Mode 2'' to a Triple Helix of University–Industry–Government

Research Policy 29Ž. 2000 109±123 www.elsevier.nlrlocatereconbase

The dynamics of innovation: from National Systems and ``Mode 2'' to a Triple Helix of university±industry±government relations

Henry Etzkowitz a,), Loet Leydesdorff b,1 a Science Policy Institute, Social Science DiÕision, State UniÕersity of New York at Purchase, 735 Anderson Hill Road, Purchase, NY 10577-1400, USA b Department of Science and Technology Dynamics, Nieuwe Achtergracht 166, 1018 WV Amsterdam, Netherlands

Abstract

The Triple Helix of university±industry±government relations is compared with alternative models for explaining the current research system in its social contexts. Communications and negotiations between institutional partners generate an overlay that increasingly reorganizes the underlying arrangements. The institutional layer can be considered as the retention mechanism of a developing system. For example, the national organization of the system of innovation has historically been important in determining competition. Reorganizations across industrial sectors and nation states, however, are induced by new technologiesŽ. biotechnology, ICT . The consequent transformations can be analyzed in terms of Ž. neo- evolutionary mechanisms. University research may function increasingly as a locus in the ``laboratory'' of such knowledge-intensive network transitions. q 2000 Elsevier Science B.V. All rights reserved.

Keywords: Mode 2; Triple helix; University±industry±government relations; Innovation

1. Introduction: From the endless frontier to an and expectations that reshape the institutional ar- endless transition rangements among universities, industries, and gov- ernmental agencies. The Triple Helix thesis states that the university As the role of the military has decreased and can play an enhanced role in innovation in increas- academia has risen in the institutional structures of ingly knowledge-based societies. The underlying contemporary societies, the network of relationships model is analytically different from the national sys- among academia, industry, and government have tems of innovationŽ. NSI approach Ž Lundvall, 1988, also been transformed, displacing the Cold War 1992; Nelson, 1993. , which considers the firm as ``Power Elite'' trilateral mode of MillsŽ. 1958 with having the leading role in innovation, and from the an overlay of reflexive communications that increas- ``Triangle'' model of SabatoÂ Ž. 1975 , in which the ingly reshape the infrastructureŽ Etzkowitz and Ley- state is privilegedŽ. cf. SabatoÂ and Mackenzi, 1982 . desdorff, 1997. . Not surprisingly, the effects of these We focus on the network overlay of communications transformations are the subject of an international debate over the appropriate role of the university in ) Corresponding author. Tel.: q1-914-251-6600; fax: q1-914- technology and knowledge transfer. For example, the 251-6603; e-mail: [email protected] Swedish Research 2000 Report recommended the 1 E-mail: [email protected]. withdrawal of the universities from the envisaged

``third mission'' of direct contributions to industry the Small Business Technology Transfer Program Ž.see Benner and Sandstrom,È this issue . Instead, the Ž.STTR , the Advanced Technology Program Ž. ATP , university should return to research and teaching the IndustryrUniversity Cooperative Research Cen- tasks, as traditionally conceptualized. However, it tersŽ. IUCRC and Engineering Research Centers can be expected that proponents of the third mission Ž.ERC of the National Science Foundation, etc. from the new universities and regional colleges, Ž.Etzkowitz et al., 2000 . In Sweden, the Knowledge which have based their research programmes on its Competency Foundation and the Technology Bridge premises, will continue to make their case. Science Foundation were established as public venture capi- and technology have become important to regional tal source, utilizing the Wage Earners Fund, origi- developmentsŽ. e.g., Braczyk et al., 1998 . Both R&D nally intended to buy stock in established firms on and higher education can be analyzed also in terms behalf of the public. The beginnings of a Swedish of marketsŽ. Dasgupta and David, 1994 . movement to involve academia more closely in this The issues in the Swedish debate are echoed in direction has occasioned a debate similar to the one the critique of academic technology transfer in the that took place in the US in the early 1980s. At that USA by several economistsŽ e.g., Rosenberg and time, Harvard University sought to establish a firm Nelson, 1994. . The argument is that academic tech- jointly with one of its professors, based on his nology-transfer mechanisms may create unnecessary research results. transaction costs by encapsulating knowledge in Can academia encompass a third mission of eco- patents that might otherwise flow freely to industry. nomic development in addition to research and But would the knowledge be efficiently transferred teaching? How can each of these various tasks con- to industry without the series of mechanisms for tribute to the mission of the university? The late 19th identifying and enhancing the applicability of re- century witnessed an academic revolution in which search findings? How are development processes to research was introduced into the university mission be carried further, through special grants for this and made more or less compatible with teaching, at purpose or in new firms formed on campus and in least at the graduate level. Many universities in the university incubator facilities? USA and worldwide are still undergoing this trans- The institutional innovations aim to promote closer formation of purpose. The increased salience of relations between faculties and firms. The ``endless knowledge and research to economic development frontier'' of basic research funded as an end in itself, has opened up a third mission: the role of the with only long-term practical results expected, is university in economic development. A ``second being replaced by an ``endless transition`` model in academic revolution'' seems under way since World which basic research is linked to utilization through War II, but more visibly since the end of the Cold a series of intermediate processesŽ. Callon, 1998 , WarŽ. Etzkowitz, forthcoming . often stimulated by government. In the USA in the 1970s, in various Western The linear model either expressed in terms of European countries during the 1980s, and in Sweden ``market pull'' or ``technology push'' was insuffi- at present, this transition has led to a reevaluation of cient to induce transfer of knowledge and technol- the mission and role of the university in society. ogy. Publication and patenting assume different sys- Similar controversies have taken place in Latin tems of reference both from each other and with America, Asia, and elsewhere in Europe. The Triple reference to the transformation of knowledge and Helix series of conferencesŽ Amsterdam, 1996; Pur- technology into marketable products. The rules and chase, New York, 1998; and Rio de Janeiro, 2000. regulations had to be reshaped and an interface have provided a venue for the discussion of theoreti- strategy invented in order to integrate market pull cal and empirical issues by academics and policy and technology push through new organizational analystsŽ. Leydesdorff and Etzkowitz, 1996, 1998 . mechanismsŽ e.g., OECD, 1980; Rothwell and Different possible resolutions of the relations among Zegveld, 1981. . the institutional spheres of university, industry, and In the USA, these programs include the Small government can help to generate alternative strate- Business Innovation Research programŽ. SBIR and gies for economic growth and social transformation. H. Etzkowitz, L. LeydesdorffrResearch Policy 29() 2000 109±123 111

