Geography Reference
In-Depth Information
for higher level entities (Sober, 1984). This means that the i tness of ideas will evolve
more quickly than the i tness of the cluster as a whole. Ideas themselves may vary in the
speed of evolution. Some ideas, such as those related to local traditions, may evolve very
slowly if errors in the replication process of traditions are rare (Shils, 1981). Others, such
as those related to collective strategy, may evolve more quickly if contestable ideas are
more prone to replication errors.
The third point concerns the relationship between the intended adaptation at the cog-
nitive level of individual actors and the adaptive capacity of the cluster as a whole. For
example, a policy of attracting resource-rich i rms into the cluster may have the perverse
ef ect of decreasing the i tness of the cluster if these i rms, through their actions, destabi-
lize the balance of power (Bathelt and Taylor, 2002). Competing ideas may help create a
diverse repertoire of potential solutions, useful for innovation if they are enacted in ways
that provide the system with freshness as opposed to complacency. A broad distribution
of competing ideas may have a more invigorating ef ect on change than a narrow dis-
tribution of nearly identical ideas. The key adaptation challenge for actors in uncertain
environments, such as those in new-media clusters, is not to imitate, codify, or routinize
what successful i rms are doing, but to create opportunities for redrawing and broaden-
ing cognitive categories by reconi guring relational boundaries (Girard and Stark, 2002).
The (in evolutionary terms) optimal cluster arrangement may be one that permits the co-
existence of contradictory logics of interpreting and evaluating ideas and actions. System
diversity, and the ambiguity that goes with it, remain a general condition for the growth
of knowledge (Pavitt, 1998).
Variation
The concern for ideas as discrete units carries over into the discussion of variation, which
is central to any explanation of evolution. Without variation, there can be no selection
and retention and, therefore, no possibility of improving the i t between the organiza-
tional form of a cluster and its external environment. If variation is to enable evolution,
it has to create new opportunities (there is more than one attempt to create change) and
must be able to fail (some variations must be forgotten, given up, or lead to dead-ends)
(Campbell, 1965). For example, new knowledge created in one cluster setting may dif use
- via labor mobility, i rm relocation, and so on - to other settings where it gets mixed
in with existing knowledge. The application of knowledge in diverse contexts will likely
produce new information as to its performance implications. Some new ideas will be gen-
erated in this process, while others will be given up. The changing distribution of variable
ideas throughout the evolutionary cycle is an issue of primary concern in discussions of
epistemic and practice communities (Amin and Roberts, 2008).
Some researchers may prefer taking an essentialist approach to the question of idea-
tional variation, by dei ning a small number of salient properties, such as ideas being
instrumental or sentimental, social or economic. But this is problematic if no concrete
guidelines are provided for distinguishing between trivial and central properties, and if
it leads researchers to highlight average features and thus to overlook the entire range
of (potential) variations. New variations of potential value may arise at the margins of
a distribution, as would be expected, for example, in clusters containing i rms with easy
access to resources located in distant locations. Some investigators may take an agnostic
approach to the question of ideational variation, in order to create sui cient analytical
