Resumen de Representation and Classification of Polyfunctional Synsemantic Words in Monolingual Dictionaries and Language Corpora: The Case of the Croatian Lexeme Dakle
The paper will discuss the central issues concerning lexicographic descriptions of synsemantic words, with special regard to those with multiple syntactic and pragmatic functions. This topic will be exemplified through a description of a representative example, the Croatian lexeme dakle (Eng. well, now; consequently; accordingly, so, then, therefore, thus). We will focus on the shortcomings of lexicographic descriptions of such words in four contemporary monolingual dictionaries of the Croatian (standard) language. We pay particular attention to the inconsistent part of speech classification in these dictionaries, as well as to the type and content of their definitions, which generally do not take into account multiple syntactic and pragmatic functions of the word. This paper will analyse the functions and the use of lexeme dakle, an analysis based on language material extracted from the Croatian web corpus hrWaC, and processed by two independent annotators. We have attained fair agreement between annotators for the first task of determining the (supra)syntactic function (Cohen’s κ is 0.4332), and poor agreement for the second task of determining the semantic-pragmatic function (Cohen’s κ is 0.2908). Ultimately, the data collected, when compared to dictionary content, can serve as a starting point for a general discussion of an adequate methodology for lexicographic description of polyfunctional synsemantic words.
