We are starting this paper with a few biographical comments intended to locate our interests in this topic. One of us, Garrick, is an academic with research interests in continuing education, staff development and Human Resource Development (HRD) more generally. Rhodes leads a dual career both in academia and in commerce and is involved in research and teaching related to radical and postmodern approaches to organisational learning. Many of the issues to do with organisational learning examined in this paper are common to each of us. In relation to the professional positions that we take, the central problematic we are examining here centres on the ways organisational learning tends to be understood in contemporary contexts which are characterised by uncertainty, unpredictability and an insatiable market appetite to develop 'knowledge workers' who will give an organisation a competitive edge. We share some scepticism about existing theoretical frameworks for promoting organisational learning and suggest that deconstruction, although not an answer in itself, offers fruitful ideas for possible future research and development. This scepticism is particularly directed towards the value of organisational learning in terms of how it relates to organisational power structures that are being challenged and changed. In part it is this scepticism that informs our views on how organisational learning might be researched, how knowledge in organisations is constructed and reproduced.
We contend that power structures are changing, decision making centres are shifting and traditional notions of knowledge construction are being radically challenged. In this milieu of change, postmodern ideas bring with them a set of epistemological challenges for researchers - a questioning of anything that suggests absolute principles of reasoning. As Walter-Busch puts it: 'under postmodern conditions, scepticism toward all kinds of dogmatism and monolithic world views...becomes unavoidable' (1995: 148). 1 This condition offers the chance to create provocative theories and research practices that challenge established knowledge. It is our intention here to challenge the dominant approach to knowledge that has been manifested in 'organisational learning', and in doing so to explore possibilities for a postmodern alternative that builds on the theory of deconstruction to formulate an epistemological position on researching organisational practice. In examining the possibilities for a postmodern organisational learning research agenda, we take heed that 'postmodern theorists challenge modernist constructions that elevate the impersonal, functional and mechanical social order over the personal' (Boje 1994: 450). 2
In exploring these possibilities we assert that organisational learning is an approach to management which is attempting to reap organisational and managerial benefits through workplace learning and that research into organisational learning runs the risk of creating knowledge which attends only to narrowly defined managerial interests. It does this by conflating the notions of learning and work such that learning is seen only to be legitimate if it results in organisationally sanctioned 'benefits'. We contest that this conflation is achieved through a research agenda paradigmatically informed by systemic thinking. Such thinking is limited in its attempt to totalise diverse experience into mechanistic and organic models (and ways of modelling). In suggesting a different take on organisational learning, we are also purposefully avoiding the temptation to come up with another theory which claims to replace the old one. This thinking rests on the idea that postmodernism is not informed by grand narratives (single truths) and thus does not offer broad and integrated explanations that are alternatives to systematic theories. Rather, postmodernism resists that form of theorising by looking for differences and instabilities and by deconstructing frameworks that try to impose overall explanations of social phenomena. Where then can research into organisational learning move to in light of conditions of postmodernity (and postmodern thinking)? Can postmodernism offer a source of knowledge that informs organisational practice? Can postmodernism be deployed in researching new possibilities for organisational action (and how)? What might the future be for research into organisational learning? This paper addresses these questions by examining organisational learning and insights from postmodern thinking. In particular it is the strategy of deconstruction that we review to offer a postmodern redefinition of organisational learning and its research implications.
The 'postmodern' describes a development in the examination of cultural transformation taking place around the world and particularly in western societies. Postmodernism is partly the consequence of the failures of modernism (which assumes a logical and ordered world whose laws can be uncovered by science) and its emancipatory project. A project which is rooted in a myth of progress which assumes that the growth of human knowledge will lead to a desirable future for all. This progress of the sciences, technologies and the arts was proposed to liberate humanity from ignorance, poverty, backwardness and despotism and produce happy, enlightened citizens who are the masters of their own destinies (Lyotard, 1992). 3
The postmodern condition however is one which questions the legitimacy of this grand narrative of progress (Lyotard, 1984). 4 Faith in the idea of emancipation has been deeply eroded by war, nuclear testing, the environmental crisis and the intractability of urban social problems. Globalisation, technological change, flexible 'free markets' characterised by labour market flexibility and unpredictability are subjects of the master discourse of economic rationalism. Zygmunt Bauman (1997: 13) 5 describes post modernity this way:
'Every single orientation point which made the world look solid, and favoured a logic in selecting life strategies: the jobs, the skills, human partnerships, models of propriety...values thought to be worth pursuing and the proved ways of pursuing them...seem to be in flux. Many games seem to be going on at the same time and each game changes its rules while played. These times of ours excel in dismantling the frames and liquidising patterns - all frames and all patterns, at random without advanced warning.