2. Triple Helix configurations

The evolution of innovation systems, and the current conflict over which path should be taken in university±industry relations, are reflected in the varying institutional arrangements of university±industry±government relations. First, one can distinguish a specific historical situation which one may wish to label Triple Helix I. In this configuration the nation state encompasses academia and industry and directs the relations between themŽ. Fig. 1 . The strong version of this model could be found in the former Soviet Union and in Eastern European countries under ``existing socialism''. Weaker versions were formulated in the policies of many Latin Amer- ican countries and to some extent in European countries such as Norway. Fig. 2. A ``laissez-faire'' model of university±industry±govern- A second policy modelŽ. Fig. 2 consists of sepa- ment relations. rate institutional spheres with strong borders dividing them and highly circumscribed relations among the institutional spheres, with each taking the role of the spheres, exemplified in Sweden by the noted Re- other and with hybrid organizations emerging at the search 2000 Report and in the US in opposition to interfacesŽ. Fig. 3 . the various reports of the Government±University± The differences between the latter two versions of Ž. Industry Research Roundtable GUIRR of the Na- the Triple Helix arrangements currently generate Ž tional Research Council MacLane, 1996; cf. GUIRR, normative interest. Triple Helix I is largely viewed 1998. . Finally, Triple Helix III is generating a knowledge infrastructure in terms of overlapping

Fig. 1. An etatistic model of university±industry±government Fig. 3. The Triple Helix Model of University±Industry±Govern- relations. ment Relations. 112 H. Etzkowitz, L. LeydesdorffrResearch Policy 29() 2000 109±123 as a failed developmental model. With too little work because of the difference between cultural and room for ``bottom up'' initiatives, innovation was biological evolutions. Biological evolution theory as- discouraged rather than encouraged. Triple Helix II sumes variation as a driver and selection to be entails a laissez-faire policy, nowadays also advo- naturally given. Cultural evolution, however, is cated as shock therapy to reduce the role of the state driven by individuals and groups who make con- in Triple Helix I. scious decisions as well as the appearance of unin- In one form or another, most countries and re- tended consequences. A Triple Helix in which each gions are presently trying to attain some form of strand may relate to the other two can be expected to Triple Helix III. The common objective is to realize develop an emerging overlay of communications, an innovative environment consisting of university networks, and organizations among the helicesŽ Fig. spin-off firms, tri-lateral initiatives for knowledge- 4.. based economic development, and strategic alliances The sources of innovation in a Triple Helix con- among firmsŽ large and small, operating in different figuration are no longer synchronized a priori. They areas, and with different levels of technology. , gov- do not fit together in a pregiven order, but they ernment laboratories, and academic research groups. generate puzzles for participants, analysts, and poli- These arrangements are often encouraged, but not cymakers to solve. This network of relations gener- controlled, by government, whether through new ates a reflexive subdynamics of intentions, strategies, ``rules of the game,'' direct or indirect financial and projects that adds surplus value by reorganizing assistance, or through the Bayh±Dole Act in the and harmonizing continuously the underlying infras- USA or new actors such as the abovementioned foundations to promote innovation in Sweden.

