Implicatures in (non-)monotonic environments

Abstract

It has been shown that some embedded under universal and existential quantifiers can locally be interpreted as some but not all. This implicature can be reliable if supported by pragmatic context. We present experimental evidence showing that this local implicature can also be understood reliably if some is embedded under non–monotonic operators like only and exactly. Moreover, these operators reliably lead to a others none interpretation. We present a decomposition analysis which splits non-monotonic operators into a positive upward entailing and a negative downward-entailing component, where the positive component is responsible for the local but not all implicature, and the negative one for the others none interpretation.