Saulerand, U. 1 , Tamura, A. . 2 , Koizumi, M. 2 & Tomlinson, Jr., J. M. 1
1 ZAS Berlin
2 Tohoku University
Disjunctions such as "or" are generally thought to give rise to exclusive interpretations through conversational implicatures (Sauerland, 2004). To better understand the processing mechanisms behind disjunction, we investigate the Japanese particle "ya". Debate surrounds whether this particle is best described as a conjunction (Kuno,1973; Ohori, 2004) or as a disjunction with a conjunctive implicature (Sudo, 2014). The response profile of conjunctive inferences with ya was tested against that of logical content and scalar implicatures with mouse-tracking. In our task, participants read single Japanese sentences followed by a two-image picture and had to decide whether the sentence-picture correspondence was good or bad (forced-choice).
Our main results are: 1) "ya" differs from lexical conjunctions corroborating Sudo?s (2014) proposal. 2) The time-course of the conjunctive implicature of "ya" argues against the details of Sudo?s (2014) implementation, and instead favors an account similar to other cases of conjunctive implicatures (e.g. Meyer 2012, Singh et al. 2015). 3) Our results show that implicatures based on substring alternatives behave in processing like the upper bound of cardinals (see also Chemla and Bott 2014). We suggest therefore that both substring and cardinal alternatives be treated as inherently associated with alternatives.