3. The Triple Helix of innovation

The Triple Helix as an analytical model adds to the description of the variety of institutional arrangements and policy models an explanation of their dynamics. What are the units of operation that inter- act when a system of innovation is formed? How can such a system be specified? In our opinion, typifications in terms of ``national systems of innovation''Ž Lundvall, 1988; Nelson, 1993.Ž ; ``research systems in transition'' Cozzens et al., 1990; Ziman, 1994.Ž , Mode 2 Gibbons et al., 1994.Ž or ``the post modern research system'' Rip and Van der Meulen, 1996. are indicative of flux, reorganization, and the enhanced role of knowledge in the economy and society. In order to explain these observable reorganizations in university±industry± government relations, one needs to transform the sociological theories of institutional retention, re- combinatorial innovation, and reflexive controls. Each theory can be expected to appreciate a different subdynamicŽ. Leydesdorff, 1997 . In contrast to a double helixŽ or a coevolution Fig. 4. The overlay of communications and expectations at the between two dynamics. , a Triple Helix is not ex- network level guides the reconstruction of institutional arrange- pected to be stable. The biological metaphor cannot ments. H. Etzkowitz, L. LeydesdorffrResearch Policy 29() 2000 109±123 113 tructure in order to achieve at least an approximation economics in which the three functional mechanisms of the goals. The issue of how much we are in are: technological innovation provides the variation, control or non-control of these dynamics specifies a markets are the prevailing selectors, and the institu- research program on innovation. tional structures provide the system with retention Innovation systems, and the relationships among and reflexive controlŽ. Nelson, 1994 . In advanced them, are apparent at the organizational, local, re- and pluriform societies, the mechanisms of institu- gional, national, and multinational levels. The inter- tional control are again differentiated into public and acting subdynamics, that is, specific operations like private domains. Thus, a complex system is devel- markets and technological innovations, are continu- oped that is continuously integrated and differenti- ously reconstructed like commerce on the Internet, ated, both locally and globally. yet differently at different levels. The subdynamics Innovation can be defined at different levels and and the levels are also reflexively reconstructed from different perspectives within this complex dy- through discussions and negotiation in the Triple namics. For example, evolutionary economists have Helix. What is considered as `ìndustry'', what as argued that one should consider firms as the units of ``market'' cannot be taken for granted and should analysis, since they carry the innovations and they not be reified. Each ``system'' is defined and can be have to compete in marketsŽ Nelson and Winter, redefined as the research project is designed. 1982; cf. Andersen, 1994. . From a policy perspec- For example, national systems of innovation can tive, one may wish to define national systems of be more or less systemic. The extent of systemness innovation as a relevant frame of reference for gov- remains an empirical questionŽ Leydesdorff and ernment interventions. Others have argued in favour Oomes, 1999.Ž. . The dynamic ``system s of innova- of networks as more abstract units of analysis: the tion'' may consist of increasingly complex collabora- semi-autonomous dynamics of the networks may tions across national borders and among researchers exhibit lock-ins, segmentation, etc.Ž e.g., David and and users of research from various institutional Foray, 1994. . Furthermore, the evolving networks spheresŽ. Godin and Gingras, this issue . There may may change in terms of relevant boundaries while be different dynamics among regions. The systems developingŽ. Maturana, 1978 . of reference have to be specified analytically, that is, In our opinion, these various perspectives open as hypotheses. The Triple Helix hypothesis is that windows of appreciation on the dynamic and com- systems can be expected to remain in transition. The plex processes of innovation, but from specific an- observations provide an opportunity to update the gles. The complex dynamics is composed of subdy- analytical expectations. namics like market forces, political power, institutional control, social movements, technological trajectories and regimes. The operations can be ex- 4. An endless transition pected to be nested and interacting. Integration, for example, within a corporation or within a nation The infrastructure of knowledge-intensive state, cannot be taken for granted. Technological economies implies an endless transition. Marx's great innovation may also require the reshaping of an vision that `àll that is solid, melts into air''Ž Ber- organization or a communityŽ Freeman and Perez, man, 1982. underestimated the importance of seem- 1988. . But the system is not deterministic: in some ingly volatile communications and interactions in phases intentional actions may be more successful in recoding theŽ. complex network system. Particularly, shaping the direction of technological change than in when knowledge is increasingly utilized as a re- othersŽ. Hughes, 1983 . source for the production and distribution system, The dynamics are nonlinear while both the inter- reconstruction may come to prevail as a mode of action terms and the recursive terms have to be ``creative destruction''Ž Schumpeter, 1939, 1966; declared. First, there are ongoing transformations Luhmann, 1984. . within each of the helices. These reconstructions can Can the reconstructing forces be specified? One be considered as a level of continuous innovations mode of specification is provided by evolutionary under pressure of changing environments. When two 114 H. Etzkowitz, L. LeydesdorffrResearch Policy 29() 2000 109±123 helices are increasingly shaping each other mutually, the analytical position of the Triple Helix model in coevolution may lead to a stabilization along a tra- relation to other nonlinear models of innovation, like jectory. If more than a single interface is stabilized, Mode 2 and national systems of innovation. the formation of a globalized regime can be expected. At each level, cycles are generated which guide the phasing of the developments. The higher- 5. Nonlinear models of innovation order transformationsŽ. longer-term are induced by the lower-order ones, but the latter can seriously be As noted, nonlinear models of innovation extend disturbed by events at a next-order system's level upon linear models by taking interactive and recur- Ž.Schumpeter, 1939; Kampmann et al., 1994 . sive terms into account. These nonlinear terms can Although this model is abstract, it enables us to be expected to change the causal relations between specify the various windows of theoretical apprecia- input and output. The production rules in the systems tion in terms of their constitutive subdynamicsŽ e.g., under study, for example, can be expected to change Leydesdorff and Van den Besselaar, 1997. . The with the further development of the inputroutput different subdynamics can be expected to select upon relationsŽ. e.g., because of economies of scale . Thus, each other asymmetrically, as in processes of negoti- the unit of operation may be transformed, as is ation, by using their specific codes. For example, the typical when a pilot plant in the chemical industry is markets and networks select upon technological fea- scaled up to a production facility. sibilities, whereas the options for technological de- By changing the unit of analysis or the unit of velopments can also be specified in terms of market operation at the reflexive level, one obtains a differ- forces. Governments can intervene by helping create ent perspective on the system under study. But the a new market or otherwise changing the rules of the system itself is also evolving. In terms of methodolo- game. gies, this challenges our conceptual apparatus, since When the selections ``lock-in'' upon each other, one has to be able to distinguish whether the variable next-order systems may become relevant. For exam- has changed or merely the value of the variable. The ple, airplane development at the level of firms gener- analysis contains a snapshot, while the reality pro- ates trajectories at the level of the industry in coevo- vides a moving picture. One needs metaphors to lutions between selected technologies and markets reduce the complexity for the discursive understand- Ž.e.g., Nelson, 1994; cf. McKelvey, 1996 . Nowa- ing. Geometrical metaphors can be stabilized by days, the development of a new technological trajec- higher-order codifications as in the case of paradigms. tory invokes the support of national governments and The understanding in terms of fluxesŽ that is, how even international levelsŽ. like the EU , using increas- the variables as well as the value may change over ingly a Triple Helix regimeŽ Frenken and Leydes- time. , however, calls for the use of algorithmic dorff, forthcoming. . simulations. The observables can then be considered We have organized this theme issue about the as special cases which inform the expectations Triple Helix of university±industry±government re- Ž.Leydesdorff, 1995 . lations in terms of three such interlocking dynamics: Innovation, in particular, can be defined only in institutional transformations, evolutionary mecha- terms of an operation. Both the innovatorŽ. s and the nisms, and the new position of the university. This innovated systemŽ. s are expected to be changed by approach allows us to pursue the analysis at the the innovation. Furthermore, one is able to be both a network level and then to compare among units of participant and an observer, and one is also able to analysis. For example, both industries and govern- change perspectives. In the analysis, however, the ments are entrained in institutional transformations, various roles are distinguished although they can while the institutional transformations themselves sometimes be fused in ``real life'' events. Langton change under the pressure of information and com- Ž.1989 proposed to distinguish between the ``pheno- munication technologiesŽ. ICT or government poli- typical'' level of the observables and the ``genotypi- cies. Before explaining the organization of the theme cal'' level of analytical theorizing. The ``pheno- issue in detail, however, we wish to turn briefly to types'' remain to be explained and the various expla- H. Etzkowitz, L. LeydesdorffrResearch Policy 29() 2000 109±123 115 nations compete in terms of their clarity and useful- which are possible within and among the sectors. In ness for updating the expectations. Confusion, how- particular, the distribution of relevant actors contains ever, is difficult to avoid given the pressure to jump an heuristic potential which can be made reflexive to normative conclusions, while different perspec- by a strategic analysis of specific strengths and tives are continuously competing, both normatively weaknessesŽ. Pavitt, 1984 . and analytically. The solution of the production puzzle typically Let us first focus on the problem of the unit of brings government into the picture shifting the dy- analysis and the unit of operation. In addition to namics from a double to a triple helix. The conse- extending the linearŽ. inputroutput models of neo- quent processes of negotiation are both complex and classical and business economics, evolutionary dynamic: one expects that theŽ. institutional actors economists also changed the unit of analysis. will be reproduced and changed by the interactions. Whereas neoclassical economics focused on markets Trilateral networks and hybrid organizations are cre- as networks in terms of inputroutput relations among ated for resolving social and economic crises. The individualŽ. rational agents, evolutionary economists actors from the different spheres negotiate and define have tended to focus on firms as the specificŽ and new projects, such as the invention of the venture bounded. carriers of an innovation process. Both the capital firm in New England in the early postwar era unit of analysis and the unit of operation were Ž.Etzkowitz, forthcoming . Thus, a Triple Helix dy- changedŽ. Andersen, 1994; cf. Alchian, 1950 . namics of university±industry±government relations LundvallŽ. 1988, at p. 357 noted that the interac- is generated endogeneously. tive terms between demand and supply in user±pro- Gibbons et al.Ž. 1994 argued that this ``new mode ducer relations assume a system of reference in of the production of scientific knowledge'' has be- addition to the market. The classical dispute in inno- come manifest. But: how are these dynamics in the vation theory had, in his opinion, referred to the role network arrangements between industries, govern- of demand and supply, that is, market forces, in ments, and academia a consequence of the user±pro- determining the rate and direction of the process of ducer interactions foregrounded by LundvallŽ. 