Technology policy in a complex world

Technology In Society, Vol. 16, No. 3, pp. 243267. 1994 Copyright Q 1994 Elsevier Science Ltd Printed in the USA. All rights reserved 0160-791X/94 $6.00 + .OO

Pergamon

0160-791X(94)00009-3

Technology Policy in a Complex World Robert W. Rycroft and Don E. Kash

ABSTRACT. President Clinton has given commercial technology policy a priority position amid signs that there is growing consensus in America that such policy is needed. However, while there is growing consensus on the need, there is widespread disagreement about exactly what initiatives we should undertake. Insight into the choice of policy options is provided by the new science of complexity. Four concepts are developed to guide technology policy: selforganization, learning, positive feedback, and emergence.

Continuous technological innovation is the driving force of our time. Technology pervades both our problems and opportunities. Commercial technology has become so vital that it now plays a part in presidential politics: The Clinton administration has stated that “technology is the engine of economic growth” and “we can promote technology as a catalyst for economic growth by directly supporting the development, commercialization, and deployment of new technology.“1 Robert W. Rycrofi is Program Director and Associate Professor in the Center for International Science and Technology Policy, Elliott School of International Affairs, The George Washington University. He holds a Ph.D. in political science from the University of Oklahoma, and previously taught at the University of Denver. Dr. Rycroft’s research interests include science and technology policy and environmental policy. He is the coauthor of eight books, including The Acid Rain Controversy, The Risk Professionals, U.S. Energy Policy: Crisis and Complacency, and Energy in the Global Arena. He has also written numerous journal articles, book chapters, and reports. Don E. Kash holds the Hazel Chair of Public Policy at the Institute of Public Policy, George Mason University. He was formerly George Lynn Cross Research Professor of Political Science and Director of the Science and Public Policy Program at the University of Oklahoma. He has a Ph.D. in Political Science from the University of Iowa. Dr. Kash is author of Perpetual Innovation: The New World of Competition, and coauthor of U.S. Energy Policy: Crisis and Complacency, Our Energy Future: The Role of Research Development, and Demonstration in Reaching a National Consensus on Energy Supply, and Energy Under the Oceans. 243

244

R. W. Rycroft and D. E. Kash

The substantive proposals that followed this declaration, however, represent little more than a shift in limited resources and the addition, in a fragmented way, of a few constrained initiatives. Although there is growing agreement that a technology policy is needed, there is widespread disagreement about what initiatives should be taken. Fear about the long-term consequences of involving the federal government in these activities permeates the whole debate.2 This conundrum is, in part, the product of a new physical reality.

The New Physical Reality Technological innovation, or the process of the commercialization of technological change, is consistently Janus-like: It routinely presents us with obstacles embedded in opportunities, and vice versa. For example, automobiles, which each have about 14,000 parts and cybernetic linkages,3 are critical to international competitiveness and economic growth and are also threats to the global environment. Given sufficient threat, some technologies are banned and replaced by other artifacts, which have their own promises and hazards. The incremental innovation of the automobile takes place in the context of dramatically altered time and cost considerations. Speed to market is critical to success. Product cycles are shrinking everywhere, and niche markets for specialized products proliferate. Traditional cost considerations often take a back seat to market share as large families of automobiles and parts, rather than specific cars, become the focus of production and profit. Automobile production and assembly has become global in scope as new collaborative arrangements are worked out across national boundaries. Temporal and financial boundaries have become blurred and difficult to assess.4 A great deal of confusion is created, and old models of understanding and traditional assumptions no longer apply. Many see these changes as chaotic and beyond governance. However, within the apparent chaos there are patterns. Evolutionary economics and the science of complexity may offer ways to understand these patterns. It is this characteristic that makes these bodies of thought potentially valuable to technology policy.5

Evolution

to Complexity

In searching for patterns that are useful for technology policy, one must note that contemporary technology did not appear fully developed. The accelerating changes in the world’s physical boundaries reflect the evolution of a production system that has taken us, over the last two centuries, from a world of scarcity to one of seemingly limitless abundance

lkhnology

Policy in a Complex World

245

and diversity. Figure 1 illustrates the evolution from simplicity through an intermediate stage to complexity for production (the conversion of knowledge, materials, and other inputs into technologies), organization (the mobilization of expertise and capabilities in ways that link production to technological innovation),6 and technology. As Figure 1 shows, production has evolved from a set of craft practices through the system of mass manufacturing or industrial production to a newly emerging “synthetic” mode.7 Production organizations have evolved from entrepreneurial concerns through conglomerates to SIMPLE

INTERMEDIATE

COMPLEX

PRODUCTION Craft;

s Manufacturing

Synthetic

Tacit knowledge, artisan/apprentice expertise, basic

Cognitive knowhow, division of labor in expertise,

materials, multi-

natural resource

synthetic ma-

purpose tools, small scale, individual as source of innovation

intensive, single purpose tools, large scale, institutionalized innovation

terials, multipurpose, adaptable tools, variable scale and scope, group innovation

Organic knowledge, constant learning key to expertise,

ORGANIZATION

RntreDreneurial

Conalomerate

Owner managed, nonhierarchical, niche creation strategy, innovation as core

Decentralized, diversified, strategy of portfolio management, control

capability, per-

of risk as core capability, complementary assets as

sonal contact as complementary asset, informal links to other organizations

specialized knowledge, R&D links to other organizations

One-of-a-K&kJ Products

Commoditv Prow

High variety, low volume, few components or architectures, value added by uniqueness

High volume, low variety, many linear-linked components and architectures, added value by efficiency

Network Flat structure, multi-functional, cross-institutional teams, collaboration as strategy, innovation as core capability, groups as complementary assets, extensive external linkages

Flexibility to deliver high/low variety/volume, large numbers of components, architectures, systems

Figure 1.Complexity in Production, Organization, and Technology

246

R. W. Rycrofi and D. E. Kash 1900

1960

I, I! II II I1 I

1992

I

synthetic Pro-

duction Industrial

Production

Figure 2. Phases of Production a new “network” type. 8 Technologies have developed from handicrafts through mass-produced commodities, to new high-value-added products and processes that quickly meet changing and diverse customer preferences by delivering new capabilities, improved quality, enhanced performance, and lower cost. 9 Technologies have evolved from having a few components (parts that embody core concepts and perform distinct functions) and architectures (ways components are integrated in a coherent whole)lo connected in a simple, linear fashion to highly integrated, nonlinear, cybernetic systems composed of many subsystems and architectures. The complexity of our new physical reality requires different and ever changing patterns of production, organization, and technology. As Figure 2 shows, until the 19th century, craft production was dominant. Industrial production was invented in the 19th century, and it became the primary source of products. In the second half of the 20th century, synthetic production became dominant.ll The process of change following World War II looks like a rope into which different fibers have been incrementally woven. Before 1940, the rope was overwhelmingly composed of two fibers: craft and mass production. Over the next five decades a new type of fiber - synthetic products and processes - was twisted into the rope. Today, the rope looks very different. It still has some of the original fibers, but if one looks only at the two ends, they appear unconnected. Production has moved toward greater complexity. But what does “complexity” mean, and why is it such an important concept for technology policy?

