Giving Computers Personality? Personality in Computers is in the Eye of the User - Coverbal Synchrony in Human-Machine Interaction - page 48

it and, as follows, the communication with it. These are discussed
in connection with the development of mistrust and uncanny valley
effects.
4. Description of the WoZ Experiment
4.1 Design issues
Our WoZ scenario was designed in such a way that many aspects that
are relevant during the mundane situations of planning, re-planning,
and strategy change (e.g. conflicting goals, time pressure, etc.) would
be experienced by the subjects (Rösner et al., 2012a). The overall
structure of the experiment was organized in modules as follows
(Frommer et al., 2012):
￿ Greetings and Purpose,
￿ Prevention of Negative Courses of Dialogs (sub-module Initial
Dialog, sub-module LAST MINUTE (see Figure 2)),
￿ Module Saying Goodbye.
The Initial Dialog served as a personalization phase, followed by the
problem-solving at the 'LAST MINUTE' sub-module. These sub-mod-
ules served quite different purposes and were further sub-structured in
a different manner (for more details, cf. Rösner et al., 2012a).
Personalization sub-module
Throughout the whole personalization phase the dominant mode of
interaction was system initiative only, i.e. the system asked a question
or provided a prompt. In other words, this sub-module was a series of
dialog turns or 'adjacency pairs' (Jurafsky and Martin, 2008) that were
made up of a system question or prompt followed by the user's answer
or reaction. In some sense, this sub-module thus more resembled
investigative questioning than a symmetric dialog.
Figure 2. LAST MINUTE: stages of the experiment (The names for the levels of communication
refer to Halliday's model of 'systemic functional grammar' (Halliday and Christian, 2004).)