1988 ? innovationŽ cf. Mowery and Rosenberg, 1979; Free- Are national systems still a relevant unit of analysis? man, 1982, p. 211. . If, however, the dynamics of Since the new mode of knowledge productionŽ Mode innovationŽ. e.g., product competition are expected 2. is characterized as an outcome, it should, in our to be different from the dynamics of the marketŽ e.g., opinion, be considered as an emerging system. The price competition. , an alternative system of reference emerging system rests like a hyper-network on the for the selection should also be specified. For this networks on which it buildsŽ such as the disciplines, purpose, Lundvall proposed ``to take the national the industries, and the national governments. , but the system of production as a starting point when defin- knowledge-economy transforms ``the ship while a ing a system of innovation''Ž. p. 362 . storm is raging on the open sea''Ž Neurath et al., Lundvall added that the national system of pro- 1929. . duction should not be considered as a closed system: Science has always been organized through net- ``the specific degree and form of openness deter- works, and to pursue practical as well as theoretical mines the dynamics of each national system of pro- interests. Centuries before ``Mersenne'' was trans- duction''. In our opinion, as a first step, innovation mogrified into an Internet site, he was an individual, systems should be considered as the dynamics of who by visits and letters, knitted the European scien- change in systems of both production and distribu- tific community together. The Academies of Science tion. From this perspective, national systems com- played a similar role in local and national contexts pete in terms of the adaptability of their knowledge from the 16th century. infrastructure. How are competencies distributed for The practical impetus to scientific discovery is solving ``the production puzzle'' which is generated long-standing. The dissertation of MertonŽ. 1938 by uneven technological developments across sectors reported that between 40% and 60% of discoveries Ž.Nelson, 1982; Nelson and Winter, 1975 ? The in- in the 17th century could be classified as having frastructure conditions the processes of innovation their origins in trying to solve problems in naviga- 116 H. Etzkowitz, L. LeydesdorffrResearch Policy 29() 2000 109±123 tion, mining, etc. Conversely, solution of practical from corruption by the Nazi doctrine of a racial basis problems through scientific means has been an im- for science and from Lysenko's attack on genetics in portant factor in scientific development, whether in the Soviet Union. Merton's formulation of a set of German pharmaceutical science in the 17th century norms to protect the free space of science was Ž.Gustin, 1975 or in the British-sponsored competi- accepted as the basis for an empirical sociology of tion to provide a secure basis for navigationŽ Sobel, science for many years. 1995. . The third element in establishing the ideology of The so-called Mode 2 is not new; it is the original pure science was, of course, the Bush Report of format of science before its academic institutional- 1945. The huge success of science in supplying ization in the 19th century. Another question to be practical results during World War II in one sense answered is why Mode 1 has arisen after Mode 2: supplied its own legitimation for science. But with the original organizational and institutional basis of the end of the war at hand and wanting to insure that science, consisting of networks and invisible colleges science was funded in peacetime, a rationale was Ž.cf. Weingart, 1997; Godin, 1998 . Where have these needed in 1944 when Bush persuaded President Roo- ideas, of the scientist as the isolated individual and sevelt to write a letter commissioning the report of science separated from the interests of society, Ž.Bush, 1980 . come from? Mode 2 represents the material base of In the first draft of his report, Bush proposed to science, how it actually operates. Mode 1 is a con- follow the then current British method of funding struct, built upon that base in order to justify auton- science at universities. It would be distributed on a omy for science, especially in an earlier era when it per capita basis according to the number of students was still a fragile institution and needed all the help at each school. In the contemporary British system of it could get. a small number of universities, the funds automati- In the USA, during the late 19th century, large cally went to an elite. However, if that model had fortunes were given to found new universities, and been followed in the US, even in the early postwar expand old ones. There were grave concerns among era, the flow of funds would have taken a different many academics that the industrialists making these course. The funding would not only have flowed gifts would try to directly influence the universities, primarily to a bicoastal academic elite but would by claiming rights to hire and fire professors as well have been much more broadly distributed across the as well as to decide what topics were acceptable for academic spectrum, especially to the large state uni- research and instructionŽ. Storr, 1953 . To carve out versities in the Midwest. an independent space for science, beyond the control In the time between the draft and the final report, of economic interests, a physicist, Henry Rowland, the mechanism for distribution of government funds propounded the doctrine that if anyone with external to academic research was revised and ``peer review'' interests tried to intervene, it would harm the con- was introduced. Adapted from foundation practices duct of science. As president of the American Asso- in the 1920s and 1930s, it could be expected that ciation for the Advancement of Science, he promoted ``the peers'', the leading scientists who would most the ideology of pure research in the late 19th cen- surely be on those committees, would distribute the tury. Of course, at the same time as liberal arts funds primarily to a scientific elite. The status sys- universities oriented toward pure research were be- tem of U.S. universities that had been in place from ing founded, land grant universities, including MIT, the 1920s was reinforced. pursued more practical research strategies. These two This model of ``best science'' is no longer accept- contrasting academic modes existed in parallel for able to many as the sole basis for distribution of many years. public research funds. Congresspersons who repre- Decades hence, Merton posited the normative sent regions with universities that are not significant structure of science in 1942 and strengthened the recipients of research funds have disregarded peer ideology of ``pure science.'' His emphasis on uni- review and distributed research funds by direct ap- versalism and skepticism was a response to a particu- propriation, much as roads and bridges are often lar historical situation, the need to defend science sited through ``log rolling'' and ``pork barrel'' pro- H. Etzkowitz, L. LeydesdorffrResearch Policy 29() 2000 109±123 117 cesses. Nevertheless, these politically directed funds little concern. The relationship between the site where support also serious scientific research and instru- knowledge is produced and its eventual utilization mentation projects. Even when received by schools was not seen to be tightly linked, even as a first with little or no previous research experience, these mover advantage. This view has changed dramati- `òne-time funds'' are typically used to rapidly build cally in recent years, as has the notion that high-tech up competencies in order to compete within the conurbations, like Route 128 and Silicon Valley, are peer-review system. unique instances that cannot be replicated. The more Indeed, when a leading school, Columbia Univer- recent emergence of Austin, TX, for example, is sity, needed to renew the infrastructure of its chem- based in part on the expansion of research at the istry department, it contracted with the same lobby- University of Texas, aided by state as well as indus- ing firm in Washington, DC as less well-known try and federal funds. schools. Through public relations advice, Columbia Less research-intensive regions are by now well relabeled its chemistry department ``The National aware that science, applied to local resources, is the Center for Excellence in Chemistry''. A special fed- basis of much of their future potential for economic eral appropriation was made and the research facili- and social development. In the USA, it is no longer ties were renovated and expanded. To hold its fac- acceptable for research funds to primarily go to the ulty, the university could not afford to wait for the east and west coasts with a few places in between in slower route of peer review, and likely smaller the Midwest. The reason why funding is awarded on amounts of funding. bases other than the peer review system, is that all Increasing competition for research funds among regions want a share of research funding. new and old actors has caused an incipient break- The classic legitimation for scientific research as down of ``peer review'', a system that could best a contribution to culture still holds and military and adjudicate within a moderate level of competition. health objectives also remain a strong stimulus to As competition for research funds continues to ex- research funding. Nevertheless, the future legitima- pand, how should the strain be adjusted? Some tion for scientific research, which will keep funding propose shrinking the research system; others sug- at a high level, is that it is increasingly the source of gest linking science to new sources of legitimation new lines of economic development. such as regional development. Newly created disciplines are often the basis for these heightened expectations. Such disciplines do not arise only from the subdivision of new disci- 6. The future legitimation of science plines from old ones, as in the 19th centuryŽ Ben David and Collins, 1966. . New disciplines have It is nowadays apparent that the development of arisen, more recently, through syntheses of practical science provides much of the basis for future indus- and theoretical interests. For example, computer sci- trial development. These connections, however, have ence grew out of elements of older disciplines such been present from the creation of science as an as electrical engineering, psychology, philosophy, organized activity in the 17th century. Marx pointed and a machine. Materials science and other fields them out again in the mid-19th century in connection such as nanotechnology that are on every nation's with the development of chemical industry in Ger- critical technology list were similarly created. many. At the time, he developed a thesis of the The university can be expected to remain the core growth of science-based industry on the basis of a institution of the knowledge sector as long as it single empirical example: Perkins researches on retains its original educational missionŽ Etzkowitz, dyestuffs in the UK leading to the development of an Webster, Gebhardt, and Terra, this issue. . Teaching industry in Germany. is the university's comparative advantage, especially The potential of science to contribute to economic when linked to research and economic development. development has become a source of regional and Students are also potential inventors. They represent international competition at the turn of the millen- a dynamic flow-through of ``human capital'' in aca- nium. Until recently, the location of research was of demic research groups, as opposed to more static 118 H. Etzkowitz, L. LeydesdorffrResearch Policy 29() 2000 109±123 industrial laboratories and research institutes. Al- more productive and cost effective to combine the though they are sometimes considered a necessary two functions. distraction, the turnover of students insures the pri- The Triple Helix overlay provides a model at the macy of the university as a source of innovation. level of social structure for the explanation of Mode The university may be compared to other recently 2 as an historically emerging structure for the pro- proposed contenders for knowledge leadership, such duction of scientific knowledge, and its relation to as the consulting firm. A consulting company draws Mode 1. First, the arrangements between industry together widely dispersed personnel for individual and government no longer need to be conceptualized projects and then disperses them again after a pro- as exclusively between national governments and ject, solving a client's particular problem, is com- specific industrial sectors. Strategic alliances cut pleted. Such firms lack the organizational ability to across traditional sector divides; governments