Some Sources of Error in the Transcription of Real Time in Spoken Discourse.