Professional Documents
Culture Documents
Cohen Det Meyer K Rahm Er 2001
Cohen Det Meyer K Rahm Er 2001
BETA, Universite Louis Pasteur, 61, Avenue de la Foret Noire, 67085 Strasbourg Cedex, France
Fraunhofer Institute for Systems and Innovation Research ISI, Breslauer Str. 48, 76139 Karlsruhe, Germany
Abstract
The purpose of the following contribution is to present the main consequences in terms of policy of firms and governments
that emerged from the theoretical and empirical work achieved under the TIPIK project. In a first part, the introduction is
devoted to a summing-up of the main theoretical dimensions used in TIPIK, that led in particular to a special issue of Industrial
and Corporate Change (June 2000). Then, the discussion envisages the main consequences of the theoretical debate in terms
of policy issues and addresses the new perspectives of a knowledge oriented policy of firms and governments. Finally, we
expose a selected series of policy implications, that are derived from the empirical studies presented in this special issue.
2001 Elsevier Science B.V. All rights reserved.
Keywords: RTD; Policy; Codification; Communities; Competencies; TIPIK project
1. Theoretical framework
1.1. The nature of the codification process
The codification of knowledge is a process that aims
at reducing and converting knowledge into messages.
These messages can then be processed as information
that will serve to reconstitute knowledge at a later
time, in a different place, or by a different group of
individuals. The main interest of this process of transformation of knowledge is to facilitate the treatment
of knowledge as an economic good, which can be exchanged, in particular on markets. 1
Corresponding author. Tel.: +33-3-90-24-21-89;
fax: +33-3-90-24-20-71.
E-mail address: cohendet@cournot.u-strasbg.fr (P. Cohendet).
1 The treatment of knowledge as a commodity is sometimes called
commodification of knowledge. A risk of confusion is to assume that codification of knowledge will automatically allow the
commodification of knowledge. Codification obviously facilitates
the exchange of knowledge, but in many situations, knowledge
can be transferred without codification (for instance by recreating
0048-7333/01/$ see front matter 2001 Elsevier Science B.V. All rights reserved.
PII: S 0 0 4 8 - 7 3 3 3 ( 0 1 ) 0 0 1 6 8 - 8
1564
(numerical, symbolic, geometrical languages and taxonomies of many kinds), standards of performance, a
vocabulary of precisely defined and commonly understood terms, and a grammar to stabilise the language.
Once these steps have been achieved, a code-book
becomes available, and agents are able to carry out
knowledge operations at low marginal costs, since
messages are reproducible. 3
It must be emphasised that knowledge is simultaneously an input and an output of the codification
process. In other words, some knowledge is needed
to codify knowledge, and furthermore, knowledge is
needed to exploit a given piece of codified knowledge. This recursive and dynamic structure of knowledge has two consequences: one is that knowledge
cannot be regarded as a mere stock resulting from
an accumulation of information. The second is that
knowledge cannot be considered separately neither
from its holder(s), nor from its time/space location.
Both arguments stress the importance of the (individual and collective) cognitive processes through which
knowledge is converted, combined, stored, retrieved,
exchanged, and interpreted.
1.2. The relationships between codified and tacit
knowledge in the codification process
The codification process alters the relationship between the codified and the tacit form of knowledge.
On the one hand, codified knowledge refers to a view
of knowledge as necessarily explicit, formal or systematic which can be expressed in words and numbers, scientific procedures, or universal principles.
This codified category of knowledge is easy to transfer, to store, to recall and to valorise. On the other
hand, the notion of tacit knowledge refers to that form
of knowledge that cannot be easily expressed. The
main forms of tacit knowledge are know-how (that
results from the accumulation of practice), the mastering of a language (that results from the accumulation
of the ability to communicate) and representations
3 As noticed by Steinmueller (2000), this simplified vision of the
process of codifying knowledge as a three step process, does not
encompass all the cognitive issues associated with the codification
process. In particular, there may be problems in aligning cognitive
understanding with the language by which models and messages
are constructed.
guage in order that codified knowledge can be reconstituted as operational and generative. Moreover, as
more knowledge gets codified, the nature of the tacit
form is also generally changing. In other words, the
process of codification is a process of knowledge creation that alters both the codified and the tacit forms of
knowledge. 7 As new knowledge is codified, new concepts and terminology will inevitably be introduced so
that the codification of knowledge inherently involves
further creation of knowledge.
