Processing implied meaning through contrastive prosody

Abstract

Understanding implicature--something meant, implied, or suggested distinct from what is said--is paramount for successful human communication. Yet, it is unclear how our cognitive abilities fill in gaps of unspecified information. This study presents three distinct sets of experiments investigating how people understand implied contrasts conveyed through prosody, which includes the accentual pattern of the sentence, known as intonation. Despite long-standing claims that intonation can convey contrastive meaning, empirical evidence for how it does so--especially at the tune level--remains scarce. An example is The cookie jar WAS full…, where a strong contrastive accent on WAS is followed by a rising tone at the end. This tune suggests that the jar is no longer full. However, to what extent do people actually perceive such meaning, and if so why? Moreover, what are the cognitive processes of interpreting such cues and meanings that arise beyond the word level meaning? The Sentence Continuation studies show that adult native English listeners often infer both contrast and implicature from this example above (i.e., the jar is now empty) and that these meanings reflect an additive effect of sub-tonal elements in the tune. The Picture Naming studies reveal that the implicature is created approximately two seconds after the sentence offset. However, this latency does not reflect a simple two-step sequential processing where implicature is delayed due to the initial computation of the literal assertion. The results instead indicate that listeners initially access alternative meanings such as a full jar and an empty jar, and then select one dominant implication. Importantly, the degree of competition between multiple meanings depends on the nature and accessibility of the alternatives. The Visual World Eye-Tracking studies show that when alternatives are subtly represented only as background information, listeners initially attend to the asserted meaning before building an implicature. However, when sentences contain contradictory predicates that easily activate binary alternatives, an implied meaning is constructed from these competing alternatives. Together, the findings impact our understanding of how the mind maps information from tune, text, experiential knowledge, and discourse context to generate implicatures.