Review of Technological Innovation as an Evolutionary Process

Technological Innovation as an Evolutionary Process. Cambridge University Press, 2000. For our third original review (published January 9, 2001), Keith Pavitt has taken on the challenging task of evaluating this important new edited book for etss.net. John Ziman (editor) brought together leading scholars on technological change (John Ziman, Eva Jablonka, Joel Mokyr, Richard Nelson, Alan Macfarlane, Sarah Harrison, Gerry Martin, David Turnbull, Paul A. David, W. Bernard Carlson, David Perkins, Walter G. Vincenti, Joan Solomon, Geoffrey Miller, Edward W. Constant II, Rikard Stankiewicz, James Fleck, Gerard Fairclough, Janet Davies Burns) to appraise the usefulness of evolutionary ideas in explaining technical change. You can download the first chapter of the book as a pdf file here. Share your thoughts on the book or the review by posting a message in the discussion forum.

Reviewed by Keith Pavitt (k.pavitt@britishlibrary.net)

"The central idea of this book is very simple. Of the innumerable inventions that are put on the market, only the few that survive the test of use are reproduced. Stone axes, bicycles, medicines, jet aircraft and other technological artifacts thus 'evolve' in much the same way as biological organisms. What can we learn about technological innovation by thinking of it as a cyclic process of variation and selection, analogous to Darwinian evolution? For the first time, leading experts from many disciplines discuss this metaphor thoroughly in non-technical language, showing how it throws completely new light on many aspects of social and economic change, with many practical policy implications."

The above blurb on the dust jacket of this important book nicely describes its nature and purpose. With support from The Epistemology Group, John Ziman has pursued further his dual career as a distinguished physicist and as a scholar of the social relations of science, and has entered the fields of biology and technology along paths traced by evolutionary theory. To address the theme defined above, he has mobilized distinguished scholars from history, education, economics, management, anthropology, biology, psychology, innovation studies and engineering. With firm editorial direction, he has produced a volume that is both accessible to a wide readership, and original in its content. It deserves to become a landmark in the development of evolutionary theory in the social sciences.

He has also produced a book that is probably too rich to be digested by any single reviewer. It is certainly the case that this one knows too little about recent developments in biology (or about biology tout court) to say anything useful about their implications for understanding technological innovation. I am also not yet comfortable with the notion of shifting from physical artifacts and innovations to include cultural artifacts and innovations. It might turn out to be useful to describe and analyze the demise of, say, Inca civilization or centrally planned economies through the same lens as failed technological innovations, but I am competent neither to undertake such an exercise nor to judge the quality of its outcome.

I shall instead restrict myself to what the book contributes to our understanding of technological innovation. The potential and limits of biological metaphors in improving such understanding is a recurring theme throughout the book. Ziman summarizes the debate in the introductory chapter and I shall not repeat it here. It suffices for the purposes of this review to agree that mechanisms of generating variety and selection are central to both biological evolution and technological innovation. However, the book helps identify continuing puzzles and challenges in applying rigorously these concepts to technological innovation.

As Mokyr points out, there are three possible units of analysis for the processes of variety generation and selection: techniques (bodies of understanding and practice), artifacts and firms. He prefers techniques, evolutionary economists often prefer firms, and this writer (and apparently the book's editor) prefer artifacts. I have argued elsewhere (Pavitt, 1998) that considerable confusion can arise from failing to be clear about the unit of analysis. For example, is greater variety in techniques a cause of higher rates of technical innovation, given uncertainty in outcomes? Or (as this writer believes) are the main directions of rapid technical change pretty easy to identify, and a cause of both greater variety (and uncertainty) in artifacts, and a higher rate of technical innovation?

A common assumption in evolutionary economics is that the main locus of selection of innovations is the competitive market for user acceptance. However, the authors of several chapters (e.g. Carlson, Vincenti, Stankiewicz, Fleck and Fairtlough) show that selection also takes place amongst inventors and innovators within business firms. Understanding these selection processes not only means getting inside the heads of individual inventors. It also means understanding the nature and implications of increasing functional and cognitive specialization in and around the business firm. For example, causes of major failures in innovation in most modern large business firms are not the inability to meet technical challenges, since they now have established R & D laboratories that are capable of mastering major new technologies. The problems arise mainly because the effective exploitation of new technologies may require radically changed organizational practices and power relations, which the firm finds difficult to implement. Thus, Levinthal (1998) has developed the notion of "speciation", and Christensen (1997) of "disruptive technologies" both to describe how established firms have difficulty coping when major new applications and users emerge unexpectedly from the technologies that they master.

Again in much evolutionary economics, variety is determined by an industry-specific, exogenous variable often called "technological opportunity". Whilst this is empirically more appealing than a narrow notion of path dependency, defined as "random events and increasing returns", it leaves the variety-generation process firmly locked in the black box. The authors of several chapters point to how it might be opened up, namely through what Constant calls recursive practice: "alternate phases of selection and of corroboration by use. The result is strongly corroborated foundational knowledge: knowledge that is implicated in an immense number and variety of designs embodied in an even larger population of devices, artifacts, and practices, that is used recursively to produce new knowledge." (p. 221). Perkins further suggests that improvements in selection involve both "code" (i.e. theories) and "construction" (i.e. prototypes). And Martin explains the non-evolution of the Japanese sword over a period of 700 years in terms of lack of theoretical understanding of the factors affecting the blade's properties, and the high cost of constructing prototypes.

All of which suggests that increases in the level of technological opportunity depend on both increases in the level of theoretical understanding, and decreases in the costs of constructing and experimenting with prototypes. Recent advances in molecular biology illustrate the former mechanism, and in computer-based simulation technology the latter. It also raises questions about how best to conceptualize how prototypes are constructed. Is the concept of a "fitness landscape", taken from biology, necessarily the best basis, given that engineers have themselves developed methodologies for decomposing the design process? Shouldn't we building on these instead?

Finally, it is worth noting as social scientists the conclusions of a group of scholars including many from the "hard" sciences. "Where have we got to? Certainly not to a model in the mechanistic sense. Technological change is, above all, a social phenomenon. As such, its categories and their interactions are too imprecise and contextual to be represented realistically by a computable algorithm. Indeed, the same holds true for any evolutionary system. A mathematical simulation of its contingent, path-dependent behaviour can never be true to life. It too is a metaphor in its representation of a real system with complex unquantifiable structural relationships between its elements" (p.312).

Christensen, C. (1997) The Innovator’s Dilemma. Boston: Harvard University Press.

Levinthal, D. (1998) "The Slow Pace of Rapid Technological Change: Gradualism and Punctuation in Technological Change", Industrial and Corporation Change, 7: 217-247.

Pavitt, K. (1998) 'Technologies, products and organization in the innovating firm: what Adam Smith tells us and Joseph Schumpeter doesn't, Industrial and Corporate Change, 7: 433-51.