Abstract: The beginning of a "New Time," the theme of this year’s academy meeting, provides a welcome opportunity to take stock of the accomplishments, open questions, and most promising research avenues of evolutionary models in management and organization theory. A panel of three leaders in evolutionary thought -- Howard Aldrich, Sid Winter, and Dan Levinthal -- will appraise the state of the art in evolutionary research and will present their views on where scholarly efforts should go in the new millennium.

Key Terms: State of Evolutionary Theory, Future Directions in Evolutionary Research, Evolutionary Theory and Entrepreneurship

 I. Overview Statement for the Symposium

As part of the open-systems revolution in organization theory, evolutionary models blossomed in the 1970s. Within a short period of time, scholars formulated evolutionary accounts to explain phenomena ranging from the micro to the macro levels of organization. Inspired by the work of Donald Campbell , Karl Weick developed a social psychological theory of how individuals coordinate their actions. Weick’s theory drew heavily on the variation, selection and retention logic of evolutionary models. What Weick did for individual level organizing processes, Howard Aldrich achieved for the level of entire organizations. Aldrich formulated a comprehensive application of the variation, selection, and retention model to the study of how organizations change over time. In Aldrich’s evolutionary model, organizations flourish or fail because they are more or less fit for the particular selection environment in which they operate. Rather than explaining organization success and failure by appealing to managerial intentions, Aldrich’s evolutionary account focused on whether organizations have the appropriate traits for a particular selection environment, irrespective whether managers intended these traits are not.

While Weick and Aldrich were mostly inspired by Donald Campbell, Nelson and Winter’s evolutionary explanations of economic organization built on the Carnegie School’s routine-based models of organizational action and Schumpeter’s body of work. In the middle of the century, Schumpeter was the most prominent advocate of the position that economic change needed to be conceptualized as an evolutionary process. Unlike most other economists, Nelson and Winter (1982) took organization theory seriously in the conceptualization of firms and made it an integral part in the explanation of how industries and their structures change over time.

Selection-based explanations of the development of organizations and industries also formed an integral part of Hannan and Freeman’s work in the population ecology of organizations and Bill McKelvey’s complementary work on organizational taxonomy. While there has been relatively little systematic research on taxonomy, a large empirical literature has developed around Hannan and Freeman’s ecological models. (For overviews, see Carroll and Hannan, , Amburgey and Rao, , and Singh and Lumsden .)

Although the work of James March’s and his students’ work on organizational learning is typically not classified under the rubric of evolutionary theory, it clearly has a strong evolutionary flavor. As is evident from his introductory essay to Baum and Singh’s edited volume on the Evolutionary Dynamics of Organizations, March has been a close observer of evolutionary theorizing and his coworkers have readily moved from work on organizational learning to work that bears directly on advancing evolutionary theory. Dan Levinthal’s work on fitness landscape, for instance, deals with the fundamental issue in evolutionary theory of how to conceptualize selection processes . Levinthal has also employed evolutionary theory to advance our understanding of competitive processes .

The beginning of a "New Time," the theme of this year’s academy meeting, provides a welcome opportunity to take stock of the accomplishments, open questions, and most promising research avenues of evolutionary models in management and organization theory. I have been able to recruit a panel of three leaders in evolutionary thought—Howard Aldrich, Sid Winter, and Dan Levinthal—who will appraise the state of the art in evolutionary research and who will present their views on where research should go in the new millennium.

The three panelist were asked to address the following kind of questions:

I have recruited three panelists with diverse backgrounds—economics, sociology, management—to cover a broad spectrum of evolutionary thought in management and organization theory. Given the wide interest in evolutionary models among members of Academy, I believe that the symposium will be of interest to a large number of people who come to the Academy of Management meetings.