The liquidising patterns that Bauman refers to include diverse and contradictory value-messages that are part of every day organisational life. For instance, through the privatisation of public services and utilities, the dispersal of authorities and deregulation, the construction of appropriate behaviour and knowledge about what organisations ought to do is constantly contested. Accompanying this contestation of organisational practice in the face of axiological and ethical dilemmas is the increasing importance of image and status consciousness. The boundaries of organisation and the practice of management are embedded in the postmodern context in which they are located. Indeed, diverse and fragmentary organisational and professional contexts characterise both the material organising for change and the discursive practices of 'organisational change'.
As described by Jeffcut (1996: 32) 6 a focus on creating order from disorder permeates the modern world with modern organisations and organisational theory and research seeking to manage and eliminate disorder through 'rationalisation of the organisational irrational, and harmonisation of the organisational dysfunctional through nostalgic and utopian visions of progress.' As modernism pursues order through a system of hierarchies which gain authority through censorship and suppression of alternatives, postmodernism is more interested in a critique of the practices which create the sense of order and suggests an approach where meaning is an articulation of provisional and plural difference.
What then is there for a postmodern research agenda for organisational practice? We believe that the challenge is to articulate an approach to research that questions absolute principles of reasoning and works through its own values. What can result is a form of research where understanding organisations, change and learning is informed by a scepticism towards conclusiveness and an embrace of the tentative.
If the postmodern is a reflection of rapid change to the assumptions that underlie our view of organisational realities, then 'organisational learning' has emerged as a way of managing through this change. It has been taken up almost unquestioningly by learning organisation specialists and consultants, Human Resource Development (HRD) practitioners, staff development 'learning facilitators' and managers as an answer to the uncertainties of contemporary organisational life. Indeed, would you rather work in an organisation that didn't learn? Organisational learning has gained academic and practice-oriented currency over the last twenty years as a preferred model for managing change. Burns (1995: 55) 7 suggests that organisations have responded to this need to 'manage change' by encouraging the involvement of all employees, 'from the shop-floor to management, in a lifelong learning process'. He goes on to assert that it is through organisational learning that companies can create the environment for this learning to occur and that 'the effective organisation is one that fosters individual, group and organisational learning as a key characteristic in order to become a learning system' (p. 61). This focus sets the scene where organisations that learn are those that facilitate the learning of all individuals within them (Pedler, Burgoyne and Boydell 1991) 8 . This links individual learning with the requirements of an organisation's performance. This is a discourse of organisational learning that conflates the education of adults and their work practices - 'learning becomes an everyday part of the job and is built into routine tasks' (Watkins and Marsick 1993: 26). 9
It is our contention that the development of the discourse of organisational learning has taken a path which seeks to fix change through a modernist project of control and that research into organisational learning has helped to lay this path. In this project, control is achieved through two of organisational learning's key focal points. First, organisational learning rests on a world view based on the ontological reification of systemic approaches to understanding organisational phenomena. This approach attempts to 'fix' the world through the application of mechanistic, systemic or organic modelling practices which are deemed to be universal and free from the influence of time and context. Second, organisational learning legitimates its practices through an unquestioned belief that learning will lead to business success and that business success is a valuable social goal in itself, or that it will lead to other valuable social goals (a variant on the 'trickle-down' effect). This legitimisation creates totalising views of organisational 'realities', values and priorities that can suppress and marginalise organisational and social activities which do not result in commercially measurable results. Each of these two points is discussed in more detail below.
