The multiple reals of workplace learning The multiple reals of workplace learning

The multiple reals of workplace learning are explored in this paper. Drawing on a Foucauldian conceptualisation of power as distributed, relational and productive, networks that work to produce particular objects and subjects as seemingly natural and real are examined. This approach enables different reals of workplace learning to be traced. Data from a collaborative industry-university research project is used to illustrate the approach, with a focus on the intersecting practices of a group of professional developers and a group of workplace learning researchers. The notion of multiple reals holds promise for research on workplace learning as it moves beyond a view of reality as fixed and singular to a notion of reality as performed in and through a diversity of practices, including the practices of workplace learning researchers.


Introduction
The distinction between workplaces as the domain of practice and Higher Education Institutions (HEIs) as the domain of knowledge production, theory and learning are no longer as clear-cut as they once might have seemed. Learning has escaped its traditionally understood setting in educational institutions and has been located in other sites, including workplaces (Billett, Fenwick & Somerville, 2006;Gherardi, 2006;Hager, Lee & Reich, 2012;Wenger, 1998). Indeed, as far back as the early 1990s, workplaces were named as 'the Learning Organisation' (e.g. Garvin, 1993;Marsick & Watkins, 1990;Senge, 1992). Similarly, work has crossed institutional boundaries and is increasingly counted as learning within the academy. For example, many undergraduate degrees incorporate a work-based learning component, professional doctorates are now available in various disciplines and work-based learning sometimes comprises entire degree programmes (Boud & Solomon, 2001;Garnett, Costley & Workman, 2009). It could be said that work has been translated into learning and learning translated into work. It is to the almost seamless translation of learning into work, as objects of knowledge move between these domains and particular modes of subjectivity are produced as seemingly natural and real, that is of interest in this paper.
The contemporary intersections between the institutions of work and Higher Education (HE) both give rise to and come out of research and theory development on learning at work and, as various authors have pointed out, this is a heterogeneous field (Corradi, Gherardi & Verzelloni, 2010;Easterby-Smith, 1997;Fenwick, 2006Fenwick, , 2008Hager, 1999). Yet, while learning at work is understood in multiple ways, the terms learning and learner are often used generically as if they had a fixed and shared meaning, both in government policy documents and in much of the literature on learning at work (Edwards & Nicoll, 2004;Fenwick & Edwards, 2011). For instance, Fenwick (2010, p. 80) points to an ongoing failure in the workplace learning literature to provide definitions of what is meant by learning and the purpose of learning, which contributes to the assumption that 'learning' is a single object, self-evident and mutually understood'.
The move to singularity and cohesion through the generic use of learning and learner can be problematic as it tends to mask the politics of learning at work and the various purposes to which discourses of learning might be put. For example, Boud and Solomon (2003) have drawn attention to the politics around naming oneself as a learner at work. They suggest that 'being a learner is a risky business as it can position one apart from the group ' (p. 330). For instance, employees may feel vulnerable identifying as a learner with their managers and, at times, with colleagues. Moreover, drawing attention to diversity and difference in workplaces, Edwards and Nicoll (2004, p. 160) also suggest that 'the notion of workplace learning itself needs to be used cautiously, lest it results in unsustainable generalizations'. For instance, variations in the organisation of workplaces such as: large, medium or small; heavily bureaucratised to 'virtual'; team based forms of organising work to more individualised modes; produce very different practices. They conclude that grouping the multiple and varied practices in workplaces under the general banner of workplace learning overlooks 'the complexities ordered in the actor-networks of specific workplaces ' (p. 172).
Furthermore, the notion of difference in terms of what learning is tends to be overlooked in the educational policy arena with a push by governments in a number of countries for the formation of HE -Industry partnerships in the design, delivery and assessment of higher education programmes (Nicoll & Fejes, 2011). For instance, recent policy documents in the UK speak of learning, both in HE and in workplaces, primarily in terms of skill development and employability and it is the desire of government, both the former Labour government and the current conservative government, for HE to be active in the provision of continuing professional development of employees (Department for Business Innovation andSkills, 2009, 2011;Wedgwood, 2006;Wilson, 2012).
While some workplace learning research has engaged with variation (e.g. Nerland, 2012), Fenwick (2008Fenwick ( , 2010 has pointed to the ongoing failure in much of the literature to recognise and work with difference and diversity. Diversity and plurality draw attention to the politics of learning at work, which raises a set of questions that tend to be overlooked in the literature including: Who is able to speak about learning at work? And what is able to be said? How is learning at work able to be known? And what are the material effects of knowing learning in particular ways? For example, what modes of worker subjectivity are produced as seemingly natural and real through particular ways of knowing learning at work (e.g. Usher & Solomon, 1999)?
