RESUMEN
Discourse understanding is hampered when missing or conflicting context information is given. In four experiments, we investigated what happens (a) when the definite determiner "the," which presupposes existence and uniqueness, does not find a unique referent in the context or (b) when the appropriate use of the indefinite determiner is violated by the presence of a unique referent (Experiment 1 and Experiment 2). To focus on the time-course of processing the uniqueness presupposition of the definite determiner, we embedded the determiner in different sentence structures and varied the context (Experiment 3 and Experiment 4). Reading time served as an index of processing difficulty in a word-by-word self-paced reading task and acceptability judgments provided hints for a possible repair of a presupposition violation. Our results showed that conflicting and missing context information lowered acceptability ratings and was associated with prolonged reading times. The pattern of results differed depending on the nature of the presupposition (Experiments 1 and 2) and whether supplementing missing context information was possible (Experiment 3 and Experiment 4). Our findings suggest that different cognitive processes come into play when interpreting presuppositions in order to get a meaningful interpretation of a discourse.
Asunto(s)
Ursidae , Animales , Humanos , Lenguaje , SemánticaRESUMEN
Discourse structure enables us to generate expectations based upon linguistic material that has already been introduced. The present magnetoencephalography (MEG) study addresses auditory perception of test sentences in which discourse coherence was manipulated by using presuppositions (PSP) that either correspond or fail to correspond to items in preceding context sentences with respect to uniqueness and existence. Context violations yielded delayed auditory M50 and enhanced auditory M200 cross-correlation responses to syllable onsets within an analysis window of 1.5s following the PSP trigger words. Furthermore, discourse incoherence yielded suppression of spectral power within an expanded alpha band ranging from 6 to 16Hz. This effect showed a bimodal temporal distribution, being significant in an early time window of 0.0-0.5s following the PSP trigger and a late interval of 2.0-2.5s. These findings indicate anticipatory top-down mechanisms interacting with various aspects of bottom-up processing during speech perception.