Reference games and the nature of exhaustification
Daniel Asherov, Danny Fox, Roni Katzir
October 2021

The literature on scalar implicatures (SIs) varies in its views on the division of labor between grammar and general reasoning in the derivation of SIs. According to the grammatical approach, the SIs of a given sentence are logical entailments of particular parses of that sentence – specifically, parses with a silent exhaustivity operator, notated as Exh, a covert counterpart of ‘only’. According to a competing view, there is no need for anything like Exh in the grammar; rather, the dynamics of conversation suffice to derive SIs. The present squib compares the two views. We will consider the Exh-free approach in the context of what we will refer to as iterated rationality models (IRMs). According to this prominent approach, SIs arise from iterated steps of reasoning by discourse participants about the goals and means available to other discourse participants. Focusing on one-shot reference games (Rosenberg & Cohen 1964, Frank & Goodman 2012), we will argue that in this setting SIs can arise only in circumstances where a speaker's utterance can be parsed with Exh. This observation, which is a direct prediction of the grammatical theory, is problematic for IRMs that are able to derive SIs from non-exhaustified representations.
Format: [ pdf ]
Reference: lingbuzz/006257
(please use that when you cite this article)
Published in:
keywords: scalar implicature, exhaustivity, iterated rationality, reference games, semantics
Downloaded:80 times


[ edit this article | back to article list ]