492 Organization Studies 25(3)

Alessandro Lomi and Erik R. Larsen (eds): Dynamics of

Organizations: Computational Modeling and Organizational
Theories
2001, Menlo Park, CA and Cambridge, MA: AAAI Press and MIT Press.
352 pages (inc. index). $39.95/£25.95, ISBN 0262621525 (pbk)

Filippo Carlo Wezel As social scientists, we are all aware that the objects of our theorizing manifest
Tilburg University, themselves at different levels of analysis. Several prominent scholars have
The Netherlands
investigated the process through which individual decisions influence
collective behaviours and vice versa. Adopting a backward approach from
macro to micro, sociologists like Durkheim (1897) have shown how suicide
rates can be interpreted in the light of macro-structural forces such as the
decline of mechanical solidarity and the rise of organic solidarity due to
industrialism. By the same token, embracing a forward-looking perspective
from micro to macro, economists like Schelling (1978) have illustrated how
local interactions among agents, following simple behavioural rules, give rise
to aggregate regularities.
Nevertheless, the micro–macro transition is complicated by the inherent
recursive dynamics of social processes (Coleman 1990). Markets, for instance,
can be envisioned as complex adaptive systems, consisting of large numbers
of agents involved in multiple simultaneous local interactions. These local
interactions, however, shape macro-economic regularities which in turn affect
the dynamics of local interactions. The notion of intricate two-way feedback,
linking micro-behaviors to macro-structure, was present in the work of
prominent social scientists a long time ago (Hayek 1948; Olsen 1965).
The editors of this book, Lomi and Larsen, make use of the introductory
essay to position their volume within this research tradition and to clarify
their way of conceiving simulations as a theoretical laboratory to explore
multi-level, recursive social phenomena. As stated on p. 9, ‘by focusing on
the multiple connections that confound or, as the case may be, decouple
different levels of action, computational organization theories invite reinter-
pretation of forward and backward views as complementary strategies’. The
wide reach of theory-building based on computational modelling is illustrated
by the nature of the problems the volume discusses, by the substantial variety
of literatures utilized, and by the heterogeneous background of its contributors,
economists, sociologists, and computer scientists. Throughout the book the
imagination of the reader is stimulated by the illustration of multiple parallel,
although equally possible, theoretical worlds. The afterword by Burton cements
the file rouge of the volume: simulations represent ‘a versatile laboratory
where we can specify relations that are complex, path dependent and involve
feedback to do experiments, generate different new and plausible worlds and
explore what might be for organizations’ (p. 443).
Although simulations are only implicitly discussed in the introductory
essay, in subsequent chapters the reader becomes aware of them as a method
placed at the intersection between induction and deduction (Axelrod 1997).
Book Reviews 493

Like deduction, simulations are developed from a set of theoretical assump-

tions, but without demonstrating theorems or forming generalizations. On the
contrary, they are designed to inductively analyse data generated from
behavioural specifications rather than from empirical measurements. Thus,
as an inductive method, simulations can be used to infer general patterns, but,
as deduction, they represent meaningful ways of investigating the implica-
tions of behavioural assumptions. The foreword by James March reminds us
that the study of organizational dynamics involves historical complexity (i.e.
multiple interactions) and temporal complexity (i.e. path dependency and
feedbacks). Under similar circumstances, simulations represent a plastic
methodology for organizational theorists, and ‘the rich simulated world can
be used to understand the limits of our “real” world; it can be extended to
investigate the limits of the possible; it can create the plausible of what might
be’ (Burton, p. 442).
To avoid superimposing convenient categories, Lomi and Larsen adopt a
weak classification to divide the 14 chapters of the volume into ‘rediscovering
problems’, ‘reframing arguments’ and ‘taking views’. According to the editors,
the first part of the volume can be viewed as attempts at problem structuring.
Chapters 1 to 5 provide compelling evidence on how multiple theoretical
concepts can be expressively represented via computational models. Harrison
and Carroll’s chapter, placed at the junction between organizational demog-
raphy and organizational ecology, is an insightful contribution. Here the
micro–macro link is articulated through a reflection on the dynamics of
organizational culture and with respect to its consequences for organizational
survival. By elaborating on the link between organizational culture and
organizational metabolism (i.e. personnel turnover) Harrison and Carroll
speculate on the conflicting perspectives on age dependence (i.e. the effect of
organizational ageing on failure) as spuriously related to cultural heterogeneity.
In a similar way, and by simultaneously modelling agents, task, and network
dynamics, Carley and Hill investigate the mechanisms through which culture
is communicated in organizations. The reader particularly interested in
organizational learning and network theories will be impressed by the far-
reaching implications of this chapter. The contribution by Strang and Macy
can be considered a highlight for institutional theorists. Moving from simple
behavioural assumption (i.e. limited and biased information) Strang and Macy
provide evidence of the two-way feedback linking micro-decisions, i.e.
rationality vs suppression of private beliefs, to aggregate-patterns, i.e. the rise
and fall of conformity and institutional norms. The contribution of Loch,
Huberman and Ülkü addresses the interplay between individual interests and
team performance. The sceptic empiricist is challenged by the development
of a set of empirically testable hypotheses regarding the variables moderating
the relationship between status-based competition and group productivity (i.e.
number of status dimensions and group size). The stark but far-reaching
modelling technique of Prietula investigates the micro-dynamics of the
stabilization and routinization of knowledge in organizations. The chapter
succeeds in bridging communication technologies, anthropological theories
of communication and social network theories of trust and advice.
494 Organization Studies 25(3)

