Workshop: Computational Reasoning in Natural Language, October 10, 2015
Workshop: Concept composition and experimental semantics and pragmatics, September 3, 2013
Project meeting: March 22, 15:30, room 007 (Drift 6).
Project meeting: March 7, 13:00, room 051.
Project meeting: January 30, 15:00-16:30, room 051.
Project meeting: December 11, 13:30 – 15:00, JKH 13, room 006.
Yoad Winter: “Event Orientated Adnominals and Compositionality” (joint work with Joost Zwarts).
Project meeting: November 15, 13:30 – 15:00, JKH 13, room 006.
First, I will present some of our sample stimuli to illustrate the task, and report pre-pilot results regarding color judgments in different linguistic and visual contexts. I will then discuss the planned changes in our task design for the main experiment, along with alternative feature categories in addition to colors (e.g., straight, many). Second, I will also discuss our ongoing replication study of Hampton’s (1996) overextension effect in the combination of colors and letter shapes.
My project explores how conceptual structure affects logical interpretation. I will present ongoing experimental work on plural predicate conjunction sentences (e.g. The women are sitting and standing). For such sentences, logical interpretations may differ systematically depending on the nature of the two conjoined predicates. Specifically, I aim to illustrate that the typicality of two particular predicates applying to one person determines to what degree subjects accept a weaker reading than intersective conjunction. Moreover, I will show how this relates to results we have already obtained on typicality in relation to reciprocal sentences.
Project meeting: October 9, 13:00-15:00, room 0.26.
We introduce a new theoretic model for elucidating textual entailment based on semantic annotation. The model accounts for inferences in a restricted fragment of English – a set of constructions that give rise to restrictive, intersective and appositive modification, and supports also lexical/syntactic inferences based on textual alignment. The model is defined as a trio: an interpreted lexicon, which specifies the inventory of symbols and semantic operators supported in the model, an annotation scheme that instructs annotators the way in which words and constructions from a given pair of premise and hypothesis are to be bound to the interpreted lexicon, and a theorem prover that composes the sentential meaning of the premise and the hypothesis, and validates the entailment relation (or lack thereof). We explore the applicability of the model to RTE data and describe a first-stage annotation scheme based on which a manual annotation work was carried out. Our choice of semantic phenomena is driven by their predominance in the RTE datasets, the ability to annotate them with high consistency and the possibility to capture their various syntactic expressions by a limited set of concepts.
Project meeting: September 10, 15:00-17:00, room 0.51.
Project meeting: August 20, 15:00-17:00, Trans 8, room 0.19.
Project meeting: June 29, 14:00-16:00, Trans 6, room 0.26.
Workshop: On Logic and Common Sense, organized on June 15, 2011.
Workshop: Properties and Optionality in Syntax and Semantics, organized on February 13-14, 2012.