Popular views of organisational learning are based on the idea that fostering learning in individuals can be transformed into more general improvements that will lead to success and prosperity for the organisation. This is firstly based on the idea that people have a natural proclivity to learning which can be fostered and developed by the organisation and secondly that attaining this learning is translatable into business success. Organisational learning is a version of the progress myth suggesting that learning can lead to business success and that such success is unquestionably a valuable social goal.
For Senge (1990) 10 , a US learning organisation guru since the early 1990's, learning is a natural and desirable state for individuals and organisations. But he suggests an irony where traditional organisations focus more on performance and control rather than curiosity and a desire to learn. As such, organisations which learn are those which facilitate the learning of all individuals within them and therefore can respond to internal and external changes (Pedler, Burgoyne and Boydell, 1991). 11
Furthermore, Argyris and Schön (1978) 12 see that organisations tend to be good at 'single loop' learning, that is, learning to detect and correct errors, but they also create learning systems which inhibit 'double loop' learning which requires questioning and modifying organisational norms, policies and objectives. They in turn advocate that organisations ought to adopt models and cultures that facilitate both single and double loop learning through a system of openness of information and an absence of defensiveness. Organisational learning is then a process of opening up an organisation to the learning potential of its members.
Describing organisations as being able to 'learn' is a metaphor for changed collective patterns of behaviour on the part of individuals in an organisation. This organisational learning metaphor is realised when individuals, acting as learning agents for the organisation, detect and correct errors in the organisation's behaviour patterns which in turn become embedded in the culture of the organisation (Argyris and Schön, 1978) 13 . Learning is seen to have occurred when organisations perform in improved ways usually as a result of requirements to adapt and improve efficiency in times of change (Dodgson, 1993) 14 .
Thus, we can interpret modernist organisational learning in terms of systemic (and 'culture') change in organisations which leads to organisational improvements. Based on this definition we then need to look at what constitutes an 'improvement'. Quoting Fortune magazine, Senge (1992) 15 illustrates that becoming a learning organisation is a way to sustained competitive advantage. This indicates that an 'improvement' is measured in terms of market position. Thompson (1995) 16 says that organisational learning drives a company's ability to increase revenues, profit and economic value. Organisational learning can also be looked at in terms of the processes of learning and the outcomes of learning. If we are to accept that market position or financial success are the true measures of organisational learning then we must also accept that the learning of individuals in the organisation is merely a process whereby people become the instrument of the organisation's objectives (ie the objectives of the organisational 'elite'). An alternative perspective would be that goals such as market share and financial success are not the true end of organisational endeavour, but that commercial success is the means by which organisations can provide society with goods and services, employment and improved standards of living (O'Toole, 1985) 17 . From this point of view, the commercial success of the organisation is a process which leads to a socially valuable outcome.
What remains is that such conceptions of learning legitimate themselves by suggesting that the process of learning is considered to be 'good' in that it leads to directly and indirectly to socially valuable outcomes on account of the business success it is said to generate. Learning, as the generation of knowledge, will only be viewed as legitimate if it results in commercial benefits and research is hence legitimate if it can create knowledge which will inform or assist this process. This process of legitimation then masquerades in a narrative where learning is the creation of social value, but in fact uses business criteria as a way of deciding whether a particular activity is, or is not, learning. It is business success that establishes what is valued and recognised in learning. This is what Lyotard calls the mercantilisation of knowledge, where the primary question asked of learning is: "Is it saleable?"' (Lyotard 1984: 51) 18 . This saleability of knowledge operates within organisations themselves, and in their complicit nexus with formal educational institutions (eg universities) that are now promoting 'knowledge for dollars' both in terms of research and teaching. What Lyotard is referring to is legitimation through what he calls 'performativity'; the domination of technical reason through ideals of efficiency and technical performance. This process of legitimation can then be seen as a hoax used (possibly unknowingly) to support dominant views of the world where performativity as a means towards a supposed social end becomes the end in itself (Alvesson and Deetz, 