While I sympathise with Alan's disappointment at not being included as a
contributor to the CMOT symposium, I hope he and others will understand
that the number that could be included was necessarily small and that some
decisions had to be made to ensure coherence and interest for the readers
of that journal.  I presume I ended up organising the symposium because
several of my papers have appeared in CMOT and I fairly regularly review
papers for it.  However, I take responsibility for my own decisions and I
regret any offence that has been caused.

Alan also raised a substantive point:

> Is this really the case that the philosophical and methodological
> bases of computational modelling are not central concerns?

Computational models of social behaviour and interaction are
descriptions.  They are, as Bruce Edmonds puts it, less expressive and
less ambiguous than verbal descriptions, but descriptions none the less.
Philosophers have (I believe) a long tradition of inventing examples and
thought experiments to explore abstract issues.  There are certainly
computational models used to implement such examples and thought
experiments.  I would have been surprised if reviewers for CMOT would have
thought that such issues were particularly germane to the readership of
that journal.  As a reviewer, certainly I would not have thought so.

The remaining issue in this context must then be whether "the
philosophical and methodological bases of computational modelling" are
central concerns in the design and implementation of computational models
**intended to serve as descriptions of observed or perceived individual
behaviour and social interaction and their consequences**.  At this stage
in the development of agent based social simulation, procedures for
representing what we observe and validating those representations should
in my view take precedence over the application of philosophical and
methodological bases and the verification of the representations in
relation to them.  (This was not, I should say, the dominant -- much less
the only -- view expressed in the published symposium.)

We have explored on this list and in the symposium whether the
verification of models with respect to sociological theory should take
precedence over the validation of models as descriptions of individual
behaviour and social interaction.  There would seem to be little point in
further discussions in the abstract as to whether verification with
respect to such theory or more generally (and abstractly) to philosophical
or methodological bases should take precedence over validation.  However,
I would be interested to see some examples of a representational,
computational model that was both verified with respect to some
theoretical or philosophical basis and was independently validated as a
plausible description of at least one specific example of social
interaction by clearly specified social entities (e.g. individuals or
organisations).

Scott

--
Professor Scott Moss
Director
Centre for Policy Modelling
Manchester Metropolitan University
Aytoun Building
Manchester M1 3GH
UNITED KINGDOM

telephone: +44 (0)161 247 3886
fax: +44 (0)161 247 6802

http://www.cpm.mmu.ac.uk/~scott