can act pursue a cumulative research program as a matter of at national, regional, or increasingly also at interna- course. The university's unique comparative advan- tional levels. Corporations adopt ``global'' postures tages is that it combines continuity with change, either within a formal corporate structure or by organizational and research memory with new per- alliance. Trade blocks like the EU, NAFTA, and sons and new ideas, through the passage of student Mercosul provide new options for breaking ``lock- generations. When there is a break in the genera- ins'', without the sacrifice of competitive advantages tions, typically caused by a loss of research funding, from previous constellations. For example, the Air- one academic research group disappears and can be bus can be considered as an interactive opportunity replaced by another. for recombination at the supra-national levelŽ Fren- Of course, as firms organize increasingly higher ken, this issue. . level training programsŽ e.g., Applied Global Uni- Second, the driving force of the interactions can versity at the Applied Materials Devices, a semicon- be specified as the expectation of profits. ``Profit'' ductor equipment manufacturer in Silicon Valley. may mean different things to the various actors they might in the future also, individually or jointly, involved. A leading edge consumer, for example, attempt to give out degrees. Companies often draw provides firms and engineers with opportunities to upon personnel in their research units, as well as perceive ``reverse salients'' in current product lines external consultants, to do some of the teaching in and software. Thus, opportunities for improvements their corporate universities. Nevertheless, with a few and puzzle-solving trajectories can be defined. Note notable exceptions, such as the RAND, they have not that analytically the drivers are no longer conceptual- yet systematically drawn together research and train- ized as ex ante causes, but in terms of expectations ing into a single framework. However, as the need that can be evaluated only ex post. From the evolu- for life-long learning increases, a university tied to tionary perspective, selectionŽ. ex post is structure the workplace becomes more salient. determined, while variation may be randomŽ Arthur, 1988; Leydesdorff and Van den Besselaar, 1998. . Third, the foundation of the model in terms of 7. Implications of the Triple Helix model expectations leaves room for uncertainties and chance processes. The institutional carriers are expected to The Triple Helix denotes not only the relationship be reproduced as far as they have been functional of university, industry and government, but also hitherto, but the negotiations can be expected to lead internal transformation within each of these spheres. to experiments which may thereafter also be institu- The university has been transformed from a teaching tionalized. Thus, a stage model of innovation can be institution into one which combines teaching with specified. research, a revolution that is still ongoing, not only The stages of this model do not need to corre- in the USA, but in many other countries. There is a spond with product life cycle theory. BarrasŽ. 1990 , tension between the two activities but nevertheless for example, noted that in ICT `à reverse product they coexist in a more or less compatible relationship life'' cycle seems to be dominant. Bruckner et al. with each other because it has been found to be both Ž.1994 proposed niche-creation as the mechanism of H. Etzkowitz, L. LeydesdorffrResearch Policy 29() 2000 109±123 119 potential lock-out in the case of competing technolo- growth. , theoretical expectations, assessment of what gies. A successful innovation changes the landscape, can be realized given institutional and geographic that is, the opportunity structure for the institutional constraints Ð have to be related and converted into actors involved. Structural changes in turn are ex- one another. The helices communicate recursively pected to change the dynamics. over time in terms of each one's own code. Reflex- Fourth, the expansion of the higher-education and ively, they can also take the role of each other, to a academic-research sector has provided society with a certain extent. While the discourses are able to inter- realm in which different representations can be enter- act at the interfaces, the frequency of the external tained and recombined in a systematic manner. interaction isŽ. at least initially lower than the fre- Kaghan and BarnettŽ. 1997 have used in this context quency within each helix. Over time and with the the term ``desktop innovation'' as different from the availability of ICT, this relation is changing. laboratory modelŽ. cf. Etzkowitz, 1999 . Knowledge- The balance between spatial and virtual relations intensive economies can no longer be based on is contingent upon the availability of the exchange simple measures of profit maximization: utility func- media and their codifications. Codified media pro- tions have to be matched with opportunity structures. vide the system with opportunities to change the Over time, opportunity structures are recursively meaning of a communicationŽ. given another context driven by the contingencies of prevailing and possi- while maintaining its substanceŽ Cowan and Foray, ble technologies. A laboratory of knowledge-inten- 1997. . Despite the ``virtuality'' of the overlay, this sive developments is socially available and can be system is not `òn the fly'': it is grounded in a improved uponŽ. Etzkowitz and Leydesdorff, 1995 . culture which it has to reproduceŽ. Giddens, 1984 . As this helix operates, the human capital factor is The retention mechanism is no longer given, but `òn further developed along the learning curves and as the move'': it is reconstructed as the system is an antidote to the risk of technological unemploy- reconstructed, that is, as one of its subdynamics. mentŽ. Pasinetti, 1981 . As the technological culture provides options for Fifth, the model also explains why the tensions recombination, the boundaries of communities can need not to be resolved. A resolution would hinder be reconstituted. The price may be felt as a loss of the dynamics of a system which lives from the traditional identities or alienation, or as a concern perturbations and interactions among its subsystems. with the sustainability of the reconstruction, but the Thus, the subsystems are expected to be reproduced. reverse of ``creative destruction'' is the option of When one opens the black-box one finds Mode 1 increasing development. The new mode of knowl- within Mode 2, and Mode 2 within Mode 1. The edge production generates an endless transition that system is neither integrated nor completely differen- continuously redefines the borders of the endless tiated, but it performs on the edges of fractional frontier. differentiations and local integrations. Using this model, one can begin to understand why the global regime exhibits itself in progressive instances, while the local instances inform us about global develop- 8. The organization of the theme issue ments in terms of the exceptions which are replicated and built upon. As noted above, this issue is organized in three Case materials enable us to specify the negative main parts, addressingŽ. 1 institutional transforma- selection mechanisms reflexively. Selection mecha- tion,Ž. 2 evolutionary mechanisms, and Ž. 3 the sec- nisms, however, remain constructs. Over time, the ond academic revolution. Each part contains five inference can be corroborated. At this end, the func- contributions. tion of reflexive inferencing based on available and In Part OneŽ. Institutional Transformations , new theories moves the system forward by drawing Michael Nowak and Charles Grantham open the attention to possibilities for change. discussion with a paper about the impact of the Sixth, the crucial question of the exchange media Internet on incubation as an institutional mechanism Ð economic expectationsŽ in terms of profit and for technological innovation. The increased complex- 120 H. Etzkowitz, L. LeydesdorffrResearch Policy 29() 2000 109±123 ity of the process induces reflexivity about the choices While the Portuguese team focuses on the re- to be made, and human capital becomes increasingly gional level, Susanne Giesecke takes the analysis to crucial for carrying the transformations. the level of comparing national governments in her The failure of the `òpening to the market'' as an contribution entitled ``The Contrasting Roles of answer to the state-dominated economies in the for- Government in the Development of the Biotechnol- mer Soviet Union, because of the neglect of the ogy Industries in the US and Germany''. She notes knowledge-intensive dimension, is discussed by test- the counter-effective policies of German govern- ing three models against each other in Judith Sedaitis' ments which have operated on the basis of assump- paper entitled ``Technology Transfer in Transitional tions about previous developments. Policies have to Economies: Comparing Market, State, and Organiza- be updated in terms of bottom-up processes and thus tional Frameworks''. The author concludes that pro- come to be understood in terms of reflexive feed- cesses of transfer in these cases can be understood at backsŽ. instead of control . the intermediate network level. Rosalba Casas, Rebeca de Gortari, and Ma. Josefa Norma Morris, in ``Vial Bodies: Conflicting In- Santos from Mexico combine the issues of regional terests in the Move to New Institutional Relation- developments and differences between the technolo- ships in Biological Medicines Research and Regula- gies involved by cross-tabling them for the case of tion'', discusses normative issues that arise when the Mexico. These authors focus on what they call ``the borders are no longer defined institutionally and building of knowledge spaces''. How is the interrela- governmentally. The case of the EU places the role tionship between knowledge-intensity, industrial ac- of safety regulation at national and transnational tivity, and institutional control shaped in terms of levels on the agenda. In a paper entitled ``The interhuman and interinstitutional relations? What is Evolution of Rules for Access to Megascience Re- the function of shared culture, values, and trust? Is search Environments Viewed from Canadian Experi- the region a habitat for the technology, or the tech- ence'', Cooper Langford and Martha Whitney Lang- nology a precondition for restructuring the region? ford document what it means for the organization of In a contribution entitled ``The Triple Helix: An Canadian science that government and industry rela- Evolutionary Model of Innovations'', Loet Leydes- tions are deeply involved in this enterprise. Are the dorff uses simulations to show how a lock-in can be Kudos-norms of MertonŽ. 1942 increasingly being enhanced using a coevolution like the one between replaced by a new set of normsŽ. Ziman, 1994 ? If so, regions and technologies. A third source of random what are the expected effects on reward systems and variation, however, may intervene, reversing the or- funding? In a contribution to the latter question, der in a later stage and leading to more complex Shin-Ichi Kobayashi argues that a third form of arrangements of market segmentationŽ that is, differ- funding can be distinguished nowadaysŽ in addition ent suboptima. . A mechanism for lock-out can also to peer recognition and institutional allocation. . The be specified. author develops the new format using the metaphor Koen Frenken takes the complexity approach one of the audition system for the performing arts. step further by confronting it with empirical data in Thus, not only the institutions themselves are the case of the aircraft industry. Using Kauffman's transformed, but also their mechanisms of transfor- Ž.1993 model of ``rugged fitness landscapes'' he mation. These evolutionary mechanisms are central shows the working of a Triple Helix in different to the second part of the theme issue. The contribu- phases of this industryŽ cf. Frenken and Leydesdorff, tion from the Aveiro teamŽ Eduardo Anselmo de forthcoming. . The model can be extended to account Castro, Carlos JoseÂ Rodrigues, Carlos Esteves, and for the additional degree of freedom in international Artur da Rosa Pires. returns to the impact of ICT on collaborations to develop new aircraft. The failure of changing the stage. How can institutional arrange- Fokker Aircraft, for example, can be explained using ments be shaped to match the options which telemat- these concepts: one cannot bet on two horses at the ics provide? How can a retention mechanism be same time, since the markets are fiercely competi- organized as a niche or a habitat for knowledge-in- tive, technological infrastructures are expensive, and tensive developments? learning curves are steep. H. Etzkowitz, L. LeydesdorffrResearch Policy 29() 2000 109±123 121