The second set of chapters, from 6 to 9, departs from existing organi-

zational theories to provide evidence of how computational modelling
improves theory testing and comparison. In this respect, the endeavour of
Bothner and White is notable. Building upon 20 years of White’s research on
markets as social structures, Bothner and White develop a set of formal
models in which firm strategies can be envisioned as contingent on the
polarization of the market, i.e. in the direction of buyers or suppliers. They
also present a set of testable hypotheses. The contribution of David Barron
is interesting for three reasons. First, it sheds light on the value of simulations
as a fruitful method of theory comparison, in this case between three different
perspectives on populations’ evolution. Second, it demonstrates the significant
value of relying on both actual and simulated data. Last, it is notable for
the elegance of its exposition. The chapter by David Krackhardt moves from
a set of simple assumptions to investigate the dynamics of diffusion of
innovations within a complex system. The thrust of his contribution is to trace
the structural conditions under which controversial innovations (i.e. inno-
vations valued through a dynamic social process) propagate. Krackhardt
intelligently nudges the reader to consider the social micro-dynamics respon-
sible for the success and failure of innovations. Lomi and Larsen participate
in their volume with an interesting contribution. By simultaneously building
on Harrison White’s research and on recent developments in evolutionary
biology, Lomi and Larsen propose a new interpretation of organizational
failures as an exit from a social structure of interconnected roles. The chapter
discusses the alternative explanations of age dependence (i.e. liability of
newness, adolescence and obsolescence) in the light of this perspective.
A comparison between the results obtained from synthetic data and empirical
estimates is presented.
In the third part of the volume, the tone of the discussion becomes more
epistemological and five chapters are presented to consolidate the relevance
of simulations for theory-building. The opening chapter by Miller adopts a
simple simulation technique to present a multi-level, interdisciplinary analysis
of the evolution of information-processing organizations. A clear contribution
of this article is to proclaim the complementarity between computational
models and formal results addressing specific theoretical concerns in
organizational design, e.g. centralization/decentralization, size. Levinthal’s
contribution departs from the joint consideration of spatial and temporal
complexity (see above) to speculate on the emergence of organizational
heterogeneity within industries. The findings of the chapter suggest that
adaptation to shifting environments hinges upon the level of epistatic
complexity of competitive landscapes, as in the K variable in Kauffman’s
(1993) NK models. From an epistemological standpoint, the contributions of
Malerba, Nelson, Orsenigo and Winter, together with that of Pólos and
Hannan, are intriguing. ‘History-friendly models’ are designed to close the
divide between formal models and ‘appreciative’ theory, i.e. discursive,
narrative. Here the potential of computation techniques as complementary
tools to mathematical formalizations and verbal theories is fully articulated.
The versatility of history-friendly models in linking micro-decisions
 Book Reviews 495

(diversification strategies) to macro-dynamics (market concentration) is

presented to the reader through a simulation of the evolution of the computer
industry. Anjali Sastry stimulates the theoretical imagination of readers by
introducing them to the fascinating world of recursive thinking. The perfor-
mance consequences of delays to environmental feedbacks are elegantly
discussed and coupled to heterogeneous organizational features, such as age
and size. The results, by and large, provide support to ecological theories of
inertia, but a set of under-explored research questions is left as food for further
thought. The last chapter of the volume meets the highest expectations.
Without recourse to simulation techniques, Pólos and Hannan provide
another excellent example of the simplification required by mathematical
formalizations. The chapter convincingly argues that divergent fragments of
theories, in this case age dependence, can be fruitfully juxtaposed by the use
of a new formal language (i.e. nonmonotonic) that relaxes the tight constraints
of first-order logic by avoiding binary rules. The elegant qualitative reasoning
of Pólos and Hannan sets a new agenda for quantitative research.
Without diminishing the relevance of the editors’ categorization, it is fair
to argue that several chapters could have been allocated to multiple sections.
As a result, the rich composite of insights contained in the volume does not
necessarily need to be read sequentially. The reader interested in technical
details will not be fully satisfied: most of the chapters do not include the codes
of the models presented. Since the relative advantage of computational
modelling relies on the ease of replicating existing findings and cumulating
them by exploring alternative interpretations, an effective adoption of this
method for theory-building purposes renders it imperative to specify the
behavioural assumptions behind the models. The accomplishment of this goal
is indeed complicated by the heterogeneous modelling procedures adopted
by the contributors. The general impression is that Lomi and Larsen
deliberately positioned their volume within cutting-edge theoretical debates,
privileging the illustrative power of computational modelling to its techni-
calities. By doing so, they give the impression of emphasizing the potential
contribution of computational techniques, leaving the reader to further
develop the instances discussed.
In summary, Lomi and Larsen’s acrobatic exploit is to effectively introduce
organizational theorists to the use of simulation techniques for theory-building
by offering a set of contributions highly worthy of ‘A’ rated publications.
This is why Dynamics of Organizations: Computational Modeling and
Organizational Theories represents an insightful contribution destined to
whet the appetite both of beginners and more experienced scholars.

References Axelrod, Robert Coleman, James S.

1997 The complexity of cooperation. 1990 The foundations of social theory.
Princeton, NJ: Princeton University Cambridge, MA: Harvard
Press. University Press.
496 Organization Studies 25(3)

Durkheim, Emile Olsen, Mancur

1897 The division of labor in society. 1965 The logic of collective action:
[1933] Glencoe, IL: Free Press. Public goods and the theory of
groups. Cambridge, MA: Harvard
Hayek, Friedrich A. University Press.
1948 Individualism and economic
order. Chicago: University of Schelling, Thomas C.
Chicago Press. 1978 Micromotives and macrobehavior.
New York: W. W. Norton & Co.
Kauffman, Stuart
1993 The origins of orders. New York:
Oxford University Press.