Professional Documents
Culture Documents
about just how and (to some extent) why of the industries to which a particular paradigm is
prevailing practice works. associated.
An important part of paradigmatic knowledge Technological paradigms identify the opera-
takes the form of design concepts, which charac- tive constraints on prevailing best practice and
terize in general the configuration of the particular the problem-solving heuristics deemed promising
artefacts or processes that are operative at any for pushing back those constraints. More gener-
time. Shared general design concepts are an ally, they are the cognitive frames shared by tech-
important reason why there is often strong simi- nological professionals in a field, that orient what
larity among the range of particular products such professionals think they can do to advance a
manufactured at any time – for example, the technology (Constant 1980). Technological para-
large passenger aircraft produced by different air- digms also encompass normative aspects, such as
craft companies or the different television sets criteria for assessing performance, and thus pro-
available at the electronics stores. Indeed, the vide ways of judging what is better as well as
establishment of a given technological paradigm identifying goals for the improvement of practice.
is quite often linked with the emergence of some Each paradigm involves a specific ‘technology of
dominant design (Abernathy and Utterback 1978; technical change’, that is, specific heuristics of
Rosenbloom and Cusumano 1987; Henderson search. So, in sectors such as organic chemicals
and Clark 1990; Utterback and Suarez 1993; for example, these heuristics relate to the ability of
Suarez and Utterback 1995; and the critical coupling basic scientific knowledge with the
review of the whole literature in Murmann and development of molecules that present the
Frenken 2006). A dominant design is defined in required characteristics, while in pharmaceuticals
the space of artefacts and is characterized both by the additional requirement is the ability to match
a set of core design concepts embodied in compo- molecular knowledge with receptors and patholo-
nents that correspond to the major functions gies. In microelectronics, search concerns
performed by the product and by a product archi- methods for further miniaturization of electrical
tecture that defines the ways in which these com- circuits, the development of the appropriate hard-
ponents are integrated (Murmann and Frenken ware capable of ‘writing’ semiconductor chips at
2006; drawing upon Henderson and Clark 1990). such a required level of miniaturization and
However, sometimes the establishment of a dom- advances in the programming logic to be built
inant paradigm is not associated with a dominant into the chip. There are numerous examples:
design. A revealing case in point is pharmaceuti- some are discussed in Dosi (1988). Here notice
cal technologies which do involve specific knowl- in particular that distinct (paradigm-specific)
edge basis, specific search heuristics and so search and learning procedures first imply diverse
on – that is, the strong mark of modes of creating and accessing novel technolog-
paradigms – without, however, any hint at any ical opportunities, and, second, entail different
dominant design. Molecules, even when aimed organizational forms suited to such research pro-
at the same pathology, might have quite different cedures. As we shall see, both properties will turn
structures: in that space, one is unlikely to find out to be central when trying to characterize dis-
similarities akin those linking even a Volkswagen tinct ‘regimes’ of technological evolution.
Beetle 1937 and a Ferrari 2000. Nevertheless, the Together, the foregoing features of technolog-
notion of ‘paradigm’ holds in terms of the under- ical paradigms provide a focus for efforts to
lying features of knowledge bases and search advance a technology and channel them along
processes. distinct technological trajectories, with advances
Whether or not the establishment of a domi- (made by many different agents) proceeding over
nant paradigm entails also the establishment of a significant periods of time in certain relatively
dominant design is of importance in terms of the invariant directions, in the space of techno-
dynamics of industry structure along the lifecycle economic characteristics of artefacts and produc-
tion processes. As paradigms embody the
Technological Paradigms and Technological Trajectories 3
identification of the needs and technical require- 2006). Some properties of trajectories are impor-
ments of users, trajectories may be understood in tant to note.
terms of the progressive refinement and improve- First, trajectories order and confine but do not
ment in the supply responses to such potential at all eliminate the persistent generation of vari-
demand requirements. ety, in the product and process spaces that inno-
A growing number of examples of technolog- vative search always produces. The paradigm
ical trajectories include aircraft, helicopters, vari- defines proximate boundaries of feasibility and
ous kinds of agricultural equipment, cars, together shapes the heuristics of search. However,
semiconductors (Gordon and Munson 1981; there continues to be plenty of possible trade-offs
Sahal 1981, 1985; Dosi 1984; Grupp 1992; between output characteristics, which different
Saviotti and Trickett 1992; Saviotti 1996). So, producers explore (Saviotti 1996) and which will
for example, technological advances in aircraft be eventually the object of (imperfect and time-
technologies have followed two distinct trajecto- consuming) market selection.