The combination and the composition of tacit and
codified knowledge depend strongly on the context
within which agents or organisations manipulate
knowledge. This means in particular that there are
contexts in which agents will be willing to invest
more on codification, and other contexts in which they
would rather use and reinforce their tacit knowledge.
Thus, the ability of a cognitive agent to exploit different categories of knowledge matters. On the one hand,
the existence of given tacit forms of knowledge (beliefs, languages, know-how), of accumulated learning
and habits, and of norms will shape the ways codified
knowledge is produced. On the other hand, the way
codified knowledge is produced (the nature of the
codes, the types of organisation, the nature of physical
carriers of knowledge) will also shape the ways learning processes are directed, focused and assimilated.
The context-dependence nature of knowledge clearly
suggests that tacit knowledge also refers to knowledge
which is not mobilised (at least consciously) when
conducting some activities in a given context. To this
extent, some codified knowledge can be made tacit by
an agent, who places part of her knowledge in a zone
of subsidiary awareness, whereas some other part is
put under her focal awareness (Polanyi, 1962).
The context-dependence of the codification process
is particularly important when considering the type
of tacit knowledge that is the point of focus of the
codification effort. The nature and the costs of the
codification process is not the same if one aims at
codifying a belief, a representation of the world, a specific know-how, or the mastering of a given language.
Cowan this issue showed that in the case of expert sys7 As Cowan et al. (2000) underlined in the case of the design
of expert systems, the act of codifying is not merely translating
the experts knowledge out of his head and onto paper, but is
typically an act of knowledge creation.
1565
1566
1567
1568
As underlined by Cohendet et al. (2001), in the traditional vision of the firm that dominated in the 20th century, there was a
clear delineation between the communities in charge of a deliberate production of knowledge (epistemic communities) and the
communities focused on the use of knowledge (these include the
hierarchical communities and autonomous communities such as
communities of practice). There is now less and less separation
between the communities in charge of producing new knowledge
and the communities in charge of using and transferring knowledge. The ability to proceed to on-line learning (as we shall see
further with the development of experimental learning) affects all
the types of communities that where not focused on the deliberate
production of knowledge.
20 In a first step, some parents gathered through a mutual commitment, by exchanging best practices, in order to improve the
situation of their children. The community (of practice) attracted
quickly new members, and succeeded in cumulating a critical mass
of knowledge that allowed in turn epistemic communities to function (in particular scientific ones) in order to continue the process
of knowledge building.
21 It is an operating system based on Minix, a small operating
system for PC inspired from Unix.
22 However, in order to cope with the flux of contributions, a
committee in charge of evaluating the incoming contributions was
created. This committee was then apart and above the bunch of
developers. There was thus the creation of a procedural authority
The development and expansion of the knowledgebased economy will enhance the role of communities. They are the nucleus of knowledge formation
and use. They encapsulate incentive mechanisms to
produce new knowledge, either through passion or
commitment to the collective action, or through the
adherence to a procedural authority as was the case
for the epistemic communities. The role of communities is also essential in the distribution and dissemination of knowledge throughout society. The process
of codification aims also at reaching one or several
communities (some codification processes aim at the
same community in which the building of knowledge
is taking place, in general codification processes are
directed towards specified communities, sometimes
the codification has a universal vocation). To a large
extent, the process of diffusion of knowledge can be
seen as a process of sequential transmission of knowledge through communities. This interactive process
supposes active means of mediation of knowledge,
but it also reveals the ambivalent role of the codification of knowledge. As noticed by a recent OECD
report (2000, p. 27), on the one hand, the production
and use of highly specialised codes or codes using
technical or local jargon would create an obstacle to
appropriation in the field by lay people or potential
users of the knowledge. On the other hand, a lack of
codification would also create an obstacle as users
would not have access to sufficiently explicit knowledge. This ambivalence indicates the importance
of designing and implementing metacodes or semicodes as mechanisms for developing compromises
between the need to make knowledge more explicit
and the need to avoid excessive technicalities and
local jargon. In this perspective, one can measure
how much the dynamics of social networks and the
co-evolution of codification methodologies and social
organisations are interrelated.
To a large extent, the process of codification of
knowledge will, thus, exhibit different modes and patterns according to the community within which the
process of codification is undertaken, and the community (or communities) which is (are) the target of the
codification process. The community at the origin of
to judge what constituted a progress toward the objective. It should
be noted that the objective evolved from the enhancement of a
student project to the building of a competitor for Windows NT.