More recently, Fenwick has proposed Mol's notion of Ontological Politics as a useful starting point for conceptualising difference and multiplicity in Workplace Learning (2010). Mol (1999) suggests that different practices produce their own material reality. Thus, rather than understanding reality as fixed and singular (an underlying premise in most social science research), there are multiple reals. Fenwick (2010) extends this concept to the field of workplace learning and suggests there may be multiple reals of learning produced in and through different practices. This is more than the claim that that there are different perspectives on learning, rather that learning is different in different locations.
This paper explores the multiple reals of workplace learning by drawing attention to alignments and contestations in a collaborative industry-HE project examining the significance of everyday learning at work. The project was conducted in Australia in the early to mid 2000s. While the research was undertaken in Australia, the industryuniversity collaboration can be understood as an example of a broader shift to the coproduction of knowledge between academics and industry partners (Antonacopoulou, 2009;Nicoll & Fejes, 2011). The first part of the paper introduces a poststructuralist analytic framework for examining actor networks and the multiple reals of workplace learning. Next, various actors in the research project are introduced. Then, the assemblages of the project are traced and the ways knowledges and power sit within networked relations to produce particular modes of subjectivity as seemingly natural and real. The final section considers the implications for workplace learning researchers and openings for workplace learning research.

Conceptual framing
The approach for examining the multiple reals of workplace learning is underpinned by a Foucauldian notion of power as relational, distributed and productive (Dreyfus & Rabinow, 1982). Detailed discussions of a Foucauldian perspective on power and what this approach enables for research on lifelong learning (e.g. Fejes & Nicoll, 2008) and workplace learning (e.g. Edwards & Nicoll, 2004) have been provided in the adult education literature and only a brief introduction is provided here.
Rather than power residing with an individual or group such as the sovereign, the church or the academy, Foucault proposes that power in modern times should be understood as distributed across social institutions and practices. There is no longer a single authority. Furthermore, rather than thinking of power as only repressive, it is also productive.
If power were never anything but repressive, if it never did anything but to say no, do you really think one would be brought to obey it? What makes power hold good, what makes it accepted, is simply the fact that it doesn't only weigh on us as a force that says no, but that it traverses and produces things, it induces pleasure, forms knowledge, produces discourse. It needs to be considered as a productive network which runs through the whole social body, much more than as a negative instance whose function is repression. (Foucault, 1980, p. 119) A Foucauldian conceptualisation of power as networked and functioning through the production of particular objects and modes of subjectivity as seemingly natural or real, underpins an actor network approach. Indeed Law (2008) refers to actor networks as scaled down, empirical versions of Foucault's discourses. Law (2008, p. 145) proposes that an actor network approach enables the 'strategic, relational, and productive character of particular, smaller-scale, heterogeneous actor-networks' to be examined. Fenwick and Edwards (2011) detail the usefulness of an actor-network approach for examining the ongoing 'press for similarity to overcome difference' (p. 709) in educational policy. Following Law, they argue it enables 'matter-ing processes' to come into view through examining 'the political negotiations going on at micro levels that mobilize particular attachments and uptakes ' (pp. 711-712). This introduces an important concept in actor-network approach, which is translation. Latour used the concept of 'the token' (1986, pp. 267-268) to illustrate translation and the ongoing renegotiation of objects in networks: the spread in time and space of anything -claims, orders, artefacts, goods -is in the hands of people; each of these people may act in many different ways, letting the token drop, or modifying it, or deflecting it, or betraying it, or adding to it, or appropriating it…the chain is made of actors…and since the token is in everyone's hands in turn, everyone shapes it according to their different projects. This is why it is called the model of translation.
What is of particular interest in this paper is the ways learning (as a token) is taken up and translated as it moves from HE to particular workgroups beyond the academy.
Rather than assuming a unified and fixed reality, an actor network approach provides the analytic space for exploring multiple reals (Law, 2004). For instance, using an actor network approach Mol (1999) directs attention to the ways different practices in the identification and treatment of anaemia (clinical, statistical and pathophysiological) produce different reals, including the real of sex difference. It is the ongoing enactment of particular realities in and through practice that work to stabilise the real and enable it to appear fixed.
