How do language dynamics -- in acquisition, processing or historical change -- relate to the structure of linguistic systems? How do social and cognitive factors interact in shaping human languages? We explore these questions using experiments, statistical analyses of large corpora, and computational modeling. Most of our work is focused on words. We look at how words behave over time from when they are first created to their eventual success or extinction in linguistic communities. We also look at how words behave at different levels of representation, from variability in their phonetics to discourse-level variability in relation to topic and social identity.
The shared vocabulary of a language community may be the ultimate public good, supporting cooperation and collective intelligence at a scale that is unparalleled in other species. The goal of the Wordovators project is to understand how complex shared vocabularies are created, negotiated and transmitted within communities. A collaboration with the New Zealand Institute of Language Brain and Behaviour, the project combines large-scale experiments in the form of computer games with mathematical and computational analysis.
Sponsor: John Templeton Foundation
The SWORDFISH project (Spoken WOrdsearch with Rapid Development and Frugal Invariant Subword Hierarchies) seeks to develop algorithms for detecting words and phrases in audio recordings of under-resourced languages. We are collaborating on this project with scientists and engineers at the International Computer Science Institute in Berkeley, the University of Washington, Ohio State University, and Columbia University. Our own effort is focused on semi-supervised and unsupervised methods for learning the most productive morphological patterns in languages with rich morphology.