ries (one civilian and one military) characterized Second, by the same token, trajectories ‘extra-
by log-linear improvements in the trade-offs polated forward’ – in so far as their knowledge is
between horsepower, gross take-off weight, cruise shared by the community of firms, practitioners,
speed, wing load and cruise range (Sahal 1985; engineers – are powerful uncertainty-reducing
Frenken et al. 1999; Frenken and Leydesdorff representations of what the future is likely to
2000; Giuri et al. 2007; and, more specifically on yield in technological terms. However, this
aircraft engines, Bonaccorsi et al. 2005). Analo- remains a far cry from any unbiased expectation
gously, in microelectronics, technical advances on the time and costs involved in ‘getting
are accurately represented by an exponential tra- there’ – wherever ‘there’ means – and, even
jectory of improvement in the relationship more so, of the probability distributions of indi-
between density of electronic chips, speed of com- vidual actors over both technological and eco-
putation and cost per bit of information (see Dosi nomic success. That is, trajectories are not
1984, but the trajectory has persisted since then). means to reduce Knightian uncertainty into
In fact, it is fair to say that trajectory-like patterns probabilizable risk. Indeed, notwithstanding
of technological advance have generally been roughly predictable trajectories of advance, both
found whenever the analyst bothered to plot over substantive uncertainty – concerning future states
time the fundamental techno-economic features of of the world – and procedural
discrete artefacts or processes, say from the DC3 uncertainty – regarding future problem-solving
to the Airbus 380, among aircrafts; or from cruci- procedures – continue to be ubiquitous.
ble to Bessemer to basic oxygen reduction among Note that there is no a priori economic reason
steel-making processes. (Admittedly, trajectories why one should observe limited clusters of tech-
in the space of processes and related input inten- nological characteristics at any one time and
sities have been studied much less than trajecto- ordered trajectories over time. On the contrary,
ries in the output characteristic space, and this is as we have already argued in Dosi
indeed a challenging research area ahead.) (1988) – given consumers with different prefer-
The emergence of relatively ordered trajecto- ences and equipment users with different techni-
ries, as already suggested, is not always associated cal requirements and different relative prices over
with the emergence of dominant designs. When it different countries, if technologies were perfectly
is, the trajectories appear to be driven by ‘hierar- ‘plastic’ and malleable – as standard economic
chically nested technological cycles’ entailing representations are implicitly suggesting – one
both relatively invariant core components improv- would tend to observe sorts of ‘isoquants’ in the
ing over time and a series of bottlenecks and space of techniques and technoeconomic charac-
‘technological imbalances’ (Rosenberg 1976) teristics and products with the familiar shape.
regarding the consistency among all the compo- And, over time, if technological recipes – both in
nents of the systems (cf. Murmann and Frenken the procedural aspects and their input
4 Technological Paradigms and Technological Trajectories
contents – could be freely added, divided, and in the related equipment-embodied technolo-
recombined, substituted and so on, one would gies is a powerful trend towards mechanization
also tend to observe an increasingly disperse vari- and/or automation of production activities. Recent
ety of technical and performance combinations in pieces of evidence are in Klevorick and col-
products, production inputs and available tech- leagues (1995), but the phenomenon has been
niques (even if not necessarily in their use, given noticed in the classical literature and plays an
relative prices). The ubiquitous evidence on tra- important role in the analyses of the dynamics of
jectories, on the contrary, suggests that technolog- capitalist economies by Adam Smith and Karl
ical advances are circumscribed within a quite Marx. Note that such a tendency holds across
limited subset of the techno-economic character- sectors and across countries characterized by dif-
istics space. We could say that the paradigmatic, ferent capital intensities, and broadly occurs
cumulative nature of technological knowledge irrespective of variations in relative prices. Due
provides innovation avenues (Sahal 1985) which to its generality, in another work (Nelson and
channel technological evolution, while major dis- Winter 1977) it was called a ‘natural trajectory’:
continuities tend to be associated with changes in of course, there is nothing ‘natural’, strictly speak-
paradigms. Indeed, here and throughout what we ing, but it is indeed a general reflection of a long
shall call ‘normal’ technical progress those term trend towards the substitution of inanimate
advances occur along a given energy to human and animal efforts, and, more
trajectory – irrespective of how ‘big’ they are recently, also of inanimate information processing
and how fast they occur – while we reserve the to human cognition and control.
