Graphics Reference
In-Depth Information
the Big Five model) on the maintenance and development of trust and
co-operation within HCI, is a topic of ongoing research (Saariluoma
and Oulasvirta, 2010; Saleem et al., 2011; Schweer, 2008).
However, aside from negative ascribing, numerous positive
outcomes were ascribed showing that the user formed the impression
that the system wanted to connect with him/her on a personal
level by asking for private information and experiences, wanting
to develop a relation with the user, wanting to adapt to the user's
individuality. It was particularly during the system's intervention
that positive ascribing arose, such as the ascriptions of understanding
or supportiveness towards the system. These experiences enhanced
the users' continued co-operation and the development of trust, even
under conditions of feasible anthropomorphization.
8. Summary and Conclusion
When discussing spoken dialog systems, Edlund et al. (2008) have
suggested and convincingly argued for distinguishing between
different metaphors that users may use during interactions with
such a system. Do users treat the system more like a tool, i.e. do
they choose the interface metaphor, or do they prefer the human
metaphor by accepting the system as an interlocutor and behave more
like they would during human-human dialogs? In the light of the
metaphor discussion, we can summarize and re-interpret our results
from Analyses on Linguistic Structures, Rapport and Politeness as
follows: those subjects whose linguistic behavior provided a strong
indication for the dominance of one of these metaphors were in the
minority within our sample. This holds for the minority group of
those that preferred technically sounding “telegrammatic structures”
(cf. above) and thus obviously preferred the interface metaphor on
the one hand, as well at the other extreme, for the group of those
that heavily employed interpersonal signals such as formal pronouns
and politeness particles, thus indicating a human metaphor at work.
Although further investigations are necessary, the majority of our
subjects—see as well the discussion of hybrids in Chapter 7—seem
to have worked with “a metaphor that lies between a human and a
machine—the android metaphor” (Edlund et al., 2008).
The results of the linguistic analyses regarding changes in speech
behavior due to the system's intervention revealed that, based on the
mean scores, the intervention had a twofold effect on the users: firstly,
it made them more involved in the task and secondly, the users were
less emotional. These effects are traceable to the characteristics of
Search WWH ::
Custom Search