Published online by Cambridge University Press: 06 December 2018
Implicature interpretation is sensitive to many contextual factors. This experimental study investigates two:
(A) instructions to think carefully about exactly what is said
(B) access to the verbatim form of what has been said
Participants encountered (1) below, which can give rise to the contradictory relevance implicature in (2), as feedback during a decoy task:
(1) I’m not suggesting that you’re responding too slowly, but it’s important to give the first response that comes to mind.
(2) (I am suggesting that) you’re responding too slowly.
When participants were questioned post-task, (B) significantly reduced rates of agreement that the speaker of (1) had said (2), whether the verbatim form provided was written (Experiment 1) or audio (Experiment 2). (A) had no such effect. In Experiment 3, we added a final task for participants: to recall (1) verbatim. One-third had forgotten it, typically substituting the implicature (2). We argue that this memory loss can explain the lower implicature rates associated with verbatim access: verbatim access reminds forgetful participants of (1)’s compositional interpretation, and that interpretation is inconsistent with the implicature in (2). Consequently, verbatim access reduces the chances of endorsing (2), thus introducing an inherent literal meaning bias in interpreting previous conversation.
We gratefully acknowledge NSF grant BCS-1349009 to Florian Schwarz for support of this research. We are also grateful for helpful discussions with Joe Bowring, Alice Hausman, Jeff Kaplan, Mandy Simons, Rob Wilder, and the participants in Florian Schwarz’s lab seminar. We thank Meaning in Flux (2016) audience members, especially Lyn Frazier and Adele Goldberg, for their incisive questions and invaluable recommendations. Any remaining errors are our own.