Page 261 -

260 O. Barreteau et al.

consequences of nonpoint source pollution are uncertain. This leads some farmers to
challenge the accusation, made by domestic water companies downstream of their
fields, that they are polluting their sources. Sometimes, disparate viewpoints do not
conflict. The gathering of these disparate pieces of knowledge is a way to reduce
uncertainty and allows the group of stakeholders involved in a participatory process
to progress, provided that they can work together.
Another characteristic of any social system which might hamper participation
is its dynamicity. Socioecological systems exhibit a range of dynamics, not only
social but also natural, which evolve at various paces. In the application developed
by Etienne and colleagues in Causse Mejan, pine tree diffusion has a typical time
step of 20 years which is long according to the typical time steps of land-use choices
and assessment (Étienne et al. 2003). In a participatory process, it might be difficult
to put these dynamics on the agenda. Simulation models are known to be good tools
to deal with dynamic systems.
Simulation models are therefore a means to gather distributed pieces of knowl-
edge among stakeholders and to cope with scenarios in the face of uncertainties.
They can also help make the participants aware of potential changes or regime shifts
generated by their interactions (Kinzig et al. 2006).

12.2.3.2 Towards Social Learning

Participation is often linked with the concept of social learning (Webler et al. 1995).
However, for social learning to occur, participants should have a good understanding
of their interdependencies as well as of the system’s complexity. Social simulation
can provide these bases, provided that the communication is well developed (Pahl-
Wostl and Hare 2004).
This learning comes from exchanges among stakeholders involved in the par-
ticipatory process but also from new knowledge which emerges in the interaction.
Externalisation of tacit knowledge in boundary objects (Star and Griesemer 1989)
is useful for both: it facilitates communication in giving a joint framework to make
one’s knowledge explicit, and it enhances individual, as well as social, creativity
(Fischer et al. 2005).
Simulation models are good candidates to become such boundary objects. Agent-
based models have long been considered as blackboards upon which various
disciplines could cooperate (Hochman et al. 1995). Through simulation outputs,
they provide the necessary feedback for reflexivity, be it individual or collective.
The question then remains whether such models constrain the format of knowl-
edge which might be externalised.