Nelson and Winter have pioneered the systematic application of evolutionary models to economics change. Their 1982 book An Evolutionary Theory of Economic Change has had a great influence in a number of fields. But arguably, its greatest impact has been in the field of business and corporate strategy. This is not all that surprising given that the book’s opening sentence places firms at the center of attention: "In this volume we develop an evolutionary theory of the capabilities and behavior of business firms operating in a market environment…" (my italics). Nelson and Winter’s (1982) work has inspired important theoretical and empirical work in the field of strategy such as Teece, Pisano and Shuen’s Dynamic Capabilities and Strategic Management , Kogut and Zander’s Knowledge and the Speed of the Transfer and Imitation of Organizational Capabilities: An Empirical Test , and Zander and Kogut’s Knowledge of the Firm, Combinative Capabilities, and the Replication of Technology . The literature around the resource-based theory of the firm also has been heavily influenced by Nelson and Winter’s (1982) evolutionary views on firm development. I am convinced that members of the BPS division would be very much interested in participating in a discussion of the important questions that remain in applying evolutionary theory to business strategy.

A key element of any evolutionary theory—whether it be in linguistics, biology, culture, technology or industry—is a mechanism for creating novel things. At the center of Howard Aldrich’s new book Organizations Evolving lies the question of how new organizations emerge. Aldrich writes in the introduction: "Organizational scholars have done an excellent job in explaining how things work in organizations that have been around for a while, but not how they came to be that way. In contrast, I am interested in the genesis of organizations, organizational populations, and communities…Without understanding why and how social units emerge, we miss the connection between on-going creative ferment in human societies and the particular realizations of it in organizations. Thus I give more attention to the early days of organizations, populations, and communities than most other organization studies’ books and articles" (p. 1). I believe that Howard Aldrich’s ideas on how evolutionary theory should inform the study of entrepreneurship will be of great interest to the members of the entrepreneurship division.

As reflected in its name, the OMT division of the academy is most explicitly concerned with advancing organization and management theory. Nelson and Winter (1982) have had a great influence on organization theory precisely because they offered a realistic portrayal of the micro behavior (routines) that can account for system level effects (e.g. market outcomes). Organization theories in general —not just its evolutionary versions—in many instances have not provided coherent and satisfactory links between micro and macro processes. I my view, members of the OMT division would find it very useful to hear the views of our three panelists on how to address this theoretically so important aggregation problem.

In his new book, Howard Aldrich argues that the evolutionary approach "serves as an overarching framework within which the value of other approaches [such as the ecological, institutional, interpretive, organizational learning, resource dependence, transaction cost economizing approach] can be recognized and appreciated"(p. 42). He elaborates that "The evolutionary approach holds us the promise of using these views to achieve an understanding—although perhaps not a theory—of organizations" (p. 42). I also believe that the members of the OMT division would find it very rewarding to discuss to what extent evolutionary theory can serve as an overarching framework for studying organizations.

Johann Peter Murmann, the organizer of the symposium, will provide a very brief introduction, talking about the purpose of the symposium. Next the three panelists—Howard Aldrich, Sid Winter, and Dan Levinthal—will offer their ideas on the state of the art of evolutionary thought and suggest promising avenues for future research. Johann Peter Murmann will then facilitate a discussion among the three panelists that tries to bring out the similarities and differences in their views. Finally, time permitting (this depends in part on whether we will get a 80 or 120 minute slot) the audience will be invited to participate in the discussion by asking questions or presenting their own views on the subject matter.

By Howard E. Aldrich

I use the term evolutionary approach to describe my perspective in my new book, Organizations Evolving. We now have the basic outlines for a theory of organizational evolution, thanks to the work of a few pioneering authors, especially Donald Campbell. In this paper, I take up two issues. First, I briefly define and explain the four generic processes that drive evolution and generate the critical events occurring in the life histories of organizational entities. Second, I critically examine the units being selected in evolutionary processes, noting points of disagreement among theorists. Some favor focusing on activities and structures on which evolution operates, such as routines, competencies, jobs, and strategies, whereas others favor bounded entities that carry activities and structures, such as groups, organizations, populations, and communities.

Evolutionary processes.