Technological

Complexity

The simplest measure of technological complexity is the number of components in a product. Figure 3 illustrates how complexity, measured by the number of components in a range of manufactured products, has

lkchrwlogy Policy in a Complex World

247 Space Shuttle

Ha,uard Mark I Calculator (2%~ pa*&..

LU

l-

Musket (51 parts, 10,000 units)

1800

1820

1840

1\111‘

+ .__o___ .....1(.....

lLr” pa*)

1860

18Ew

1900

1920

Mass Produced Producls Small Batch Production Unique Products

1940

1960

1980

Year

Figure 3. Trends in Component Complexity (Source: R. U. Ayres, VIM: A Challenge to ‘Ikchnology Management,” International Journal of Technology Management, Special Publication on Strengthening Corporate and National Competitiveness Through Technology, 1992, p. 19.)

increased over the past 180 years. For example, the musket that Eli Whitney produced for the U.S. government had 51 components. By comparison, the space shuttle has some ten million. If one looks only at components, however, one misses a key part of complexity: the cybernetic contribution made by architectures that integrate component parts and subsystems. 12 Increasingly, technologies have self-adjusting and self-correcting attributes. The Ford Taurus, for instance, has sensors and computers that adapt the performance of its engine to the temperature, altitude, and various other conditions. The Model A Ford had no such capability. On the process side, Ford has produced its 4.6-litre modular V-8 using concurrent engineering, which is the simultaneous involvement of research, design, development, and other types of expertise in the production process, to maximize feedback and interaction.13 Both product and process innovation now emphasize adjustment and adaptation through continuous feedback. The trends toward complexity are not uniform, however. Greater complexity in process technology can reduce complexity in the product. For example, process changes in Ford’s engine manufacturing system reduced engine components by 25%. On balance, however, the movement

248

R. W. Rycroft and D. E. Kash

is towards higher levels of complexity as technologies become aggregations of complex subsystems.14 Stephen Kline offers insight into this process. He views technological innovation as the output of “sociotechnical” systems. For both technologies and the sociotechnical systems that produce or use them, Kline has developed an index of complexity by aggregating three measures: the number of ways changes can occur within technological systems; the number of items that must be decided to design the technological system; and the number of control modes in the technological system plus those that connect the system to its surrounding environment. Control modes refers to feedbacks that result in adaptation. Such cybernetic systems can quickly acquire very large complexity indices. Kline, for example, estimates that an automobile’s complexity index is lo6 and the index for the sociotechnical systems that innovate cars is not less than 1013.15 We have used Kline’s formulation to examine the 30 most valuable product technologies traded in the international market between 1970 and 1990. Figures 4 and 5 distribute these technologies among the cells of a matrix with two axes: process/product and simple/complex. This taxonomy sorts technological sectors by: simple processes and simple products, simple processes and complex products, complex processes and simple products, and complex processes and complex products. When the results for 1970 and 1990 are compared, two differences stand out. First, only about 75% of the most valuable technologies for 1970 are listed as such for 1990. Significant changes have occurred that redefine which technologies are the most valued. Second, both the complexity of technologies and the proportion of total value represented by complex technologies increased markedly over the two decades. For instance, technologies in the simple process/simple product cell represented 58% of the value of the top 30 exports in 1970, but only 12% 20 years later. Alternatively, the shift between the two periods for the complex process/simple product sectors was from 12% to 36%. Complex process/complex product went from 31% to 51% (Figure 6). A revolution has occurred quietly and incrementally. l6 Another way of viewing complexity is to distribute technologies along a spectrum (Figure 7). The technologies divide, rather distinctly, into one set containing simple technologies, such as pharmaceuticals and petrochemicals, and another that contains increasingly complex technologies, such as aircraft and telecommunications equipment. The division is three-quarters complex and one-quarter simple.17 From the perspective of the science of complexity, there are two fundamental problems with our characterization of complexity First, it is too analytical and mechanistic. As W. Bechtel and R. C. Richardson point out, “mechanistic explanations,” or the accounting for system behavior in terms of the functions performed by parts and the interactions between these parts, demand a combination of “analytical” and “synthetic” strategies.

lkchnology Policy in a Complex World

Simple Process/Simple Product

249

Simple Process/Complex Product

qoneleotrio Machines, Crude Petroleum, Copper, Clothing, Cron L Steel, Power Machinery, Paper C Paper Board, Organic Zhemicals, Non-Cotten Woven Pextiles, Ships & Boats, Pextile Yarn C Thread, Meat, Zoffee, Iron C Steel Shapes, Pextile & Leather Machinery, #heat, Nonferrous Base Metal Ire, Pulp 6 Waste Paper, Chemicals Petroleum Products, Plastics,

Road Motor Vehicles,

Machines for Special

Electrical Machinery,

Industries, Medicinal Products

Telecommunications Equipment, office Machines, Aircraft, Switchgear for Electrical Power Machinery, Instruments

Complex Process/Simple Product

Complex Process/Complex Product

Figure 4. Most Valuable Product Technologies in World Trade, 1970 (Source:Don E.Kash andRobert W. RycrofX,“NurturingWinners withFederalR&D," Tkhnology Review, November/December 1993,~. 62.)

By analytical, Bechtel and Richardson mean the isolation of system components and the determination of what each does to reconstruct how the system as a whole operates. By synthetic, they mean “to conjecture how the behavior of the system might be performed by a set of component operations, and then to identify components within the system responsible for the several subtasks.“lS Both strategies are needed to understand the new physical reality that we face. Although Kline’s formulation provides an analytical base, more recent work in complexity science and evolutionary economics provides perspective into the synthetic dimensions of the puzzle.

