Page 125 -

122 J.M. Galán et al.

7.2 Three Symbolic Systems Used to Model Social Processes

Modelling is the art of building models. In broad terms, a model can be defined
as an abstraction of an observed system that enables us to establish some kind of
inference process about how the system works or about how certain aspects of the
system operate.
Modelling is an activity inherent to every human being: people constantly
develop mental models, more or less explicit, about various aspects of their daily
life. Within science in particular, models are ubiquitous. Many models in the
“hard” sciences are formulated using mathematics (e.g. differential equation models
and statistical regressions), and they are therefore formal, but it is also perfectly
feasible—and acceptable—to build non-formal models within academia; this is
often the case in disciplines like history or sociology, consider, e.g. a model written
in natural language that tries to explain the expansion of the Spanish Empire in the
sixteenth century or the formation of urban “tribes” in large cities.
We value a model to the extent that it is useful—i.e. in our opinion, what makes a
model good is its fitness for purpose. Thus, the assessment of any model can only be
conducted relative to a predefined purpose. Having said that, there is a basic set of
general features that are widely accepted to be desirable in any model, e.g. accuracy,
precision, generality, and simplicity (see Fig. 7.1). Frequently some of these features
are inversely related; in such cases the modeller is bound to compromise to find a
suitable trade-off, considering the perceived relative importance of each of these
desirable features for the purpose of the model (Edmonds 2005).
Some authors (Gilbert 1999; Holland and Miller 1991; Ostrom 1988) classify
the range of available techniques for modelling phenomena in which the social
dimension is influential according to three symbolic systems.
One possible way of representing and studying social phenomena is through
verbal argumentation in natural language. This is the symbolic system traditionally
used in historical analyses, which, after a process of abstraction and simplification,
describe past events emphasising certain facts, processes, and relations at the
expense of others. The main problem with this type of representation is its intrinsic
lack of precision (due to the ambiguity of natural language) and the associated
difficulty of uncovering the exact implications of the ideas put forward in this way.
In particular, using this symbolic system, it is often very difficult to determine the
whole range of inferences that can be obtained from the assumptions embedded in
the model in reasonable detail; therefore it is often impossible to assess its logical
consistency, its scope, and its potential for generalisation in a formal way.
A second symbolic system that is sometimes used in the social sciences,
particularly in economics, is the set of formal languages (e.g. leading to models
expressed as mathematical equations). The main advantage of this symbolic system
derives from the possibility of using formal deductive reasoning to infer new facts
from a set of clearly specified assumptions; formal deductive reasoning guarantees
that the obtained inferences follow from the axioms with logical consistency. Formal
languages also facilitate the process of assessing the generality of a model and its