Left-periphery effects in Spanish wh-questions
Professor Nicholas Henriksen, a new faculty member in Romance Languages and Literatures, and a welcome new addition to the phonetics/phonology community at the University of Michigan, will talk about his research in Phondi this week. Nick is a specialist in Spanish sociolinguistics, with a focus on prosody and the syntax/phonology interface. His presentation will address some of these issues. The title of his talk, as well as an abstract, is given below.
Left-periphery effects in wh-questions: Supporting the pitch accent analysis for Spanish
Within intonational analysis it is known that initial peaks are particularly susceptible to late alignment when in contact with the utterance-initial boundary (Silverman & Pierrehumbert, 1990). A confounding issue across languages is that multiple options may be available for seemingly identical utterance-initial rises (i.e., phrasal peak or accentual peak) (Frota, 2003; Haan, 2001). In this presentation I argue that a single analysis is possible for left-periphery F0 rises in Spanish wh-questions. Specifically, a single L*+H analysis accounts for a variety of phonetic implementation effects.
For this experiment, twenty speakers of Peninsular Spanish produced 72 wh-questions and 30 declarative statements each in a reading task. The wh-questions were produced under three manipulation conditions. Acoustic analysis was conducted in Praat (Boersma & Weenink, 2011). In utterances comprised of two unstressed initial syllables before the
wh-word, the rise magnitude was greater (i.e., lower F0 height for the L point) than in utterances initiated by the wh-word alone, indicating truncated L in the latter condition. The H peak did not align relative to the initial boundary, precluding an accentual peak analysis.
Furthermore, within-speaker comparisons for wh-question and statement peaks did not reveal F0 height differences, precluding an upstep analysis (cf. Prieto, 2004). As for AM notation, stricter alignment of the L tone with the wh-word indicates L* of bitonal L*+H with truncation when the wh-word is adjacent to the left-periphery. To conclude the talk, we will discuss potential drawbacks to the L*+H proposal and consider the applicability of non-alignment approaches to intonational phonology.