Introduction

The knowledge management and knowledge economy literature often refers to knowledge development as a result of interrelations between different knowledge forms, in particular the interrelation between codified and uncodified knowledge (Nonaka and Takeuchi 1995; Spender 1997; Amin and Cohendet 2004). Furthermore, enterprise and workplace development literature often takes as a point of departure an emphasis on locally communicated and intersubjective generated knowledge (Gustavsen et al. 1998). The intersubjective argument (in its principal form) implies that knowledge is locally constructed and has no or less meaning outside the local context. The communicative argument holds that all knowledge is communicated through (everyday) language that is rooted in the lifeworld of the individual. Knowledge thereby only has its full meaning in a local, lifeworld context (Pålshaugen 2002). In line with this Peter Reason and Hillary Bradbury argue that:

Action research is a participatory, democratic process concerned with developing practical knowing in the pursuit of worthwhile human purposes, grounded in a participatory worldview which we believe is emerging at this historical moment. It seeks to bring together action and reflection, theory and practice, in participation with others, in the pursuit of practical solutions to issues of pressing concern to people, and more generally the flourishing of individual persons and their communities (Reason and Bradbury 2001, p. 1).

In Action Research and Action Science, knowledge is locally constructed and co-generated through a reflective process between researchers and practitioners (Argyris et al. 1985; Greenwood and Levin 1998). This has been the guiding methodology in the two Norwegian programmes on innovation that form the background for our reflection in this paper.Footnote 1

In this paper we will refer to the above position as the dualist approach to knowledge—it is an approach where the argument is made meaningful by contrasting it to its opposite (dualism). In the discussion of knowledge, the predominant method has been to make such dualisms. As we argue, this method tends to lead to a misunderstanding of the concept of knowledge, as it is not a phenomenon that can be defined along one dimension.

Based on experience with Action Research in enterprise development programmes, we discuss the dualist approach to knowledge, in relation to local knowledge generating demands. We present some critical remarks on the validity and relevance of the dualist approach to knowledge, and argue that both context and underlying philosophical positions have to be clarified in order for the concept to give meaning, and in order to understand innovation as knowledge creation processes.

Part one: the dualist approach to knowledge

Background: Action Research in enterprise development

It has been a central theme in two large enterprise and workplace development programmes in Norway (Enterprise Development 2000 and Value Creation 2010) to enhance local practical knowledge (Gustavsen et al. 1998; Gustavsen et al. 2001; Johnsen 2002; Levin 2002; Gustavsen 2004). These programmes have had Action Research as their main methodology. Action Research in the Scandinavian tradition has been inspired by two major trends: a position on social science based on an intersubjective, constructivist perspective on knowledge, (Toulmin and Gustavsen 1996; McKelvey 2002; Gustavsen 2004), and over the last decade, the communicative turn (Gustavsen 2004).

Our experiences from working within these national funded enterprise developmental programmes indicate that the results and outputs from these programmes do not match the expectations and goals that were indicated in the startup phase of these programmes (Arnold et al. 2005). The main goal of these programmes has been to enhance workplace development and innovation, that is to develop genuinely new, relevant, and local knowledge in co-operation between practitioners and researchers (Gustavsen et al. 1998). Even though there has been a considerable amount of learning in these programmes, the nature of this learning has not been of such a character that one unarguably can state that it has resulted in new jobs and products. To some extent there have been results regarding change in work processes, but most of these are local adaptations of internationally well-known organization principles (Gustavsen et al. 2001; Levin 2002; Arnold et al. 2005).

We refer to this as a frustration. The frustration relates to the difficulty of being able to identify genuine new local knowledge creation in these programmes. This frustration has its parallel with frustration in Action Research of this Scandinavian work life kind, in order to produce and communicate to the scientific community research results in this field (Greenwood 2002; Gustavsen 2004).

We have increasingly focused our research on trying to understand the type of knowledge generation process we have been involved in, in these programmes. Starting out from a common sense perspective on how to enhance co-ordination, reflection and learning in networks of enterprises and researchers, we have increasingly been asking ourselves if we really understand the knowledge generation processes we are involved with. When we have tried to unroll the underlying assumptions on this issue in enterprise development processes, the picture still is that there is a lack of clarity and consistency. So what is our experience?