R. W. Rycroft and D. E. Kash

250

Simple Process/Simple Product

Simple Process/Complex Procuct

Crude Petroleum, Iron & Steel, Meat, Furniture, Precious Pearls & Semi-Precious Stones, Footwear, Fish Nonelectric Machines,

Road Motor Vehicles, Office

Clothing, Organic Chemicals,

Machines, Electrical

Petroleum Products, Plastics,

Machinery, Power Machinery for

Paper 61Paper Board, Medicinal

Telecommunications,

Products, Chemicals, Non-

Instruments & Apparatus,

Cotton Woven Textiles,

Aircraft, Switchgear for

Aluminum, Plastic Articles

Electrical Power

Machinery,

Machines for Special Industries, Sound Recorders & Producers,

Metalworking

Machinery

Complex Process/Simple Product

Complex Processs/Complex Product

Figure 5. Most Valuable Product Technologies in World Trade, 1990 (Source: Don E. Kash and Robert W. Rycroft, “Nurturing Winners with Federal R&D,” Technology Review, November/December 1993, p. 63.)

Organizational

Complexity

The second problem with our investigation of technological complexity thus far is the lack of attention devoted to the organizations that innovate technologies. In general, organizational systems have received little attention from those who study complexity. This reflects the fact that mechanical interactions are easier to specify than social ones. However, as Kline’s estimates of complexity indices in the automobile sector indicate, there is an assumption that sociotechnical or organizational systems are, for the most part, more complex than the technologies they produce. The synthetic nature of complexity science and evolutionary economics offers insight into organizational complexity that analytical approaches cannot provide. Before examining these new concepts, however, the evolution from simple to complex production sketched in Figure 1 should be reviewed. Specifically, it should be noted that there has always been a link and

Technology Policy in a Complex World

Simvle Procees/S~le

Cmlex

Product

Simple Procese/Complex

251

Product

1970 = 58% ($86,708,435)

1970 = 0%

1990 - 12% ($224,699,631)

1990 = 1% ($25,549,954)

1970 - 12% ($17,906,225)

1970 = 31% ($46,021,270)

1990 = 36% ($644,454,846)

1990 - 51% ($919,266,926)

Process/Simple

Product

Simple ProcessKcmplex

Product

Figure 6. Thirty Most Valuable Exports: A 1970,199O Comparison interdependence among production, organization, and technology. That interdependence has become much stronger recently; so strong, in fact, that it is nearly impossible to separate the three trends. Today, competitive success in many sectors requires product integrity. Achieving product integrity requires an intimate union of organizational systems and the technological artifacts that they use and produce. Components, architectures, and subsystems must be integrated so that quality and customer satisfaction are delivered consistently. Integrity requires the synthesis of technologies and organizational systems because “every product reflects the organizations and the development process that created it.“19 But this was not always the case. Entrepreneurial and conglomerate organizations did not need integration to accomplish innovation. Figure 8 describes their pattern of innovation and why these earlier organizational types could be less concerned about coherence: As long as most valuable technologies were simple, a linear pattern of innovation worked well. For most of the post World-War-II period, the linear model dominated the conceptual landscape. The linear model assumes that research is the precursor to everything else. Innovation proceeded through a sequence of stages. There may be overlapping among stages, but typically only if the project was on a fast track. The implications for organizational patterns are obvious. The linear model was the conceptual description of most mass manufacturing, and not only in the private sector. In many defense and space programs, in which responsibility is required for each part of the delivery operation, the linear model provided the theoretical framework for technological innovation.20 Unfortunately, the linear model provides little insight into

252

R. W. Rycroft and D. E. Kash

the organizational complexity needed for much of the present technological innovation. When research is not at the core of the innovation process and innovation does not take place sequentially, the linear model is of little value. In today’s complex technological innovation processes, constant risk and uncertainty, multiple feedback loops, surprises, and midcourse adjustments are ubiquitous. Success in this context requires users, suppliers, and assemblers to be intimately linked to manufacturing, design, research and development, and servicing. More accurate conceptual models are needed for this kind of innovation, and Figures 9 and 10 illustrate two of the most useful. These figures show the “chain-linked” and “concurrent” models. Both are nonlinear and reflect the dynamic missing in the linear model. The chain-linked model displays aspects of linearity as well as tight linkages. The concurrent model displays the compression of innovation in which activities proceed in parallel. As with the linear model, both these conceptualizations have organizational implications. Both models emphasize the organizational need for external linkages with other firms, university or government laboratories, and the marketplace. They also suggest the necessity for interorga-

75% CoMrrPLEX Of World Trade AIRCRAFT MOTORVEHICLES OFFICE MACHINES TELECOMMUNICATIONS EQUIPMENT SCIENTIFIC INSTRUMENTS ELECTRICAL MACHINERY PHOTOGRAPHIC EQUIPMENT

__-____________--___--\\\\ Simplicity Threshold 1111 ____-________----___--INDUSTRIAL COMPONENTS PHARMACEUTICALS PETROCHEMICALS u

SIMPLE

25% Of World Trade

Figure 7. Most Valued Products in World Trade (Source: Don E. Kash and Robert W. RycroR, “Two Streams of Technological Innovation: Implications for Policy,” Science and Public Policy, February 1993, p. 29.)

L?vzhnology Policy in a Complex World

253

Research -e Development * Design * Manufacturing * Marketing * Servicing

Figure 8. Linear Model nizational communication. Whether the purpose is to monitor outside research or constantly changing user needs, maintaining information flows is crucial. It is in the information context, however, that even the chain-linked and concurrent models fall short: They are too analytical in their conceptualization of complex organizational and technological change. It is in this context that evolutionary economics and complexity science provide guidance. It is the synthetic aspects of complexity science and evolutionary economics that help us understand the Asian experience. Indeed, much of the evidence for the new ideas comes from Japan and other East Asian countries. The real power of complexity and evolutionary concepts lies in the fact that most have been formulated from either empirical work in the physical or life sciences or advanced computer simulations of those sciences. Taken together, this research provides a framework with which we can begin to move beyond the mechanistic and highly reductionist Newtonian view that defines the world in terms of a clockwork towards something more than the Darwinian explanations of evolution based on natural selection.