In the third part of the issue, we turn to the While a ``hands off'' may have been functional to second academic revolution. In their contribution previous configurations, the exigencies of today de- entitled ``The Place of Universities in the System of mand a more intensive interrelationship. As noted, a Knowledge Production'', BenoõtÃ Godin and Yves Triple Helix arrangement that tends to reorganize the Gingras argue against the thesis that the university knowledge infrastructure in terms of possible over- would have lost its salient position in the lays, can be expected to be generated endogenously. university±industry±government relations of Mode 2. Using scientometric data, they show that collabo- ration with academic teams is central to the opera- Acknowledgements tions of the networks which transform this knowledge infrastructure. Although based on Canadian We acknowledge support from the US National data, the argument is made that this holds true also Science Foundation, the European Commission DG for other OECD countries. XII, the FundacËaoÄ Coppetec in Brazil, the CNRS in From another world region, Judith Sutz reports France, the Netherlands Graduate School for Sci- about university±industry±government relations in ence, Technology and Modern Culture WTMC, the Latin America. These young democracies, on one Center for Business and Policy Studies in Stock- hand, wish to free themselves from the limitation of holm, the State University of New York SUNY, and the so-called ``import substitution'' regime by open- our respective departments. We thank Alexander Et- ing up to the market. On the other hand, the connec- zkowitz for assistance with graphics. tions are then established through the world system, and regional infrastructures tend to remain underde- References veloped. The issue will be central to the Third Triple Helix Conference to be held in Rio de Janeiro, Alchian, A.A., 1950. Uncertainty, evolution, and economic the- 26±29 April 2000. How can social, economic, and ory. Journal of Political Economy 58, 211±222. scientific developments be networked at the regional Andersen, E.S., 1994. Evolutionary Economics: Post-Schumpe- terian Contributions. Pinter, London. level? What does niche management mean in an Arthur, W.B., 1988. Competing technologies. In: Dosi, G., Free- open system's environment? man, C., Nelson, R., Silverberg, G., Soete, L.Ž. Eds. , Techni- In a contribution entitled ``Institutionalizing the cal Change and Economic Theory. Pinter, London, pp. 590± Triple Helix: Research Funding and Norms in the 607. Academic System'', Mats Benner and Ulf SandstromÈ Barras, R., 1990. Interactive innovation in financial and business services: the vanguard of the service revolution. Research take a neo-institutional approach to the transforma- Policy 19, 215±237. tion of the university system in Europe. How does Ben David, J., Collins, R., 1966. Social factors in the origins of the system reactŽ. resist and embody institutional new science: the case of psychology. American Sociological transformation and neo-evolutionary pressures? In a Review 3, 45±85. further article, Eric Campbell and his colleagues Benner, M., Sandstrom,È U., this issue. Institutionalizing the Triple Helix: research funding and norms in the academic system, raise the question of how this affects research prac- Research Policy. tices in terms of ``Data Withholding in Academic Berman, M., 1982. All that is Solid Melts into Air: The Experi- Medicine''. Can characteristics of faculty denied ac- ence of Modernity. Simon and Schuster, New York. cess to research results and biomaterials be distin- Braczyk, H.-J., Cooke, P., Heidenreich, M.Ž. Eds. , Regional Inno- guished? vation Systems. University College London Press, London. Bruckner, E., Ebeling, W., JimenezÂÄ Montano, M.A., Scharnhorst, In a final article, Henry Etzkowitz, Andrew Web- A., 1994. Hyperselection and innovation described by a ster, Christiane Gebhardt, and Branca Terra substan- stochastic model of technological evolution. In: Leydesdorff, tiate their claim that the transformation of the univer- L., Van den Besselaar, P.Ž. Eds. , Evolutionary Economics and sity system is a worldwide phenomenon. In addition Chaos Theory: New Directions in Technology Studies. Pinter, to research and higher eduction, the university nowa- London, pp. 79±90. Bush, V., 1980. The Endless Frontier: A Report to the President, days has a third role in regional and economic reprinted by Arno Press, New YorkŽ. 1945 . development because of the changing nature of both Callon, M., 1998. An essay on framing and overflowing: eco- knowledge production and economic production. nomic externalities revisited by sociology. In: Callon, M. 122 H. Etzkowitz, L. LeydesdorffrResearch Policy 29() 2000 109±123

