Page 259 -

258 O. Barreteau et al.

stage of a modelling process will aid better tuning of the model with its intended use:
either through interactions with people represented in the model, or with potential
users. Both cases have a major concern with making viewpoints explicit.

12.2.2.1 Case of Increasing Knowledge

The case of use for knowledge increase builds upon the previous subsection. The key
element treated here deals with the uncertainty of social systems. The involvement
of stakeholders represented in the simulation model is a way to improve its
validation or calibration. Participants may bring their knowledge to reduce or better
qualify some uncertainties. The simulation model is then expected to give back to
the participant’s simulation outputs based on the interactions between their pieces
of knowledge. On the other hand, this feedback is sometimes difficult to validate
(Manson 2002). Its presentation and discussion with stakeholders represented in
the simulation model is a way to cope with this issue. This approach has been
explored by Barreteau and colleagues to improve the validation of an agent-based
model of irrigated systems in Senegal River valley (Barreteau and Bousquet 1999).
The format of this feedback, information provided and medium of communication,
might make the model really open to discussion.
This joins another expectation which is probably the most common in work
that has so far implemented such participatory approaches with a social simulation
model: making each participant’s assumptions explicit, included the modellers (Fis-
cher et al. 2005;Mossetal. 2000; Pahl-Wostl and Hare 2004). This is a requirement
from the simulation modelling community: making stakeholders’ beliefs, points of
view and tacit knowledge explicit (Barreteau et al. 2001; Cockes and Ive 1996;
D’Aquino et al. 2003; McKinnon 2005). Moreover, so that participants might
become part of the model, the assumptions behind the model should be made
explicit in order to be discussed, as should the outputs of the simulations so that
they can also be discussed, transferred and translated in new knowledge. This is to
overcome one major pitfall identified with the development of models which is the
underuse of decision support models because of their opacity (Loucks et al. 1985;
Reitsma et al. 1996). This concern of making explicit assumptions in the modelling
process is also at the heart of the participatory approach community. One aim of
gathering people together and making them collectively discuss their situation in a
participatory setting is to make them aware of others’ viewpoints and interests. This
process involves and stimulates some explanation of tacit positions.
This means that the interactive setting should allow a bidirectional transfer of
knowledge between stakeholders and the simulation model: knowledge elicitation
in one direction and validation and explanation of simulation outputs in the other
direction.