Getting implicatures wrong

Chris Cummins, University of Edinburgh, UK

Background. Implicatures, particularly quantity implicatures, have been extensively studied in experimental semantics and pragmatics. Experimental tasks often involve testing whether participants have inferred specific pragmatic enrichments from the materials provided (that “some” means “not all”, “or” means “not and”, “like” means “not love”, and so on). However, across paradigms, participants tend to exhibit variability in their responses. More specifically, although adults (and children) share awareness that it is underinformative to say “some” if “all” is the case, they do not necessarily go as far as inferring that “some” means “not all” (Katsos and Bishop 2011).

In principle, this arguably shouldn’t be happening: implicatures are intentionally communicated and thus, according to the speaker’s intention, should either be available or unavailable. Of course, in an experimental setting, the problem is artificial: there is no real communicative intention being conveyed by artificial materials. However, if the kind of variability documented in experimental tasks is manifest in real interactions, it suggests that miscommunication is widespread in the domain of implicature – even though such miscommunication may not be apparent to interactants. This also suggests that studying failure in implicature should represent a possible line of attack on the question of what determines whether implicatures are “recovered” (or indeed erroneously inferred) by a hearer.

Presentation. I plan to discuss this problem with reference to numerical expressions such as “more than n” which have been shown to give rise to pragmatic enrichments (Cummins, Sauerland and Solt 2012). In particular, I examine experimentally which aspects of context are crucial to determining the correct pragmatic enrichments, and how adept hearers are at obtaining correct interpretations when context is not entirely shared. I expect to show that there is indeed widespread miscommunication in this domain, at least to some extent, although this may not be sufficiently obvious to trigger any repair processes. I will argue that this has potential implications for the way we represent the semantic and pragmatic meaning of quantity expressions, and is coherent with recent work suggesting the presence of typicality effects in this domain (Geurts and van Tiel 2013). I will also consider whether this constitutes a reason to adopt a less crisply defined notion of implicature, and whether we can conceive of this process as somehow probabilistic, despite intuitive arguments to the contrary.

References

Cummins, C., Sauerland, U., and Solt, S. (2012). Granularity and scalar implicature in numerical expressions. Linguistics and Philosophy, 35: 135-169.

Geurts, B., and van Tiel, B. (2013). Embedded scalars. Semantics and Pragmatics, 6(9): 1-37.

Katsos, N., and Bishop, D. V. M. (2011). Pragmatic tolerance: Implications for the acquisition of informativeness and implicature. Cognition, 20: 67-81.

Principal labs

Collaborators