Ž.Ed. , The Laws of the Market. Macmillan, London, pp. Academy Press, Washington, DC. -http:rrbob.nap.edur 244±269. readingroomrbooksrnstr) Cowan, R., Foray, D., 1997. The Economics of Codification and Gustin, B., 1975. The emergence of the German chemical profes- the Diffusion of Knowledge. MERIT, Maastricht. sion, 1790±1867. PhD dissertation, University of Chicago. Cozzens, S., Healey, P., Rip, A., Ziman, J.Ž. Eds. , 1990. The Hughes, T.P., 1983. Networks of Power: Electrification of West- Research System in Transition. Kluwer Academic Publishers, ern Society 1880±1930. Johns Hopkins Univ. Press, Balti- Boston. more, MD. Dasgupta, P., David, P., 1994. Towards a new economics of Kaghan, W.N., Barnett, G.B., 1997. The desktop model of innova- science. Research Policy 23, 487±522. tion. In: Etzkowitz, H., Leydesdorff, L.Ž. Eds. , Universities in David, P.A., Foray, D., 1994. Dynamics of competitive technol- the Global Economy: A Triple Helix of University±Industry± ogy diffusion through local network structures: the case of Government Relations. Cassell Academic, London, pp. 71±81. EDI document standards. In: Leydesdorff, L., Van den Besse- Kampmann, C., Haxholdt, C., Mosekilde, E., Sterman, J.D., 1994. laar P.Ž. Eds. , Evolutionary Economics and Chaos Theory: Entrainment in a disaggregated long-wave model. In: Leydes- New Directions in Technology Studies. Pinter, London, pp. dorff, L., Van den Besselaar, P.Ž. Eds. , Evolutionary Eco- 63±78. nomics and Chaos Theory: New Directions in Technology Etzkowitz, H., 1999. Bridging the gap: the evolution of industry± Studies. Pinter, London, pp. 109±124. university links in the United States. In: Branscomb, L., Langton, C.G.Ž. Ed. , 1989. Artificial Life. Addison Wesley, Red- Kodama, F.Ž. Eds. , Industrializing Knowledge: University±In- wood City, CA. dustry Linkages in Japan and the United States. MIT Press, Leydesdorff, L., 1995. The Challenge of Scientometrics: The Cambridge, MA. Development, Measurement, and Self-organization of Scien- Etzkowitz, H., forthcoming. The Second Academic Revolution: tific Communications. DSWO Press, Leiden University, Lei- MIT and the Rise of Entrepreneurial Science. Gordon & den. Breach, London. Leydesdorff, L., 1997. The non-linear dynamics of sociological Etzkowitz, H., Leydesdorff, L., 1995. The triple helix± reflections. International Sociology 12, 25±45. university±industry±government relations: a laboratory for Leydesdorff, L., Etzkowitz, H., 1996. Emergence of a Triple knowledge-based economic development. EASST Review 14 Helix of university±industry±government relations. Science Ž.1 , 14±19. and Public Policy 23, 279±286. Etzkowitz, H., Leydesdorff, L.Ž. Eds. , 1997. Universities in the Leydesdorff, L., Etzkowitz, H., 1998. The Triple Helix as a model Global Economy: A Triple Helix of University±Industry± for innovation studies. Science and Public Policy 25Ž. 3 , Government Relations. Cassell Academic, London. 195±203. Etzkowitz, H., Gulbrandsen, M., Levitt, J., 2000. Public Venture Leydesdorff, L., Oomes, N., 1999. Is the European monetary Capital: Government Funding Sources for Technology En- system converging to integration?. Social Science Information trepreneurs. Harcourt-Brace, New York. 38Ž. 1 , 57±86. Freeman, C., 1982. The Economics of Industrial Innovation. Leydesdorff, L., Van den Besselaar, P., 1997. Scientometrics and Pinter, London. communication theory: towards theoretically informed indica- Freeman, C., Perez, C., 1988. Structural crises of adjustment, tors. Scientometrics 38, 155±174. business cycles and investment behaviour. In: Dosi, G., Free- Leydesdorff, L., Van den Besselaar, P., 1998. Competing tech- man, C., Nelson, R., Silverberg, G., Soete, L.Ž. Eds. , Techni- nologies: lock-ins and lock-outs. In: Dubois, D.M.Ž. Ed. , cal Change and Economic Theory. Pinter, London, pp. 38±66. Computing Anticipatory Systems, Proceedings of the Ameri- Frenken, K., this issue. A complexity approach to innovation can Institute of Physics 437. American Institute of Physics, networks. The case of the aircraft industryŽ. 1909±1997 . Woodbury, NY, pp. 309±323. Research Policy. Luhmann, N., 1984. Soziale Systeme. Grundriû einer allgemeinen Frenken, K., Leydesdorff, L., forthcoming. Scaling trajectories in TheorieŽ. Suhrkamp, Frankfurt am Main .w Social Systems civil aircraftŽ. 1913±1997 . Research Policy. Ž.Stanford University Press, Stanford 1995x . Gibbons, M., Limoges, C., Nowotny, H., Schwartzman, S., Scott, Lundvall, B.-A.,Ê 1988. Innovation as an interactive process: from P., Trow, M., 1994. The New Production of Knowledge: The user±producer interaction to the national system of innovation. Dynamics of Science and Research in Contemporary Societies. In: Dosi, G. Freeman, C., Nelson, R., Silverberg, G., Soete, L. Sage, London. Ž.Eds. , Technical Change and Economic Theory. Pinter, Lon- Giddens, A., 1984. The Constitution of Society. Polity Press, don, pp. 349±369. Cambridge. Lundvall, B.-A.Ê Ž. Ed. , 1992. National Systems of Innovation. Godin, B., 1998. Writing performative history: is this a new Pinter, London. Atlantis?. Social Studies of Science 38Ž. 3 , 465±483. MacLane, S., 1996. Should universities imitate industry?. Ameri- Godin, B., Gingras, Y., this issue. The Place of Universities in the can Scientist 84Ž. 6 , 520±521. System of Knowledge Production. Research Policy. Maturana, H.R., 1978. Biology of language: the epistemology of Government±University±Industry Research Roundtable GUIRR, reality. In: Miller, G.A., Lenneberg, E.Ž. Eds. , Psychology and 1998. National Science and Technology Strategies in a Global Biology of Language and Thought. Essays in Honor of Eric Context. Report of an International Symposium. National Lenneberg. Academic Press, New York, pp. 27±63. H. Etzkowitz, L. LeydesdorffrResearch Policy 29() 2000 109±123 123

