Bohemica Olomucensia

Nominalization of Adjectives as an Issue in Machine Analysis of Czech Language

We focused on units with adjectival inflection, for which the primary function is expressing an attribute and its secondary function is expressing a subject. Analysis of corpora data showed that units with adjectival inflection which perform primarily the function of a subject (hajný, pokojská) are reliably identified in the dictionary. We also found that such units with adjectival inflection which serve primarily as an attribute and only secondarily as a subject are in most cases assigned two interpretations in the dictionary, however, when disambiguating, the interpretation is selected incorrectly (strážný, výchovný). This type of expression is interpreted considerably erratically in the corpora. We propose such a solution to this problem that would enable proper disambiguation of units with adjectival inflection, but would at the same time help to avoid increasing homonymy in the morphological dictionary.