Fall Research Expo 2020

Using Prosody to Resolve Syntactic Ambiguity in Speech Production

Prosody is ”the linguistic structure expressed in the suprasegmental properties of utterances” (Cutler, 2012). Suprasegmental properties include pitch, tempo, loudness and timing patterns. Suprasegmental cues are also referred to as prosodic cues. Prosodic cues are crucial to parse sentences and resolve syntactic ambiguity. There is limited literature on prosody research because prosody is poorly integrated into speech processing models. In our study, we asked participants to mimic human conversation while going through PowerPoint presentation via BlueJeans. Audacity was used to record the conversations and PRAAT was used to annotate and analyze the recordings. In this study, we focused on comparing duration in early and late juncture sentence pairings. We completed a qualitative analysis of the data.

PURM - Penn Undergraduate Research Mentoring Program
College of Arts & Sciences 2023
Advised By
John Trueswell
Join Michelle for a virtual discussion
PURM - Penn Undergraduate Research Mentoring Program
College of Arts & Sciences 2023
Advised By
John Trueswell

Comments

Hi Michelle! I find it incredible that you were able to find a way to analyze human conversation despite the COVID restrictions and having to do it all over Zoom. It would be fascinating to see how this impacts people in different age groups. What interested you in specifically looking at duration rather than pitch?

Incredible job Michelle! I was very intrigued by the general application of this research when regarding human conversation. It is interesting to see how duration impacts the meaning and understanding of an ambiguous sentence. I think it would be interesting to explore other populations, such as babies. 

Hi Michell! I really enjoyed your project. You did a lot in one summer! I'd be interested in hearing about the final analyses you run. What was the reasoning for you initiating the conversation? To have this standardized across participants?

Michelle, this research is amazing! I was wondering if you could talk a bit more about why you had the participants "act out" each script instead of reading the slides? What is missing when the participants just reads the slides?

Hi! Your project is super fascinating. It is really interesting how someones meaning can change just by pauses in sentences and emphasizing certain words in sentences. It would be really interesting to see this data compared to the equivalent data for non-native english speakers.