Self-Organization

A significant contribution made by complexity science is the idea of selforganizing capability. Self-organization refers to the ways in which material and living systems reorder themselves into ever more complex structures.21 Nobel Laureate Ilya Prigogine and his colleagues first advanced this concept, which is based on research concerning the thermodynamic behavior of physical systems thought to be dissipative (i.e., chemicals whose behavior was thought to be determined by boundary conditions, especially the entropy law). Instead of finding that all these systems inevitably experience disordering processes, Prigogine found many examples of dynamic systems that are characterized by increasing levels of complex organization. These thermodynamic systems, which are far from equilibrium, are able to counteract entropy by achieving a level of complex self-organization, which Prigogine called “autopoiesis,” to enable evolution. For these systems, initial conditions are more important than boundary conditions. 22 Later research moved beyond the study of thermodynamic chemical flows to the study of biological and then

254

R. W. Rycrojl and D. E. Kash

Chain-linked model showing flow paths of information Symbol.3 on arrows: C = centralchain-af-innovation; f = feedback loops; F = particularly important feedback. K-R D: I: S:

and w-operation.

links through knowledge to research and return paths. If problem solved at node K. link 3 to R not activated. Return from research (link 4) is problematic - therefore dashed line. Direct link to and from research from problems in invention and d&an. Support of scientific research by instrbments. machines. tools. snd p&adures of technology. Support of research in sciences underlying product area to gain Information directly and by monitoring outside work. The information obtained may apply anywhere along the chain.

Figure 9. Chain-Linked Model (Source: Stephen J. Kline, “Styles of Innovation and Their Cultural Basis,” Chemtech, August 1991, p. 474.)

socio-economic systems to examine the role of complex information exchanges as the basis for self-organization. It is this work that is most relevant to social systems, and it closely parallels the theories of many present-day evolutionary economists.23 Self-organization would be a key synthetic concept if for no other reason than helping to explain how some social systems are able to adjust to physical reality and to devise and select solutions that take them to higher possibilities. But it also suggests that social systems overcome physical obstacles, especially the inevitable disorganization and entropy of the second law of thermodynamics, or create new physical opportunities, or achieve higher evolutionary fitness levels, in large part by devising technological innovations. In fact, it can be argued that most of the

lkchrwlogy Policy in a Complex World

I

255

___--_ __.

(@iGiG ________

*\ ,’

,__.______

I

’

I

Sharing of Information

Figure 10. Concurrent Model (Source: Ministry of International Trade and Industry, Issues and Trends in IndustriallScientifzc Fxhnology: Tbwards Techno-Globalism, Tokyo, Japan: Ministry of internation Trade and Industry, September 1992, p. 12.)

increase in complexity in self-organizing social systems is due to the need for more sophisticated skills and structures to support the development and use of new technologies to overcome physical limitations or to identify new physical pathways.a4 Self-organization challenges strong orthodoxies. It has long been the conventional wisdom that most internal organizational actions such as development of production routines, recruitment of expertise, or choice of research and development (R&D) targets were a direct response to the external context or environment, including market forces. It was asserted that organizations operated with little or no discretion or autonomy; markets would create whatever organizational arrangements were appropriate. Neoclassical economics, for example, treated the firm as a “black box.“25 As long as organizational development and interaction with the environment was relatively simple and predictable, it was true that any organizational type had a chance to survive working in a reactive fashion. Increased complexity, however, has shattered the distinction between the internal and external dynamics of organizational change. A major idea put forth in the literature concerning self-organization is that “environmental influences are as much a part of organizations today as are organizational structures, administrative activities or production facilities.“26 In other words, complex organizational systems feature a great many agents, such as individuals, groups, and suborganizations, that are involved in rich interaction with each other and with outsiders, including suppliers and customers, in ways that enable continuous evolution. Interactions of this intricacy demand very sophisticated interpretations and response capabilities. This is the realm of organizational strategy; strategy does matter, especially in relation to the facilitation of intelligence gathering and information processing. Organizations that

256

R. W. Rycroft and D. E. Kash

become proficient at monitoring and acting upon rapid and complex information flows are the most successful in reforming themselves by self-organization. The managers of these organizations come to understand a key point: self-formation is never complete or perfect. Nonetheless, it is a powerful capability. These organizations also accept what Gareth Morgan calls “systematic wisdom,” rather than attempting to manipulate artificial organizational “causesn and “effects.“27 The abundance of temporary networks of collaborators that quickly unite to exploit a specific technological opportunity - virtual corporations - are examples of the process of complex self-organization. An example is TelePad Corporation, which collaborated with more than two dozen partners and suppliers to bring its pen-based computer to the market.28 Virtual corporations tap into diverse sources of expertise and capabilities to gain synthesis. These networks are in perpetual flux, continuously adding and dropping organizational elements. One expert has called this the “Terminator II” organization, because it resembles the metal monster faced by Arnold Schwarzennegger in the movie of the same title. The monster had the ability to liquefy, then harden again into a new shape.29 The objective of self-organization is to find appropriate forms of connectivity, which does not exclude patterns of structure or behavior adopted from earlier craft and mass production modes. For instance, self-organization in the network may include a certain amount of hierarchy normally associated with older mass manufacturing systems. After all, complex systems and subsystems sometimes imply hierarchy. In the process of self-organization, hierarchy, or vertical integration, or any other form, is a pattern of interaction that may emerge from strategy and experience.30

Learning

The reason for interaction and information to be at the heart of selforganization is that complex organizations must continuously acquire new knowledge or skills to reorder themselves: They must learn. A learning organization is skilled at creating, acquiring, and transferring knowledge and in modifying its behavior and structure to reflect new knowledge and insights. 31 Learning is at the core of the capabilities of today’s networks. Learning is self-organization in practice, and synthetic, incremental technological innovation is an organizational learning process. Both tacit and cognitive knowledge and ideas fuel organizational learning, but only if communication is continuous and enhanced by trust and reciprocity. Tacit learning, including shop floor experience, and the intuitions, insights, and hunches of all workers, has been underestimated

lkchnology Policy in a Complex World

257

and is critical for self-organization. 32 Often tacit understanding resides work groups that have multiple skills and diverse understanding, and are able to substitute for each other and for other groups. Tacit learning of this type often provides the connectivity among bodies of cognitive knowledge in the organization. Because of tacit learning, teamwork and reliance on group intelligence are requirements for effective communication and self-formation.33 Self-organization based on changing knowledge, ideas, and learning are messy and far from rational. For this reason, complex network organizations engage in what Daryl McKee calls “error-embracing” behavior. Discoveries go untapped, previous experiences must be “unlearned,” and mistakes must be viewed as natural by-products of uncertain operating environments.34 Learning is essential but imperfect. Network organizations accept that “not knowing” is a constant. As one study states: “An effective network can exist only if the organizational thinking advocates that it is not bad not to know, it is only bad not to learn.“35 Complex organizational learning involves continuous change in systems, norms, and values in nonlinear ways under intense timing constraints. Faster product cycles and shrinking appropriability of new ideas demands more sophisticated collaborative inquiry, organizational self-reflection, and extensive scanning capabilities, because they are involved in what amounts to an ongoing crisis. Patricia Meyers summarizes the dilemma faced by complex innovative organizations as they try to steer innovation while experiencing pervasive uncertainties: in autonomous

