Geography Reference
In-Depth Information
must be successful in resolving the problem that may have triggered it. People are subject
to social biases, they engage in struggles over control, and they strike political compro-
mises. In their concern for institutions, researchers should study not only processes that
reproduce routines successfully but also those that cause routines to be abandoned or
forgotten. These processes include also 'failed' ideas that, by opening up new spaces,
contribute to the dif erential reproduction of behaviors across time and space.
While evolutionary theory cannot predict
exactly
how actions and patterns will evolve,
it
can
make predictions that guide future research. The task for evolutionary theorists is
to explain how the variation, selection, and retention of ideas - as well as other relevant
units - have led to the patterns that we recognize as clusters. The analysis of change is
then a matter of testing hypotheses that link the content of ideas to those features of the
agents and their environment which increase or decrease the likelihood that they will be
repeated and that they will displace those previously selected. This requires data sets with
sui cient detail to capture relevant variations in actors, ideas, practices, and social mech-
anisms, across levels of aggregation and in all relevant domains (Holmes and McKelvey,
2005). It also requires dynamic data structures that enable the separation of random
noise from true underlying time trends and regularities. Of obvious interest are histori-
cal data, including those collected through qualitative methods, in order to explain how
things currently are, what social actors think they are doing, and what it means to them.
Research on clusters would benei t greatly from temporally sensitive data-gathering
and analysis methods, as well as more process-oriented qualitative methods, such as
process-analytic narratives (Grii n, 1993) and process-tracing methods (Bennett and
Elman, 2006). To be sure, the data requirements of such methods are so severe that many
researchers may resort to simulation studies instead (Chang and Harrington, 2005).
At a meta-theoretical level, the social-evolutionary approach proposed here draws on
a variety of perspectives employed in cluster research, such as interpretivism, resource
dependence, and learning theory. Dif erent perspectives shed light on unique aspects of
variation, selection, and retention in a way that no single approach can do. When com-
bined, dif erent perspectives reveal the complexity, multiplexity, and probabilistic nature
of the forces that shape the life of clusters. By contemplating multiple rationalities of
actors and by examining action at multiple levels, dif erent perspectives provide useful
raw material for the further development of the evolutionary approach. The objective
of an evolutionary research program, one grounded in society rather than biology, is
not to replace other perspectives but to of er a platform from which the insights of other
perspectives can be interpreted, in order to improve our understanding of how regional
clusters came to be the way they are.
Notes
* I am grateful to Ron Boschma and an anonymous reviewer for their critical comments on a i rst draft of this
chapter. Any remaining errors are mine.
1. I refer to clusters here as a geographically concentrated set of related populations of organizations with suf-
i cient unit character to be recognizable as a more or less integrated 'community', having what Wittgenstein
(1958, p. 32) called 'family resemblance'. Depending on theoretical focus or policy interest, such communities
are variously referred to as regional learning systems, industrial districts, innovative milieux, or hot spots.
2. Adaptations are any traits that increase an entity's chances of survival, as opposed to advantages that
improve its general 'well-being'. Many cluster studies focus on advantages, such as interi rm collaboration,
institutional support, or shared identity, which may or may not have adaptive value for the entity.
3. An algorithm is a set of step-by-step instructions, like the building rules that govern the activities of
