Project Details
SFB 732: Incremental Specification in Context
Subject Area
Humanities
Computer Science, Systems and Electrical Engineering
Computer Science, Systems and Electrical Engineering
Term
from 2006 to 2018
Project identifier
Deutsche Forschungsgemeinschaft (DFG) - Project number 22956010
Taking as a starting point the pervasive presence of ambiguity that characterises all levels of linguistic analysis, the task of the Collaborative Research Centre is to achieve a better understanding of the mechanisms that lead to "ambiguity" control/disambiguation. We regard ambiguity as the result of underspecification and hence disambiguation processes as processes of specification of an underspecified input. Indeed a wide variety of linguistic processes can be captured along these lines (speech perception and recognition, morphological, syntactic and semantic disambiguation etc.) and constitute the research agenda of the Collaborative Research Centre.
By a specification process we understand any linguistic process that can transform one linguistic representation into any one of several alternative more specific ones and which, in doing so, has to make a choice between these on the basis of the evidence taken from some relevant context. Specification processes are always then per definition situated in a particular context which provides constraints and triggering conditions, and they refer to two types of representations: those involving forms of underspecification and more specified ones.
The research programme deals with two main questions:
(1) What is the nature of the transformation ("incremental specification process") from underspecified to more specified representations?
(2) What is the role of the context in this process, what kind of information does it provide, and when does this become relevant? Concerning our understanding of the notion "context", four aspects are relevant for the investigations in this Collaborative Research Centre: linguistic versus non-linguistic contexts, local versus global contexts, dynamic versus non-dynamic contexts, and the cross-linguistic (in)stability of contexts.
The Collaborative Research Centre, with its many parallel explorations of specification in context, provides a unique opportunity for coming to a better general understanding of how exactly specification works in contexts and in particular languages. And we think that successful transfer between theoretical linguistics and computational linguistics not only gives us a better understanding of incremental specification in language, but also leads to a closer collaboration in methods and results of these two branches of linguistics. In the first six months, this newly established Collaborative Research Centre has mainly been concerned with recruiting a team of qualified researchers, getting the work in the individual project(group)s started and in November held its comprehensive opening colloquium.
By a specification process we understand any linguistic process that can transform one linguistic representation into any one of several alternative more specific ones and which, in doing so, has to make a choice between these on the basis of the evidence taken from some relevant context. Specification processes are always then per definition situated in a particular context which provides constraints and triggering conditions, and they refer to two types of representations: those involving forms of underspecification and more specified ones.
The research programme deals with two main questions:
(1) What is the nature of the transformation ("incremental specification process") from underspecified to more specified representations?
(2) What is the role of the context in this process, what kind of information does it provide, and when does this become relevant? Concerning our understanding of the notion "context", four aspects are relevant for the investigations in this Collaborative Research Centre: linguistic versus non-linguistic contexts, local versus global contexts, dynamic versus non-dynamic contexts, and the cross-linguistic (in)stability of contexts.
The Collaborative Research Centre, with its many parallel explorations of specification in context, provides a unique opportunity for coming to a better general understanding of how exactly specification works in contexts and in particular languages. And we think that successful transfer between theoretical linguistics and computational linguistics not only gives us a better understanding of incremental specification in language, but also leads to a closer collaboration in methods and results of these two branches of linguistics. In the first six months, this newly established Collaborative Research Centre has mainly been concerned with recruiting a team of qualified researchers, getting the work in the individual project(group)s started and in November held its comprehensive opening colloquium.