  • It has been difficult to specify the role of researcher: is he a facilitator, somebody who brings in theoretical knowledge, or somebody with superior (practical and/or theoretical) knowledge?

  • It has been difficult to identify the knowledge generation process: is it genuine knowledge creation, imitation or diffusion?

  • It has been difficult to find out if there is a “growth” of knowledge or merely a change of knowledge.

In short, we do not know what kind of knowledge generation process we have been part of, because we can not distinguish between genuine new knowledge and knowledge diffusion, between good and bad knowledge, between social processes, power play and learning. Action research gives little guidance to solve this problem. A partial reason for that, is in our view, that the dualist conception of knowledge is a too narrow framework. We think this is a general problem, and one that is important to understand in order to enhance workplace development through Action Research. We also believe that a cause of this problem has to do with the lack of conceptualization of what knowledge is, how it develops and what it does.

It is our belief that this experience refers to limitations with the dualist approach to knowledge. Action Research emphasizes the duality of knowledge, and is based on an approach to knowledge as socially and communicatively constructed. Action Research therefore seemingly seems to give practical knowledge, locally constructed knowledge, a status as most important and most relevant. At the same time Action Research tends to challenge the dichotomy between theories and practice (Argyris et al. 1985; Greenwood and Levin 1998; Gustavsen 2004). This antagonism and contradiction is our point of departure for discussing and challenging the dualist approach to knowledge generation.

The dualist approach to knowledge

If the dualist approach to knowledge is right, one should expect that new knowledge would be created when research (theory, codified knowledge) and practitioners (tacit, uncodified knowing and practice) meet. The duality of knowledge relates to the conception that knowledge can exist as theory and practice (Ryle 1949; Polanyi 1966; McKelvey 2002). For the purpose of our discussion, we will use the two terms local and universal.

In its ideal form, it has been argued that local knowledge or practice is subjective, it belongs to the lifeworld, it is often as knowing how, it is often tacit, and is produced in action (Ryle 1949; Polanyi 1966; Geertz 1983). Scientific knowledge, on the other hand, is often presented as knowing that, as knowledge developed from reflection-on-action, presented in codified form as theory and belongs to the “system-world”, and produced in a separated, epistemic environment (Merton 1973; Knorr-Cetina 1999).

In Table 1, we have tried to identify the use of the dualist conception and have clustered the different conceptions in two columns along some dimensions. We use the term local and universal in order to identify the two clusters of theories and positions, which we call knowledge regimes. This gives us the following two regimes.

Table 1 The “conventional” dualist approach to knowledge
Full size table

The table represents a rough summing up of some of the “conventional” dualist arguments found in the knowledge economy literature. Some argue based on philosophical positions, others argue in a more applied manner, such as how to organize knowledge creation processes, and how to get tacit knowledge more codified. Literatures about these issues are found in many fields, such as geography (Cooke 2002), management and organization theory (Nonaka and Takeuchi 1995; Tsoukas 2005), philosophy (McKelvey 2002), anthropology (Geertz 1983), work-sociology (Gustavsen 2004), and communicative theory (Pålshaugen 2002). The perspective on knowledge economy and the dualism between local and global knowledge has a strong standing in applied economic geography, i.e. in influencing policies in Europe, in particular through the Lisbon process (Lundvall 2002; Rodrigues 2002) and in the current university/region debates (Nowotny et al. 2001).

Critique of the dualist approach

Common to the literature mentioned above is an emphasis on knowledge and knowledge generation as a key force in economic development. However, most knowledge generation processes will by definition be slower than the developmental demands in their environment. This has many causes. Firstly, there are individual constraints, due to limitations embedded in social structures and the social reality of the knowledge creators. Individuals do not normally change their perception of the world very fast. Secondly, knowledge generation will be embedded in power structures, in cognitive frames, in traditions and culture and in institutional norms, in other words the social reality that make up the local practices of the knowledge creators. This makes the study of these processes difficult. When knowledge creation is interwoven with many other social processes, it is difficult to distinguish between processes that lead to development of genuine new knowledge, knowledge diffusion, and knowledge imitations. However, our argument goes beyond these more methodological difficulties. We believe (in line with Deetz 1996) that there are other limitations to the dualist approach to knowledge creation, and that these can partly explain the problems we have met on a more practical level when trying to apply the dualist approach, and which we have referred to above as frustration.