‘lb meet the challenge of technological acceleration firms must learn faster and more effectively because the most successful innovative organizations are also high performance learning systems. Moreover firms must create different ways of learning and of “learning how to learn” under these unfamiliar, dynamic environmental conditions.36

Learning under conditions of great external uncertainty often generates a vulnerability that can become a network’s strength. In fact, synthetic innovation is a delicate process, and outside pressure mandates reciprocal obligations in the learning process. Thus, even factories can become “learning laboratories” that develop into complex organizational systems that integrate problem solving, internal knowledge, innovation, experimentation, and external information. Chaparral Steel, cited by Fortune as one of the ten best managed factories in the U.S., has developed an organic learning system that is heavily dependent on international linkages. A “virtual research organization,” incorporating expertise from Italy, Germany, Mexico, and Japan is crucial in helping the company learn how to develop specialized steel products.37 The managers of factories, as well as those of other types of learning organizations, have to understand that an increase in individual learning

258

R. W. Rycroft and D. E. Kash

does not automatically lead to an increase in organizational learning. One of the major factors in the closing of the acclaimed Uddevalla automobile plant was the absence of a strategy for facilitating continuous learning across the organization as a whole. This was a serious liability in competing with Japanese automobile manufacturing systems, which explicitly focus on strategies for organizational learning. The Japanese approach has been extended to transplant facilities abroad. For instance, the Toyota-General Motors joint venture, the New United Motor Manufacturing Inc. (NUMMI) plant in California manages skill development plans for individual workers as components of a system-wide learning capability.38 Learning among work groups in organizations like NUMMI or Chaparral Steel requires long-term commitments; yet, there must be substantial group autonomy. As noted above, it all comes back to communication, the glue of the organization. The central advantage of teams lies in the reduction of organizational barriers to the communication of articulated knowledge. Free-flowing information is the new raw material of innovation. Valuable technologies are not as information-intensive as are the organizational networks that produce them. Knowledge, however, ultimately resides in groups of people. Facilitating the flow of information depends on eliminating fear and fostering a common vision, sense of direction, and understanding of values.39

Positive Feedback

The rewards of communicating the learning experience are enormous. Learning is one of the major sources of positive feedback, the effects of which include increased returns. Brian Arthur has argued for some time that one of the major differences between natural resource-based commodity sectors, a category that includes most of the simple technologies listed in Figures 4 and 5, and knowledge-based, high-technology sectors, which include those categorized as complex in our formulation, is that the latter provide many more opportunities for learning and applying the new understanding to the production of the next iteration of technologies. Moreover, he says that greater organizational understanding of how to make one product technology can enhance the ease with which other products that incorporate similar or related technologies can be innovated.40 Of course, complex technologies require initial investments in R&D and tooling, which are much greater both in terms of specialized expertise and capital, and the design and manufacturing tasks, which are much more complicated than those for simple sectors. However, the positive feedback from learning leads to enormous improvements in organizational capabilities in more knowledge-intensive sectors. Enough positive feedback, in combination with high and specialized initial costs, can lock-in a dominant process and its products and make it

!l’echnology Policy in a Complex World

259

very difficult for competitors to displace them. In fact, if learning and positive feedback take place rapidly enough, the first product technology may be the only one used. Thus, the choice of technology for use in the initial prototype can be critically important.4l The power of self-reinforcement as a barrier to a competitor not experiencing comparable positive feedback is illustrated by the capture of virtually the entire video cassette recorder market by a network led by JVC that produces the VHS, which is by most estimates inferior to Sony’s Betamax. 42 The degree of increasing returns for the JVC group enhanced the probability of locking in the market. Moreover, lock-in can be an advantage as new products are introduced. The manufacturing and marketing expertise learned in and applied to the production of the VHS has had great payoff for JVC and its partners in the camcorder market. It may do the same for future high-definition television linkages.43 Expanding the range of positive feedbacks means engaging expertise and capabilities wherever they can be found. Organizations engage in purposeful search for and construction of technological capabilities that are built on past experiences of success and failure, as well as future expectations. The innovative organization cannot be passive in the process of reconfiguration. Advantage goes to those organizations that reduce uncertainty by learning and adapting through continuous positive reinforcement. In the words of David Teece: Successful technological innovation requires complex forms of business organization. To be successful innovating organizations must form linkages, upstream and downstream, lateral and horizontal. Advanced technological systems do not and cannot get created in splendid isolation. The communication and coordination requirements are often quite stupendous, and in practice the price system does not suffice to achieve the necessary coordination.44 By whatever name, the organizational key to positive feedback is the network (Figure 11). Network organizations connect the expertise needed for technological innovation. The groups that hold much of the expertise are linked in all kinds of ways, ranging from face-to-face contact to long distance communication. Expert groups may reside in a myriad of institutional settings, such as government laboratories and corporate manufacturing plants. Finally, links are both formal - contracts, licenses and informal - personal relationships. Such collaborative connections often appear anarchistic or random. As traditional organizational boundaries erode, the competitive environment becomes increasingly turbulent, and some networking takes place out of the purview of managers. Order is difficult to detect when linkages are informal and cross-organizational, but complexity science suggests that neither total chaos nor equilibrium is likely to be found in these systems. To expect either is to miss the final key concept.

