We report an experiment designed in order to study whether models of human-human voice dialogues could be applied successfully to human-computer communication using spoken natural language. Two groups of 6 subjects were asked to collect information about air travel, through dialoguing with a remote interlocutor. Subjects of the machine group were led to believe they were talking to a computer, subjects of the operator group were told they were interacting with a human operator. Both groups actually talked to the same human experimenter. The study focuses on subjects' representations of interlocutor's skills and knowledge, and differs from previous analogous experiments in several respects: - the task is more complex, giving rise to structured exchanges in natural language instead of question/answers pairs in a restricted language; - specific attention has been paid to the experimental protocol, in order to avoid the possible bias of other studies (in particular, conditions are strictly identical for both groups); - evolution over time has been considered (3 sessions, with a one-week interval). Some results confirm those of the literature, namely that subjects of the machine group tend to control and simplify their linguistic expression more than those in the operator group. However, most observations are either original or in contradiction with previous results: - subjects of the machine group are more talkative; no significant difference can be detected between the two groups with respect to most aspects of linguistic practice; - concerning evolution over time: both groups tend to get more concise; with respect to scenario processing (i.e. problem solving), the strategies of the operator group evolve significantly, whilst those of the machine group remain stable. These differences in the behaviors of the two groups are ascribed to different representations of the interlocutor's abilities.
In order to set up a list of libraries that you have access to,
you must first login
or sign up.
Then set up a personal list of libraries from your profile page by
clicking on your user name at the top right of any screen.