Semantically-Independent but Contextually-Dependent Interpretation of Contrastive Accent

Abstract

A pair of eye-tracking experiments compared the effect of prominent pitch accent on pre-nominal intersective color adjectives and subsective size adjectives. Because subsective size adjectives are inherently contrastive, they may prompt comparison among a subset of referents regardless of the presence of prominent accent. In contrast, intersective color adjectives may require prosodic marking in order to be interpreted contrastively. Accent-driven contrast interpretation was tested within a real-world object manipulation paradigm, where participants followed pre-recorded instructions to decorate holiday trees. In both the Color and the Size experiments, a prominent accent (L + H*) on the adjective facilitated the detection of a contrastive target (e.g., Hang a red/medium star. → Next, hang a YELLOW/LARGE star.). When L + H* was infelicitously used in non-contrastive sequences (e.g., Hang a red/medium tree. → Next, hang a YELLOW/LARGE ball.), a reliable ‘garden-path’ increase in fixations to the incorrect contrastive competitor (e.g., yellow/large tree) was found in the Size but not in the Color experiment. As a result, the fixations to the correct target were visibly delayed in the Size experiment. In the Color experiment, the bias toward a contrastive interpretation was found regardless of accent type on the adjective. We argue that this was due to differences in the salience of visual contrast. While the present results confirm that L + H* on pre-nominal modifiers evokes an anticipatory contrast interpretation, they also suggest that the effect of accentual prominence is modulated by the discourse and referential context, rather than by the inherent semantics of accented words.