LINGUIST List 26.4892
Tue Nov 03 2015
Diss: Computational Ling, Pragmatics, Semantics: Noortje Venhuizen: ' Projection in Discourse: A data-driven formal semantic analysis'
Editor for this issue: Ashley Parker <ashleylinguistlist.org>
Noortje Venhuizen <njvenhuizen
Projection in Discourse: A data-driven formal semantic analysis E-mail this message to a friend
Institution: Rijksuniversiteit Groningen
Program: Center for Language and Cognition
Dissertation Status: Completed
Degree Date: 2015
Author: Noortje Venhuizen
Dissertation Title: Projection in Discourse: A data-driven formal semantic analysis
Dissertation URL: http://www.rug.nl/research/portal/files/25460491/Complete_thesis.pdf
Linguistic Field(s): Computational Linguistics
This thesis presents a data-driven formal semantic analysis of projection phenomena, which include presuppositions, anaphoric expressions, and conventional implicatures (as defined by Potts, 2005). The different contributions made by these phenomena are explained in terms of the notion of information status, which describes how content relates to the unfolding discourse context. This unified analysis forms the basis for the development of a semantic formalism called Projective Discourse Representation Theory (PDRT). PDRT is an extension of traditional Discourse Representation Theory (Kamp, 1981; Kamp and Reyle, 1993), which directly implements the anaphoric theory of presuppositions (van der Sandt, 1992) by means of the introduction of projection variables. It is shown that this formalism captures the differences, as well as the similarities between the contributions made by presuppositions, anaphora and conventional implicatures. In order to verify its robustness, PDRT is implemented as part of the NLP library 'PDRT-SANDBOX', and forms the basis for the meaning representations in the Groningen Meaning Bank (GMB), a semantically annotated corpus of texts. Based on the data from the GMB, a data-driven analysis of the information status of referential expressions is presented. This analysis shows how token-based, contextual, and deep semantic features from the PDRT representations can be used to predict the information status of a referential expression. Taken together, the results presented in this thesis pave way for a more integrated formal and empirical analysis of different aspects of linguistic meaning.
Page Updated: 03-Nov-2015