name of ‘radical innovations’ for those innova- There is another relatively common feature of
tions linked with paradigm changes. trajectories of innovation (even if we still do not
A change in the paradigm generally implies a know how common – a task for future empirical
change in the trajectories. Together with different research), namely learning curves (see Thompson
knowledge bases and different prototypes of arte- 2010). This original statement of the ‘law’ comes
facts, the techno-economic dimensions of innova- from Wright (1936), in reference to aircraft
tion also vary. Some characteristics may become manufacturing (see also Alchian 1963). Similar
easier to achieve, new desirable characteristics regularities appear in various energy-producing
may emerge, some others may lose importance. technologies, in computers, light bulbs, and
Relatedly, the engineers’ vision of future techno- many other artefacts and processes: for
logical advances will change, together with a technology-specific evidence and surveys see
changing emphasis on the various trade-offs that Conley (1970), Baloff (1971), Dutton and
characterize the new artefacts. So, for example, Thomas (1984), Gritsevskyi and Nakicenovic
the technological trajectory in active electrical (2000), MacDonald and Schrattenholzer (2001),
components based on thermionic valves had as Neij (1997), Yelle (1979), Argote and Epple
fundamental dimensions heat-loss vacuum- (1990) and Thompson (2010). Semiconductors
parameters, miniaturization and reliability over offer an archetypical example of a trajectory
time. With the appearance of solid state compo- driven by miniaturization efforts yielding the
nents (the fundamental building block of the so-called Moore’s Law involving the doubling of
microelectronic revolution) heat loss became rel- the density of elementary transistor-per-chip and
atively less relevant, while miniaturization later microprocessors every 2–3 years (more
increased enormously in importance. Similar details in Gordon and Munson 1981; Dosi 1984;
examples of change in the dimensions of the Jovanovic and Rousseau 2002; Nordhaus 2007).
design space can be found in most transitions Interestingly, a steady fall in unit labour inputs
from one paradigm to another. seems, at least in some circumstances, to appear
Are there some features which most technolog- even when holding the equipment constant. This
ical trajectories share? A common feature which is the so-called Horndahl effect, named after a
characterizes trajectories in process technologies Swedish steel mill (Lundberg 1961), an effect
Technological Paradigms and Technological Trajectories 5
which contributed to Arrow’s (1962) inspiration designs (and nowadays with the advances incor-
for the concept of learning by doing. Notice that porated into new vintages of capital equipment
learning effects appear at the levels of industry, and the ability of using these). On the other
firms and plants, even if distinct rates and hand, understanding can advance rapidly when
intertemporal variabilities with micro learning there are fields of science dedicated to that objec-
displaying higher irregularities over time than tive. Several studies (see, e.g., Klevorick
industry-level rates of progress (for some discus- et al. 1995; Nelson and Wolff 1997) have shown
sion of the evidence see Auerswald et al. 2000). that the fields of technology which, by a variety of
The interpretation of learning mechanisms under- measures, have advanced most rapidly are associ-
lying the observed performance trajectories of ated with strong fields of applied science or engi-
their differences across different paradigms are neering. Moreover, firms operating in these fields
important research tasks for future evolutionary also tend to have levels of R&D intensity that are
analyses of innovation. higher than average. In fact, in a secular perspec-
Together with differences across paradigms in tive, the evidence is in tune with Mokyr’s general
the rates of technological advance, one observes conjecture that the ‘epistemic’ elements of tech-
major differences in the processes through which nological knowledge – that is, those elements
such advances occur. In fact, significant progress associated with an explicitly casual knowledge
has been made in the conceptualization of what of natural phenomena – are of crucial (and
different technological paradigms have in com- increasing) importance in modern technological
mon and how they differ in terms of the sources advances (Nelson and Wolff 1997; Mokyr 2002,
of knowledge upon which they draw – the tech- 2010; Nelson and Nelson 2002; Nelson 2003).
nological opportunities which they tap – the Since the Industrial Revolution, the contribu-
mechanisms through which such opportunities tion of science to technology has been increasing,
are seized, and the possibilities they entail for and, in turn, such a science base has been largely
innovators to extract economic benefit from their the product of publicly funded research, while the
technological advances – that is, the knowledge produced by that research has been
appropriability conditions. generally open and available for potential innova-
tion to use (more in David 2001a, b, 2004; Pavitt
2001; Nelson 2004).
Technological Opportunities, The This, however, is not sufficient to corroborate
Processes of Knowledge Accumulation any simple ‘linear model’ from pure to applied
and Their Cumulativeness science to technological applications.
