Scalar implicatures in a signed language

This paper tests the calculation of scalar implicatures in American Sign Language (ASL) in one of the first experimental pragmatic studies in the manual/visual modality. Both native signers of ASL and native speakers of English participated in an automated Felicity Judgment Task to compare implicatures based on two traditional scales as well as “ad hoc” scales in their respective languages. Results show that native signers of ASL calculate scalar implicatures based on a prototypical scale <all, some> in ASL in the same pattern as native speakers of English, within the same experimental paradigm. There are similarly high rates of exact interpretations of numbers <three, two> in ASL as in English, despite the iconicity of the numerals in ASL. Finally, an ad hoc scale was tested showing fewer implicatures in English than on the conventionalized scales. In ASL, there was a trend toward increased implicatures on the ad hoc scale which made use of the unique ability of ASL to convey spatial information using the classifier system. Taken together, these results show that conventionalized scales in ASL have the same semantic/pragmatic scalar properties as in spoken languages, although in non-conventionalized scales the inclusion of additional information such as spatial location may affect pragmatic interpretation.