This fall, an amazing team of researchers is hoping to begin the process of creating an automatic pronunciation assessment tool to help identify both speech disorders and second language learning difficulties in children, thus bridging a gap between speech–language pathology and second language instruction. On the team, we have three professors, Evgeny Chukharev-Hudilainen, John Levis, and Gary Ockey, in the Department of English who specialize in learner modeling, pronunciation teaching, and assessment. The project will also be informed by the expertise of the fourth co-PI, Dr. Carrie Richie, a speech–language pathologist and clinical lecturer in the Department of Psychology.
They plan to record children with speech disorders, with L2-related pronunciation errors, and with a combination of both problems to help train a mispronunciation detection system to recognize and categorize the types of errors that children in these different groups may make. This will help them develop a prototype tool that can be used to screen children for possible language problems on a much larger scale than is currently possible in schools, hopefully helping children to receive the best language support for their needs as early as possible. In addition to the early stages of development, they will work with local speech–language pathologists and ESL teachers to develop a strong needs analysis to support further research on this project.