1569
The above developments have deliberately emphasised two main dimensions of the transformation of
knowledge, the tacit/codified dimension and the individual/collective dimension. However, as underlined
by Malerba and Orsenigo (2000), the multiplicity and
the contextuality of other features of knowledge are often underestimated. There are many other dimensions
of knowledge (declarative knowledge versus procedural, procedural knowledge versus generic knowledge)
that require a careful examination which is beyond the
scope of this contribution.
1.4. Benefits and costs of the codification process
As it was mentioned in the introduction, the main
expected benefit from the process of codification
is that knowledge is picking up some properties of
a (private) commodity. Codification facilitates the
commodification of knowledge. Knowledge becomes transferable independently of the transfer of
other things, such as people, in which the knowledge is embedded (Cowan and Foray, 1997). This
implies that market transactions are facilitated by
codification, reducing uncertainties and information
asymmetries in transactions involving knowledge. 24
It was shown for instance that the production of
quality assurance standards (Gunby, 1996) reduced
the asymmetric information on the related markets.
In a similar perspective, codification also facilitates knowledge externalisation and allows a firm to
24 It should be emphasised that the use of knowledge is at the
origin of very specific asymmetries. As Bessy and Brousseau
(1998) noted: Regarding its value, knowledge is a highly uncertain
resource when it is transferred between two economic units. This is
partly due to the well-known information asymmetry about quality
inherent in every market transfer of information (Arrow, 1962).
But it is also due to the radical uncertainty over the potential use
of knowledge. Two different phenomena are in question: (1) the
receiving party can use knowledge in ways that are not easily
observable by the prior holder. This is typically a moral hazard
problem (hidden action), but in a specific informational context
where it is impossible to build ex ante incentive schemes or
(infaillible) supervision mechanisms because all the potential uses
of knowledge cannot be anticipated and made observable, and (2)
knowledge is itself an input in the process of knowledge creation.
The unit that benefits from a transfer can be conducive to the
creation of new knowledge, that will increase the value of the
initial knowledge. The question is then how to remunerate each
party for its contribution to the knowledge creation process. Again,
due to the uncertainty of such a process, it is very difficult to
solve it ex ante by an optimal incentive scheme. . . .
1570
1571
1572
of the traditional research, technology and development (RTD) policies. These traditional policies
were inspired by the vision of linear models of innovation, where the production of new knowledge
could be assimilated to mere information. The rationale for designing policy instruments in this traditional vision is based on supplying market failures due to externalities in the production of new
knowledge-reduced-to-information. Several ways of
intervention have been proposed, and often tested, in
order to move more closely to the optimal level of
research. David (1993) speaks about the three Ps:
public patronage (prizes, research grants, subsidies,
etc.), state procurement (or production) and the legal exclusive ownership of intellectual property that
shape the nature of government interventions to supply the failure of the market. The focus of traditional
RTD policies is thus essentially on the conditions of
production of new knowledge, and not on the ways
knowledge is assimilated and diffused through society, since, according to the traditional vision, agents
are supposed to be able to assimilate new knowledge
without significant costs.
As the knowledge based economy is growing, there
is a crucial need to think in terms of knowledge oriented policies (KOP) which could take into account
the specific characteristics of knowledge that have
been examined, in particular the specific impacts of
the processes of codification. Contrary to traditional
RTD policies, KOP refer to the non-linear interactive model of innovation, and assume that knowledge
and competencies are crucial factors for production,
innovation and competitiveness rather than physical
resources. Highlighting interactions that allow the generation and diffusion of knowledge, this approach differs from linear models (science-push, market-pull) for
which knowledge constitutes only one input among
others intervening in a sequential process. In particular, a key difference is that, for KOP, the way knowledge is assimilated and acquired by agents (the ways
competencies are built), is as important as the conditions of production of new knowledge (the ways innovations are produced).The KOP should promote ways
of codifying knowledge that contribute to enhancing
the generation of knowledge externalities, since the
process of knowledge generation is cumulative and integrative. Cumulative forms of knowledge are those
in which todays advances lay the foundation for suc-
1573
1574
fusion process relies, are isotropic. As in a market mechanism, one can refer to a representative
agent, who benefits from the knowledge spillovers
emitted by the producer of knowledge.
