Paper

An Incremental RSA Model for Adjective Ordering Preferences in Referential Visual Context

Authors
  • Fabian Schlotterbeck (University of Tübingen)
  • Hening Wang (University of Tübingen)

Abstract

We report data from a preference rating experiment that tested for conflicting effects of subjectivity and discriminatory strength on adjective ordering preferences in referential visual context. Results indicate that, if the communicative efficiency of an adjective is low in a given context, it is preferred later in a multi-adjective expression. To account for qualitative aspects of these data, we propose a novel computational model of incremental processing in the Rational Speech Act framework. What sets the model apart from previous approaches is that it assumes fully incremental interpretation, without the need to anticipate possible sentence completions.

Keywords: Adjective Ordering, Preference Ratings, RSA Model, Incremental Processing

How to Cite:

Schlotterbeck, F. & Wang, H., (2023) “An Incremental RSA Model for Adjective Ordering Preferences in Referential Visual Context”, Society for Computation in Linguistics 6(1), 121-132. doi: https://doi.org/10.7275/1why-1g73

Downloads:
Download PDF

187 Views

55 Downloads

Published on
01 Jun 2023