First, the point made elsewhere by Rosenberg
Prevailing technological paradigms differ over (1982), Kline and Rosenberg (1986), Pavitt
time and across fields regarding the nature of the (1999) and Nelson (1981) continues to apply:
knowledge underlying the opportunities for tech- scientific principles are helpful but are rarely
nical advances. Relatedly, they differ in the extent enough. An enlightening case in a ‘science-
to which such knowledge has been gained through based’ area – medical innovation – is discussed
operating experience, as opposed to scientific in Rosenberg (2009). Semiconductors technology
research. is another good example. For many decades,
While in most fields there is a mix, in the fields efforts to advance products and process
generally thought of as ‘high tech’ a more signif- technology – crucially involving the ability to
icant contribution is nowadays grounded in the progressively make circuits smaller and
specialized fields of science or engineering. smaller – have taken advantage of the understand-
Where operating experience and learning by ings in material science and the underlying solid
doing and using are the primary bases for profes- state physics. However, much more pragmatic and
sional understanding, the learning trajectory will tacit elements of technological know-how have
advance, paced by experience with actual new been persistently crucial.
6 Technological Paradigms and Technological Trajectories
governance and peer evaluation, (ii) a shared cul- differences in the strength and richness of techno-
ture of scientists emphasizing the importance of logical opportunities. More generally, let us sug-
motivational factors other than economic ones gest that the widespread view that the key to
and (iii) an ethos of disclosure of search results increasing technological progress is in strength-
driven by ‘winner takes all’ precedence rules. In ening appropriability conditions, mainly through
Nelson (2006), David and Hall (2006), and Dosi making patents stronger and wider, is deeply
et al. (2006a) the dangers from the erosion of open misconceived. Obviously, inventors and innova-
science institutions are discussed. We have tors must have a reasonable expectation of being
already mentioned above the importance of able to profit from their work, where it is techno-
(free-flowing) advances in pure and applied sci- logically successful and happens to meet market
ences as a fundamental fuel for technological demands. However, in most industries this is
advance – albeit with significant variation across already the case. And there is no evidence that
technologies, sectors and stages of development stronger patents will significantly increase the rate
for each technological paradigm. However, the of technological progress. (See, further,
major share of inventive activities finalized to Mazzoleni and Nelson 1998; Granstrand 1999;
economically exploitable technologies that goes Jaffe 2000; Dosi et al. 2006b; and the growing
on in contemporary capitalist societies is made in literature cited therein.) In fact, in many instances
profit-seeking organizations with the hope and the opposite may well be the case. We have noted
expectation of being economically rewarded, if that, in most fields of technology, progress is
that work is successful. In turn, the very existence cumulative, with yesterday’s efforts, both failures
of a relation between economically expensive and successes, setting the stage for today’s efforts
search efforts by private agents and (uncertain) and achievements. If those who do R&D today are
economic rewards from successful innovations, cut off from being able to draw from and build on
entails the fundamental what was achieved yesterday, progress may be
incompatibility – originally pointed out by Marx hindered significantly. Historical examples, such
and Schumpeter – between any sort of zero-profit as those presented in Merges and Nelson (1994)
general equilibrium and any incentive to endoge- on the Selden patent around the use of a light fuel
nous innovation (that is, endogenous to the pri- in an internal combustion engine to power a car or
vate, ‘capitalist’, sector of the economy). the Wright brothers’ patent on an efficient stabi-
Granted this is the case, however, two major lizing and steering system for flying machines, are
sets of questions arise. good cases in point, showing how the intellectual
First, how profound is such a trade-off, if any, property rights (IPR) regime probably slowed
between monopolistic departures from competi- down considerably the subsequent development
tive (zero profit) conditions and incentives to of cars and aircrafts, due to the time and resources
innovate? More precisely, what is the evidence, consumed by lawsuits against the patents them-
if any, on the monotonic relation between (actual selves. The current debate on property rights in
and expected) returns from innovation, on the one biotechnology suggests similar problems,
hand, and innovative efforts, on the other? whereby granting very broad claims on patents
Such a monotonic relation is, in fact, built-in as might have a detrimental effect on the rate of
one of a core assumption within most ‘neo- technical change, insofar as they preclude the
Schumpeterian’ models of growth, while the lim- exploration of alternative applications of the pat-
ited ability to appropriate returns to invention and ented inventions.
innovation is often offered as the reason why the This is particularly the case when inventions
rate of technological progress is very slow in some concerning fundamental techniques or knowledge
industries. The aforementioned studies on the are concerned, as with genes or the Leder and
nature and sources of technological opportunities Stewart patent on the achievement of a genetically
suggest that this is unlikely to be the primary engineered mouse that develops cancer. This is
reason. The far more likely reason lies in clearly a fundamental research tool. To the extent
8 Technological Paradigms and Technological Trajectories
that such techniques and knowledge are critical private investors, possibly because of some
for further research that proceeds cumulatively on awareness of the anti-commons problem, have
the basis of the original invention, the attribution preferred to avoid claiming patents and to delib-
of broad property rights might severely hamper erately operate in a weak IPR regime somewhat
further developments. Even more so if the patent similar to that of open science, involving the free
protects not only the product the inventors have disclosure of inventions to one another (see Allen
achieved (the ‘onco-mouse’) but all the class of 1983, and Nuvolari 2004 on blast furnaces and the
products that could be produced through that prin- Cornish pumping engine, respectively). Interest-
ciple, that is, ‘all transgenic non-human mam- ingly, these cases of ‘collective invention’ have
mals’, or all the possible uses of a patented been able to yield rapid rates of technical change.
