Accenting Givenness and Syntactic Role By E G Bard and M P Aylett Presented by David Vespe Presentation Overview Summary of authors work Authors results Analysis Previous Work Broadcast monologues interviews Elicited descriptions What about spontaneous speech Main Idea Compare repeated mentions In single task Across many tasks Test Setup Directions given imaginary map Speakers encouraged to contribute fully Characteristics Examined Intelligibility Loss Accent deaccented reaccented Structure Conversational Move Results Within a single dialogue Just 18 of repeated words deaccented For dialogues in general Second use of a word tends to be less intelligible regardless of accent Structure not significant for predicting deaccenting Conclusions Givenness does not imply deaccenting Introduction of a term tends to maintain structure across dialogues repeated mention does not Controlled experiments don t generalize to spontaneous speech Observations Results are for Glaswegian Southern Scottish English Did their own labeling Observations Directions based on imaginary map Everything is new Penalty for bad directions may lead to overaccenting Small numbers 48 cases of repetition across tasks 3 of these are deaccented QuickTime and TIFF LZW decompr are needed to see this