The concept of multiple reals problematises the notion of a single truth produced through cohesive representations in research accounts, thus opening these truths to examination. While other conceptual approaches such as perspectivalism and constructivism work with plurality, Mol (1999) proposes a useful distinction between these approaches and ontological politics. Ontological politics directs attention to action rather than observation. It enables 'a semiotic analysis of the way reality is done, from studying performances, from making a turn to practice' (p. 87). In drawing attention to performativity, the performative practices of workplace learning researchers are able to come into view in accounts of workplace learning.

The industry-university research collaboration
A three year, industry-university research collaboration undertaken in an Australian workplace provides one site for empirically exploring the multiple reals of workplace learning. In this research partnership, a cross-institutional and multidisciplinary research team of workplace learning academics and professional development practitioners set out to examine the significance of learning embedded in practice in a large public-sector workplace. The organisation, referred to as PSE in this paper, provided post-secondary education and was the workplace of both the professional development unit coresearching workplace learning and four work groups participating in the research project. The four workgroups were a group of senior managers, a group of HR administrators, a group of trade teachers and a group of business teachers providing training and development in various workplaces. The author was a member of the research team and the project provided the research site for her doctoral research.
The project was conducted in two stages. Stage One involved initial interviews conducted with individual members from each of the four workgroups. Twenty three interviews were conducted, each approximately one hour in duration, where employees were asked to speak about the challenges and changes involved in their work. The interviews were transcribed and analysed and the findings were reported back to each of the workgroups at the completion of Stage 1. During Stage 2 of the project, a series of meetings were conducted with two of the workgroups to explore particular learning themes arising from the Stage 1 interviews. The project findings were reported in a range of documents including a project report for the organisation and various journal articles.
One way of understanding the project is that the 'truth' of workplace learning was established through following a typical model of social science research where researchers enter the research site, collect data, analyse the data, and then report the findings to research partners, academic communities and other relevant stakeholders. An actor network analysis, however, enables the project to be understood as a site for exploring assemblages connected with the production, circulation and consumption of learning knowledges; and more specifically in this paper, an analysis of the intersecting practices of adult education and professional development.

The professional developers
A representative from the professional development unit was a member of the crossinstitutional research team and participated as a co-researcher on the project. This involved participating in the project planning meetings, collecting and analysing data and contributing to the preparation of various publications from the project. The professional development unit were part of central services at PSE and provided professional development knowledge and expertise to each of the colleges throughout the state. The actual provision of staff training was generally conducted at the college level, usually by professional development staff within the college. The professional developers described their role in PSE, in the following way on the PSE intranet: The [Professional Development Unit] is committed to supporting [PSE] staff in acquiring and maintaining the skills and capabilities essential for [PSE] to maintain its position as the leading [post secondary] education and training provider in Australia.
Thus, the professional developers were interested in the skill development of PSE employees for the purpose of enhancing organisational performance.

The workplace learning academics
The workplace learning academics were located in a department of Adult Education in a metropolitan university in Australia. The academics had an interest in learning in and through practice, which was, in part, connected with the provision of Work Based Learning programmes in the department. These programmes had been discontinued at the university, however, by the time the project commenced. While they were a crossdisciplinary group, there was a strong Adult Education ethos in circulation whereby learning is understood as lifelong, and taking place in multiple sites, including beyond the walls of the academy. The workplace learning academics were interested in mapping and making visible the everyday learning of workers in the PSE workplace.
Re-configuring workplaces as sites of learning Law (2004, p. 21) proposes that 'Realities don't exist without their matching inscription devices and such inscription devices (and their particular products) are elaborate and networked arrangements that are more or less uncertain, more or less able to hold together and more or less precarious'. Thus, taking a Foucauldian perspective that knowledges are never neutral, the knowledge products of the project can be examined for the objects and subjects they brought into effect. For example, the inscriptions of workplace learning produced throughout the project provided a language for talking about work and workers, and making workplaces and workplace learning imaginable in particular ways (and not others) (Rose, 1999). And it is in this sense that the knowledge products were not mere reflections of a pre-existing reality in the PSE workplace but instead worked to produce particular reals. This theme is explored by examining the alignments of the project as well as contestation and resistance.
The textual products of the project, both written and spoken, were multiple and varied and included an initial project proposal for the funding body, the project planning meetings of the research group, feedback sessions with PSE representatives, written reports to PSE, a brochure for distribution in the PSE workplace, everyday conversations of the researchers, the meetings with the four PSE workgroups, journal and conference papers, including this paper,…. the list goes on. In many of these texts the PSE workplace was re-presented as a site of learning and workers as workplace learners.
