The Linear Model Is Dead— Long Live the Linear Model

Karl Grandin; Nina Wormbs; Sven Widmalm, editors. The Science-Industry Nexus: History, Policy, Implications. Sagamore Beach, MA: Science History Publications, 2004. xvii + 457 pp. $54.95.

Reviewed By Arthur Daemmrich

Considering the size and extent of industrial research today—in the United States, for example, it now exceeds nondefense government-sponsored research—it ought to occupy a prominent role in historical, sociological, and policy-oriented accounts of science and technology. A promising wave of studies came out in the 1980s and early 1990s, including a detailed study of DuPont R&D by David Hounshell and John K. Smith and a biography of Charles Steinmetz of GE by Ronald Kline. But the field seems to have drifted since, with comparatively fewer studies in recent years. On the one hand, scholarship in history of science has returned to easier pickings in academic and government archives that are accessible without fragile and laborious negotiations with companies; on the other, sociologists of science today seem more interested in markets and ethics than in industrial research laboratories. Economists and business historians are advancing the field, but with a different set of questions about macroeconomic outcomes, competitiveness, and business organization.

Industrial research has itself undergone some significant transitions in the last three decades. Quasi-government labs like Bell Labs, supported by monopoly positions, have closed or been transformed into completely different enterprises. Many chemical companies have moved research into their business units, looking more for process improvements and innovation in implementation than for the next nylon-type transformational discovery. Big pharmaceutical firms run central research sites with thousands of scientists, but their main challengers today come from small biotech start-ups, often spun out of university labs, and manufacturers of off-patent drugs. This is a rich terrain to explore, and there ought to be a lot more research into industrial research.

Trying to make up for lost time, 22 contributors in this weighty 450-page text largely focus on the question: is it meaningful to understand industrial research as a component of a linear path from basic science on the one end to marketed products on the other? This linear model has come under fire before from economists, historians, and policy experts; in fact, the literature depicting it as a straw man greatly exceeds that of its defendants. Comparatively recent “chain-link” models (Stephen Kline and Nathan Rosenberg) and the “total process” model (Mark B. Myers and Richard S. Rosenbloom) show complex feedback loops between research, communities of practice, and markets. Yet the linear model will not go away. In fact, so much of our regulatory apparatus (e.g., to determine pharmaceutical safety and efficacy or environmental and worker impacts of chemicals) and national infrastructure (e.g., government funding for “pure” science) is built on the notion that products are discrete and fully formed entities that come out of laboratories that to take the attacks on the linear model seriously would require a radical restructuring of the National Science Foundation, the National Institutes of Health, and most likely also the Department of Homeland Security.

Intriguingly, the private sector has found a solution. Research management tools like “stage-gate” retain a linear sequence with specific points of managerial intervention to promote or cut projects. At the same time, companies increasingly send research scientists or the occasional corporate anthropologist to “live in the market.” Sending scientists to work side by side with computer chip manufacturers led Dow Chemical to introduce a new semiconductor dielectric resin. Anthropologists from Intel traveling through Europe recognized that European households are structured differently and would have different ways of accepting and integrating technology from those in the United States. Nobody in industry thinks either“producer-push” or “market-pull” offers a way to innovate products that will sell.

So what do the remarkable set of historians and sociologists assembled in this book suggest? As several of the authors are at pains to point out, the relationship of history to policy has been particularly strained in the area of science policy and industrial R&D policy. While it is not unusual to see historians and sociologists affecting national policies on education and welfare, science and technology policies still rely most heavily on experts with scientific credentials, not on analysts.

Intriguingly, papers in the volume then examine the ways in which research itself has been industrialized. The “triple helix” of relations linking academe, industry, and government is at the core of new regimes for generating knowledge and commodities, and knowledge as a commodity.

In their introduction, the editors explain that the book originated in a symposium sponsored by the Nobel Foundation. As with any edited volume of this size, the quality of individual papers varies, and some are only distantly connected to the central theme of the linear model. The book has a broad geographic coverage and spans from the 1930s to the present, with important contributions by scholars in Europe and North America and a notable contribution from South America. The book is elegantly designed and has a good index for an edited volume. Let us hope that it will provoke further research and additional publications on the topic, including some sharpening of arguments and clearer statements of outcomes that can carry the day in the policy realm.