Evolution results from the operation of four generic processes: variation, selection, retention and diffusion, and the struggle over scarce resources (Campbell, 1969). Evolutionary theory is not a set of deductively linked law-like statements (Sober, 1984). Instead, it is "a concatenated system of loose, but apparently true and heuristic propositions ... it poses interesting questions, provides clues to their solution and, perhaps most crucially, generates testable hypotheses" (Langton, 1984: 352). The four generic processes comprising evolutionary theory are necessary and sufficient to account for evolutionary change. If processes generating variation and retention are present in a system, and that system is subject to selection processes, evolution will occur. Most importantly, as Alland (1967: 196) noted: "There is no requirement that these mechanisms be specifically biological in nature." The principles I draw upon are generic ones, applicable to social as well as biological systems.

Over the past few decades, Boyd and Richerson (1985), Dawkins (1986), and Dennett have provided lucid explanations of evolutionary thinking, and Nelson has applied evolutionary ideas to economic change. Many researchers have used evolutionary principles in their investigations. For example, McPherson applied an explicitly Darwinian evolutionary model in a series of projects investigating the growth and decline of voluntary associations . Lomi and Larsen used computational models to analyze the dynamics of localized competition in organizational populations. Such empirical projects demonstrate the gains that follow from exploiting the natural affinity between evolutionary principles of explanation and a substantive focus on organizational- and population-level change.

The Unit of Selection Issue in Evolutionary Theories of Organizational Change.

Theorists have proposed at least two possible units of selection in an evolutionary analysis: (1) routines, competencies, and strategies (Axelrod and Cohen, 2000; Nelson and Winter, 1982), taken singly or in bundles, and (2) organizations. Some theorists add populations and communities to this list. When analysts are interested in the founding or disbanding of whole organizations, organizations are the unit of analysis. However, are they a natural unit of selection? If organizations are, in fact, inseparable bundles of routines and competencies, then selection forces affect entire organizations as units. Founding and disbanding analyses could treat all units as essentially the same, except for basic demographic characteristics such as age and size. Bundles of routines and competencies, as tied up in an organizational form, would rise and fall together. Entrepreneurs who create new organizations would thus be the primary source of change in populations, rather than managers who engage in futile attempts at organizational transformation.

However, routines, competencies, and strategies may persist separately, carried by organizations but not locked into a single possible configuration. If so, then an evolutionary analysis must consider the relative contribution to fitness (for growth and survival) of each. The founding and persistence of organizations would drive evolution insofar as organizations act as hosts for those routines and competencies that, in various combinations, make organizations viable. Changes within existing organizations, as well as the founding and disbanding of organizations, would accordingly play a major role in evolution. Selection forces would thus affect the distribution and mix of routines and competencies in organizations, as well as their diffusion throughout a population. In this view, entrepreneurs play their most important role as creators of the units that will act as incubators for the generation of new routines and competencies.

Organizational learning theorists, in contrast to population ecologists, have favored routines, competencies, and strategies as units of selection. Nelson and Winter (1982), March , Axelrod and Cohen and others have proposed models of evolutionary change in which organizations are portrayed as learning systems. Some studies framed in ecological terms have also implicitly depicted organizations as carriers of routines and strategies, rather than units of selection. Research on organizational transformation, for example, has treated "strategy" as an organizational characteristic that diffuses throughout a population. In one such study, Haveman studied the diversification of thrift institutions into new lines of business, as well as the consequences of adopting a new strategy for thrift survival .

Empirical evidence with which to evaluate these arguments is limited. First, a synthesis of ecological and institutional theorizing in organization studies has elevated the concept of organizational forms as social constructions to center stage. When organizations are treated as bounded entities, researchers tend to overlook their internal diversity. Second, McKelvey’s (1982) arguments about populations as polythetic groupings heavily influenced Baum and Singh (1994), but the wide-ranging program he advocated never took root in the field. Systematic study of large numbers of organizations over long periods of time would tell us much about the independence of routines and competencies, but such research is costly and time consuming. Third, some evidence on routines, competencies, and strategies has come from simulations and laboratory studies by cognitive psychologists and evolutionary economists over the past decade. However, this work is proceeding slowly and somewhat independently of other research in organization studies.

