Geography Reference
In-Depth Information
used without limits by anyone, at any time, and anywhere across geographical bounda-
ries. From such a perspective, it is apparent that knowledge and information are treated
as being largely synonymous. However, once tacit knowledge is taken into considera-
tion, it becomes clear that technology as a whole cannot be easily traded or exchanged,
and only the potentially public knowledge component is liable to be assessed in terms of
transaction cost analysis.
When narrowing the notion of technology to something akin to information, and con-
centrating on the organisation of the exchange of such information, there is a tendency
to overemphasise the appropriability issue (Winter, 1987, 1993). The return on innova-
tion to a i rm may well be mainly a return on its creation of tacit capability, a process
supported by, but not reducible to, the generation of new potentially public knowledge.
In addition, knowledge can be at the same time both 'sticky' within the organisation or
i rm boundaries, while also being 'leaky' or mobile, generating outl ows in the environ-
ment external to the i rm (von Hippel, 1994; Wernerfelt, 1984). Ideas, inventions and
practices that are unable to move within organisations, in some circumstances may prove
to be quite capable of moving outside of them (Seely Brown and Duguid, 2001; Winter,
1987), thereby putting in question the centrality of the appropriability issue.
The main reason for knowledge to be coni ned to certain geographical contexts is
assumed to be its inherent complexity - particularly with regard to technical knowledge
- that may make it dii cult to share among dif erent interacting actors or organisations.
Such complexity may prevent knowledge from being codii ed and made explicit and
mobile, and thereby stored and transmitted by way of information. These arguments
underlie the knowledge 'i lter' hypotheses (Acs, 2002), and the wider notion of knowl-
edge may also embrace cultural and institutional dif erences, which shape the spatial
patterns of its functional dimensions, that is, knowledge production, absorption and
dif usion.
By viewing knowledge-creation processes as complex, systemic, cumulative, par-
tially tacit and sticky (whether codii ed or not) phenomena, there are strong grounds
for arguing that innovation is very likely to stay highly concentrated across space,
organisations and hierarchies, thereby giving rise to rather distinctive growth patterns.
Technological innovation is in fact generally more 'sticky' than production, and this
may be explained through the distinction - introduced by Henderson and Clark (1990)
and developed within spatial systems by Phene and Tallman (2002) - between
component
and
architectural
knowledge. The i rst encompasses specii c resources, assets, skills and
technical systems that refer to particular constituents of the organisational system rather
than to the whole. Architectural knowledge is instead related to an organisational system
as a whole, which, through its institutional structure and routines, arranges the compo-
nents for productive use (Phene and Tallman, 2002). Such architectural knowledge is
cluster- specii c, has a path-dependent and evolutionary nature and is developed through
relationships at both inter-i rm and intra-community level. Thus, it provides a common
framework through which the creation, adaptation, and dif usion of component knowl-
edge is highly facilitated, raising intra-cluster absorptive capacity and extra-cluster
linkages - therefore attractiveness to MNEs - vis-à-vis that of other locations.
Regarding the third limit of the transaction cost view, we must also consider how
each of the issues discussed here is related to the base and the sources of innovation in
the industry. On this point, works based on Pavitt's seminal taxonomy (1984) - used to