The more fundamental critique of the dualist approach is that it is contradictory to apply arguments from the two knowledge regimes as if they are complementary. If you really hold a “local” knowledge position, you will not accept a “universal” knowledge position. So any dualism is then false. You need to take the one stand, or the other. An exception will be if you have a meta-perspective, and from that look at the to kinds or regimes of knowledge as two practices. The meta-perspective could be taken from knowledge sociology (Mannheim 1936; De Greé 1943; Merton 1973), or it could be a communicative perspective (Habermas 1997). In both cases, it is not the knowledge as such, but the practice of using knowledge that is compared. However, we still need to ask the question whether we have to know what knowledge is, in order to compare different regimes of knowledge?

Part two: beyond the dualist approach to knowledge, what is knowledge?

The argument here is that knowledge has a contradictory nature, and that this is one of the reasons for the difficulty in dealing with knowledge development. We use a quotation from Michel Foucault to illustrate this. Foucault has argued that:

It is for this reason that in Nietzsche we find the constantly recurring idea that knowledge is at the same time the most generalized and most particular things. Knowledge simplifies, possesses over differences, lumps things together, without any justification in regard to truth. It follows that knowledge is always a misconstruction [méconnaissance]. Moreover, it is always something that is aimed, maliciously, insidiously, and aggressively, at something like a single combat, a tête-à-tête, a dual is set up, contrived, between man and what he knows. There is always something in knowledge that is analogous to the dual and accounts for the fact that it is always singular. That is the contradictory character of knowledge, as it is defined in the Nietzsche texts that seem to contradict on another—generalizing and always singular, (Foucault 1973, p. 14).

This Nietzschean conception of knowledge implies that knowledge “avoids” any attempt to define or limit it. Following this, any attempt to group, define, categorise, etc. knowledge will fail and not have any legitimacy outside the construction itself. We cannot identify true and universally valid knowledge categories. The Nietzschean conception of knowledge implies that knowledge is decoupled from morality, and has been interpreted (by Foucault among others) as a argument for structuralism. As we will argue later, when we refer to this perspective, we do not regard it as a threat to reason (Habermas 2003; Putnam 2004), but rather as an insight to a challenge.

How can we research a field where the phenomenon we research is so contradictory? One answer and implication is to develop a humble attitude towards knowledge, acknowledging its complex and contradictory character (Rorty 2000). Another approach could be to argue that all knowledge is political or value laden in a certain sense, so objectivity is out of reach. One could of course (pragmatically) argue that, although there is no universal definition of knowledge, we could try to develop one. Or, as we will try here, argue that there are different episteme or domains of knowledge, and that these domains or regimes as we have called them, have different validity claims.

Epistemological positions in the discourses of knowledge economy

One way forward from our critique is therefore to identify fundamental positions on knowledge, what we have referred to as episteme or knowledge regimes. Our position is to stress the importance of a consequent position on ontology and epistemology in the way the concept of knowledge is used. The concept of knowledge cannot be separated from the fundamental episteme, neither can it be separated from its ontology. This is the problem with the dualist position. The dualist position presupposes a meta-position that in our view unsuccessfully attempts to bridge two or more incomparable epistemic positions.

An obvious approach to different knowledge regimes is to identify positions on ontology and epistemology. Along these two dimensions there are a multitude of positions; however, as a rough approach we might identify the distinction between a realist and a constructivist position on ontology, and a subjective and an objective position on epistemology.

An objectivist epistemology will argue that we can form true opinions of phenomenon around us, and that this is merely a question of method and not of principle. A subjectivist epistemology rejects this belief. Truth is never available to us. A realist ontology holds that there is an identifiable world “out there”, while a constructivist position holds that whatever is “out there” is subject to our subjective or intersubjective construction. On this basis we simplify and construct four different knowledge paradigms (Kuhn 1962; Burrell and Morgan 1979; Deetz 1996; Arbnor and Bjerke 1997), which individually overstep the categorization, but are located on the basis of the nature of core components (Table 2).