In my new book, I followed the spirit of McKelvey’s contribution, but I also was governed by my goal of accurately reflecting the contours of existing research. When I used the term "form," I wished to convey an image of organizations similar enough to be called a population but different enough to display unmistakable variety. I also believe that both modes of selection – routines and competencies, and entire organizations – have a place in evolutionary and entrepreneurial analyses, and it would be premature to close off investigation along either line.

Conclusions.

Evolutionary theory applies to many levels of analysis: groups, organizations, populations, and communities. Variation, selection, retention, and struggle are processes occurring within all social units and across all levels of analysis. Populations, for example, are only reproduced insofar as individual organizations within them persist or are replaced, and environmental conditions continue to sustain them. Simultaneously, the reproduction of individual organizations depends upon the continued contributions of members and an organization’s degree of fit with the selection forces acting on its population. Evolutionary theory takes what it needs from the various approaches on organizations available today, as befits its eclectic nature. Whether promiscuous borrowing will corrupt its products is not yet clear, but theoretical eclecticism has not seemed to harm other long-lived perspectives, such as "contingency theory."

Although evolutionary explanations give great weight to external selection forces, they are not deterministic. Indeterminacy is a key feature of evolutionary analysis, and human agency is very much a part of the explanations I have reviewed. Major issues are still under debate, including the question of what is being selected in evolutionary processes: (1) activities and structures on which evolution operates, such as routines, competencies, and jobs, or (2) bounded entities that carry activities and structures, such as groups, organizations, populations, and communities.

Evolutionary theory does not deny that organizations are constructed by people who have strong interests in understanding what they are doing, and who invest much of what they do with symbolic meaning. People forge their own meaning systems as they respond to their social locations and confront the material implications of their positions. But opportunities to construct new worlds are limited because of the historical accumulation represented by existing organizations, populations, and social structures. When the social world becomes a taken for granted reality, people often decline to challenge it. Evolutionary theory thus posits a world where people are intendedly rational, can’t always get what they want, and certainly don’t always get what they need.

By Sidney G. Winter

In this presentation I offer a perspective on the recent progress and future prospects of evolutionary thinking in economics and management – with an excessive but understandable emphasis on the branches that have at least some of their roots in my 1982 book with Richard Nelson, An Evolutionary Theory of Economic Change.

What Nelson and I set out to do was to propose a workable and appealing program of fundamental reform in economic theory, a program that is in significant respects an extension of the contributions of the Carnegie School. The need for reform is apparent to most management scholars. For a long time now, standard economic theory has been committed to an account of economic actors that is determinedly and self-consciously fictive. Whether the fictions are more extreme for individuals or organizations is hard to say, but there are added complications in the case of organizations because the distortions associated with extreme behavioral assumptions are compounded by their linkage to more fundamental problems of ontology (e.g., "What is a firm?").

A famous methodological paper by Milton Friedman provided the philosophical apologetics for this approach; he essentially made the case that success in understanding events at one level of analysis (market outcomes) did not necessarily depend on accurate characterization of constituent causal processes at lower levels (behavior of organizations and individuals). The record of progress in economic research makes it clear, I think, that there is considerable validity in this view – but, I hasten to add, far from enough validity to justify the overwhelmingly dominant position that the rational actor paradigm actually occupies in economic theory. There are many purposes for which the burden of fiction involved in that paradigm is excessive; this is true for economics and it seems obvious that it is even more true when the questions at hand relates to not to economics but to management.

