Semantic Layer of the Valence Dictionary of Polish Walenty

Semantic Layer of the Valence Dictionary of Polish Walenty El˙zbieta Hajnicz, Anna Andrzejczuk, Tomasz Bartosiak Institute of Computer Science, Polish Academy of Sciences ul. Jana Kazimierza 5, 01-248 Warszawa, Poland hajnicz@ipipan.waw.pl, anna.andrzejczuk@gmail.com, tomasz.bartosiak@gmail.com Abstract This article presents the semantic layer of Walenty—a new valence dictionary of Polish predicates, with a number of novel features, as compared to other such dictionaries. The dictionary contains two layers, syntactic and semantic. The syntactic layer describes syntactic and morphosyntactic constraints predicates put on their dependants. In particular, it includes a comprehensive and powerful phraseological component. The semantic layer shows how predicates and their arguments are involved in a described situation in an utterance. These two layers are connected, representing how semantic arguments can be realised on the surface. Each syntactic schema and each semantic frame are illustrated by at least one exemplary sentence attested in linguistic reality. The semantic layer consists of semantic frames represented as lists of pairs ⟨semantic role, selectional preference⟩ and connected with PLWORDNET lexical units. Semantic roles have a two-level representation (basic roles are provided with an attribute) enabling representation of arguments in a flexible way. Selectional preferences are based on PLWORDNET structure as well. Keywords: valence, semantic frames, semantic roles, selectional preferences, Polish 1. Introduction Walenty (Walenty, 2016), a comprehensive valence dic- tionary of Polish, is developed at the Institute of Com- puter Science, Polish Academy of Sciences (ICS PAS) within several projects, but mainly within CLARIN-PL.1 The dictionary is ment to be both human- and machine- readable; in particular, it is employed by two parsers of Polish, ´Swigra2 (Woli´nski, 2004) and POLFIE3 (Patejuk and Przepiórkowski, 2012). For this reason, lexicon entries have a strictly defined formal structure. Additionally, syn- tactic and semantic phenomena represented therein should be attested in linguistic reality. Walenty is composed of two main layers: syntactic and semantic. The syntactic layer was described in (Przepiórkowski et al., 2014c; Przepiórkowski et al., 2014a), whereas Przepiórkowski et al. (2014b) focuses on its phraseological component. In this paper we want to fo- cus on the semantic layer. While extending Walenty with the semantic layer, we had semantic parsing and textual entailment in mind. There- fore, unified representation of utterances bearing similar in- formation is crucial. The dictionary comprises over 15 000 entries (composed of about 84 000 schemata), with about 12 000 verbs,4 2 000 nouns, 950 adjectives and 200 adverbs. Each entry is iden- tified by its lemma. The semantic layer of Walenty is strictly connected with PLWORDNET (Piasecki et al., 2009), one of two Polish wordnets. PLWORDNET describes the meaning of a lex- ical unit by placing this unit in a network of relations (such 1http://www.clarin-pl.eu/en/ 2http://zil.ipipan.waw.pl/´Swigra 3http://zil.ipipan.waw.pl/LFG 4Walenty verbal entries cover 99.8% verb tokens in balanced NKJP subcorpus of 300M tokens, based on frequency list of auto- matic tagging. as synonymy, hypernymy, meronymy, etc.). 2. Related works There exist valence dictionaries connecting syntactic and semantic information about predicates and their arguments. The most famous is FrameNet5 (Fillmore et al., 2003; Rup- penhofer et al., 2006) …

Content truncated to 500 words