5. The producer of knowledge is not supposed to have
emitting capacities, that is the ability to tune the
disclosure/secrecy dimension. He is just supposed
to try and avoid the loss of the integrity of the piece
of new produced knowledge.
6. All the agents of the economy have the full capability to absorb the innovative idea emitted by the
producer of knowledge. Any buyer of the knowledge can effectively destroy the market, since he
can reproduce the knowledge at very low cost.
7. The epistemic content of knowledge does not
matter. The content of knowledge exhibits a
cognitive equivalence, which means that in such
a context it is impossible for instance to distinguish between the generic and specific forms of
knowledge.
From these hypotheses, Arrow showed that the
characteristic of knowledge-reduced-to-information
as a public good implies the existence of some major positive externalities that prevent the effective
functioning of a market for knowledge. Any buyer
of the knowledge can effectively destroy the market,
since he can reproduce the knowledge at very low
cost. If the producers of knowledge cannot appropriate the benefits of knowledge, then they have no
incentive to produce it. Without an external intervention, the incentives for doing research are not high
enough and the level of research in society will be
under-optimal. The consequences of this broad traditional vision, that assimilates knowledge to mere
information, were considerable. Since the knowledge generated by the different research activities
possesses the generic properties of public good, it
cannot be optimally produced or distributed through
the workings of competitive markets. Here is the
justification for government subsidisation of science,
technological and engineering research, and for innovative activity more generally (the3P). This vision
shaped the conception of public intervention in R&D
during decades. It justified the role and creation of
public laboratories, of research centres, of public
programmes of R&D, of public institutions (patent
offices, for instance), of public infrastructure for trans-
1575
1576
producer of knowledge. First, firms cannot assimilate, absorb knowledge without any effort. 39 In
order to absorb new external information, the firm
needs to develop what Cohen and Levinthal (1990)
called an absorptive capacity. The principle is that
we cannot understand something if we know nothing. Firms need to build a knowledge background,
based on knowledge previously acquired, to be able
to absorb the external knowledge the two faces of
R&D. It not only creates new knowledge, it also
helps firms to assimilate external knowledge. Thus,
Cohen and Levinthal show that spillovers do not
necessarily have a negative impact on R&D.
7. The epistemic content of knowledge matters. As
noted by Callon et al. (1999), we should distinguish
the case of knowledge having a high degree of generality (knowledge which can be potentially used
in various contexts by a large variety of agents)
from the case of very specific forms of knowledge that can be absorbed and used by few other
agents.
This reconsidered set of hypotheses leads to the following remarks in terms of KOP.
2.1.3. New interpretations of traditional hypotheses
It must be emphasised that the above hypotheses of
the conditions of production of innovative ideas in a
KOP context do not invalidate the traditional hypotheses. In fact, the traditional context appears as a particular case of the more general one. This implies that the
traditional policy instruments (3P) elaborated within
the traditional context, will still be valid as instruments
in a KOP context. However, the ways to interpret them,
39 As Saviotti (1998) mentioned Traditionally, knowledge was
considered a public good because it is impossible for its creator to
prevent it being used by economic agents who do not pay anything
in exchange of it. However, even a completely codified piece of
knowledge could not be used at a zero cost by anyone. Only
agents who know the code can use the piece of knowledge at
zero (imitation) cost. Agents who do not know the code, if they
realise the economic value of a given piece of knowledge, have
to know the code before being able to retrieve and to imitate. The
cost of learning has to be included among the costs of imitation. . .
the concept of absorption capacity tells us that in order to be able
to access a piece of knowledge, we must have done R&D on
something similar. As a consequence, we have included among the
variables determining the degree of appropriability, not only the
degree of codification, but also the fraction of agents knowing the
code and the distribution of agents with respect to the frontier.
1577
Fig. 1. The role of incentive mechanisms: differences in Romers and DasguptaDavids approaches concerning the production of knowledge
(Callon, 1999).
1578
1579
1580
44
In the domain of competencies, the firm holds significant pieces of knowledge, but needs to access
complementary forms of knowledge held by other
firms to be able to develop and use the knowledge
efficiently. The firm has, thus, incentives to participate to networks, in order to proceed to the mutual
exchange of complementary forms of knowledge.
As it has been underlined, networks offer precisely
a way to share and exchange knowledge complementarities.