invention (say, a gene sequence), even though Similar phenomena of free revelation of innova-
they are not named in the application. In this tion appear also in the communities of users inno-
respect, Murray and colleagues (2009) offer a vators (see von Hippel 2005).
striking illustration of how ‘opening up upstream’ The second set of questions concerns the char-
(again, in the case of the mouse) – in such an acteristics of the regimes stimulating and guiding
instance, a discrete change in the IPR regime in technological advance in a field of activity – that
the US – yielded more search and more diverse is, how inventors appropriate returns. The conven-
rates of exploration of ‘downstream’ research tional wisdom has long been that patent protection
paths. is the key to being able to appropriate returns. But
In general, today’s efforts to advance a tech- this is the case only in some fields of technology.
nology often need to draw from a number of Pharmaceuticals is an important example. How-
earlier discoveries and advances which painstak- ever, a series of studies (Mansfield et al. 1981;
ingly build upon each other. Under these circum- Levin et al. 1985; Cohen et al. 2002, among
stances, IPRs are more likely to be a hindrance others) has shown that in many industries patents
than an incentive to innovate (see Merges and are not the most important mechanism enabling
Nelson 1994; Heller and Eisenberg 1998). If past inventors to appropriate returns. Thus Levin
and present components of technological systems et al. (1985) find that, for most industries, ‘lead
are patented by different parties, there can be an time and learning curve advantages, combined
anti-commons problem (the term was coined by with complementary marketing efforts, appear to
Heller and Eisenberg). While in the standard com- be the principal mechanisms of appropriating
mons problem (such as an open pasture) the lack returns to product innovations’ (p. 33).
of proprietary rights is argued to lead to Patenting often appears to be a complementary
overutilization and depletion of common goods, mechanism for appropriating returns to product
in instances such as biotechnology the risk may be innovation, but not the principal one in most
that excessive fragmentation of IPRs among too industries. For process innovations (used by the
many owners may well slow down research activ- innovator itself) secrecy is often important, pat-
ities because each owner can block the other. ents seldom so. These findings were largely con-
Further empirical evidence on the negative effects firmed by a follow-on study by Cohen
of strong patent protection on technological pro- et al. (2002). David Teece (1986) and a rich sub-
gress is in Mazzoleni and Nelson (1998); and, at a sequent literature (cf. the special issue of Research
more theoretical level, see the insightful discus- Policy, 2006; taking stock of the advancements
sion in Winter (1993) showing how tight since his original insights) have analysed in some
appropriability regimes in evolutionary environ- detail the differences between inventions for
ments can deter technical progress (cf. also the which strong patents can be obtained and
formal explorations in Marengo et al. 2009). Con- enforced, and inventions where patents cannot
versely, one can document, well before the con- be obtained or are weak, in the firm strategies
temporary movement of ‘open source’ software, needed for reaping returns to innovation. A basic
cases in which groups of competing firms or and rather general finding is that, in many cases,
Technological Paradigms and Technological Trajectories 9
David, P.A., and B. Hall. 2006. Property and the pursuit of Jovanovic, B., and P.L. Rousseau. 2002. Moore’s law and
knowledge: IPR issues affecting scientific research. learning by doing. Review of Economic Dynamics 5:
Research Policy 35: 767–771. 346–375.
Dosi, G. 1982. Technological paradigms and technological Klevorick, A.K., R.C. Levin, R.R. Nelson, and
trajectories: A suggested interpretation of the determi- S.G. Winter. 1995. On the sources and significance of
nants and directions of technical change. Research interindustry differences in technological opportuni-
Policy 11: 147–162. ties. Research Policy 24: 185–205.
Dosi, G. 1984. Technical change and industrial transfor- Kline, S.J., N. Rosenberg, S.J. Kline, and N. Rosenberg.
mation. London: Macmillan. 1986. An overview of innovation. In The positive sum
Dosi, G. 1988. Sources, procedures and microeconomic strategy: Harnessing technology for economic growth.
effects of innovation. Journal of Economic Literature Washington, DC: National Academy Press.