While this might seem unsurprising, the aim of the project was after all to produce knowledge about workplace learning, it is to the almost seamless translation of the PSE workplace into a site of learning through the knowledge products of the project that I seek to draw attention.

Aligning institutions
The project was undertaken in the early to mid 2000s and was funded through the Strategic Partnerships with Industry -Research and Training (SPIRT) programme. At that time, higher education institutions in Australia were increasingly being called on by the federal government to operate along commercial lines (eg. Fullerton, 2005;Gallagher, 2000), and the government had implemented the SPIRT programme to encourage industry-university research partnerships (eg. Nelson, 2002;"Strengthening Australia's Higher Education System", 2004). The aims of SPIRT included the development of 'long-term strategic research alliances between higher education institutions and industry in order to apply advanced knowledge to problems...', as well as providing 'industry-oriented research training to prepare high-calibre postgraduate research students' in order to 'produce a national pool of world-class researchers to meet the needs of Australian industry' (Linkage -Projects, 2005).
The SPIRT programme could be read as a programme of government, which aimed to shape worker conduct in particular ways (Edwards & Nicoll, 2004;Miller & Rose, 1993;Nicoll & Fejes, 2008). For example, it attempted to shape the conduct of academics through producing enterprising subjects and more commercial modes of operation in universities. The programme also potentially opened up a position for industry partners as co-researchers and knowledge producers. Furthermore, the aim of the project, which was to produce knowledge about everyday learning at work, can be understood as interconnected with the ongoing desire by workplace managers to govern worker conduct through the alignment of individual employee goals with broader organisational goals and objectives (see Rose, 1999).

Aligning individual and organisational goals
The problem of employee alignment has been an ongoing theme in managerial literature for decades, with top-down managerial techniques no longer understood as effective strategies of government. For instance in 1985, Walton wrote an article in the Harvard Business Review called 'From control to commitment in the workplace'. This trend has led Rose (1999), amongst others, to suggest that the successful government of workplaces will be achieved through harnessing worker subjectivity rather than through its suppression. This is evident in the trend in managerial texts to call on learning as a technique for producing alignment, with communities of practice being the latest in a number of versions of this theme (e.g. Garvin, 1993;Senge, 1992;Wenger, McDermott & Snyder, 2002).
A community of practice discourse was a dominant learning discourse in circulation at the time of the project (Contu & Willmott, 2003;Gee, Hull & Lankshear, 1996) and this discourse was drawn on by both the academics and the professional developers at the outset of the project to speak about learning at work. From a communities of practice perspective (Wenger, 1998), learning at work is understood as identity work, whereby workers move from peripheral to full participation in communities of practice and in so doing take up an occupational identity such as the teacher, the manager, the researcher, and so on. In other words, learning is understood as the socialisation of workers through their participation in communities of shared practice in workplaces.
A community of practice discourse appealed to the academics as it enabled workplaces to be understood as sites of learning through directing attention to learning in and through practice. It also appealed to the professional developers as it was a discourse that potentially put learning on the managerial agenda as a strategy for managing the conduct of employees. For example, the benefits of communities of practice are described by Wenger et al. (2002, p. 18) in the following way: The ability to combine the needs of organizations and community members is crucial in the knowledge economy, where companies succeed by fully engaging the creativity of their employees. The multiple and complex ways in which communities of practice deliver value to both members and organizations is the reason they are fast becoming a central part of the management agenda.
In (some) communities of practice texts, workplace learners are re-presented as active and emotionally engaged knowledge-producers in the contemporary workplace and work is re-presented as the mechanism through which a creative and passionate self can be produced. For example, Mitchell (2003, p. 5), a change management consultant contracted by the Australian government at the time to advise on the successful implementation of change across the PSE sector, described communities of practice as: 'groups of people bound together by common interests and a passion for a cause, and who continually interact'. According to Mitchell, the 'new' worker was one who placed an emphasis on relationships, for example, PSE practitioners needed to be 'more clientfocused by establishing improved relationships with both enterprise clients and individual students' (p. 8). Workers were to become knowledge producers where: 'The development of practice involves a balance between exploring ideas together and producing documents and tools' (p. 6). The members of the community of practice, as described by Mitchell, 'communicate regularly and continuously in an atmosphere of trust, enabling collective enquiry about issues of importance to the members' (p. 6).
The communities in this change management text had great appeal (at least to those concerned with managing employee conduct) and these were the texts that the professional developers drew on to guide their everyday practices. And perhaps the most appealing aspect of this document was that the communities of practice described were all aligned (seemingly unproblematically) with broader organisational goals.

