Local and global hyperarticulation

10 Abstract 11 In this paper we examine the two-way relationship between hyperarticulation and evidence of misrecognition of computer-directed 12 speech. We report the results of an experiment in which speakers spoke to a simulated speech recognizer and received text feedback about 13 what had been ''recognized''. At pre-determined points in the dialog, recognition errors were staged, and speakers made repairs. Each 14 repair utterance was paired with the utterance preceding the staged recognition error and coded for adaptations associated with hyper- 15 articulate speech: speaking rate and phonetically clear speech. Our results demonstrate that hyperarticulation is a targeted and flexible 16 adaptation rather than a generalized and stable mode of speaking. Hyperarticulation increases after evidence of misrecognition and then 17 decays gradually over several turns in the absence of further misrecognitions. When repairing misrecognized speech, speakers are more 18 likely to clearly articulate constituents that were apparently misrecognized than those either before or after the troublesome constituents, 19 and more likely to clearly articulate content words than function words. Finally, we found no negative impact of hyperarticulation on 20 speech recognition performance. 21 � 2007 Published by Elsevier B.V. 22 Keywords: Hyperarticulation; Clear speech; Speaking rate; Adaptation in speaking; Speech recognition; Spoken dialog 23