Relatively complex or relatively simple? Toward new ways of analyzing language variation
Project description
Pilot research carried out in the PI’s lab in Leuven (Gardner et al. 2021) indicates that grammatical variation (e.g., Give me one vs. Give one to me) does not complicate language production, as choice contexts do not attract dysfluencies (i.e., long pauses and um’s and uh’s) across a large corpus of spoken English. This is astonishing given customary conjectures in theoretical linguistics. Building on these preliminary findings, the project will establish a foundation for large-scale follow-up research. Specifically, we will extend the dataset used in Gardner et al. (2021) and expand the analysis to address several urgent unresolved questions. For example, do non-canonical dysfluencies (e.g., discourse markers such as like and you know) behave differently to um and uh? Are different types of choice contexts more or less prone to attracting dysfluencies? Are more restricted choices “easier’’ for speakers than freer choices? Our results will have major implications for theorizing in linguistics.
Members involved
- PI Benedikt Szmrecsanyi
- Executing Thomas Van Hoey
- Collaborator Matt H. Gardner
- Collaborator Ruiming Ma
My role in the project
I joined this project as a postdoctoral fellow. My role consists/ed of
- annotating data
- overseeing students who annotate data
- analysis
- writing
Outputs
Basic info
BOF C1
Number: C14/22/045
KU Leuven