utorak, srijeda…) and sentence elements (rečenica: pprox., predikat, pprox…). However, the reasons for moving away from this idea and using a different approach to meronyms will be explained. The meronymy-holonymy relation is reflected in the collocation block (introduced by
the collocational question Što x ima? ‘What does x have?’), but sometimes also in definitions beginning with the word dio (‘part’) or član (‘member’). However, the authors will show that in some cases it was difficult to show this relation consistently for all parts and sometimes even to distinguish it from the hypernymhyponym relation.
The investigation of semantic relations in Mrežnik has raised several questions: Are semantic relations important to the users and if so, why? Are there semantic relations between functional word classes (syntactic synonyms)? Is synonymy only possible between members of the same word class? These questions will be answered in the presentation. The advantages of a born-digital dictionary in representing semantic relations as compared to printed dictionaries will be highlighted and the benefits of this approach to semantic relations for the dictionary user and the dictionary compiler will be explained. The linking of synonyms makes it possible to insert a larger number of words. It also encourages the compiler to consider whether the same or a modified definition should be used. The same applies to antonyms. If two words are antonyms, they usually have a similar definition and differ only in a single element. All co-hyponyms have the same hypernym. The advantages for language users are that they can become aware of words that belong to the same semantic field, expand their vocabulary, become aware of the lexical system, and systematize their linguistic knowledge. Some results and solutions used in Mrežnik were presented to the target groups of users (especially non-native speakers learning Croatian, Hudeček, Mihaljević & Pasini, 2024) and their suggestions were taken into account. Further research with non-native speakers of Croatian is planned. The semantic network of relations and its lexicographic treatment in Mrežnik is compared with the semantic relations in Croatian Linguistic Terminology – Jena (Mihaljević, Hudeček & Jozić, 2023), a terminological database created simultaneously in the same institution and partly by the same authors. The research of semantic relations in Mrežnik, already applied to the Croatian Linguistic Terminology – Jena project, could be a model for future dictionaries of (South) Slavic languages.
