Logo ÚFAL Logo MFF

SynSemClass (formerly CzEngClass)


Contextually-based synonymy and valency of verbs in a biling

ÚFAL
SynSemClass

Contextually-based synonymy and valency of verbs in a bilingual setting

This project is supported by the grant No. GA17-07313S of the Grant Agency of the Czech Republic https://ufal.mff.cuni.cz/grants/czengclass, and it uses resources hosted by the LINDAT/CLARIN Research Infrastructure, project No. LM2015071, supported by the Ministry of Education, Youth and Sports of the Czech Republic, and of the LINDAT/CLARIAH-CZ Research Infrastructure since 2020 (project LM2018101, MEYS CZ).

The main results (on top of publications) are the SynSemClass 1.0 lexicon available from the LINDAT repository (http://hdl.handle.net/11234/1-3125) and a web-based browser of the lexicon at http://lindat.mff.cuni.cz/services/synsemclass .

PI: Zdeňka Urešová

Duration: 2017-2019

Research team

PhDr. Zdeňka Urešová, Ph.D. (70%) is in charge of the project.

Mgr. Eva Fučíková (30%) provides technical support.

Prof. PhDr. Eva Hajičová, DrSc., dr. h.c. (10%) provides expert support and professional consultations.

 

The project focuses on contextually-based synonymy and valency of verbs in a bilingual setting. The analysis of semantic ‘equivalence’ (synonymy or near synonymy) of verb senses, and their valency behavior in parallel Czech-English language resources is the core of the proposed research. Using the translational context supports more language-independent specification of properties of verb sense classes of synonyms and leads towards generalization across languages. An initial sample bilingual verb lexicon of classes representing synonym or near-synonym pairs of verbs (verb senses) based on richly annotated corpora and existing lexical resources will be created. The main contribution of the project will be a deeper insight from the bilingual perspective into the topic of verb meaning in context based on the Functional Generative Description theory, thus extending it towards appropriate description of contextually-based verb synonymy.

 

 

The project is closely related to the project A comparison of Czech and English verbal valency based on corpus material (theory and practice) https://ufal.mff.cuni.cz/czengvallex and to the Prague Czech-English Dependency Treebank (PCEDT) http://ufal.mff.cuni.cz/prague-czech-english-dependency-treebank.

 

 

  • Topic: verbal synonymy in translation (bilingual context) based on the Functional Description Theory to explore semantic ‘equivalence’ of verb senses of different verbal lexemes in a Czech-English setting
  1. focus on valency behavior and semantic roles
  2. assumption:  bilingual context (translation) enables to delimit  synonymous verbs and so specify the verb senses more precisely than monolingual text 
  • Goal: to group verbs used as synonyms in Czech and English into (cross-lingual) synonym classes
  • Novelty: use of a richly annotated bilingual corpus to get more insight into the usage of verbs (together with their arguments in translation) synonym classes + “bottom-up” approach, starting with evidence in corpus (vs. “topdown”, with predefined set of semantic classes)
  • Result: CzEngClass: lexicon of verb synonyms with valency mapped to semantic roles and linked to existing lexical resources

The project is based on the valency theory of the Functional Generative Description and on its application to a corpus, namely to the Prague Czech-English Dependency Treebank (PCEDT; http://hdl.handle.net/11858/00-097C-0000-0015-8DAF-4).

The project uses the following lexical resources: 

CzEngClass S​chema