Prosodic cue weighting in disambiguation: Case ambiguity in German

Anja Gollrad, Esther Sommerfeld, Frank K├╝gler, Potsdam University

Previous work has shown that speakers and listeners efficiently exploit prosodic information to make the meaning of syntactically ambiguous sentences explicit. However, quantifiable phonetic properties of prosody in speech production (segmental duration, pause duration and fundamental frequency (f0)) stand in a complex relationship to the percept they invoke in the auditory domain. Not all measurable prosodic differences are actually used in sentence parsing. This study investigates the prosodic cues used by speakers to disambiguate a German case ambiguity in order to examine to which degree the individual cues contribute to disambiguation in perception. In a series of perception experiments sentences were consecutively manipulated to verify whether segmental duration, pause duration or pitch was one of the cues used by listeners in assigning a syntactic structure. Our findings show that durational cues are sufficient for listeners to identify the reading speakers assigned to the structures, whereas solely f0 information does not allow listeners to disambiguate the structures.