Geography Reference
In-Depth Information
indicators of social contact, as we do. It is dii cult, however, to rule out all other factors
that the proximity measures might rel ect.
Similarly, our interdependence measure may capture not only the complexity of an
item of knowledge but also its breadth of applicability. Our results might then rel ect a
process in which low-
k
knowledge is broadly applicable and dif uses widely; moderate-
k
knowledge is of particular interest to select groups who tend to be socially proximate to
the inventor; and high-
k
knowledge is of such narrow application that it dif uses very
narrowly. This would produce a pattern in which actors socially proximate to a source
of knowledge most frequently receive and build on it if the knowledge has moderate
k
. The driving force under this alternative interpretation is not the relative ability of
dif erent actors to search in the face of complexity but the relative interest that dif er-
ent actors have in obtaining knowledge. The alternative interpretation raises the ques-
tion of precisely what makes an item of knowledge broadly or narrowly applicable.
Knowledge becomes broadly applicable in part because it is modular and therefore can
mix and match with other pieces of knowledge across a wide range of circumstances.
Applicability, then, may capture the interdependence of a piece of knowledge (especially
if one dei nes interdependence broadly and not in a narrow technological sense). To the
extent that applicability rel ects interdependence, we return to our original core hypoth-
esis: individuals proximate to the source of some knowledge have the greatest advantage
in receiving and building on knowledge of moderate interdependence/applicability.
6. Discussion
The analysis of patent citation patterns supports our basic theoretical perspective on
knowledge dif usion: search in the space of possible combinations of ingredients of ers
a useful lens for understanding the l ow of knowledge. Recipients socially proximate
to the source of the knowledge have preferential access to the original success, which
serves as a template during ef orts to receive and build on the knowledge. All recipients,
socially near and far, compete on equal footing when receiving and extending simple
knowledge; incremental search sui ces to reproduce simple knowledge, so guidance
from a prior success has little value. Highly complex knowledge, on the other hand,
equally resists dif usion to both classes of would-be recipients. Hence, at both extremes
of complexity, the close recipient has no lasting advantage over the distant. In con-
trast, for knowledge whose ingredients display a moderate degree of interdependence,
superior but imperfect access to the template translates into greater success in receiving
and building on pre-existing knowledge. The close recipient can complete its initially
imperfect replica via local search, but local search alone cannot guide the distant recipi-
ent to an accurate replica. Thus in our patent data, the largest gap between the ability
of a close recipient to receive and build on prior knowledge relative to the ability of a
distant recipient arises when the cited patent involves moderate interdependence. This
result appears when social distance is measured by proximity in a collaboration network
as well as when geographic and - to a lesser extent - organizational proximity proxy for
social distance.
Our i ndings have an array of practical and theoretical implications, especially for
the issue of knowledge inequality across social borders. Consider the graph of a typical
social network. It is quite common in such a graph to observe patches of actors with
dense connections amongst themselves and areas of sparse connections between patches