The project proposal
The initial uptake of the government funded programme by both the academics and the professional developers involved the construction of allied interests and the development of a shared language for talking about work and learning (Miller & Rose, 1993). One way this was established was through the development of a shared problem space in the text of the project proposal. The proposed project outcomes were identified as: • Improved recognition of the learning to be found in the organisation, to the benefit of both the organisation and individual employees. • Improved understanding by key personnel in the organisations of the ways in which organisational culture and procedures encourage or inhibit learning, and the issues which need to be resolved in developing the learning organisation. • Improved learning systems and learning strategies in the organisation that will more effectively facilitate learning embedded in practice.
The research proposal text can be examined for the ways learning was spoken about in this document, the meaning of learning embedded in this text and the subjects this version of learning worked to produce as seemingly natural. The reference to 'the learning to be found in the organisation', where learning is used as a noun, works to reinforce the view that learning is a thing that exists, and has a fixed and accepted meaning. Also, the notion that learning can be 'encourage(d)' or 'inhibit(ed)' again implies that learning has a fixed and unified meaning as does the promise that 'systems' and 'strategies' will be developed to enhance (this particular understanding of) learning. Moreover, the reference to 'the learning organisation', already re-writes the workplace as a seemingly natural site of learning. The representation of learning in the anticipated project outcomes as benefiting both individuals and the organisation at one and the same time works to reinforce a view of workplaces as sites where the goals of individual employees are necessarily aligned with organisational goals and objectives. This is referred to in the industrial relations literature as a 'unitarist' assumption of employee relations (see Fox, 1974). In other words, learning was re-presented as a solution to the longstanding managerial problem of worker alignment and commitment. Furthermore, there was not just the assumption of alignment between employees and organisations in this text, but the promise of providing instances of alignment. For example, an anticipated outcome was 'improved recognition' of learning that benefits 'both the organisation and individual employees'.
Learning was named in the proposal as 'embedded in practice', which suggests the circulation of a community of practice discourse and similarly to much of this literature, there was little space for resistance and contestation in the representation of learning provided in the proposal. The language of learning used in the proposal proved persuasive as the funding application was successful and an alliance was forged between the workplace learning academics and the professional developers at PSE. Who, after all, could possibly be 'against learning' (see Contu, Grey & Ortenblad, 2003)?

Translating learning
The academics could be understood as authorities on workplace learning and, indeed, it was because of their expertise in this knowledge domain that they were able to enter into a collaborative relationship with their industry partner in order to investigate learning in the PSE workplace. The cross-institutional alignment suited the professional developers as it provided the project with the stamp of the academy and the power associated with more traditional forms of knowledge production.

Stage 1 interviews
While the Stage 1 project texts, which included interviews with members of the participating workgroups and analysis of the transcripts, can be understood as simply reflecting a pre-existing workplace learning reality at PSE, they can also be examined for the ways the PSE workplace was re-inscribed as a site of learning and the subject positions produced for PSE employees in these texts. For example, the interview texts had been translated by the researchers into a typology of learning where it was proposed (in a handout given to each of the workgroups) that 'we' learn at work through the following processes: 'mastering organisational processes', 'negotiating relationships' and 'dealing with atypical situations'.
The typology of learning at work proposed by the researchers encompassed a broad range of work practices. Indeed, it is difficult to think of particular work practices that fall outside this categorisation of learning. And through inscribing these generic work practices as learning, most work, and workers, were re-inscribed as workplace learners. The sense that learning at work was universal, and that being a learner was a characteristic shared by all workers, was emphasised by the use of unifying words in this document such as 'we'. 'We' suggested the collective and that the knowledge about workplace learning produced by the researchers applied to all workers. In these texts the researchers were attempting to persuade the workgroup members that they were, in fact, workplace learners.
While proposing that these categories encompass a broad range of practices, thereby reinforcing a sense of universality, it is also useful to consider what they excluded. For example, if 'mastering organisational processes' is recognised as learning, does this suggest that lack of mastery denotes a failure to learn? Moreover, who decides when mastery is achieved? Workgroup managers? Senior managers? Workgroup members? Service recipients? Can learning at work be understood in ways other than skills development and 'mastery'? Similarly, when dealing with 'typical' situations at work, rather than atypical, is there nothing to be learnt? Is learning absent from the mundane, everyday labour that forms a part of many workplace practices? And if so, who and what does this re-presentation exclude as learners and learning? This is not to suggest, however, that the knowledges produced in the project were false, and that the truth could have been established through better research methods, but rather to direct attention to the ways categories work to produce exclusions and more specifically to examine the particular realities produced in this inscription of workplace learning.