Table 2 Knowledge paradigms
Full size table

What can we learn from a definition and structuring of knowledge regimes? Firstly, some of the controversies in the debate of knowledge generation and knowledge transfer can be traced back to such differences in fundamental positions (paradigms). There is for instance a tendency in economic geography to use critical realism as the basic philosophy when discussing the knowledge economy and regional innovation system (Sayer 1984) and it is also used in normative science. There is also a tendency to use a constructivist position in some of the Action Research literature on enterprise development (Pålshaugen 2002), a bounded rationality positioning in some of the knowledge management literature (Mintzberg et al. 1998) and interpretivist positioning in some of the communicative development literature (Habermas 2001).

Our argument is that comparing these literatures is complicated because of their differences in underlying assumptions, and thereby often strong differences in the definition of the concepts that are used in the different literatures. We cannot go directly from the one to the other.

However, most of the applied research in the knowledge generation area (knowledge economy, innovation, enterprise development, learning organization and knowledge management) does not take a very strong stand on such fundamental issues, but rather positions themselves in between the four positions. We can argue that they hold an integrationist rather than a purist position on these issues. As we have argued, such a position requires a meta-position where different knowledge regimes are defined as different, and comparable practices.

This is how we have experienced the enterprise development discourse as referred to earlier. It is at least a hypothesis that some of the problems relating to reporting and comparing knowledge development, refer to the differences in epistemological and ontological positions. At least this is our interpretation, and on that might explain our frustration on trying to implement the dualist approach in our Action Research practice. Our practical approach did not enable us to deal with the underlying complicity and incomprehensibility of the knowledge concepts.

The practical arguments in the discourse on knowledge, innovation and development might for this reason, be easy to compare, but that does not imply that we have understood the theoretical foundation or the underlying assumptions on knowledge that are implied in these arguments. Again, it is our interpretation and experiences that the theories on knowledge generation, knowledge transfer, learning, and innovation are often discussed as “common sense” issues, without any fundamental discussion of the entities (concepts) that are discussed. It is the social practice, rather than the phenomenology of the concepts, that we observe in this type of research.

In the integrationist approach to knowledge, knowledge can take different forms that refer back to such underlying (pure) positions, that is, knowledge can, according to the situation it appears in, be both constructed, and real, both bounded rational and rational. This is in line with the dualist position on knowledge, and it also illustrates its limitations. We manage, through this approach, to say something about practice, but do not thereby manage to penetrate into innovation as a knowledge development process. We believe that a deeper investigation of the concept of knowledge will bring some fruitful insights into these processes.

How to handle different validity claims?

The problem with different knowledge paradigms is that they have different validity claims. That is, what counts as true and relevant knowledge differs between the paradigms. To illustrate this, the Table 3 (inspired by Habermas 1997) indicate how social sciences produce different types of knowledge with different validity claims.

Table 3 Research perspectives, forms of knowledge and validity claims
Full size table

The argument that can be developed from Table 3 is that different researches, or rather different level and units of analysis in research, requires different validity claims. This adds to the complexity of understanding knowledge development. One is left with the question of how to compare and how to utilize different insights and kinds of knowledge from different paradigms and different levels of analysis.

Part three: overstepping regimes: innovation and transformation of knowledge

Given the problem we have discussed in part two related to the complicity of knowledge, what can be a strategy for research? This part is not a comprehensive discussion, rather an attempt to indicate some perspectives forward in this discussion. If the argument that we develop here is correct or at least meaningful, it implies that in order to understand innovation, we need to understand how different kinds of knowledge are transformed between different knowledge regimes or paradigms. Radical new knowledge is probably developed when we go beyond a regime and overstep regimes (refer Kuhn’s scientific revolutions). But how is such a knowledge transfer possible? This is in our view a huge question and one that we only briefly will investigate into here.