26: 1120–1171. Kuhn, T. 1962. The structure of scientific revolutions.
Dosi, G., and R.R. Nelson. 2010. Technical change and Chicago: University of Chicago Press.
industrial dynamics as evolutionary processes. In Levin, R.C., W.M. Cohen, and D.C. Mowery. 1985. R&D
Handbook of the economics of innovation, vol. 1, ed. appropriability, opportunity and market structure: New
B.H. Hall and N. Rosenberg. Burlington: Academic. evidence on some Schumpeterian hypotheses. Ameri-
Dosi, G., P. Llerena, and M. Sylos Labini. 2006a. Science- can Economic Review, Papers and Proceedings 75:
technology-industry links and the ‘European Paradox’: 20–24.
Some notes on the dynamics of scientific and techno- ^
Lundberg, E. 1961. Produktivitet och RûEntabilitet. Stock-
logical research in Europe. Research Policy 35: holm: Studieförbundet Näringsliv och samhälle.
1450–1464. MacDonald, A., and L. Schrattenholzer. 2001. Learning
Dosi, G., L. Marengo, and C. Pasquali. 2006b. How much rates for energy technologies. Energy Policy 29:
should society fuel the greed of innovators? On the 255–261.
relations between appropriability, opportunities and Malerba, F., and L. Orsenigo. 1996. The dynamics and
rates of innovation. Research Policy 35: 1110–1121. evolution of industries. Industrial and Corporate
Dutton, J.M., and A. Thomas. 1984. Treating progress Change 5: 51–87.
functions as a managerial opportunity. Academy of Mansfield, E., M. Schwartz, and S. Wagner. 1981. Imita-
Management Review 9: 235–247. tion costs and patents: An empirical study. The Eco-
Frenken, K., and L. Leydesdorff. 2000. Scaling trajectories nomic Journal 91: 907–918.
in civil aircraft (1913–1997). Research Policy 29: Marengo, L., C. Pasquali, M. Valente, and G. Dosi. 2009.
331–338. Appropriability, patents, and rates of innovation in
Frenken, K., P.P. Saviotti, and M. Trommetter. 1999. Vari- complex products industries, LEM working paper
ety and niche creation in aircraft, helicopters, motorcy- series, 2009/05. Pisa: Scuola Superiore Sant’Anna.
cles and microcomputers. Research Policy 28: Mazzoleni, R., and R.R. Nelson. 1998. The benefits and
469–488. costs of strong patent protection: A contribution to the
Giuri, P., C. Tomasi, and G. Dosi. 2007. L’industria current debate. Research Policy 27: 273–284.
aerospaziale. Innovazione, tecnologia e strategia Merges, R.P., and R.R. Nelson. 1994. On limiting or
economica. Milan: Il Sole 24 Ore e Fondazione Cotec. encouraging rivalry in technical progress: The effect
Gordon, T.J., and T.R. Munson. 1981. Research into tech- of patent scope decisions. Journal of Economic Behav-
nology output measures. Glastonbury: The Future ior & Organization 25: 1–24.
Group. Mokyr, J. 2002. The gifts of Athena: Historical origins of
Granstrand, O. 1999. The economics and management of the knowledge economy. Princeton: Princeton Univer-
intellectual property. Cheltenham: Edward Elgar sity Press.
Publishing. Mokyr, J. 2010. The contribution of economic history to
Gritsevskyi, A., and N. Nakicenovic. 2000. Modeling the study of innovation and technical change. In Hand-
uncertainty of induced technological change. Energy book of the economics of innovation, vol. 1, ed.
Policy 28: 907–921. B.H. Hall and N. Rosenberg. Burlington: Academic.
Grupp, H. 1992. Dynamics of science-based innovation. Murmann, J.P., and K. Frenken. 2006. Toward a systematic
Berlin: Springer. framework for research on dominant designs, techno-
Heller, M., and R. Eisenberg. 1998. Can patents deter logical innovations, and industrial change. Research
innovation? The anti-commons in biomedical research. Policy 35: 925–952.
Science 280: 698–701. Murray, F.E., P. Aghion, M. Dewatripont, J. Kolev, and
Henderson, R.M., and K.B. Clark. 1990. Architectural S. Stern. 2009. Of mice and academics: Examining the
innovation: The reconfiguration of existing product effect of openness on innovation. NBER Working Paper
technologies and the failure of established firms. Series, 14819. Cambridge, MA.
Administrative Science Quarterly 35: 9–30. Neij, L. 1997. Use of experience curves to analyse the
Jaffe, A.B.. 2000. The U.S. patent system in transition: prospects for diffusion and adoption of renewable
Policy innovation and the innovation process. Research energy technology. Energy Policy 25: 1099–1107.