Stage 2: workgroup meetings
The workgroup meetings conducted during Stage 2 of the project provided another site for re-inscribing the PSE workplace as a site of learning. Interestingly, Stage 2 of the project had initially been named as the 'Intervention Stage'. This name, however, was later dropped by the researchers because of its methodological implications. The academics wanted to distance themselves from an action research style of methodology whereby an intervention is introduced and the effects of the intervention are systematically examined. While the name 'intervention' went out of circulation, the Stage 2 meetings can still be understood as performing the function of an intervention as they provided an obligatory point of passage (Callon, 1986) for the research subjects, whereby those who participated were automatically positioned as 'the workplace learner'.
Furthermore, through organising the Stage 2 workgroup meetings as sessions where workgroup members reflected on their workplace experience in order that the truth might be revealed about learning at work, it could be said that a dominant learning discourse of reflection on practice was being enacted by the researchers. This has become a powerful learning discourse, particularly within the field of professional development (Bradbury, Kilminster, Frost & Zukas, 2009;Edwards, 1998;Edwards & Nicoll, 2006). In this discourse, workplace learning is understood as a process of rational reflection on experience for the purpose of enhancing work practices.
During Stage 2 of the project the language of learning was introduced by the researchers as a theme for exploring particular practices in the workgroups. Through this process particular work events and practices that had been described by workgroup members (but not necessarily in relation to learning, nor named as learning) were selected and named by the researchers as connected with learning and thus reconfigured as learning. For example, for the senior managers it was: 'learning through transition' and for the trade teachers: 'learning through the challenge of new students'.
The workgroup meetings can be understood as part of a network that contributed to re-producing 'the workplace learner' as a seemingly natural subject in the PSE workplace. Through re-inscribing work as learning, and workers as learners, the subject position of 'the workplace learner' was constructed and workers became thinkable and knowable in this way. While this is not necessarily a problem, as 'the workplace learner' subject can be constructed in multiple ways, the ways this subject is constructed in research accounts requires closer examination as it plays a part in the ongoing struggle over worker subjectivity (Weedon, 2004). The spaces made available to workplace learners, in naming learning this way were 'learning through transition' and 'learning through the challenge of new students'. And as will be discussed later, neither of these learning spaces seemed to appeal to workgroup members participating in the project.
Shaping the token While the project provided a site for alignment, there was also struggle between the workplace learning academics and the professional developers where each attempted to shape learning in ways that suited their own purposes. Latour's concept of 'the token' (1986) is useful here for examining the ongoing renegotiation around the knowledge object 'workplace learning', and the implications in terms of the subjects particular notions of learning brought into effect.

The brochure
The academics attempted to introduce reflexive methods throughout the project, whereby the researchers considered their own learning in the cross-institutional project and differential relations of power between the academy and the PSE workplace. The push for reflexive texts can be linked with the poststructuralist leanings of the group and an interest in relations of power and the ongoing struggle over subjectivity. The academics were not interested in using learning as a technology for bringing about workplace change and sought to disrupt a programmatic approach with a focus on interventions.
However, the reflexive approach adopted by (some of) the researchers became a site of struggle and was considered, by some, not to be the point of the project. In other words, the purpose of the project was not to produce texts about us (the researchers), but texts about them (the researched). The gaze of the researchers was to be firmly directed on the other employees in the workplace and this, I suggest, was connected with the desire by the professional developers to know 'the workplace learner' in order that their conduct might be governed. This was what counted as knowledge about workplace learning for the industry partners and this was the type of knowledge outcome that the professional developers (and other groups in the organisation) expected from the project.
In a final attempt to achieve an outcome that (might) contribute to enhancing workgroup and organisational performance, the professional developers requested the production of a brochure about learning in the PSE workplace. This was to be a document that could be used by multiple groups in the organisation with the aim of enhancing learning in that workplace. As such it needed to appeal to professional developers, workgroup managers and workgroup members. While the academics collaborated on the production of the brochure, the production of this text was not a seamless and cohesive recording of the seemingly transparent findings ('matters of fact') of the research project. Rather, it involved a protracted struggle between the academics and the professional developers over the way learning in the PSE workplace might be re-presented.
The production of the brochure was a drawn out exercise where the text was passed backwards and forwards between the professional developers and the academics, with each group working on it in a way that suited their own interests. For instance, the professional developers insisted on the inclusion of text that re-presented them as active players in learning in their workplace, so sections such as: 'Structured professional development plays a vital role in skilling the organisation' were included. The academics, in wanting to avoid an overly programmatic approach wrote of multiplicity and 'the different languages of learning' in this organisation.
