From meaning to inference: Exploring the semantics-pragmatics interface through online processing and development

Theories of language often make a distinction between SEMANTICS (linguistically-encoded meaning) and PRAGMATICS (inferences about the speaker's communicative intentions). The boundary between these representations can be unclear and counter-intuitive. For example, theorists have argued that the semantic meaning of some encompasses the meaning of all while the intuition that some implies not all results from a pragmatic inference called a scalar implicature. This thesis explores the comprehension of these inferences as a test case for exploring semantics-pragmatics interface during processing and development. In critical trials, participants' heard commands like "Point to the girl that has some of the socks" and their eye-movements were recorded to a display in which one girl had 2 of 4 socks and another had 3 of 3 soccer balls. Critically, these utterances contained an initial period of ambiguity in which the semantics of the quantifier some was compatible with both characters. This ambiguity could be immediately resolved by a scalar implicature which would restrict some to a proper subset.;Papers 1 and 2 found that following the onset of some, adults were initially fixated on both critical characters, suggesting an initial lag between semantic and pragmatic processing. Nevertheless, adults subsequently began excluding referents compatible with all, indicating that they had calculated the scalar implicature during real-time comprehension. Finally, adults were able to quickly resolve the referent when presented with competitors that were inconsistent with the semantics of some (girl with socks vs. girl with no socks). This suggests that previous slowness were specifically linked to delays in pragmatic analysis. Paper 3 found that children hearing some were also delayed in their reference restriction. However unlike adults, children continued to fixate on both critical characters until the final disambiguating phoneme, indicating a failure to generate the implicature. Furthermore, while children quickly rejected competitors inconsistent with the semantics of some, they failed to distinguish between referents that were inconsistent with the scalar implicature. Altogether, these results support the distinction between semantics and pragmatics and demonstrate that even routine and robust pragmatic inferences only occur after initial semantic processing during comprehension and acquisition.