In the peripheral zone, the firm does not hold any
specific advantage in terms of knowledge and the
risks for it in terms of asymmetries of information is very high. This is a zone of quasi market,
where the degree of supplierbuyer interdependence
is generally low. Products are standardised ones, and
require few interactions with other inputs. Contracts
are arm-length contracts, the duration of which depends on the classical transactional parameters. 46
In this zone, firms organise the process of on-going
allocation of resources and adaptation to the environment. The latter requires governance mechanisms that are well analysed by the transaction cost
approach. 47
The building of competencies by firms constitute
one of the major sources of heterogeneity between
firms and their respective competitiveness. It has
strong impacts on the processes of codification of
knowledge within and between firms:
In the core domain, firms will favour those processes allowing both a strong absorptive capacity
and a strong selecting emission, understandable
only by the firm and the small nucleus of partners.
It is a domain, where part of the codification effort
is devoted to code productive and organisational
46 For a given firm, in terms of supplier management practices,
this zone requires a minimal assistance to suppliers, with single
functional interfaces (sales to purchasing for instance), and the
practice of price benchmarking.
47 This hypothesis implies that there are two main governance
mechanisms to be considered by the firm: the mechanism for governing competencies, defined by the need to co-ordinate knowledge, and the mechanism for governing transactions, reliant on
the need to manage transactions (see also Amin and Cohendet,
2000). The organisation of the firm requires a dual structure of
governance, one for governing the domain of the core activities
of the firm, the other one to govern the remaining activities (the
peripheral domain).
1581
1582
policy implementation can be thought of for the realisation of KOP. Networks, clusters and other communities of firms and/or individuals may be starting-points
for knowledge-oriented policy measures. The aim of
KOP in this respect is the generation, transformation and diffusion of knowledge through learning processes.
3.1. KOP for firms
The building of competencies is the main driver
of the ways knowledge is used as an input by firms.
It will determine the ways they acquire, absorb and
use knowledge. In particular, it will explain the constitution of their absorptive capabilities, of the degree
and nature of interaction with others, and of their signalling strategies. KOP are characterised by a vision
of the firm in which knowledge-based competencies
constitute the main source of competitive advantages.
These advantages are based on learning processes
within the firm and beyond the boundaries of single
firms, namely, in inter-firm relationships. In this respect, RTD policies distinguish between competition
and co-operation whereas KOP consider opportunities
induced by situations of co-opetition.
Traditional RTD policies focusing on manufacturing industries are dominated by thoughts of specialisation. KOP on the contrary pay particular attention
to interfaces and overlapping between manufacturing
and service activities (e.g. special focus on knowledge
intensive services (KIBS). KOP should pay attention
of the fact that each firm, whatever its size or sector, could have the possibility to build and develop
competencies. This raises in particular the problem of
SMEs that do not always have the critical mass of internal knowledge to build a process of accumulation
of competencies, or cannot afford the fixed costs incurred by the building and managing of competencies.
This certainly encourages the setting up of new agents
of knowledge such as KIBS to make SMEs competencies explicit and to assist the further development
of these competencies. However, in this perspective,
public policy should progressively support absorptive
capabilities inside the small units themselves.
Considering policy instruments, grants and subsidies constitute typical tools of traditional RTD policies, according mainly to the picking the winners
principle. Policy instruments of KOP could be initia-
1583
1584
born. Allowing community members to access knowledge from other sectors or from academic research
may help, although policy measures in this direction
may be in contrast with the appropriability measures
and staff management practices of the employees of
the episternic community members. It should be noted
that many technology transfer actions which currently
target existing SMEs as potential innovators, could
be instead directed at giving some members of local
epistemic communities the chance to found their own
start-up.
More generally, within a KOP concept one should
distinguish between traditional technology transfer
and knowledge transfer. The target groups of technology transfer are users and practical communities primarily, the target groups of knowledge transfer mostly
epistemic communities. Technology transfer policies,
which focus on specific sectors and locations, but do
not arise from an agreement with local members of
the existing epistemic communities, are very likely to
end up offering very generic, and possibly irrelevant
services (as many assessment of technology transfer
policies actually show). Since knowledge circulates
within a number of relatively close networks, policy
initiatives have to focus on access to knowledge and
inter-personal networks, the degree of geographical
dispersion of the relevant epistemic community, and
the extent to which knowledge can be considered
as public (i.e. shared by different communities) or
semipublic (i.e. circulating only within one community). Some of the links between SMEs and larger
firms which many technology transfer policies try to
set in motion, are already in place within the existing
epistemic communities.