260

R. W. Rycroft and D. E. Kash

Emergence

By “emergence,” we mean that complex systems generate new “properties.“45 In organizational systems, these properties may be structures that form and interact in new ways or manifest new behaviors. Complexity is related to emergent properties. If the formulations found in the science of complexity are correct, in the case of technological innovation, one should expect to see the persistent evolution of emergent organizational systems such as the web of incentives, constraints, and connections that characterizes networks and emergent behaviors, including innovative ways of interacting to develop new configurations of processes and products, which lead to emergent concepts such as positive feedback.45 As an example, consider the recent emergence of R&D collaborations in biotechnology. These are among the most fluid networks, and they are often interorganizational, crossing university and corporate boundaries. It is obvious that many biotechnology networks begin as informal linkages, then evolve into more formal structures. To date, the only rational explanation for the transformation from casual relationships to more highly organized ones is the integrating power of the market. Currently, however, the concepts of self-organization, learning, and positive feedback lead to emergence as an alternative explanation for cases such as the following from the Danish biotechnology industry: [In the beginning] the actors are loosely coupled and their number is variable. Their interaction is free, driven more by accidental opportunities than by precise intentions and organizational strategies. When occasionally shared projects emerge, interactions crystallize into somewhat more durable, committing forms of relationships. On such occasions, elements of structure and order can be recognized, if only temporarily. While they exist, these crystallized relationships in the sense that they seem to attract more act as “centres of gravity,” researchers, more often than not resulting in quite complex crystals of collaboration. However, it is important to add that first of all many such centres of gravity exist at any point in time, and secondly being part of one such crystal of collaboration does not seem to prevent (in terms of time or loyalty) anybody from participating fully in the current stream of interaction . . . . The accidental connections described above bring hitherto separate parts of the field to mesh temporarily, reshuffling links of communication and collaboration. Through this, highly unlikely partners sometimes find themselves sharing projects that nobody would consciously have designed.47

The strength of the emergence concept is its integration of the internal and external variables, the synthesis of organization and environment. As the above quote illustrates, complex networks are not always created, nor do they always evolve strictly as a function of either selection by the environment or conscious choices by network members. Complex network organizations are continually emerging. John Holland, a scientist

261

!t?xhnology Policy in a Complex World

TECHNOLOGY NElWORK

d BASICRESEARCH d APPLIEDRESEARCH d DEVELOPMENT 4 DESIGN +- PRODUCTION # MARKETING / SERVICING /

MARKETPLACE

O=WORKGROUP

TECHNOLOGY COMPLEX

Figure 11. The Network Organizational System (Source: Robert W. RycroR and Don E. Kash, “Complex Technology and Community: Implications for Policy and Social Science,” Research Policy, forthcoming.)

involved with complexity, states that these innovative organizations “never get there” in traditional performance terms, which include efficiency and optimality, because they are always “in the process of becoming” rather than reaching end-points.48 As Holland states, there is nothing efficient or optimal about the evolving structures or behaviors of complex organizational systems, even if efficiency or optimality could be defined for systems as a whole. Those roles and rules that do not promote adaptation and emergence by way of learning and positive feedback will be systematically weeded out by successful organizations. Adaptation of this kind is anticipatory and knowledge-based. In adapting to changing circumstances, complex systems alter those surroundings ahead of time, if possible. Finally, the process of anticipatory adaptation leading to emergence by any one network organization influences every other organization with which it interacts. Each complex organizational system’s strategy or action that is aimed at technological innovation alters the context in which others develop their own innovative strategies and actions. This is

262

R. W. Rycrofi and D. E. Kash

the process of coevolution, a central concept of evolutionary economics, and is the basis for the ever-shifting landscape of alliances, rivalries, and supplier-producer-customer relationships.49 Perhaps more significant is the fact that coevolution blurs the distinctions between internally fueled and externally generated organizational change. Self-organization and selection by market forces are becoming two sides of the same coin.

Implications for Technology Policy

All policy makers have an understandable preference for dealing with simple, linear phenomena. They prefer that policy be made under conditions of relative stability, continuity, and testability. Technology policy is no exception. Until recently, U.S. technology policy simplified physical reality by treating technology as a black box, an exogenous variable external to the political economy. Arthur has characterized the simplification of technology as follows: [Tlhe notion was that technologies came at random out of the blue, fell from heaven in celestial books of blueprints for making process steel, or silicon chips, or anything like that. And those things were made possible by inventors - smart people like Thomas Edison who sort of got these ideas in their bathtubs and added a page to their book of blueprints.50

Defining technology and the organizations that create it in simple terms has meant that highly reductionist analysis could continue to be applied to the policy process. Under such circumstances, theoretically elegant frameworks could be developed, even if they were detached from physical reality. Over time, techniques and analytical procedures came to be emphasized more than conformance with things in the world. The key assumptions of the policy context were equilibrium and stability. Continuity was given much greater significance than change, even in the dynamic area of continuous technological innovation. In part, the emphasis on equilibrium, except for the occasional crisis that could not be ignored, reflected the desire for standardized policy interventions and stable projections in order to reduce political uncertainty and ease administrative and accountability concerns.51 Above all, simplicity granted primary importance to market forces. In the language of complexity science, the emphasis is clearly on the power of selection. There is more than a good deal of social Darwinism in the policy community’s search for simplicity. Policy models based on simplicity also dovetail nicely with the dominant American ideology of laissez-faire. Simplicity in studying and formulating technology policy guaranteed that for most of the postwar period America’s passion for free-market principles would not be challenged. A muddled collection of largely macro-level initiatives intended to facili-

lkhrwlogy

Policy in a Complex World

263

tate the technology selection process were put in place. It was also legitimate at times to subsidize technologies that had dual-use characteristics. Some of these interventions had results consistent with Asian experience and European theory. Anything of this sort undertaken by civilian U.S. agencies, however, ran the risk of being labeled picking winners.c2 Slowly, however, complexity has been creeping into the American technology policy debate. The proliferation of complex products in the international marketplace has made it more difficult to see innovation through the lens of simplicity. There are some good general descriptions of these trends and even a few policy assessments that make explicit reference to complex technologies and organizational networks as part of the changing global marketplace. However, even the analyses that recognize the importance of complexity do not translate this into their policy prescriptions for the future. The conventional wisdom still suggests that tinkering at the margin of the existing policy apparatus will suffice. That complexity demands radically different policy capabilities and actions has little currency. In the words of George Cowan, former president of the Santa Fe Institute, considered the center of complexity science, “The moment you depart from the linear approximation, you’re navigating on a very broad ocean.“53 Consider some proposals from the U.S. National Academy system. After making only the most casual mention of increasing complexity, one study emphasizes greater public-private sector cooperation, with the pivotal caveat that “as the federal government seeks to promote technological leadership more actively, it has available only the economic policy levers that permit market forces to play a major role.“54 Another study mentions complexity in passing, calling for policies to enhance investments in human capital, promote collaboration in “precompetitive” research, and support infrastructure development.55 A third analysis, the only one to be explicit about growing complexity, suggests the need for “an institutional focus” within the federal government to “monitor, harness and supplement” the existing federal technology programs. Otherwise, the proposals are mostly modifications of traditional market-oriented initiatives: R&D tax credits, government procurement changes.56 And yet, these studies devote more attention to complexity than do any others. Their weakness flows not from the fact that their prescriptions are too general. A major lesson to be drawn from complexity science is that policies cannot be specified apart from physical reality, ‘lb continue Cowan’s metaphor of navigation, consider the following comment from Arthur: Actually, you’re just the captain of a paper boat drifting down the river. If you try to resist, you’re not going to get anywhere. On the other hand, if you quietly observe the flow, realizing that you’re part of it, realizing that the flow is everchanging and always leading to new complexities, then every so often you can stick an oar into the river and punt yourself from one eddy to another. What’s the

