Codification and Modes of Innovation
Department of Business Studies Aalborg University
Department of Business Studies,University of Aalborg
LATAPSES, University of Nice Sophia-Antipolis
Morten Berg Jensen
Last modified: May 25, 2004
DRUID summer conference 2004, Abstract
In an earlier paper (Johnson, Lorenz and Lundvall, 2002) we analysed the role of codification of knowledge and took as our starting point a critical review of a paper produced by Cowan, David and Foray (Cowan et al, 2000). In this paper we will argue that the different perspectives on codification reflect that we (JLL) and the others (CDW) focus on different modes of innovation that co-exist and co-evolve. There is one mode, which is dominated by science and scientific research and one, which emanates from more everyday economic activities. We can see the same split taking different forms in much of the literature on innovation and knowledge: The information society versus the learning economy, the linear model of innovation versus the interactive model, etc. Furthermore, the split is observable on different levels: Research, policy and knowledge management.
The first mode of innovation can be called the STI-mode. It is highly influenced by a style of work taken from the research laboratory and it is dominated by systematic search for solutions of reasonably well-defined problems. The outcome of the process is epistemological knowledge and serious attempts will normally be made to document this outcome. The transparency of the outcome often makes it necessary to protect the knowledge legally.
The second mode of innovation can be called the DUI-mode. It is rooted in everyday day routines, which lead to learning, by Doing, Using and Interacting. It is focused on technological opportunities and user needs but is opportunistic rather than systematic. The outcome of the process is tangible or intangible new products as well as new or improved competence and know-how shared within networks.
The STI-mode and the DUI-mode of innovation are connected to each other and both are further developed (co-evolving) in the learning economy. For both modes the motive for and the consequence of change is an attempt to speed-up knowledge production, diffusion and utilization. The STI-mode tends to be combined with on-line experimenting while the DUI-mode tends to be combined with the conscious implementation and use of learning organizations. But these developments do not necessarily point toward convergence or peaceful co-existence between the two modes.
While the STI-mode, including the on-line version, strives for complete codification of everything, the DUI-mode, including the learning organization-version, can only thrive in a world of incomplete codification. Trust, unspoken rules, dialectical concepts and discursive rationality are fundamental for the DUI-mode. While there in the STI-mode is an attempt to translate insights gained into ‘global and explicit codes’ the DUI-mode is one where ‘local and implicit codes’ predominate. In the first case the aim of codification is to ease communication of explicit know-why type of knowledge. In the second case the aim of the ‘codification’ is to ease the communication of implicit know-how types of knowledge and the codes developed tend to remain ‘local’.
The two modes of innovation co-exist and compete in all market-oriented activities in the learning economy. The STI-mode is used most widely in science-based activities while it is much less used in more competence-based activities. The DUI-mode is of fundamental importance for all activities – also within science-based activities it is necessary to establish informal communication based upon trust. Linking to the needs of users is equally important. Crises in ICT-sectors and in genetic engineering applications reflect an underestimation of the DUI-mode. On the other hand the extension of the STI-mode to ‘traditional sectors’ may give very high rates of return since there is often too little systematic documentation and codification in these sectors create new opportunities supported by ICT.
Recognizing the co-existence and co-evolution of the two modes may represent major progress in innovation theory. It might correspond to how the ‘innovation as an interactive process’-perspective did overcome the traditional split between those who argued that supply-side factors were most important and those arguing that it was demand that dominated when it comes to determine the rate and direction of innovation. It will make it clearer why there is a need for developing context-specific theories. Many of the most interesting analytical issues may be connected to understanding how the two modes meet and create conflict and tension.
Specifically we believe that knowledge management suffers from a similar split among analysts and practitioners at the enterprise level, as has been the case within academic innovation research. The most difficult challenge for management today might be how to combine the two modes in such a way that they do not undermine but support each other. Some departments in the firm will make attempts to take things toward the STI-mode (especially the Accounting and the R&D-department) while other departments will insist on giving the DUI-mode a stronger position (especially the production and the market departments).
Thinking in terms of the two modes and their evolution in the learning economy may also have rich implications for policy and institution building. Education may prepare students to work with specialized global codes in the different disciplines as well as involve them in learning to develop and use local codes through problem-based learning. The design of intellectual property rights and of labor contracts might need to strike a balance between the two modes. The over all organization of innovation policy in terms of ministries and departments needs to consider how to balance the two modes.
The outcomes of the two modes are more or less accessible to different categories of users of knowledge and have different implications for what strategies should be applied. For catching-up economies with a low per capita income it is important to consider this issue. Identifying innovation with the STI-mode is a common and potentially very costly misinterpretation in the south. The STI-mode calls for technological licensing, university training, scientific expertise and global mobility of scientists and engineers. Reverse engineering and imitation might give access to the outcomes of the DUI-mode. For this mode it is also crucial to build social capital and know-who within and outside the national innovation system.
We see codification primarily in the context of communication (including communication over time – as memory). There are codes aiming at keeping communication secret and codes aiming at making communication accessible to everybody. In between we find local codes that have developed as an outcome of learning and which facilitate future learning by interacting. While the STI-mode results in global explicit codes the DUI-mode is based upon local implicit codes. To develop global codes is costly. While ICT makes codification a more attractive option, the acceleration of the rate of change in the learning economy works in the opposite direction. In a world characterized by complexity and rapid change the DUI-mode with its local implicit codes will remain fundamental for economic performance.
The relation between the ICT-mode and the DUI-mode of innovation is similar to the relation between codified and tacit knowledge. They are both fundamental dimensions of the learning economy. The further development of the learning economy will not lead to the substitution of the ICT-modes for DUI-modes any more than codified knowledge is substituted for tacit knowledge. Many of the most important issues at the research level, the knowledge management level (the enterprise level) and the policy level are to be found in the tensions and contradictions between the two modes.
Cowan, David and Foray (2000), ‘The explicit economics of knowledge codification and tacitness’, Industrial and Corporate Change, 9, 211- 253.
Johnson, Lorenz and Lundvall (2002), ‘Why all this fuss about codfied and tacit knowledge’, Industrial and Corporate Change, 11, 245-262