Policy 29: 531–577.
Technological Paradigms and Technological Trajectories 11
Nelson, R.R. 1981. Research on productivity growth and Teece, D.J. 1986. Profiting from technological innovation:
productivity differences: Dead ends and new depar- Implications for integration, collaboration, licensing
tures. Journal of Economic Literature, American Eco- and public policy. Research Policy 15: 285–305.
nomic Association 19: 1029–1064. Teece, D.J., R. Rumelt, G. Dosi, and S.G. Winter. 1994.
Nelson, R.R. 2003. On the uneven evolution of human Understanding corporate coherence: Theory and evi-
know-how. Research Policy 32: 909–922. dence. Journal of Economic Behavior & Organization
Nelson, R.R. 2004. The market economy, and the scientific 23: 1–30.
commons. Research Policy 33: 455–471. Thompson, P. 2010. Learning by doing. In Handbook of
Nelson, R.R. 2006. Reflections on ‘The Simple Economics the economics of innovation, vol. 1, ed. B.H. Hall and
of Basic Scientific Research’: Looking back and N. Rosenberg. Burlington: Academic.
looking forward. Industrial and Corporate Change Tushman, M.L., and P. Anderson. 1986. Technological
15: 145–149. discontinuities and organizational environments.
Nelson, R.R., and K. Nelson. 2002. On the nature and Administrative Science Quarterly 31: 439–465.
evolution of human know-how. Research Policy 31: Utterback, J.M., and F.F. Suarez. 1993. Innovation, com-
719–733. petition, and industry structure. Research Policy 22:
Nelson, R.R., and S.G. Winter. 1977. In search of a useful 1–21.
theory of innovation. Research Policy 6: 36–76. Von Hippel, E. 2005. Democratizing innovation: The
Nelson, R.R., and E.N. Wolff. 1997. Factors behind cross- evolving phenomenon of user innovation. Journal f€ ur
industry differences in technical progress. Structural Betriebswirtschaft 55: 63–78.
Change and Economic Dynamics 8: 205–220. Winter, S.G. 1993. Patents and welfare in an evolutionary
Nordhaus, W.D. 2007. Two centuries of productivity model. Industrial and Corporate Change 2: 211–231.
growth in computing. Journal of Economic History Wright, T.P. 1936. Factors affecting the costs of airplanes.
67: 128–159. Journal of Aeronautical Sciences 10: 302–328.
Nuvolari, A. 2004. Collective invention during the British Yelle, L.E. 1979. The learning curve: Historical review and
Industrial Revolution: The case of the Cornish comprehensive survey. Decision Sciences 10:
pumping engine. Cambridge Journal Economics 28: 302–308.
347–363.
Pavitt, K. 1999. Technology, management and systems of
innovation. Cheltenham and Lyme: Edward Elgar. Further Reading
Pavitt, K. 2001. Public policies to support basic research: Beardsley, G., and E. Mansfield. 1978. A note on the
What can the rest of the world learn from US theory and accuracy of industrial forecasts of the profitability of
practice? (and what they should not learn). Industrial new products and processes. Journal of Business 51:
and Corporate Change 10: 761–779. 127–135.
Rosenberg, N. 1976. Perspectives on technology. Cam- Bresnahan, T.F., and M. Trajtenberg. 1995. General pur-
bridge: Cambridge University Press. pose technologies: Engines of growth? Journal of
Rosenberg, N. 1982. Inside the black box: Technology and Econometrics 65: 83–108.
economics. Cambridge: Cambridge University Press. Bush, V. 1945. Science: The endless frontier. Washington,
Rosenberg, N. 1994. Exploring the black box: Technology, DC: GPO.
economics, and history. Cambridge: Cambridge Uni- Castaldi, C., R. Fontana, and A. Nuvolari. 2009. ‘Chariots
versity Press. of Fire’: The evolution of tank technology, 1915–1945.
Rosenberg, N. 2009. Some critical episodes in the progress Journal of Evolutionary Economies 19: 545–566.
of medical innovation: An Anglo-American perspec- Chataway, J., J. Tait, and D. Wield. 2004. Understanding
tive. Research Policy 3: 234–242. company R&D strategies in agro-biotechnology: Tra-
Rosenbloom, R.S., and M.A. Cusumano. 1987. Techno- jectories and blind spots. Research Policy 33:
logical pioneering and competitive advantage: The 1041–1057.
birth of the VCR industry. California Management Cohen, W., and R. Levin. 1989. Empirical studies of inno-
Review 29: 51–76. vation and market structure. In Handbook of industrial
Sahal, D. 1981. Patterns of technological innovation. New organization, vol. 2, ed. R. Schmalensee and R. Willig.
