I'm looking for English corpora that include frequent formulaic sequences used in telephone conversation between strangers, such as calling to make appointments, ask for information, etc., especially beginnings and endings. Or if this doesn't exist, I'd be grateful for ideas on how to construct such a corpus, given the privacy and permissions issues. The information I want is excluded from Switchboard, as those recordings didn't start until after the initial exchanges had occurred. <br>
Linda Koza<br>University of San Francisco<br><br>