Information Technology Reference
In-Depth Information
Simulation methods are not only advantageous, they are necessary (Ahrweiler and
Gilbert
1998
).
We use an agent-based approach in which every agent or actor can investigate
a world of experiments and communication with other agents. Agents are defined
by their capacity to interact with this world and not solely by inherent or 'personal'
features. This follows from the view of science studies and with an insight of Herbert
Simon. He remarked that 'we cannot explain the path of an ant without reference
to properties of the terrain (such as chemical messages left by other ants)'. He goes
on to argue that human intellectual processes may, in fact, be relatively simple
in that most of the complexity of human behaviour may be drawn from a created
environment of objects designed to assist our intellectual endeavours (Simon
1981
,
p. 159). Research into how cultures provide for a cognitive environment bears out
this insight (Baird
2004
; Heintz
2004
; Hutchins
1995
) in a way, oddly enough, that
Simon and Newell's implementation of it (Newell and Simon
1973
) failed to do
(Collins
1990
; Gorman
1992
; Ahrweiler and Wörman
1998
). It is clear that to design
the material environment of science we must also add the social environment of
other actors.
Our simulated actors are therefore characterized by their responsiveness to opinion
(receptivity), responsiveness to data (flexibility), access to experiments, ability to
communicate with other actors, and by a belief profile. This profile specifies an actor's
confidence in each of a range of given hypotheses and allows us to represent bias.
Beliefs are formed by an actor's interaction with other actors and (
via
procedures)
with instruments that produce observations. Each actor revises its view of the world
on the basis of the data and opinions it encounters. What an actor does will depend
on which hypotheses it believes. Thus we treat beliefs as Peirce did, as dispositions
to act; that is, to make an experiment, or consult another actor, self-consult or does
nothing.
Like Simon's ants, scientists do not encounter a predictable, ready-made world:
they shape and enhance the world to make it more conducive to science. Whereas
ants must use chemical messaging, scientists cannot always apply ready-made terms,
concepts or procedures to interpret new or surprising features of the world. Studies
of visualization and modelling show that where science is producing new knowl-
edge, people are dealing with a world that is only partially described, using images
or concepts whose meaning is being worked out according to methods that were
investigated (Gooding
2004
; de Chadarevian and Hopwood
2004
; Lynch and Wool-
gar
1988
). Traditional views of science have emphasized the established, accepted,
finished product—the clearly expressed, predictable, experimentally proven knowl-
edge of textbooks and monographs. This approach hides the extent to which scientists
invent and negotiate ways of representing aspects of the world they are investigating
(Kuhn
1961
). Engaging with the natural world is both an
adaptive
and an
inherently
a social
process (Gorman et al.
2005
).
We have identified several aspects of science that should be included in a model of
inference: in particular, plasticity of representations and the inherently social char-
acter of these tokens of meaning. Plasticity can be handled according Wittgenstein's
analysis in the
Philosophical Investigations
. The meaning of a term is not given by a