The long-run promise of the evolutionary reform program, as Nelson and I tried to evoke it, is the opportunity for economics to come to terms with behavioral realities and henceforth live in productive partnership with its sister social sciences from which it has long been estranged. Evolutionary thinking can explain how relatively "dumb" myopic processes involving plausible, boundedly rational actors can produce system-level results that are coherent, functional and in many cases quite similar to what standard economic theory predicts. True, it will not explain that everything is always Pareto optimal. It will not provide assurance of the immediate recovery of coherence after a major shock. It might well illuminate in new ways how economies can sometimes malfunction, thus perhaps (it is feared) encouraging new government policy initiatives that could prove counterproductive. And it will certainly obsolete a lot of existing intellectual capital invested in equilibrium models based on rational actors. Perhaps these considerations collectively explain why the evolutionary reform program is not very popular in the economics discipline.

In any case, it should come as no surprise that much of the recent progress in what I call "evolutionary economics" has come from the contributions of researchers whose focal interest in not in economics but in strategic management, organization theory, technology studies and the other social sciences. For most of these scholars, it is the behavior of organizations and their individual members that is in the foreground, while markets constitute a part of the context of action. Unlike economists, they do not have a tradition of resistance to realism in their conceptualization of economic actors, nor are they committed to viewing events in the perspective of an equilibrium theory.

The contributions I survey range across theoretical and empirical inquiry, and relate to both foundation issues and specific application. While many are explicitly framed as contributions to a broad evolutionary account of organizations and their environments, in some the evolutionary orientation is implicit or absent. (But evolution is opportunistic; it takes its gains where it can find them – and evolutionary research should do the same.) In the concluding section I shift attention to prospects for the future, and address in particular the question of what evolutionary thinking has to offer for scholars focused on the normative concern of improving management practice. The following paragraphs highlight some portions of the survey.

Variety, complexity and learning. Evolutionary theory emphasizes the importance of variety – in the present context, variety in the "ways of doing things" displayed by organizations. Where all contenders display the same traits, selection can do nothing to "improve" the average. There is an abundance of scattered evidence that in fact there is a great deal of variety in behavior both within and among organizations. The origins of variety have long been somewhat mysterious, but in recent years the mystery has cleared up a good deal. In particular, the roles played by complexity and path dependence have become a good deal clearer .

The nature of routines. The term "routine" was used quite expansively at some points in the Nelson-Winter book, and I think largely to constructive effect. It has, however, proved fruitful to sort things more carefully. One narrow construction of "organizational routine" term makes it the multi-person counterpart of individual skills, a multi-person coordinated performance developed through learning and at least partly tacit in nature. Given that interpretation, foundation-level understanding of the nature of organizational routines has been greatly advanced by experimental studies of coordination in dyads . My own effort to sort out the concept placed that sort of routine in the broader context of a typology of quasi-genetic traits (QGTs). In the broadest sense, a QGT is any trait displaying the sort of continuity and stability in time that is needed to make the evolutionary story work. Among the QGTs of organizations are, for example, their cultures – which are quite different things from skill-like routines.

Normative guidance from evolutionary theory. It must be admitted that there is something essentially pessimistic about the evolutionary perspective on management action. It does emphasize the limited understanding of actors and the important role that luck often plays. However, as the foregoing makes clear, it also emphasizes the importance of routines and capabilities acquired through learning – and often through learning undertaken by deliberate managerial decision. The important kind of "luck" from a strategic point of view is a matter of being in the right place at the right time with the right capabilities in hand. Deliberate choice plays a role in positioning an organization in all three ways to benefit from this kind of "luck" – but most obviously so in the case of acquiring capabilities.

By Daniel Levinthal

Academic fields differ in the degree to which they are "solution" or "problem" driven. Economics, at least in its Neo-classical formulations, provides the quintessential example of a solution-driven enterprise. While nominally defined by the agenda of examining how society allocates scarce resources, in practice the field is defined by a set of modeling practices --- representations of individuals as maximizing agents whose aggregate properties are characterized by some equilibrium condition. The solution-driven nature of the analysis has allowed economic analysis to become a central element in political science and management research.

The success of the Neo-classical paradigm in penetrating new academic niches stems in large part in the ability of the framework to address the basic elements in social science analysis. It provides a characterization of individual behavior and, through the equilibrium concept, provides a means of linking these "micro" actions to aggregate, or "macro" behavior.

