This was a nice talk from yesterday. I’ve attempted to write a shorter description, but more details can be found in the long abstract that the authors submitted. See http://www.unm.edu/~labfon12/program.html
The relationship between speech errors and prosodic phrase boundaries. (Choe & Redford).
Intonational units (IUs) have psychological reality. Listeners use IUs for word finding. Strengthening, lengthening and pausing are evidence IUs are important for planning. If IUs are psychologically real, they should be related to speech errors.
Method: speakers read sentences, some of them tongue twisters. Authors coded distance between source of error (word with the same sound) and the error, choosing the nearest possible source. Most error sources were from adjacent or non-adjacent word in sentence. Authors also coded IUs (important: IUs are coded post-hoc from the speech records!)
Results: Errors tended to occur at the end of a prosodic boundary. When the data was broken down for IUs at the beginning or in the middle of the sentence versus IUs at the end of the sentence, errors still occurred at the end of the IU boundary.
Summary: There was a cumulative error pattern (progressively more errors as you get to the end of an IU): IU is a coherent planning unit in a hierarchically organized speech plan. But the structure of tongue twisters may lead to the cumulative error pattern. There was also a final position effect (I didn’t quite follow this one, but I think it refers to a drop in errors on the last word of the IU). Serial ordering of language means you inhibit the past, activate the present, and plan for the future.
Q: Could it be an error in motor planning rather than phonological planning? A: Mary Beckman says that work on perseverating in stuttering (which occurs at the beginning of an IU) may be relevant.
Q: Focusing on perseverative errors. How could they be a product of weakening of the IU over time? Why should perseverative errors occur if IU unit’s activation weakens over time? A: from Redford, there’s planning at two levels. Activation of the IU but also activation within the IU. So within-IU activation should account for perseveration.