Processing strategies for scalar implicatures

Dimka Atanassov, University of Pennsylvania


This dissertation investigates the processing of scalar implicatures. I report results from a series of studies using different scalar expressions (modals and quantifiers), which show that (a) scalar implicatures are generally accessed rapidly; and (b) access to scalar implicatures is delayed in special cases (namely when priming with another implicature occurs). Such variability in processing times mirrors the existing state of affairs in the scalar implicature processing literature, and merits a novel perspective on the processing of scalar implicatures. In the central experiment, an eye tracking study in the visual world paradigm, I investigate the effect of priming scalar implicatures triggered by quantifiers (“some”) with scalar implicatures triggered by epistemic modals (“might”). I compare the time course of scalar implicatures triggered by “some” when priming occurs, to their time course when no priming occurs. The results show that scalar implicatures triggered by “might” are rapidly accessed, and those triggered by quantifiers are generally processed rapidly as well, except when they are preceded by a modal-based implicature. This is interpreted as an interference effect. I argue that the existence of this effect is evidence that scalar implicature processing is similar across different types of triggers, and can furthermore be interpreted as evidence that at least part of scalar implicature processing depends on a pragmatic process. The variability in processing times replicates the existing state of affairs in the literature within a single study (where some studies show evidence for delayed access to scalar implicatures while other studies show evidence for rapid access). Such variability is difficult to explain within existing implicature processing frameworks, and therefore merits a novel perspective on scalar implicature processing. One such perspective is proposed. In this perspective participants select between two different processing strategies (one rapid, learned through repeated exposure, and one slow, involving a complete Gricean computation). In an adult learning study I furthermore investigate the effect that familiarity with the scale has on scalar implicature computation.

Subject Area


Recommended Citation

Atanassov, Dimka, "Processing strategies for scalar implicatures" (2014). Dissertations available from ProQuest. AAI3635466.