In the final product, however, the re-presentation of everyday learning was not dissimilar to the later communities of practice literature (e.g. Wenger et al., 2002), and similarly to that literature, the brochure drew on a unitarist assumption of alignment in an effort to persuade. For example, in an attempt to persuade employees as to 'Why everyday learning matters' [for them], one of the section headings used in the brochure, the following argument was presented: There are several reasons why everyday learning matters for workers. For example everyday learning enables workers to master organisational processes, negotiate particular aspects of their work and deal with atypical situations (both individually and collectively).
In other words, everyday learning is critical in the day-to-day jobs of workers and is significant because it helps them address local issues, do their jobs more effectively, and to respond more quickly to the problems that arise at the coalface.
Thus, what was able to count as learning in the above text was framed by a managerial account of productivity, whereby learning necessarily contributes to enhanced organisational performance. Learning was re-presented as a means of enabling workers at PSE to be more 'effective' and efficient in solving workplace problems and workplace learners were re-presented as productive and aligned employees. There was little in the brochure to disrupt a view of learning as skills development and 'mastery', and more specifically the development of skills that contribute to enhanced organisational performance. And unsurprisingly, considering the purpose of the document, there was no space for representations of resistance to managerial objectives and workers who may have alternative perspectives.
It is important to emphasise, however, that this was only one knowledge product (and one reality of workplace learning). There were other publications, with some drawing attention to differential relations of power in workplaces and the politics of learning at work. Furthermore, how effective the representation of learning provided in the brochure actually was requires further investigation. While beyond the scope of this paper, a Foucauldian analysis suggests examining how the brochure was taken up, by whom and for what purposes?
Moreover, the language of learning as skills development, which at times was shared by the academics and the professional developers, had little appeal within the workgroups in this workplace. The re-inscription of the workplace as a site of learning by the learning experts was often challenged by workgroup members and there was contestation around the positioning of the researched as 'the workplace learner' throughout the project. For example, the senior managers resisted being positioned as workplace learners, even though there was a discourse of Organisational Learning in circulation in their workgroup. 1 As the director of the senior managers work group indicated when asked about the importance of demonstrating confidence and certainty at work: ... The term I use quite a lot is 'maintaining the ascendance'. I use ascendance as a metaphor because we just simply cannot afford, if we get knocked off, then the institute, the whole structure, there's questions on the whole way of the organisation structure and its development. So we can't afford in a more public forum to be seen to be small 'L' players... 2 Furthermore, the circulation of an apprenticeship discourse amongst the trade teachers also made it difficult for them to take up a learner identity. An apprenticeship discourse works to produce the subject positions of 'master' and 'apprentice' as seemingly natural, and the trade teachers tended to understand themselves as masters rather than apprentices, particularly in relation to their students. For example: When I've got students around me, I don't seem like I'm learning now, I'm the one doing the teaching. As far as I'm concerned I'm the one in control. I'm the one with the knowledge that's being passed over. I've got the experience...

Networks, knowledge, power and multiple reals
The alignments of the project draw attention to a web of relations between professional developers, workplace learning academics, communities of practices texts, the workplace learner, masters, apprentices, government policy on HE, research proposals, workplace documents and how they were more or less precariously held together. For example, (some of) the inscriptions of the project enabled the PSE workplace to be known as a site of seemingly natural learning, with all workers re-presented as learners, and 'the workplace learner' subject re-presented as an effective and efficient problem solver in the workplace, contributing to enhanced organisational performance. And in casting this particular grid of visibility over the workplace learner, only certain types of learning at work were able to become visible (Rose, 1999, p. 270). This assemblage contributed to the durability of 'the workplace learner' as necessarily aligned with organisational goals and objectives. Thus, rather than understanding power as residing with a particular group, for example with the State, or with management or with the Academy, the project can be understood as a site where power was distributed yet interconnected to re-produce a taken for granted managerial notion of employees as necessarily aligned with organisational goals and objectives.
The account also problematises the notion of a singular object 'workplace learning' and directs attention to the ongoing enactment of different workplace learning reals by the professional developers and the workplace learning academics. At times there was alignment in terms of what counted as learning at work and at times contestation and difference. While the shared language of the professional developers and the academics in relation to the productiveness of learning at work enabled the development of alliances at the project's inception, as the project played out it became evident that there were different ways for thinking and talking about learning at work, and these were integrally interrelated with practice. The project was a space where the discourses of adult education, embodied and enacted in the practices of the workplace learning academics, and the discourses of human resource development, embodied and enacted in the practices of the professional developers, intersected. The account also suggests different workplace learning reals produced in and through the practices of the senior managers and the trade teachers, although space precludes a more detailed exploration in this paper.
