Abstract

Abstract If we can model the cognitive and communicative processes underlying speech, we should be able to better predict what a speaker will do. With this idea as inspiration, we examine a number of prosodic and timing features as potential sources of information on what words the speaker is likely to say next. In spontaneous dialog we find that word probabilities do vary with such featues. Using perplexity as the metric, the most informative of these included recent speaking rate, volume, pitch height and range, time since start of utterance, and time since end of the interlocutor's last utterance. Using simple combinations of such features to augment trigram language models modestly reduced perplexities on both Switchboard and Verbmobil II.