Retour vers liste

Détail de la contribution

Auteur: Joseph TYLER

Rising Pitch as Incompleteness, with Discourse Structuring Effects

Abstract/Résumé: Rising pitch has been analyzed for its ability to create questioning force on declarative syntax (Gunlogson 2008), as part of listing intonation (Ladd 2008), as well as indicating discourse relationships (Pierrehumbert & Hirschberg 1990, Jasinskaya 2010, Nilsenova 2006). Sometimes it is discussed simply as a final rise, like in the above work; other times as part of a rise-fall-rise contour, sometimes called contrastive pitch accent (Büring 2003, Constant 2012). Recent experimental results in Tyler (2012) show that a rise can bias the interpretation of ambiguous discourses like (1): “I sat in on a history class [S1]. I read about housing prices [S2]. And I watched a cool documentary [S3].” This discourse could be interpreted such that the narrator read about housing prices and watched a cool documentary in history class (Subord interpretation) or separate from history class (Coord interpretation). A rise at the end of sentence 1 was found to bias listeners towards the Coord interpretation (t=2.743, p=.006). Participants were also more confident when their interpretation matched the prosody (Coord with rise/Subord with fall) (F=11.657, p=.001). This biasing effect, like rises that create questioning force on declarative syntax, signals an incomplete contribution to the relevant parts of the discourse. A question introduces a question under discussion which needs an answer. A coordinating relation suggests partial answerhood to some dominating question under discussion. To flesh this proposal out more, I will convert the coordination/subordination ambiguity in (1) to a QUD structure (Roberts 1996; Büring 2003). In the Coord interpretation, S1-S3 of (1) are partial answers to the dominant QUD “What did you do yesterday?”. In the Subord interpretation, S2-S3 answer a subquestion “what did you do in the history class?” The rise communicates local incompleteness at the current hierarchical level, while a fall allows a shift. The non-categorical effect in Tyler (2012) can be due to differences in congruence, where incongruent discourses are “defective in (intonational) form” (Büring 2003). For Büring, incongruence is probabilistic, not categorical. Evidence of this incongruence can be seen in the higher confidence participants showed in congruent interpretations (Tyler 2012). This account is compared to Nilsenova’s (2006) discussion of rises as intonational adverbs of uncertainty and Jasinskaya’s (2010) discussion of intonation as a trigger away from default discourse relations.