The study by Amesse, Cohendet (this issue) views
the process of technological transfer as a process that
depends on the ways firms and other institutions deal
with knowledge. On the one hand, they underline the
role of absorptive capacities as essentially active along
the perspective suggested by Cohen and Levinthal
(1990). They show that the more groups, teams and
communities within the firm are receptive to new ideas
the higher are the chances of an efficient absorption
of technologies from outside. On the other hand, the
quality of the process of technology transfer is also
fundamentally dependent on the firms capabilities to
emit knowledge outside its frontier. When firms provide significant assistance to their strategic partners,
through multiplying functional interfaces and investment in knowledge sharing routines for instance, they
in fact deliberately contribute to enhancing the absorptive and emittive capacities of their key suppliers.
The authors also show, when negotiating the rights
of access to complementary forms of knowledge that
they need within networks, they carefully, and permanently assess the absorptive and emitting capacities
of the other members of the network. In other words,
the management of the technology transfer process is
essentially bi-directional. What matters is more the
co-evolution of the mutual absorptive and emitting capacities between partners, than the mere observation
of the technology flow between an emitter and a receiver.
Similar conclusions can be drawn in the case of
university/industry linkages (see Section 4.4) and in
the case of the dense interactions between knowledge
intensive services and their customers (see Muller
and Zenker and Creplet et al., this issue). The study
of DAdderio (this issue) on software development
demonstrates that standardised, coded procedures
and models are of little use unless they are locally
appropriated and effectively transformed into actional
routines and prototypes. Diffusion of standardised
practices, models and methodologies run the risk of
seriously miss-stating the organisational costs and
productivity effects of software adoption processes.
As a consequence the study shows that software
producers need to build greater flexibility and customisation potential into their systems in order to
facilitate the process of adaptation of generic systems
to local, context specific, circumstances and requirements. These emittive and absorptive capabilities lead
to specific requirements, to dynamic learning, translation routines etc. All these cases demonstrate that
policy has to go far beyond R&D and to focus more
on competencies.
4.2. Shifting from R&D to competencies: the specific
problems of small and medium-sized enterprises
The study by Muller and Zenker (this issue)
comes to the following conclusions: compared to
medium-sized manufacturers, small enterprises are
characterised by a lower level of knowledge-intensive
interactions. This means mainly that small firms acquire less innovation-related information from com-
1585
1586
1587
target group. On the other hand, in order to stimulate co-operations between KIBS and other types of
firms, the visibility of the former firm type should be
raised, especially for small and medium-sized firms
that often lack information concerning co-operation
partners. A helpful means could be a special type of
certification of KIBS in order to label their competencies. This could be useful for KIBS marketing as well
as for manufacturing SMEs to get information about
KIBS offers. Furthermore, innovation policy can increase collaboration by giving incentives for using
KIBS services. The benefit from this kind of support
cannot only be expected in innovation activities of
manufacturing SMEs, but also in KIBS internal innovation that are nurtured by knowledge they gain
from co-operation with their manufacturing partners.
The emittive capacities of KIBS, the absorptive capacities of SMEs and the level of interaction between
both are the main targets of policy.
4.4. Redesign of incentives: dynamics of communities
The promotion of university/industry relationships
is a well-established science and technology policy
area. A long tradition of different institutional settings,
financial incentives and other tools of technology
transfer are used. Although there is a broad consensus
that the linear model of innovation is inadequate, the
concept of the one-way bridge from public research
to industrial research is still widespread in the discussion on technology transfer. The contributions of this
volume support the need to refocus the traditional
policy of technology transfer the establishment of
transfer institutions and incentives within the public
research world and to convert it into a concept
supporting a two-way bridge. A survey at German
universities (Meyer-Krahmer and Schmoch, 1998) has
shown for all fields examined that the central linking
element in the co-operation between universities and
industrial firms is the exchange of knowledge in both
directions. Although the institutional orientation of
academic and industrial researchers is different, the
exchange of knowledge can be considered a common
denominator where both interests meet. The mutual
exchange of knowledge in techno-scientific communities is obviously a broad phenomenon that is not
limited to some exceptional cases, but applies to
whole disciplines and sub-disciplines.
