Organizer(s) | Anton Benz & Aarti Joshi |
Start of event | 07.04.2025, 13.30 o'clock |
End of event | 07.04.2025, 19.45 o'clock |
In recent years, Bayesian approaches that model communicative processes through the computation of probabilities by rational agents have achieved significant success in explaining language use in simple domains with limited sets of alternatives. Traditionally, linguistic semantics has relied on logical analysis to represent sentence meaning. The challenge now is to integrate these symbolic representations with probabilistic frameworks that account for context, uncertainty, and interlocutors’ goals. The new frontier is thus the extension of such probabilistic models to phenomena requiring the computation of probabilities over large sets of alternatives or broad meaning spaces, bridging the gap between purely logical approaches and context-sensitive, probabilistic analyses of language.
To explore the latest advancements in the (statistical) modelling of pragmatic meaning in contexts with large meaning spaces, we will host an online workshop on Monday, the 7th of April 2024 with 30-minute talks that are followed by a 20-minute discussion.
(the members of the LMBayes project)
The workshop topic is a key focus of the new project “LMBayes” (Linguistic Meaning and Bayesian Modelling), which is based at ZAS Berlin in collaboration with WIAS (Weierstrass Institute for Applied Analysis and Stochastics) Berlin and the University of Tübingen (https://www.leibniz-zas.de/de/forschung/forschungsbereiche/semantik-pragmatik/lmbayes).
13:30 – 13:45 Introduction
13:45 – 14:45 Jakub Szymanik (U Trento)
The Shifting Sands of Meaning: How Quantifier Interpretation Varies
14:45 – 15:45 Fausto Carcassi (ILLC U Amsterdam)
Program synthesis in probabilistic pragmatics
15:45 – 16:20 Hening Wang & Michael Franke (U Tübingen)
Rational Speech Act Model with Argumentative Goals
Break
17:00 – 18:00 Katrin Erk (UT Austin)
Composing word meaning representations: probabilistic constraints on word meaning in context
18:00 – 19:00 Daniel Lassiter (U Edinburgh)
Why we still need small-world models in computational pragmatics
19:00 – 19:35 Anton Benz (ZAS)
From pictures to quantifiers: Modelling complex production pathways