Derived words often share certain characteristics with their base words, which leads to the idea that identical properties are inherited from the base words. These properties also cover valency. Valency inheritance has not been used to automatically build lexical resources providing information on valency, the manual annotation of which requires significant human effort. In this paper, we propose a procedure for generating valency frames of selected semantic categories of Czech nouns and adjectives exhibiting a significant level of valency inheritance, thus covering the productive and systemic core of the lexicon. Based on a semiautomatic comparison of the noun and adjectival valency frames from NomVallex and the verbal valency frames from VALLEX, rules describing valency changes in the valency frames of noun and adjectival derivatives are formulated. The conditions imposed by the rules on valency frames identify individual base lemmas in these lexicons for which direct noun and adjectival derivatives are searched in DeriNet. Based on the changes in valency determined in the rules, more than 23,000 valency frames assigned to more than 10,000 noun and adjectival derivatives were derived, achieving high accuracy. These valency frames were included in DeriVallex, a database providing a solid basis for extending current lexical resources.

We present NomVallex, a manually annotated valency lexicon of Czech nouns and adjectives. The lexicon is created in the theoretical framework of the Functional Generative Description and based on corpus data. In total, NomVallex 2.0 is comprised of 1027 lexical units contained in 570 lexemes, covering the following part-of-speech and derivational categories: deverbal and deadjectival nouns, and deverbal, denominal, deadjectival and primary adjectives. Valency properties of a lexical unit are captured in a valency frame which is modeled as a sequence of valency slots, supplemented with a list of morphemic forms. In order to make it possible to study the relationship between valency behavior of base words and their derivatives, lexical units of nouns and adjectives in NomVallex are linked to their respective base words, contained either in NomVallex itself or, in case of verbs, in a valency lexicon of Czech verbs called VALLEX. NomVallex enables a comparison of valency properties of a significant number of Czech nominals with their base words, both manually and in an automatic way; as such, we can address the theoretical question of argument inheritance, concentrating on systemic and non-systemic valency behavior.