Relation between morphological and associative structure of Hungarian words

• Dániel Czégel ^[3] ; Zsolt Lengyel ^[1] ; Csaba Pléh ^[2]
  1. [1] University of Pannonia

     University of Pannonia

     Hungría

     
  2. [2] Central European University

     Central European University

     Hungría

     
  3. [3] Institute for Cross-Disciplinary Physics and Complex Systems, Palma de Mallorc

  Mostrar afiliaciones +
• Localización: Italian journal of linguistics, ISSN 1120-2726, Vol. 28, Nº. 1, 2016, págs. 7-24
• Idioma: inglés
• Texto completo no disponible (Saber más ...)
• Resumen

  • This paper aims to relate two types of presently available large scale databases corresponding to the organization of the mental lexicon, namely, text corpora and association dictionaries, from the point of view of morphological and associative variability. We used a paper-and-pencil based Hungarian verbal association dictionary containing 188 stimulus words and about 400 responses for each stimulus, collected from children (age 10-14) and young adults (age 18-24), to compute the associative entropies of the stimulus words. Associative entropies are then compared to the morphological entropy of the lexemes corresponding to stimulus words, determined from a large web-based Hungarian corpus. We found that for both age groups, more variability in morphological forms corresponds to higher associative variability in the case of nouns (r=0.20 and 0.18). For verbs, however, the opposite relation holds: morphological and associative entropies are negatively correlated (r= 0.19 in both age groups). In order to investigate this phenomenon in more depth, corpus frequency and the fraction of paradigmatic associations are also taken into consideration, showing that the relation between associative entropy and the fraction of paradigmatic associations is highly responsible for the different behaviour of nouns and verbs.