264

R. W. Rycrofi and D. E. Kash

connection with economic and political policy? Well, in a policy context, it means that you observe, and observe, and observe, and occasionally you stick your oar in and improve something for the better. It means that you try to see reality for what it is, and realize that the game you are in keeps changing, so that it’s up to you to figure out the current rules of the game as it’s being played. It means that you observe the Japanese like hawks, you stop being naive, you stop appealing for them to play fair, you stop adhering to standard theories that are built on outmoded assumptions about the rules of play, you stop saying, ‘Well, if we could only reach equilibrium we’d be in fat city.” You just observe. And where you can make an effective move, you make a move. Note that this is not a recipe for passivity, or fatalism. This is a powerful approach that makes use of the natural nonlinear dynamics of the system.57

Current assessments of technology policy options in the U.S. continue to be built on the “outmoded assumptions” of simplicity. Heavily analytical and reductionist, these attempts are futile whether in a general form - an “institutional focus” will “monitor, harness, and supplement” existing programs - or are made more specific - a civilian technology corporation will support development of critical technologies.58 Terms like “monitor” or “critical technologies” are meaningless unless they are linked to physical reality and to particular technologies and the organizations that innovate them. Our own four-cell categorization of simple and complex products and processes illustrates the point. Monitoring the evolution of complex product technologies is a more demanding task than monitoring changes in simple ones. Similarly, infrastructure or worker training investments differ across the simple-complex continuum. The new physical reality demands, at a minimum, that policy for complex technologies be finetuned to the level of the family of processes and products that together comprise the sector. It is at the sector level that one observes and navigates the complex flow of organizational and technological evolution. It is also at this level that one has an opportunity to be effective. Thus, while it may be possible to successfully develop and implement policies for broad categories of simple technologies, the appropriate level of analysis for complex technology policy is those items displayed in the lower half of Figures 4 and 5, not “pre-competitive,” “generic,” or “critical” technologies. Moreover, complex technology policy must take into account the organizational dimension of innovation as much as it does market selection factors. While it may be adequate to rely on market forces for the innovation of simple products and processes, complex technological innovation requires the development of policy-making capabilities to enhance selforganization, learning, positive feedback, and emergence on a sector-specific basis. Many, perhaps most, Americans will find this new physical reality discomforting. The simple metaphors by which we have lived for so long the “invisible hand, n “the rugged individual” - are no longer applicable

l%chnology Policy in a Complex

World

265

to the commercialization of the most valuable technologies in the global marketplace. Nonetheless, as Heinz Pagels argues: I am convinced that the societies that master the new sciences of complexity and can convert their knowledge into new products and forms of social organization will become the cultural, economic, and military superpowers of the next century. While there is great hope in this development, there is also the terrible danger that this new salient in knowledge will aggravate the differences between those who possess it and those who do not.5g

Pagels’ warning about the mixture of hope and danger is compelling. Countries not able or willing to embrace complexity as the concept that links Asian experience with evolutionary economic theory are very likely to find themselves at great risk in the global political economy. America is no exception.

Notes 1. President W. J. Clinton and Vice President A. Gore, Jr., Z&hnology for America’s Economic Growth: A New Direction to Build Economic Strength (Washington: Executive Offme of the President, February 22, 19931, p. 7. 2. E. Marshall, “Technology Boosting: A Checkered History,” Science, March 26, 1993, p. 1817. 3. R. E. McGinn, Science, Technology, and Society (Englewood Cliffs, NJ: Prentice-Hall, 19911, p. 32. 4. M. A. Cusumano and K. Nobeoka, “Strategy, Structure and Performance in Product Development: Observations from the Auto Industry,” Research Policy, 1992, pp. 265-293. 5. Waldrop, op. cit., pp. 327330. 6. E. B. Skolnikoff, The Elusive Transformation: Science, Technology, and the Evolution of International Politics (Princeton, NJ: Princeton University Press, 19931, p. 14. 7. D. E. Kash, Perpetual Innovation: The New World of Competition (New York: Basic Books, 1989), pp. 16-37. 8. J. B. Bush, Jr., and A. L. Frohman, “Communication in a ‘Network’ Organization,” Organizational Dynamics, Autumn 1991, pp. 23-25. 9. R. Nagel and R. Dove, 21st Century Manufacturing Enterprise Strategy: An Industry-Led View, Volume I, (Bethlehem, PA: Iacocca Institute, Lehigh University, 1991), p. 2. 10. R. M. Henderson and K. B. Clark, “Architectural Innovation: The Reconfiguration of Existing Product Technologies and the Failure of Established Firms,” Administrative Science Quarterly, March 1990, pp. 9-30. 11. R. Florida and M. Kenney, Beyond Mass Production: The Japanese System and its ZFansfer to the United States (New York: Oxford University Press, 1993); M. Piore and C. Sabel, The Second Industrial Divide (New York: Basic Books, 19841, pp. 19-104. 12. C. R. Morris and C. H. Ferguson, “How Architecture Wins Technology Wars,” Harvard Business Review, March-April 1993, pp. 86-96. 13. D. Woodruff and J. B. Levine, “Miles Traveled, More to Go,” Business Week, Special Issue on Quality, 1991, p. 71. 14. W. B. Arthur, ‘Why Do Things Become More Complex?” Scientific American, May 1993, p. 144. 15. S. J. Kline, “A Numerical Index for the Complexity of Systems: The Concept and Some Implications,” Proceedings of November 1990 Conference, Association for Computing Machinery on Managing Complexity and Modeling Reality (New York: ACM Press, 1991). 16. D. E. Kash and R. W. Rycroft, “Nurturing Winners with Federal R&D,” !BxhnoZogy Review, November/December 1993, p. 61. 17. D. E. Kash and R. W. Rycroft, “‘l%o Streams of Technological Innovation: Implications for Policy,” Science and Public Policy, February 1993, p. 30.