As in the case of economics, evolutionary theory can be applied in a relatively context free manner. That is, processes of variation, selection, and retention can be analyzed in the original biological context, but may also be effectively applied to examine organizational populations. The success and struggles of evolutionary theories in the management world can be understand in terms of how this framework is able to satisfy the basic requirements of a model of social processes of specifying a well-defined lower-level entity and the basis of action for these entities, and characterizing how these lower-level entities interact and lead to more aggregate phenomena.

The appeal of evolutionary theories, particularly for researchers wishing to depart from the norms of Neo-classical economics such as Nelson and Winter (1982), is that they provided an analytically rigorous framework with which to characterize social processes without the assumption of rational choice at the individual level and equilibrium at the macro level. The behavioral tradition of the Carnegie School provides the central building block for the representation of individual/organizational action. At the macro level, there is a clear recognition of importance of competitive processes. However, these competitive dynamics are viewed as taking place in real time, not in a "twinkling of an eye" or the calls of the mythical Walrasian auctioneer of General Equilibrium Theory.

In empirical research, this examination of competitive dynamics as an activity that takes place across time and among existing forms, not necessarily some possible best fit form, has had tremendous success in population ecology research pioneered by Hannan, Freeman, and Carroll .

Less well settled is the issue of an appropriate and useful characterization of the micro entities of evolutionary theories of organizations. Building on the work of March and Simon (1958) and Cyert and March (1992), Nelson and Winter have proposed the notion of organizational routine as a candidate to fill this rule. This has been a powerful theoretical conception and researchers are beginning to identify empirical proxies for this construct.

Even apart from the challenge of identifying empirical measures of routines, its use as a characterization of the "micro entity" sits uncomfortably with many as it violates a strong norm, particularly in management research, of reductionism. We struggle for characterizations of individual behavior that offer some modicum of behavioral realism (i.e., are not characterized by the solution to a constrained optimization problem), yet recognize a role for individual choice and action.

Fortunately, if we return to the original notion of bounded rationality , we have a strong suggestion of a possible solution to this problem. The idea of bounded rationality has been a basis for two rather distinct intellectual lineages. One is a perspective focusing on organizational learning , especially ideas of local search (Cyert and March, 1992) and the evolution of relatively stable organizational routines (Nelson and Winter, 1982). Such routines reflect experiential wisdom in that they are the outcome of trial and error learning and the selection of and retention of prior behaviors. It is this "lineage" that has been primarily incorporated into current models of evolutionary processes.

Bounded rationality, however, does not negate the possibility of action based on a logic of consequences . Indeed, it has helped spawn a second research tradition, which models individuals as explicitly considering the possible consequences of the choices they make (March and Simon, 1958; Simon, 1991). In this tradition, bounded rationality is manifest primarily in the limited or imperfect cognitive representations that actors use to form mental models of their environment . While there have been some important exceptions, such as Weick (1979), this modeling of individual cognition, what Gavetti and Levinthal refer to as forward-looking intelligence, has generally not been incorporated into our evolutionary models.

In addition to postulating a more micro, micro entity for evolutionary processes (i.e., individuals), we have under-recognized the importance of multiple levels of selection in evolutionary processes in the management context. The work of Bower and Burgelman provide a strong suggestion of how to incorporate issues of power and incentives into our evolutionary models of organizations and organizational populations. Perhaps as we build more effective representations of individual actors into our evolutionary models, researchers will more readily engage such issues.

In sum, evolutionary theories have found a receptive niche in the Academy --- an audience for whom the analytical rigor of economic analysis holds some attraction but for whom the assumptions regarding individual maximization and equilibrium at the aggregate level hold less appeal. For this audience, evolutionary theory offers a fair dose of rigor with a more appealing set of primitives. Ultimately, to speak to the full range of concerns of management scholars, the theory must develop firmer micro foundations and address the multi-level nature of selection processes in organizational populations. These are not trivial challenges, but they are issues the framework of evolutionary theory is capable of incorporating and challenges that they community of scholars is capable of addressing.