However, innovation is not the same as basic science. It can easily be seen that if a
scientist is collaborating with an entrepreneur, say, in writing a paper that exploits a
patent, they may or may not have to speak a kind of ‘pidgin’ scientific language to
each other. But invention, or discovery may well be expected to require the greater
cognitive precision associated with epistemic communities. Even Seely Brown &
Duguid (2000) who refer to the importance of communities of practice recognise,
from their long experience at Xerox PARC in Silicon Valley, that:
‘A firm, then, will almost always intersect multiple networks of practice. In
Silicon Valley, for example, some firms will have cross-cutting networks of
engineering, manufacturing, sales and marketing, and customer service.
Networks of computer engineers, for example, will run through all the firms
manufacturing computers’ (Seely Brown & Duguid, 2000)
They imply that these lateral professional links are cognitively less dissonant and
more smoothly connected even than the distinctive ‘capabilities’ linkages within the
firm. Hence the superiority of networking in a clustered environment rather than a
stand-alone competitive posture:
‘Knowledge seems to flow with particular ease where the firms involved are
geographically close together. Being in the same area allowed the Apple and
PARC scientists to meet and exchange ideas informally, paving the way for
more formal links. Relations between PARC scientists and the Dallas
engineers were in every sense far more distant’ (Seely Brown & Duguid,
2000)
Thus it seems that proximity in a cluster is fundamentally important to innovation,
that is, the stage at which deeply embedded knowledge is being confronted with
processes of knowledge exploitation and commercialisation.
When scientific method was more disciplinary than it now seems to be, that is in the
era of Mode 1 knowledge production as Gibbons et al. (1994) refer to it, it is probable
that specialisation was more important than diversification. But now there is strong
evidence, in biosciences at least, that it has become more transdisciplinary, as we have
seen, even at the exploratory R&D point in the ‘knowledge value chain’ (Cooke,
2002b) let alone the exploitation point in the same chain. Thus even basic research is
likely to contain a higher incidence of the need for ‘pidgin’ among diverse
professional scientists and engineers than used to be the case. Orsenigo et al. (2001)
date this shift in biosciences from about 1992. Thus Griliches and Glaeser were
publishing their results about specialisation at about the same time that it seems the