It has been argued that the professional developers used workplace learning for programmatic purposes, where learning was understood as a thing that could be managed and controlled in the PSE workplace and where learning interventions could be designed to enhance workplace performance. The professional developers understood learning at work as skills development and, more specifically, skills development for the purpose of enhancing organisational performance. Furthermore, and similarly to the change management texts that were in circulation in their workplace, their uptake of a communities of practice discourse was for the purpose of changing practices rather than the recognition of knowledges produced in existing practices.
In contrast, the workplace learning researchers drew on a practice discourse for the purposes of recognising the learning in workplaces that takes place in existing practices. The workplace learning academics were interested in legitimating sites other than the academy as sites of knowledge production and learning. While harnessing the power of the academy, the academics also sought to subvert it through re-writing learning in terms that disrupt traditional knowledge hierarchies and the privileging of academic knowledge.
The ongoing contestations during the project in relation to who could, and who could not, be named a learner and who could, and who could not, do this naming, suggest that workplace learning is anything but a generic term. Thus, the naming of learning at work and the uncovering of its truth is perhaps more complex than many accounts provided in the workplace learning literature and importantly, more than a matter of different perspective. This raises difficult questions for workplace learning research in terms of who is able to speak about learning at work, what is able to be said, and how might workplace learning be known?
A Foucauldian reading of power as distributed, relational and productive enables an account of workplace learning that introduces resistant rather than necessarily aligned subjects in workplaces. The multiple nodes that work to hold networks in place and reproduce seemingly durable objects and subjects are also potential sites for resistance and fracture. Thus, rather than understanding power in the project as emanating from a single site, with the academics as the agents of government and only producing knowledge that suited the purposes of their industry partner, or the industry partners as necessarily subservient to the power and authority of the academy, or workers as cultural puppets passively taking up the subject position of 'the workplace learner' during the project, another view of power is provided here. There was no ultimate authority. Instead, power was distributed across various social institutions and practices. This directs attention to the pervasiveness of power but also to its potential fragility and the multiple sites for resistance and renegotiation.
Openings for workplace learning researchers and workplace learning research An actor network approach provides a useful analytic device for directing attention to differences in what learning is, both within and across workplaces, as well as across institutions. Rather than trying to bring together different versions of the real together in a single representation, it provides a methodological tool for exploring partial connections (Law, 2004). This provides a useful opening for workplace learning researchers as it enables a more reflexive approach to examining the part played by our own knowledge products in accounts of workplace learning and the objects and subjects these accounts work to produce as seemingly natural. The notion of partial connections provides a useful opening to tired debates over theory versus practice, where theory is understood as only residing in the academy and practice only residing in workplaces.
Partial connections work to hold networks together but they provide space for resistance. An actor network approach provides the analytic space to explore resistance, thereby enabling accounts of workplace learning as other than alignment. The approach opens up potentially fruitful areas for the exploration of resistance in workplaces and its connection with learning as identity work. For example, when might it be useful to name oneself as a learner at work and use a workplace learner identity to resist other positionings? The uptake by the trade teachers of a learning discourse but using it for their own purposes is suggestive in this respect. The analysis also indicates the dominance of a discourse of learning as mastery and skills development but are there other ways of being a learner at work? Following on from Boud and Solomon (2003), are there ways of disrupting this dominant discourse in workplaces in ways that do not make employees vulnerable?
The theme of translation is particularly useful for examining the ways learning is translated as it moves across different sites. It enables attention to be directed to the active part played by knowledge products in contemporary translations around work and learning. It also points to the inevitable failures when to attempting to make two worlds equivalent (Law, 2004). The focus in this paper was on the translations associated with the movement of 'learning' into workplaces and the parts played by a network of actors, including workplace learning academics, in these translations. The approach also opens space for examining translations as work moves from workplaces into HE and becomes learning.
In summary, an actor network approach, underpinned by a Foucauldian conceptualisation of power as distributed, relational and productive, provides a useful analytical tool to examine the multiple reals performed in and through various learning practices as well as the relations between these reals -the partial connections. The approach enables networks that work to produce very real, material effects to be mapped, including the part played by academics and their knowledge products in these networks. However, rather than understanding networks as fixed, it is an approach that enables the potential fragility of power to be exposed.