te Rietmolen, N. & Astésano, C.
U.R.I Octogone-Lordat (E.A. 4156), Université de Toulouse, UTM, Toulouse, France
In this project, we use EEG to investigate prosodic information processing in French. As French has so-called post-lexical and non-distinctive accentuation, it is traditionally viewed as a 'language without accent'. This makes it difficult to integrate the language in models of speech processing in which metrical structures play a central role. However, if one considers Di Cristo's metrical model in which underlying lexical representations are marked with both a primary final accent (FA) and a secondary initial accent (IA), it becomes possible to envisage the use of stress in French speech comprehension. Here we present an ERP-study in which isolated trisyllabic nouns and pseudowords were presented with and without IA. ERP-results show a larger N325 to stimuli without IA, revealing the automaticity of stress extraction and a preference for stress-templates with IA. To better understand the role of accentuation in connected speech processing, an N400-study in which IA is manipulated within sentences is currently under way. Additionally, we will investigate whether and how neural oscillations track the metrical and accentual characteristics in the speech signal. Altogether, we question the belief of French as a language without accent and aim to show that accentuation plays a valuable role in speech comprehension.