Knowledge and innovation

If knowledge basically is something subjective, it is likely that context means something for the way individuals relate to each other and exchange knowledge. Different challenges are facing us in different contexts: social structures, power and mental frames—all these factors plays different roles in different contexts. With reference to Habermas (1997), we will argue that there are specific structural and motivational conditions that have to be met in order to promote reflection and communicative action. If context means something, and influences knowledge development (as it probably does) we need to link these contexts to the assumptions in the dualist approach. Example: to be creative probably requires an environment for challenging established structures and cognitive frames. Imitation and implementation is probably taken care of within established structures.

It is a possible hypothesis that this explains some of the difficulties we have experienced in creating new knowledge in co-operation between research and practitioners as referred to earlier. The difficulty might refer to the difficulty of establishing a sort of environment or situation where co-generation of knowledge (Habermas 1974; Geertz 1983, Elden and Levin 1991) is possible. In most practical situations, researchers will meet environments with a strong tendency towards instrumentality or functionality, and strategic behaviour and not communicative action, where the conditions for cogeneration of knowledge are not present. However, our argument goes beyond even this.

Given this insight in different regimes, or paradigms of knowledge, and also given our interpretation of knowledge as a contradictory and complex phenomenon, how do we proceed in order to discuss knowledge development and innovation? The point here is that we have on the one hand to acknowledge this complexity and incommensurability of knowledge, and on the other hand develop better and deeper understanding of what happens when different types and kinds of knowledge are applied. That is, instead of assuming that knowledge can be compared or used as if it was a consistent and unified phenomenon, we need to focus on how knowledge is transferred between different regimes and transformed when it goes from one regime or application to the other. We believe that insight into this transformation process will give better insight into how innovation happens as a knowledge development process.

An implication of this way of approaching the complexity of knowledge is to argue that discourses within a regime has limited validity outside that regime. This would for instance imply that Action Research critique of convectional research has limited validity (Johnsen 2005). This means that Table 4 does not imply that the categories are discrete variables. This is because local knowledge will involve universal knowledge (Habermas 2003).

Table 4 Innovation and knowledge: different knowledge translation (transformation) processes
Full size table

Knowledge transfer and Action Research

In order to make this general argument a little bit more specific and to show that it might have an impact, we will argue how different conceptions of Action Research influence different perspectives on transfer of knowledge. One can argue that Action Research is divided between the two lines of thought, both with roots in pragmatism. Both lines of thought share many of the same insights into communicative and learning processes within communities, but they still have some fundamental philosophical differences.

The first line of thought can be described as pragmatism as relativism and contextual bound practice. Dewey/Wittgenstein/Heidegger/Rorty can be said to have inspired this line of thought: the focus here is on doing, making practical results, improving democracy through practice, and finding good solutions through practice that is embedded in a certain local context (Rorty 2000). Action Research as practice implies taking part in practice, in order to improve practice in a democratic sense, by helping find local solutions to practical problems and engage in local learning processes. This position can be linked to the theory of science position of Kuhn (1962) and to Lyotard (1979) and the theory of language games.

The second line of thought can be termed pragmatism in a communicative theoretical sense. Pierce/Apel/Putnam/Habermas can be said to have inspired this line of thought. The key perspective here is searching for truth, justification and principles (universals) for a just society as a platform (theory) for acting (Putnam 2004). Action Research in a pragmatic, communicative perspective implies that the social world is seen as a communicative practice that we need to understand through engagement. The communicative practice gives us insight into social structures (that is: issues like dominance, power, perceptions, etc.). Dialogue is seen as a source of life world experiences, that is, insight into underlying perceived reality. Taking part in communicative practice also gives insight into collective developed norms and value systems, and thereby systems of interpretations, i.e. how facts are interpreted and converted into new practices. Taking part in dialogue in communicative communities implies possibilities to engage in learning processes, i.e. see if changing communicative practices is able to reveal new perspectives on social facts, and/or see new possibilities in existing facts. Interpretation and understanding in this sense give us a point of departure to discuss local practice in light of ethical theory (normative aspects of social practice), and understanding of self as a subject (own life world) in perceiving social reality.