1996) 19 . Where on the one hand organisational learning can be seen as legitimated by, and (re)productive of, values based on efficiency, profitability and market success, organisational learning also seeks to impose a view of organisations and of social life informed by an unquestioned systemic perspective. One of the central themes of organisational learning is a view of organisations as complex systems where subsystems interact to convert inputs to outputs, and where the system changes over time to react to changes in the environment (Coopey, 1995) 20 . Systems theory, as the paradigmatic bedrock of organisational learning is seen as the 'discipline of disciplines' which assigns 'common properties to physical, mechanical, social and mental phenomena and produces a dynamic model of the interrelated workings of the world' (Mirvis, 1996:14) 21 . The learning that takes place therefore is one which allows this systemic change to occur. As described by Addleson (1996: 57) 22 , the approach to organisational improvement adopted through the mainstream discourse of organisational learning is based on 'an understanding of organisations as social structures - complete systems made up of interrelated parts.' The object of research, then, is to understand the system and its processes and to apply this knowledge to solve organisational problems. Organisational learning is about learning how things work by understanding the rational order of the organisation. This 'modernist substance' of organisational learning is a focus on systems - for instance, Argyris and Schön's (1978) 23 single and double loop learning and Senge's (1990) systems thinking - and that by mastering these systems organisational problems can be solved. What Addleson points out is that the reification of systemic approaches does not account for differences in interpretation as systemic consensus is imposed as the theoretical super goal. This approach suggests that there is a 'true nature' of organisational problems and a right way of doing things which can be identified through the application of organisational learning processes as identified by research into organisational learning. These processes in turn leading to socially valuable outcomes as measured by the commercial success of the organisation. Coopey concludes that using learning as a metaphor for understanding organisations has the potential to be 'used manipulatively by managers with a long pedigree of instrumental interest in social science as a means of solving industrial problems' (1995: 212) 24 . Our analysis so far has suggested that research into organisational learning is a dubiously legitimated and conceived project which focuses on control, the maintenance of orthodoxy and (although the rhetoric usually suggests otherwise) the suppression of difference. The issue that remains is that of how can organisational learning emerge as being of any value in light of this critique. To address this issue, we now turn our attention to deconstruction. In formulating a postmodern approach to research, our intention here is to examine the notion of 'deconstruction' and to look for ways that it can be applied to a formulation of organisational learning. Deconstruction has developed as one of the principle tools of postmodern thinkers to accomplish a critique of the foundationalism of modernist epistemology (Derrida, 1978) 25 . Deconstruction can be read as a form of ethical practice that is concerned with 'what happens to ethics' as knowledge frameworks are increasingly challenged. Deconstruction seeks to uncover contradictory and historically conditioned assumptions within a discourse. It challenges the distinction between representation and 'real', assuming that discourses produce rather than simply represent. Deconstruction analyses human behaviour as a textual production - a kind of writing (Kilduff, 1993) 26 . Based on the work of Jacques Derrida, deconstruction sees social life as made up of texts which are constantly being read in different ways. Through such multiple readings social texts do not have a fixed meaning as any text can be read in more than one way. Where a modernist perspective sees text as a reflection of a pre-given social reality, postmodernism, as informed by deconstruction, sees text as ways of understanding the world. Each reading of a text however creates a new text which in turn results in an intertextual weaving where texts dissolve into each other, are located within other texts and are built up on other texts. This textual multiplicity has the affect of eroding (appropriately) modernist certainties. Deconstruction emphasises how words, texts and stories have multiple meanings, but that each of these meanings is arbitrary (Boje, 1995) 27 . To do this, deconstruction takes as its starting point a critique of dualistic or binary ways of thinking or theorising. This suggests that texts are ordered around oppositional categories (such as reason/emotion, objective/subjective, masculine/feminine) where one side of the binary is privileged over the other. This privileging is not a natural state but rather results through defining one term in terms of its opposite. Deconstruction does not, however, attempt to reconcile these dualisms into a newly privileged structure of meaning, rather it opposes the constraining hierarchies by disrupting a closure of meaning (Knights, 1997) 28 . Deconstruction is then