Supertag-based linguistic theories specify the extension of elementary structures through the application of linguistic principles (e.g. extended projection principle). Recent evidence from psycholinguistic experiments supports idea that syntactic relations built during processing are based on a fully connected structure. In particular, existing experimental data can be accounted for by assuming a strong connectivity hypothesis, i.e. that the supertag of each word is inserted into a single fully connected syntactic structure. By assuming the strong connectivity hypothesis and the extended projection principle, we have devised a dynamic version of Tree Adjoining Grammar (called DVTAG). DVTAG supertags exhibit an asymmetrical behavior with respect to the their possibilities for combination with other supertags on the left and the right, respectively, and one DVTAG supertag can license more than one semantic unit. After introducing the formalism, we describe the extraction of a wide-coverage DVTAG from parsed corpora, and give a detailed description of the form of the DVTAG supertags in comparison with LTAG.

Constraining the Form of Supertags with the Strong Connectivity Hypothesis

MAZZEI, Alessandro;LOMBARDO, Vincenzo;
2010-01-01

Abstract

Supertag-based linguistic theories specify the extension of elementary structures through the application of linguistic principles (e.g. extended projection principle). Recent evidence from psycholinguistic experiments supports idea that syntactic relations built during processing are based on a fully connected structure. In particular, existing experimental data can be accounted for by assuming a strong connectivity hypothesis, i.e. that the supertag of each word is inserted into a single fully connected syntactic structure. By assuming the strong connectivity hypothesis and the extended projection principle, we have devised a dynamic version of Tree Adjoining Grammar (called DVTAG). DVTAG supertags exhibit an asymmetrical behavior with respect to the their possibilities for combination with other supertags on the left and the right, respectively, and one DVTAG supertag can license more than one semantic unit. After introducing the formalism, we describe the extraction of a wide-coverage DVTAG from parsed corpora, and give a detailed description of the form of the DVTAG supertags in comparison with LTAG.
2010
Supertagging. Using Complex Lexical Descriptions in Natural Language Processing
The MIT Press
407
428
9780262013871
Formal Grammars; Computational Linguistics
Alessandro Mazzei; Vincenzo Lombardo ; Patrick Sturt
File in questo prodotto:
File Dimensione Formato  
mazzei11mit-printVersion-3.pdf

Accesso riservato

Tipo di file: POSTPRINT (VERSIONE FINALE DELL’AUTORE)
Dimensione 4.61 MB
Formato Adobe PDF
4.61 MB Adobe PDF   Visualizza/Apri   Richiedi una copia

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/2318/101228
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 0
  • ???jsp.display-item.citation.isi??? ND
social impact