McKelvey, M.D., 1996. Evolutionary Innovations: The Business Pasinetti, L., 1981. Structural Change and Economic Growth. of Biotechnology. Oxford Univ. Press, Oxford. Cambridge Univ. Press, Cambridge. Merton, R.K., 1938. Science, Technology and Society in Seven- Pavitt, K., 1984. Sectoral patterns of technical change: towards a teenth Century England. Sainte Catherine Press, Burges. theory and a taxonomy. Research Policy 13, 343±373. Merton, R.K., 1942. Science and technology in a democratic Rip, A., VanderMeulen, B., 1996. The post-modern research order. Journal of Legal and Political Sociology 1, 115±126. system. Science and Public Policy 23Ž. 6 , 343±352. Mills, C.W., 1958. The Power Elite. Oxford Univ. Press, New Rosenberg, N., Nelson, R.R., 1994. American universities and York. technical advance in industry. Research Policy 23, 323±348. Mowery, D.C., Rosenberg, N., 1979. The influence of market Rothwell, R., Zegveld, W., 1981. Industrial Innovation and Public demand upon innovation: a critical review of some empirical Policy. Pinter, London. studies. Research Policy 8, 102±153. Sabato,Â J., 1975. El pensamiento latinoamericano en la Nelson, R.R.Ž. Ed. , 1982. Government and Technical Progress: A problematicaÂÂ ciencia±technologõa±desarrollo-dependencia. Cross-Industry Analysis. Pergamon, New York. Paidos,Â Buenos Aires. Nelson, R.R.Ž. Ed. , 1993. National Innovation Systems: a Com- Sabato,ÂÂÂ Jorge, Mackenzi, M., 1982. La Produccion de Technologõa. parative Study. Oxford Univ. Press, New York. AutonomaÂ o Transnacional. Nueva Imagen, Mexico. Nelson, R.R., 1994. Economic growth via the coevolution of Schumpeter, J., 1964. Business Cycles: A Theoretical, Historical technology and institutions. In: Leydesdorff, L., Van den and Statistical Analysis of Capitalist Process. McGraw-Hill, Besselaar, P.Ž. Eds. , Evolutionary Economics and Chaos The- New Yorkwx 1939 . ory: New Directions in Technology Studies. Pinter, London, Schumpeter, J., 1966. Invention and Economic Growth. Harvard pp. 21±32. University Press, Cambridge, MA. Nelson, R., Winter, S., 1975. Growth theory from an evolutionary Sobel, D., 1995. Longitude. Penguin, Harmondsworth. perspective: the differential productivity growth puzzle. Amer- Storr, R., 1953. The Beginnings of Graduate Education in Amer- ican Economic Review 65, 338. ica. University of Chicago Press, Chicago. Nelson, R.R., Winter, S.G., 1982. An Evolutionary Theory of Weingart, P., 1997. From ``Finalization'' to ``Mode 2'': old wine Economic Change. Belknap Press, Cambridge, MA. in new bottles?. Social Science Information 36Ž. 4 , 591±613. Neurath, O., Carnap, R., Hahn, H., 1929. Wissenschaftliche Ziman, J., 1994. Prometheus Bound: Science in a Dynamic Steady Weltauffassung-Der Wiener Kreis. VeroffentlichungenÈ des State. Cambridge Univ. Press, Cambridge. Vereins Ernst Mach, Vienna. OECD, 1980. Technical Change and Economic Policy. OECD, Paris.