an analysis of the relation between dualities to show the ambiguity that is embedded in them and to demonstrate how language is used to discipline particular meanings. Deconstruction is a questioning of the 'fallacy of misplaced concreteness' where arbitrary ways of categorising 'reality' are taken as universal and independent of the people who use them (Knights, 1997) 29 . Deconstruction, then, sees knowledge and discourse as being constructed from a changing world into a fixed text. Deconstruction attempts to expose the inherent contradictions in that text using the strategies of overturning and metaphorisation. Overturning starts with the idea that texts are structured around terms which are defined in terms of polar opposites, and where one term dominates the other. Overturning is based on changing this hierarchy so that the dominant term becomes marginal and the marginal term becomes dominant. Merely overturning however, only works to create a different structure of domination; the process of metaphorisation is based on the idea that the two opposites are defined in terms of each other such that each one continually threatens the hegemony of the other. It is in this way that deconstruction not only exposes 'hidden metaphysical first principles in literary and philosophical texts, but generally the deconstructive stance is directed toward the claim for foundations in objective knowledge in general' (Söderqvist 1991: 148) 30 . Deconstruction takes the position that meaning cannot be recovered unambiguously from text or discourse. Texts contain contradictions and elements that are textually suppressed and excluded and an exposition of the sub-text of how the narrative was constructed undermines the agenda of the explicit text (Kilduff, 1993) 31 . Authors cannot express to have privileged access to the meaning of the text they write, as deconstruction reveals a tension and contradiction behind the organising process and attests to the ambiguous nature of our understanding of organisational experience (Chia, 1996) 32 . Deconstruction shows how 'the understanding of organisations is inseparable from the organisation of understanding' (Jeffcut 1994: 241) 33 and how attention must be paid to the 'organisation of writing rather than the writing of organisation' (Cooper 1989: 501) 34 . In a deconstructive stance on research we are suspicious of the 'misplaced concreteness' that sees theories as being independent of theorists and creates truth effects by confirming certain ways of thinking about organisations (Knights, 1997) 35 . Each text we use to understand the world produces new texts and new understandings as the texts dissolve into one another and we become aware that 'we use the same means of making sense of our world as our "subjects" use in making sense of their world, and in both our making sense of their world, and our making sense of their sense making' (Linstead and Grafton-Small 1989: 292) 36 . Reality cannot be separated from the way it is reconstituted as a representation (Jeffcut, 1994) 37 and how that representation is fragmented, provisional and arbitrary. As a starting point to developing a postmodern research perspective on organisational learning, we are suggesting that rather than relying on modernist systemic views of organisations, a postmodern view of organisational learning begins with an understanding of organisations as being constructed in language. Here, analysis can turn away from techniques taken from the physical sciences and to language and literary theory for insights. Learning is, therefore, not analysed from a perspective of behavioural action, but rather from the domains of interpretation, knowledge and language. Postmodern 'organisational learning' approaches 'organisational knowledge' from the point of view that nothing can be taken for granted and that all claims to truth must be deconstructed so as to disrupt privileged ideas and critically re-examine taken for granted ways of seeing the world. Such an organisational learning recognises that language is the tool and the repository of learning and that language is a social phenomenon where 'to learn is to use language to communicate, both at the interpersonal and the intrapersonal level' (Weick and Westley, 1996: 446) 38 . Deconstruction as a process for understanding learning offers an approach which is related to learning as an activity of language within a postmodern perspective. In examining the contribution of deconstruction to a theorisation of organisational learning however, we need to be mindful that deconstruction is not reducible to a set of techniques or mechanical operations that can be applied to a text. It is a way of reading that can enable the reader to see patterns of conflicting relationships. Deconstruction is not, therefore, another tool to be added to the managers' kit bag; it is a way of thinking and a way of reading the organisation. As such deconstruction can offer a critical way of generating organisational knowledge which overturns assumptions by leaving them open to multiple interpretations so as to overcome the domination of ideas. For organisational learning, deconstruction then questions organisational change which is used as a vehicle for domination where people such as key power-brokers and managers 'conspire to enact the world for others' (Hatch 1997: 367) 39 . In this scenario, learning is a narrowly focussed process by which powerful sub-groups dominate and oppress through attempts to imprint pre-determined truths and behaviours on to other people (Rhodes, 1996) 40 . A postmodern epistemology that emerges from deconstruction of scientific and systemic approaches to organisational knowledge and learning includes:
