SpeakerMarketa Lopatkova
DateOct 21, 2011
Time11:00AM 12:30PM
TitleSemantic Similarity and Syntactic Variability: Valency Lexicon of Czech Verbs.

Information on syntactic and semantic properties of verbs, which are traditionally considered to be the center of a sentence, plays a key role in many rule-based NLP tasks such as automated semantic role labeling, semantic parsing, machine translation, etc.

The talk will focus on a lexicographic description of syntactic-semantic features of verbs. I will present the valency lexicon of Czech verbs VALLEX (Lopatková et al., 2007), which is closely related to the Prague Dependency Treebank project (Hajiè et al., 2006). VALLEX provides information on the combinatorial potential of verbs in their individual senses: for each verb sense, the lexicon lists a number of its syntactic-semantic complementations labeled with their semantic roles and their possible morphological forms.

Prototypically, a single syntactic structure corresponds to a single meaning of verb. However, in many cases semantically related uses of verbs can be syntactically structured in different ways, as in the notoriously known sentences with the verb `to load’:
      Peter loaded the truck with hay. – Peter loaded hay on the truck.
      The truck was loaded with hay. – Hay was loaded on the truck.
Such changes in syntactic structure are usually referred to as `alternations’, see esp. (Levin, 1993). In the second part of my talk, I will address two basic types of alternations for Czech verbs and I will present a method of their representation in a valency lexicon (Kettnerová, Lopatková, 2011).

Key words: syntactic structure of verbs, semantic labels, alternation, lexicon

Levin, B. (1993) English Verb Classes and Alternations: A Preliminary Investigation. University of Chicago Press, Chicago, IL
Lopatková, M., et al. (2007) VALLEX 2.5 - Valency Lexicon of Czech Verbs, version 2.5., ÚFAL MFF UK (http://ufal.mff.cuni.cz/vallex/2.5)
Hajiè, J., et al. (2006) Prague Dependency Treebank 2.0. LDC, Philadelphia, USA,
Kettnerová, V., Lopatková, M. (2011) In Slovko 2011 (accepted) Slovenská akadémia vied, Bratislava, Slovakia


Markéta Lopatková is an Associate Professor in the Institute of Formal and Applied Linguistics, Faculty of Mathematics and Physics, Charles University in Prague.
Her research interests include dependency syntax, treebanking, tectogrammatical representation, lexical semantics, and formal modeling of natural languages. Her recent research is aimed at building valency lexicon of Czech verbs VALLEX and its further enhancing with semantic classification. She has also conducted extensive studies of changes in syntactic structure of Czech verbs and their lexicographic description. She has been also actively involved in the project of syntactic analysis of Czech complex sentences.
She is a local coordinator of the CLARA Marie Curie Initial Training Network, the project supporting PhD students and their training activities. She also actively participates in the LCT study programme (European Masters Program in Language and Communication Technologies).
She is involved in teaching, she gives a practical course on mathematical analysis (bachelor level) and a course on Prague Dependency Treebank (master/PhD level).

Previous Next