In the first perspective on pragmatism, there is no transfer of knowledge between different local realities (language games). In the second, different realities are interwoven into its other. How can we in this latter case perceive knowledge transfer?

Communicative theory as a meta-position: transformation through communication

Habermas argues that language has two functions: links to reality and inter-subjective meaning, that is, because we both in our speak acts wants to be understood (take part in collective speak language system (locutionary act), express something and make commitments (illocutionary acts), and make claims in order to convince others (argue: perlucutionary act). Thereby, as we propose something to others, they will investigate the validity so that communicative action becomes a way of testing truth (objectivity).

The language is like a mirror: we see us selves in the language, and link our consciousness to thought. We believe the language tells us something about reality around us. Both mirroring (the mirror of out consciousness and the mirror of outside reality) is connected with uncertainty. It is a continuous process that requires a continuous dialogue, to link our inner, subjective world (the life world) to the inter-subjective social world and to real facts (the objective world/system world).

Language and communication thereby acts upon both a reality level: expressing something about truth, and a cognitive function: develop inter-subjective categories that allow developing a mutual understanding of truth. Language and linguistic communities thereby have an locking in tendency, a tendency to create their own conditions for validity and reality interpretation.

We meet social reality and language in the form of local communicative communities. These local communicative communities represent links between how subjects perceive reality and how they share this as socially and inter-subjectively developed meaning constructions. Also local communicative communities tell us about shared interpretations of local acts and local practices. In communicative communities, criteria for truth and validity claims are established through reflection. Reflecting implies considering all possible alternative explanations. Ability in critical reflection will thereby be a feature for a communicative community.

Communicative communities are interlocked in meaning constructions, but at the same time constantly “invaded” by bureaucratic or marked based institutions. They thereby become mediators of the local world and the larger system of the social world. Communicative communities are “truth seeking” to the extent that they allow open, critical dialogue. However, dominance, power, normative structuring, lack of openness, etc., destroy truth seeking processes. Studying communicative communities thereby tells us something about dominant social structures.

This is merely an attempt to indicate a start to investigate into the complexity of knowledge and understanding knowledge development. Our argument ends up as a very general critique of some of the predominant innovation literature. On the one hand we have argued that we cannot “jump” from one knowledge regime to the other that we need to understand that different knowledge regimes are based on different assumptions. The underlying philosophical position means something. We will develop more fruitful innovation theories if we acknowledge the complexity and contradictions in the phenomenon of knowledge. On the other hand we have argued that the focus should shift in order to investigate the transformation process when knowledge from one regime is transferred to another. We believe that the key to understanding innovation is to understand this transformation of knowledge. If we can proceed in our discussion and understanding of knowledge, we will also make Action Research, and research in general, more useful, also towards workplace development.

Conclusion

Action Research is predominantly occupied with developing local knowledge as a form of counter-knowledge to universal knowledge. This definition of the task of Action Research preconceives a dualist perspective on knowledge. We think this is wrong. Knowledge is a far more complex phenomenon than one that can be framed in dualist terms. So, if one wants to proceed with knowledge development, one has to move beyond dualism.

The argument in this article is that knowledge is a complex and contradictory phenomenon, and at the same time it is so interwoven in social practice, in communication and in social structures, that it is difficult to research. When we say that we study knowledge, the likelihood that we in fact study is mainly those other things, is predominant.

This article has aimed to show that the dualist approach to knowledge generation is unsatisfying. Working with the aim of developing new knowledge in practice quickly reveals that the process of knowledge development is interwoven with many other social processes. Our argument is that the dualist approach at the present stage in development is not able to sort between these processes. This limits the perspective when we try to enhance workplace development and innovation.

We argue that although dualism is a common approach in all these theories, they are probably not as consistent and mutually compatible as it might seem. We argue that dualism on a practical level does not necessarily correspond with a clear and detailed understanding of knowledge at a philosophical level. Furthermore, the contextual dimension in knowledge development has to be taken into account, that is, we need to do more research into what happens in the transformation processes in different contexts. We need to, in our developmental work, aim at understanding knowledge generation in other ways than those implied through a dualist perspective.