York: Addison-Wesley. Amsterdam: Elsevier.
Sahal, D. 1985. Technological guideposts and innovation Consoli, D. 2005. The dynamics of technological change in
avenues. Research Policy 14: 61–82. UK retail banking services: An evolutionary perspec-
Saviotti, P.P. 1996. Technological evolution, variety and tive. Research Policy 34: 461–480.
the economy. Cheltenham: Edward Elgar Publishing. Dasgupta, P., and P.A. David. 1994. Towards a new eco-
Saviotti, P.P., and A. Trickett. 1992. The evolution of nomics of science. Research Policy 23: 487–521.
helicopter technology, 1940–1986. Economics of Inno- Dawid, H. 2006. Agent-based models of innovation and
vation and New Technologies 2: 111–130. technological change. In Handbook of computational
Suarez, F.F., and J.M. Utterback. 1995. Dominant designs economics, Vol. 2: Agent-based computational
and the survival of firms. Strategic Management Jour- economics, ed. L. Tesfatsion and K.L. Judd. Amster-
nal 16: 415–430. dam: Edward Elgar.
12 Technological Paradigms and Technological Trajectories
Dew, N. 2006. Incommensurate technological paradigms? Mina, A., R. Ramlogan, G. Tampubolon, and
Quarreling in the RFID industry. Industrial and Cor- J.S. Metcalfe. 2007. Mapping evolutionary trajectories:
porate Change 15: 785–810. Applications to the growth and transformation of med-
Dosi, G., and M. Egidi. 1991. Substantive and procedural ical knowledge. Research Policy 36: 789–806.
uncertainty: An exploration of economic behaviours in Needham, J. 1962–3. The pre-natal history of the steam
changing environments. Journal of Evolutionary Eco- engine. Transactions of the Newcomen Society 35:
nomics 1: 145–168. 3–58.
Dosi, G., K. Pavitt, and L. Soete. 1990. The economics of Nelson, R.R., and S.G. Winter. 1982. An evolutionary
technical change and international trade. Brighton/ theory of economic change. Cambridge, MA: Harvard
Wheatsheaf/New York: New York University Press. University Press.
Freeman, C., and C. Perez. 1988. Structural crises of Nightingale, P. 1998. A cognitive model of innovation.
adjustment: Business cycles and investment behavior. Research Policy 27: 689–709.
In Technical change and economic Perez, C. 1985. Microelectronics, long waves and world
theory, ed. C. Freeman, R.R. Nelson, G. Silverberg, structural change: New perspectives for developing
and L. Soete. London: Pinter Publishers. countries. World Development 13: 441–463.
Freeman, C., and L. Soete. 1997. The economics of indus- Perez, C. 2010. Technological revolutions and techno-
trial innovation, 3rd ed. London/Washington, DC: economic paradigms. Cambridge Journal of Econom-
Pinter. ics 34: 185–202.
Gary, M.S., G. Dosi, and D. Lovallo. 2008. Boom and bust Polanyi, M. 1962. Personal Knowledge: Towards a Post-
behavior: On the persistence of strategic decision bias. Critical Philosophy. Chicago: University of Chicago
In The Oxford handbook of organizational decision Press.
making, ed. G.P. Hodgkinson and W.H. Starbuck. Possas, M.L., S. Salles-Filho, and J.M. Silveira. 1996. An
Oxford/New York: Oxford University Press. evolutionary approach to technological innovation in
Geuna, A., A. Salter, and W.E. Steinmuller (eds.). 2003. agriculture: Some preliminary remarks. Research Pol-
Science and innovation: Rethinking the rationale for icy 25: 933–945.
funding and governance. Cheltenham: Edward Elgar. Soete, L. 1979. Firm size and inventive activity: The evi-
Hall, B.H., and N. Rosenberg (eds.). 2010. Handbook of dence reconsidered. European Economic Review 12:
the economics of innovation. Burlington: Academic. 319–340.
Kerker, M. 1961. Science and the steam engine. Technol- Starbuck, W., and J.M. Mezias. 1996. ‘Opening Pandora’s
ogy and Culture 2: 381–390. Box’: Studying the accuracy of managers’ perceptions.
Merton, R.K. 1973. The sociology of science: Theoretical Journal of Organizational Behavior 17: 99–117.
and empirical investigations. Chicago: University of Stoneman, P. 1995. Handbook on the economics of inno-
Chicago Press. vation and technical change. Oxford: Blackwell.