266

R. W. Rycroft and D. E. Kash

18. W. Bechtel and R. C. Richardson, Discovering Complexity: Decomposition and Localization as Strategies in Scientific Research (Princeton, NJ: Princeton University Press, 1993), p. 18. 19. K. B. Clark and T. Fujimoto, “The Power of Product Integrity,” Harvard Business Review, November-December 1990, p. 107. 20. D. J. ‘Ibece, “Inter-Organizational Requirements of the Innovation Process,” Managerial and Decision Economics, Spring 1989, p. 35. 21. M. M. Waldrop, Complexity: The Emerging Science at the Edge of Order and Chnos (New York: Simon and Schuster, 19921, p. 102. 22. I. Prigogine and I. Stengers, Or&r Out of Chaos (New York: Bantam Books, 1985); G. Nicolis and I. Prigogine, Self-Organization in Non-Equilibrium Systems (New York: Wiley Interscience, 1977). 23. J. Foster, “Economics and the Self-Organization Approach: Alfred Marshall Revisited,” The Economic Journal, July 1003, pp. 985-987. 24. M. J. Radzicki, “Institutional Dynamics, Deterministic Chaos, and Self-Organizing Systems,” Journal ofEcorwmic Issues, March 1990, p. 82. 25. R. R. Nelson, “Why Do Firms Differ and How Does It Matter?” Strategic Management Journal, Winter 1991, p. 64. 26. G. Romme, “The Formation of Firm Strategy as Self-Organization,” in C. Freeman and L. Soete (eds.), New Explorations in the Economics of !l&hnical Change (New York: Pinter Press, 1990), p. 40. 27. G. Morgan, Images of Organization (Beverly Hills, CA: Sage Publications, 1986), p. 254. 28. J. A. Byrne, R. Brandt, and 0. Port, “The Virtual Corporation,” Business Week, February 8, 1993, pp. 98-102. 29. T. A. Stewart, “The Search for the Organization of the Future,” Fortune, May 18,1992, p. 98. 30. G. Morgan, op. cit., p. 103. 31. D. A. Garvin, “Building a Learning Organization,” Harvard Business Review, July-August 1993, p. 80. 32. I. Nom&a, “The Knowledge-Creating Company,” Harvard Business Review, November-December 1991, p. 98. 33. C. F. Wojslaw, “Teamwork and Community,” IEEE Technology and Society Magazine, Winter 1992l93, pp. 23-27. 34. D. McKee, “An Organizational Learning Approach to Product Innovation,” Journal ofProduct Innovation Management, 1992, pp. 235-240. 35. J. B. Bush, Jr., and A. L. Frohman, “Communication in a ‘Network’ Organization,” Organizational Dynamics, Autumn 1991, p. 33. 36. I? W. Meyers, “Non-Linear Learning in Large Technological Firms: Period Four Implies Chaos,” Research Policy, April 1990, p. 97. 37. D. Leonard-Barton, “The Factory as Learning Laboratory,” Sloan Management Review, Fall 1992, pp. 33-35. 38. P. S. Adler and R. E. Cole, “Designed for Learning: A Tale of Two Auto Plants,” Sloan Management Review, Spring 1993, pp. 92-93. 39. A. M. Webber, “What’s So New About the New Economy?” Harvard Business Review, January-February 1993, pp. 20-42. 40. W. B. Arthur, “Positive Feedbacks in the Economy,” Scientific American, February 1990, pp. 92-93. 41. R. Cowan, “‘Ibrtoises and Hares: Choice Among Technologies of Unknown Merit,” The Economic Journal, July 1991, pp. 809-810. 42. W. B. Arthur, “Self-Reinforcing Mechanisms in Economics,” in P W. Anderson, K. J. Arrow and D. Pines (eds.), The Economy as an Evolving Complex System (New York: Addison-Wesley Publishing Company, 1989), pp. 10-17. 43. Y. Baba and K.-I. Imai, “A Network View of Innovation and Entrepreneurship: The Case of the Social Science Journal, February 1993, Evolution of the VCR Systems,” International pp. 23-34. 44. D. J. Teece, “Competition, Cooperation, and Innovation: Organizational Arrangements for Regimes of Rapid Technological Progress,” Journal of Economic Behavior and Organization, June 1992, p. 22. 45. P W. Anderson, “Is Complexity Physics? Is It Science? What Is It?” Physics Zbday, July 1991, p. 9. 46. W. B. Arthur, “Pandora’s Marketplace,” New Scientist (Complexity Supplement), February 6, 1993, pp. 6-7.

Technology Policy in a Complex World

267

47. K. Kreiner and M. Schultz, %aformal Collaboration in R&D: The Formation of Networks Across Organizations,” Organization Studies, 1993, p. 200 (emphasis added). 48. J. H. Holland, “Complex Adaptive Systems,” Daedalus, Winter 1992, p. 20. 49. S. A. Kaufman, “Principles of Adaptation in Complex Systems,” in D. L. Stein (ed.), Lectures in the Sciences of Complezity (New York: Addison-Wesley Publishing Co., 19891, pp. 675-676. 50. Quoted in Waldrop, op. cit., p. 118. 51. L. D. Kiel, “The Nonlinear Paradigm: Advancing Paradigmatic Progress in the Policy Sciences,” Systems Research, 1992, pp. 31-36. 52. R. W. Rycroft and D. E. Kash, “Technology Policy Requires Picking Winners,” Economic Development Quarterly, August 1992, pp. 227-240. 53. Quoted in Woldrop, op. cit., p. 66. 54. Committee on Science, Engineering and Public Policy, National Academy of Sciences/National Academy of Engineering/Institute of Medicine, Science, Technology and the Federal Government: National Goals for a New Era, (Washington: National Academy Press, 1993), p. 6. 55. M. C. Harris and G. E. Moore (eds.), Linking Dade and Ilkchnology Policies: Prospering in a Global Economy, (Washington: National Academy Press, 19931, pp. 8-10. 56. Committee on Technology Policy Options in a Global Economy, National Academy of Engineering, Mastering a New Role: Prospering in a Global Economy (Washington: National Academy Press, 19931, pp. 2-6. 57. Quoted in Waldrop, op. cit., pp. 330-331. 58. Committee on Science, Engineering and Public Policy, National Academy of Sciences/National Academy of EngineeringAnstitute of Medicine, The Government Role in Civilian Txhnology: Building a New Alliance (Washington: National Academy Press, 1992). 59. H. Pagels, The Dreams of Reason: The Computer and the Rise of the Sciences of Complexity (New York: Simon and Schuster, 19881, p. 53.