Presentation
Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Lightning Talks
Relevance of prosodic information for spoken communication at the lexical and discourse levels: Evidence from psychometric and electrophysiological data
Poster C41 in Poster Session C, Wednesday, October 25, 10:15 am - 12:00 pm CEST, Espace Vieux-Port
Hatice Zora1, Helena Bowin2, Mattias Heldner3, Peter Hagoort1; 1Max Planck Institute for Psycholinguistics, 2Voice Emerge & Språkfokus, Sweden, 3Department of Linguistics, Stockholm University
Prosody underpins various linguistic domains ranging from semantics to discourse. Pitch accent, one of the most notable features of prosodic phenomena, not only distinguishes lexical meaning as in the Swedish words anden (with Accent 1) ‘the duck’ and anden (with Accent 2) ‘the spirit’, but also sets focus for the most critical constituent of the discourse. The present study concerns with the relevance of lexical and focal functions of pitch accent for a coherent interpretation of linguistic message in Swedish. Using psychometric and electroencephalographic (EEG) measures, we investigated how listeners judge pitch accent violations in isolation and combined. Experimental stimuli consisted of one hundred sets of short dialogues, each having three sentences: an informative background sentence (e.g., It is a water bird with broad bill), a wh-question inquiring a specific information (e.g., What did Sven see?) and a target sentence with critical information (e.g., Sven saw the duck). The pitch accent on the critical information was either congruent or incongruent in relation to its lexical (indicated by accent numbers) and focal (indicated by square brackets) functions, creating four experimental conditions: i) correct discourse (e.g., Sven saw [the duck1]), ii) lexical accent violation (e.g., Sven saw [the duck2]), iii) focal accent violation (e.g., [Sven] saw the duck1), and iv) combined violation (e.g., [Sven] saw the duck2). The dialogues were presented auditorily, and the native Swedish speakers were asked to judge the correctness of the target sentences, actively in the psychometric paradigm and passively in the EEG paradigm. Psychometric data from forty participants (21 male, 19 female; age range 22–38 years) indicated that all violations (ii-iv) interfered with the coherent interpretation of message, and were judged as incorrect by the listeners. However, there was also a statistically significant difference in perceived correctness of pitch accent violations depending on the level they were occurring on. Focal accent violations led to a lower correctness score compared to lexical accent violations, and the effect of combined violation was additive. Put differently, the listeners were more sensitive to focal accent violations than lexical accent violations. Interim results from an ongoing EEG data collection (13 native speakers; 4 male, 9 female; age range 20–35 years) are in accordance with the psychometric data, and indicated the largest N400 response to the combined violation, followed by the focal accent violation and the lexical accent violation. These findings show that an accent mispronunciation impairs understanding of the semantic content and discourse relevance of critical information, and that the language comprehension system reveals different sensitivities to lexical accent and focal accent anomalies. Accent mispronunciations at the discourse level result in higher costs for spoken communication than accent mispronunciations at the lexical level. This pattern of results provides evidence that the brain does not only extract the lexical and focal aspects indicated by pitch accent but also gives different weights of relevance to them for a functioning spoken communication.
Topic Areas: Prosody, Speech Perception