Embeddedness, Organizations, and Language Games
Paper
for presentation at SUMMARY Modern-day
economists are taught to take it for granted that, for the smooth running of an
economic system, each economic agent needs to know all there is to know and no
agent should know something another doesn’t. The slightest bit of imperfect information is enough to
derange the elegant optimality of the Arrow-Debreu equilibrium, and even a
touch of asymmetric information can induce all manner of partial-equilibrium
pathology. In mainstream economic
theory, knowledge is like a Big Mac: you want it to be the same everywhere. It is perhaps well to remember, however, that the founding conception
of economics was entirely otherwise.
The central point of the Wealth of
Nations is that the progressive fragmentation and differentiation of
the knowledge agents possess is precisely what fuels the engine of economic
growth. The division of labor
implies — indeed, to Smith, it causes — the division of knowledge. As the extent of the market expands,
each agent knows a smaller and smaller fraction of all there is to know, and
each agent comes to know things that are quite different from what his or her
neighbor knows. Far from impeding
the smooth functioning of the economic system, such division is responsible for
the overall growth of knowledge on which a prosperous economic system
depends. In real life, knowledge
is like fine wine: you want it to be complex and differentiated in time and
place. Of course, modern-day economists understand this point, at least
within an isolated partition of the brain. Many have read F.
A. Hayek (1945) on the marvel of the price system, which is able to
coordinate tacit and widely dispersed knowledge far more effectively than any
system of central planning. Some
may actually try to convey Hayek’s vision to their introductory students —
right before explaining, rather incongruously, that economics is all about the
optimal allocation of known and given scarce resources. “The most significant fact” about the
price system, in Hayek’s view, “is the economy of knowledge with which it
operates, or how little the individual participants need to know in order to be
able to take the right action” (Hayek 1945, p. 527). This may be a marvel, as Hayek put it; but it’s not magic. What allows the decentralized system to
coordinate dispersed knowledge and action is that some knowledge is in
fact shared. “The whole acts as
one market,” says Hayek, “not because any of its members survey the whole
field, but because their limited individual fields of vision sufficiently
overlap so that through many intermediaries the relevant information is
communicated to all” (Hayek 1945, p. 526). How does this happen?
In modern terminology, we might say that the market system is a
relatively well decomposed modular system (Langlois 1999)
in which participants interact largely (though by no means exclusively) through
the standardized “interface” of the price system. The price system provides an abstract, anonymous, and
minimalist apparatus of shared knowledge, and it is this apparatus that provides
the necessary “overlap.” But pointing to the price system does not completely resolve the
puzzle of coordination in a world of disparate knowledge. At least since Ronald Coase (1937)
pointed it out, economists have understood that much economic coordination
takes place not through a minimalist price system but through (sometimes large)
organizations like firms. As
Alfred Chandler (1977) famously put it, such large organizations replace the
invisible hand of the market with the visible hand of managerial control. Surely these large organizations make
use of the division of knowledge.
But how do they achieve coordination in the absence of the price
system? One implicit answer,
encouraged by Coase himself, has been to see firms (or other organizations) as
instances of small-scale central planning.[1] That is, the firm, unlike the market,
is an example of conscious coordination:
the constituent minds within the firm are brought together and
harmonized by some single mind, or at least by some tightly overlapping set of
cognitive frames within the minds of a handful of managers. If we are speaking about the origins of the firm, this may be
exactly the right picture. There
is a recent but growing literature suggesting that entrepreneurship consists in
the successful imposition of one individual’s cognitive frame on cooperating
others through persuasion, leadership, and the exercise of what Max Weber
called charismatic authority (Langlois 1998;
Witt 1998; Yu 1999). Such
entrepreneurial behavior is called for whenever a systemic reorganization of
production or a novel recombination of capabilities is necessary to respond to
a profit opportunity (Langlois 1992; Langlois
and Robertson 1995). But this
does not describe the mature “going concern.” One could well argue that large organizations are not
unified by a single cognitive frame, but are in fact systems energized by
disparate knowledge not concentrated in any center. Such organizations are arguably as much “spontaneous orders”
as is the market (Langlois 1995).
Like Harriet Beecher Stowe’s Topsy, nobody made them — they just growed. This paper is an attempt to make progress toward understanding how
disparate bits of knowledge are used and coordinated not only within the larger
price system but also within the organizations that inhabit the interstices of
the price system. We do this by
attempting to generalize the properties of the price system — or at least to
look at them in a different way — and to apply that alternative view to both
firm and market (and to other structures in between). We do this by drawing on a concept we call language games (Koppl and
Langlois 1994; Koppl 2000), an idea cribbed from the philosopher Ludwig
Wittgenstein (1953). A language
game is a set of rules about how to talk, think, and act in various situations. At the level of action (what we call agent-practice), agents follow rules or routines. But agents also have theories or models to explain their own
actions to themselves (agent-theory) as well as ways of explaining their
actions to others and persuading those others to cooperate (agent-rhetoric). As rule-like patterns of behavior,
individual language games are subject to a selection process within the larger
society, and that process determines how the various distinctive language games
fit together into a coordinated whole.
It is here that the question of knowledge overlap becomes central. Drawing on the theory of modular
systems, we explore the ways in which different patterns of knowledge can be
fit together in a coordinated system.
Although this effort is both preliminary and conceptual, we do offer
one clear — and perhaps surprising — conclusion. Contrary to the standard view, the coordination of knowledge
within a firm does not require that all agents share a common “mental model” or
interpretation of the world.
Indeed, coordination is sometimes maintained by persistent differences
of interpretation. Coordination
without agreement becomes more important as firms grow. With growth, relatively formal internal
governance structures supplant charismatic authority. And, using principles of modular design analogous to but
different in detail from those of the market, the growing firm is able to
profit from the wide variety of language games its members play.
the MPI/LINK Workshop on “Cognition and
Evolution in the Theory of the Firm,” September 25-27, 2000, Jena, Germany,
and at
the SCANCOR conference on “Crossing
Boundaries: Economics, Sociology and Organization Theory,” September
30-October 1, 2000, Stanford, California.
[1] “[I]n economic theory we find that the allocation of factors of production between different uses is determined by the price mechanism. … Yet in the real world, we find that there are many areas where this does not apply. If a workman moves from department Y to department X, he does not go because of a change in relative prices, but because he is ordered to do so. Those who object to economic planning on the grounds that the problem is solved by price movements can be answered by pointing out that there is planning within our economic system which is quite different from the individual planning mentioned above and is akin to what is normally called economic planning.” (Coase 1937, pp. 387-88.)