1588
Another major conclusion underlines a characteristic of a national innovation system which one
could call structural absorptive capacity. We call
the absorptive capacity (of firms, research institutes
or universities) structural, because it largely depends on meso-level characteristics. This structural
absorptive capacity depends, according to empirical
evidence, on the science intensity of technologies,
the industrial life cycle, and the firm size structure
of the industrial sectors. Furthermore, the absorptive
and emittive capacities depends on micro-level factors: internal R&D capacity of firms, and interaction
patterns to relevant technologies outside traditional
linkages, as well as formal co-operation and informal
networks.
The emergence of national systems of innovation
has to be considered as a path-dependent evolutionary
process where various economic, technological, social and cultural factors interlock and strengthen each
other mutually. The cognitive structures of universities are strongly influenced by the co-evolution of
industry structure, technology and institutional factors
(see also Nelson, 1994), e.g. the close interaction of
application-oriented university departments and industrial firms in the less science-based area of mechanical
engineering in Germany has to be interpreted in this
theoretical perspective. For many years, it has been a
successful model of co-operation, but in the present
situation, it implies risks of lock-in effects which
can be overcome only by deliberate efforts. There is
broad evidence that in addition to the factors already
mentioned, the specific cognitive structure in this area
supports the relatively strong orientation of universities on application and hampers the openness towards
new, more theory-based technologies. However, this
aspect has not yet been studied in a systematic way. So
we suggest that more research on cognitive structures
of different technologies is undertaken, because it can
contribute to an improved understanding of national
systems of innovation, but also of related problems
such as the organisation of inter-disciplinarity or the
relations between science and technology.
The paradigm of inter-disciplinarity affects these
instruments by a shift from inter-organisational
to intra-organisational linkages. The paradigm of
science-based technologies calls for a renewal
of traditional instruments to improve university/
industry co-operation through new instruments link-
ing academic communities with research communities in industry. Interaction and communication are
more important than financial issues. This leads to
the following set of conclusions.
An important consequence for science technology policy is the necessary change of organisation,
communication, interaction and incentives within
the public research world. The consequences for
science-based technologies are the need for new ways
of linkage between basic research and applied research. Also trans- and inter-disciplinarity need better
horizontal linkages between disciplines. Often in the
literature inter-disciplinarity is a misleading concept
of integration and mixing disciplines. It is essential to
understand that first-class trans- and interdisciplinary
research is highly dependent on first-class disciplinary
quality of the scientists involved in interdisciplinary
research. Therefore, an efficient linkage between (and
not the integration of) disciplines is crucial for the
establishment and dynamics of such epistemic communities. Possible tools and mechanisms of such
linkages are
Organisation of research: problem orientation in the
case of well definable social or industrial-technical
problems. This requires, in contrast to the currently
predominant internal objectives of science, new
ways of project organisation and management.
A better linkage of the long-term applicationoriented basic research with applied research would
meet future requirements. This could be achieved
for example by a renewed institutional network,
co-operation research with specific subjects, new
models of financing, improved communication and
other assessment criteria.
Team research: besides the currently predominant
orientation of academic research towards individualised research setting, interdisciplinary team
research must be strengthened by appropriate incentives.
Improved intra- and inter-sectoral mobility of researchers: on an international level and also between
science and industry.
Increased flexibility of research structures: more
rapid taking-up of new developments by flexibilisation of the present rigid public service rules and
budget laws, deregulation of the academic administration, and networking of research institutions for
1589
1590
Economie
de la Science, Revue dEconomie Industrielle, No.
79. Leconomie Industrielle de la Science, pp. 7794.
Knorr-Certina, K., 1981. The Manufacture of Knowledge, An
Essay on the Constructivist and Contextual Nature of Science.
Pergamon Press, Oxford.
Lave, I., Wenger, E., 1990. Situated Learning: Legitimate
Peripheral Participation. Cambridge University Press, New
York.
Machlup, F., 1983. Knowledge, Its Creation, Distribution and
Economic Significance. Princeton University Press, Princeton,
NJ.
Malerba, F., Orsenigo, L., 2000. Knowledge, innovative activities
and industry evolution. Industrial and Corporate Change 9 (2),
289314.
Meyer-Krahmer, F., 1997. Public research/industry linkages
revisited. In: Barre, R., Gibbons, M., Maddox, J., Martin, B.,
Papon, P. (Eds), Science in Tomorrows Europe. Economica
International, Paris, pp. 155173.
Meyer-Krahmer, F., Schmoch, U., 1998. Science-based
technologies: universityindustry interactions in four fields.
Research Policy 27, 835851.
1591