DFG Programme
Collaborative Research Centres
International Connection
France
Completed projects
- A01 - Incremental Specification of Focus and Givenness in a Discourse Context (Project Heads Dogil, Grzegorz ; Kamp, Hans )
- A02 - Exemplar-Based Speech Representation (Project Heads Dogil, Grzegorz ; Kuhn, Jonas ; Möbius, Bernd ; Padó, Sebastian ; Schütze, Hinrich )
- A03 - Incremental Specification in Speech (Project Heads Dogil, Grzegorz ; Wokurek, Wolfgang ; Yang, Bin )
- A04 - Phonetic Convergence in Spontaneous Speech (Project Heads Dogil, Grzegorz ; Schweitzer, Antje )
- A06 - Encoding of Information Structure in German and French (Project Head Reyle, Uwe )
- A07 - Cross-linguistic interactions in second language prosody (Project Head Zerbian, Sabine )
- A08 - Investigating the Interaction between Speech and Language Processing for Spoken Language Understanding: A Case Study for Sentiment Analysis (Project Head Vu, Ngoc Thang )
- B01 - The form and interpretation of derived nominals (Project Heads Alexiadou, Artemis ; Fischer, Silke )
- B02 - Funktionsweise und diachrone Entwicklung deverbaler Nominalisierungsverfahren im Französischen und Italienischen (Project Heads Becker, Martin Gunter ; Stein, Achim )
- B03 - Disambiguation of nominalizations in corpus data extraction: Morphologically related words (Project Heads Heid, Ulrich ; Kuhn, Jonas )
- B04 - The role of lexical information in word-formation, and the semantics of sentence and discourse (Project Heads Kamp, Hans ; Roßdeutscher, Antje Eva Regine )
- B05 - Polysemy in a Conceptual System (Project Head Stein, Achim )
- B06 - Underspecification in Voice Systems and the syntax-morphology interface (Project Heads Alexiadou, Artemis ; Hole, Daniel )
- B07 - Conversion in French and Italian between syntax and lexicon (Project Head Marzo, Daniela )
- B08 - Alternations and binding (Project Head Hole, Daniel )
- B09 - Distributional Characterization of Derivation (Project Head Padó, Sebastian )
- C01 - The Syntax of Nominal Modification and its Interaction with Nominal Structure (Project Head Alexiadou, Artemis )
- C02 - Case and Referential Context: Argument Realisation and Referential Context (Project Heads Alexiadou, Artemis ; von Heusinger, Klaus )
- C04 - Discourse particles. Incremental interpretation in context (Project Heads Ebert, Cornelia ; von Heusinger, Klaus )
- D01 - Representation of Ambiguities and their Resolution in Context (Project Head Reyle, Uwe )
- D02 - Combining Contextual Information Sources for Disambiguation in Parsing and Choice in Generation (Project Head Kuhn, Jonas )
- D04 - Modular Lexicalization of Probabilistic Context-Free Grammars: Improved Parameter Estimation and Clustering Methods for Statistical Parsing (Project Head Schmid, Helmut )
- D05 - Biased Learning for Syntactic Disambiguation (Project Head Schütze, Hinrich )
- D06 - Lexical-semantic factors in event interpretation (Project Heads Padó, Sebastian ; Schulte im Walde, Sabine )
- D07 - Contextual Sentiment Analysis (Project Head Schütze, Hinrich )
- D08 - Data-driven Dependency Analysis - Context Factors in Parsing and Anaphora Resolution (Project Head Kuhn, Jonas )
- D10 - Incrementality in Compositional Distributional Semantics (Project Head Padó, Sebastian )
- D11 - A crosslingual approach to the analysis of compound nouns (Project Heads van der Plas, Lonneke ; Schulte im Walde, Sabine )
- D12 - Sense Discrimination and Regular Meaning Shifts of German Particle Verbs (Project Head Schulte im Walde, Sabine )
- INF - Information Infrastructure (Project Heads Dogil, Grzegorz ; Kuhn, Jonas ; Padó, Sebastian )
- MGK - Integrated Research Training Group (Project Heads Schulte im Walde, Sabine ; Zerbian, Sabine )
- Z - Central Tasks (Project Heads Alexiadou, Artemis ; Kuhn, Jonas )
Applicant Institution
Universität Stuttgart
Spokespersons
Professorin Dr. Artemis Alexiadou, until 9/2015; Professor Dr. Jonas Kuhn, since 10/2015