In these conditions, organisational learning is not to be viewed as a set of timeless prescriptions designed to portray, control and master the systemic nature of organisational behaviour. Rather, learning is an activity of indeterminate pragmatism inseparable from the intimate linkages between knowledge and power. Learning is no longer a process which can be applied 'equally' to everyone and every organisation - learning is not a binding objective truth about performance. Instead a theorisation of learning must be based on pragmatic indeterminacy, and diversity across all levels of thinking. Learning must move from the general to the local. As Polkinghorne (1992: 149) 42 argues: Each movement of change is the consequence of the coming together of a unique set of multiple forces at a particular place and time. Knowledge should be concerned with these local and specific occurrences, not with the search for context-free laws. In times of rapid change and uncertainty, there is a seductiveness that accompanies discourses that promise greater certainty, prediction and control. But in the view of a postmodern epistemology of organisational practice, and following Foucault's (1982) 43 hypothesis of the intimate link between the development of scientific discourse and the tightening of surveillance and control, scientific and systemic ways of thinking that provide the vocabulary (and world view) for popular organisational learning practices need to be viewed with some apprehension and a great deal more scepticism than is presently the case. Rather than dutifully hanging on to knowledge frameworks that promise (falsely) empowered self-directing performers, there is now a possibility to re-think and re-articulate the knowledge bases of organisational change and learning. In summary, our starting point in this paper was sparked by a scepticism towards the 'truths' and 'best practices' presented in conventional approaches to organisational learning and the knowledge practices that it entails. Particularly we have critiqued the systemic thinking and embedded progress myths where organisational learning creates a discursive formation equating commercial success with social 'good' and creates a research agenda that is justified by corporate interests alone. This scepticism has not however lead us to reject the concept of organisational learning all together. Instead we have offered deconstruction as a possible strategy for developing a postmodern epistemology of practice from which research can be based. This practice is intended to destabilise power/knowledge hierarchies by operating within an indeterminate arena of difference which does not need to be reconciled. Learning, knowledge and research in such a context is not concerned with the desire to establish organisational 'truth' in which the practices of 'learning organisations' represent objective (indeed measurable) realities. Instead it is concerned with having to make pragmatic choices and contingent decisions where no stable criteria are available to assist in the decision making. Deconstruction cannot, however, be entered in to unproblematically; it is not value neutral, nor is it content free. It is a strategy that can be used on itself. A question inevitably prompted by our discussion is: on what bases ought we choose deconstruction as a strategy to inform research? The art of handling multiple and conflicting discourses is one of the principal challenges facing the postmodern researcher. The question of how this art is shaped is of paramount interest. We have suggested the idea of multiple and competing discourses opens the door to deconstruction as a strategy related to understanding and contesting organisational learning. This in turn opens up new potentials for productive diversity and resistance to dominant discourses. A part of the challenge for us has been to articulate an epistemology of practice that questions absolute principles of reasoning and works through its own values, that is, an ethics of postmodern practice. In this paper we have suggested that one of the principal philosophical methods available to accomplish this is deconstruction. Deconstruction can be read as a form of ethical practice concerned with what happens to ethics as knowledge frameworks are increasingly challenged and firm epistemological ground is nowhere to be found. As mentioned earlier, it seeks to uncover contradictory and historically conditioned assumptions within a discourse, challenging the distinction between representation and 'real'. But there is no guarantee that users will not in turn find themselves deconstructed by engaging in the very act of employing deconstruction as a strategy; a strategy that itself be read as a discursive outcome of power relations. What remains of our argument is a pragmatic choice, a choice that suggests that being a researcher, learner and an effective actor in this world does imply being a discursive creator-one who contributes to and challenges dominant discourses by establishing some powerful allegiances and speaking up, and not through passivity, docility and compliance.