Summary in English

The thesis describes a project where possibilities for IT-support in work knowledge development has been investigated in an area where the confrontation between information technology and work knowledge is highly evident – process control work. Production disturbances of varying degree is a continuous element in process control work and quick changes in level of risk in the production situation is an important aspect of the complexity of the work. Automation reduces the operators abilities to maintain and develop the skills involved in handling production disturbances requiring manual intervention. In the long run this might lead to an erosion of this aspect of the operators work knowledge. The study has been carried out at an partly automated chemical plant with batch-oriented production. The plant produces binders and adhesives in ten separate reactor systems. A tool for collaborative learning in this setting has been developed that uses an information format named TIPS – thin interactive problem scenarios – for describing problem situations in production. The tool has been evaluated with process operators in a series of video recorded meetings.

Frames of reference

There are three major theoretical frames of reference for the thesis: skill and technology; participatory design and social constructionism. A research tradition referred to as »skill and technology« has evolved at the Centre for work life studies in Stockholm where different work contexts have been studied with a particular focus on how the introduction of new technology affects skill. Since the beginning of the 1970:s a number of case studies of computer use in different work contexts have been documented. The studies has focused on long term effects of computer use. Since then the focus in this research tradition has shifted towards the nature and development of work knowledge, and it is this part of the tradition that has been a main influence on the thesis work. Kjell S. Johannessen’s development of concepts for describing work knowledge (Johannessen 1990, 1992, 1997, 1999a, 1999b), based on Ludwig Wittgenstein’s notions of praxis and rule following, has been a major influence on the epistemological perspective in the thesis work.

      My investigation of work knowledge development and collaborative learning in process control work has been as a researcher but also as interaction designer, setting focus on the possible roles interactive digital systems can play in collaborative learning in a work setting. There are two connected lines of development in informatics that forms the second frame of reference for the thesis: the development of a Scandinavian tradition in participatory design of computer artefacts; and the growing interest for ethnographically inspired methods in participatory design. The former has influenced the design process that has been part of the thesis work, while my methodological approach in the research has been grounded in the latter. Research in both areas have influenced my understanding of the concepts of praxis and rule following.

      Social constructionism is a growing field of theories constituting the third frame of reference for the thesis. The main concern is to explicate the processes by which people come to describe, explain or in other ways account for the world in which they live including themselves (Gergen 1995). Social constructionist theories share a position taken against a number of underlying assumptions in a large part of the research community: a dualist view of the world where people and the world can be studied as separate entitities; an objectivist perspective on knowledge where meaning and properties of the world that we can perceive are assumed to be inherent in the world itself; an individualist perspective on knowledge where the individual is viewed as the owner of knowledge about the world; and the assumption that language is an objective representation of reality. In contrast social constructionist researchers render objective descriptions of reality impossible since such descriptions are influenced by our specific cultural, historical and linguistic understanding (Sandberg 1999). Instead reality is constructed in a social interchange with continuous negotiation between people on what constitutes the world we live in. This means that knowledge about reality also is developed in a social interchange and that meaning in the language used to describe reality is constituted by its use in meeting reality (Sandberg 1999).

Problem domain and research questions

A number of studies show that automation in process control reduces the operators possibilities for maintaining and developing work knowledge since the number of production disturbances requiring operator intervention decrease (Bainbridge 1978, 1987; Brehmer 1993; Ellström & Gustafsson 1996; de Montmollin & de Keyser 1985; Skorstad 1988). An investigation of psychological demands in process operator work (Sandén & Johansson 1987) showed that 50% of the operators perceived insufficient knowledge for handling production disturbances as a problem. One suggested strategy to reduce negative effects of automation is to switch to manual control on a regular basis during normal operation in order to maintain manual control skills (Bainbridge 1978; de Montmollin & de Keyser 1985). Another strategy is to work with simulations of production disturbances. According to Bainbridge (1978) operators on automated plants need frequent breakdown training exercises, and therefore simple test »games« might be an alternative to complex simulation to remind operators of process structure and relationships. However, unexpected problems cannot be simulated, so therefore only general strategies can be trained. A third strategy is to stimulate communication between operators. According to Sandén & Johansson (1987) reporting and discussing incidents on a regular basis can contribute to safer and more efficient operation, and increased operator competence. In an ethnographic study of another domain, photo-copier repair work, Julian Orr (1990) has described the importance of community memory for maintaining work knowledge. By exchanging »war stories« about difficult repairs the technicians share work knowledge and contribute in developing community memory as a basis for developing experience in repairing copying machines.

      The main reason for choosing process control work as a domain for studying computer supported collaborative learning has been that the confrontation between the »exact language« of computing and the experience based knowledge of workers is particularly evident in this domain. The main goal has been to investigate possibilities for introducing computer supported collaborative learning in process control work as an intervention to compensate for negative consequences of automation regarding workers possibilities of maintaining and developing work skills. From the perspective of the research tradition in skill and technology, there is no general theory of knowledge. Each praxis has its own forms for skill development and has to be studied separately. Studying new forms for collaborative learning in a praxis therefore requires a deep understanding of work knowledge and skill development in this praxis. Therefore, the first research question investigated can be stated as: what characterises the practice of operator process control work in the plant studied?

      Since it is the tension between the »exact language« of automation and work practice that has led to the need for maintaining and developing work knowledge, the consequences of automation for the work practice in question becomes important. Therefore the second research question investigated can be stated as: what role does the automated process control system have in the work practice of the process operators? However, the time available in the case study has not allowed a deeper investigation of long term effects of computer use in line with the methodological approach in the research tradition of skill and technology. Instead the focus has been on identifying examples of the role of automation in process control work, particularly in connection with production disturbances, in order to point out what aspects of work knowledge that needs to be developed.

      The two research questions stated provide the necessary ground for the third and main research question: what interventions are needed to create possibilities for collaborative learning supported by computer based tools in the work practice studied, and what role can such tools have in the learning process? The computer based tool presented in the thesis was developed in line with the third strategy presented above for developing process control knowledge, allowing operators to share knowledge about how to handle production disturbances in an automated chemical plant. The goal has been to investigate new possibilities for collaborative learning in the work practice studied by designing a computer based tool for the construction and use of interactive problem scenarios in work practice, and to study such a tool in use. 

      When a computer based tool is introduced the way it will be appropriated and used in a praxis can never be anticipated and the question of whether a computer based tool can contribute to collaborative learning in a particular setting cannot be answered with a simple »yes« or »no«. Instead the process operators use of the tool for constructing and working with interactive problem scenarios has been studied and documented. The documentation has then been analysed in search for examples illustrating how the tool might support or inhibit collaborative learning.

      Finally, the use of interactive scenarios raises the question of how a scenario should be represented in interactive form. When Johannessen (1990, 1992) describes development of intransitive understanding through examples and imitation in a praxis, based on the language philosophy of Ludwig Wittgenstein, these examples are assumed to be presented by an experienced member of that praxis. The question of the degree to which interactive examples represented in interactive media can be understood is therefore crucial.

Work knowledge and learning

The epistemological perspective formulated in the thesis takes as its starting point Johannessens development of a conceptual framework for work knowledge based on Ludwig Wittgenstein’s notions of praxis and rule following (Johannessen 1990, 1992, 1997, 1999a, 1999b). This perspective is compared and combined with other theoretical perspectives on professional knowledge: Charles Taylor’s concept of praxis (Taylor 1971), Schön’s notion of the reflective practitioner (Schön 1987), Dreyfus’s perspective on expert knowledge development (Dreyfus 1990), situated learning theories (e.g. Lave & Wenger 1991), and Kuhn’s notion of seeing in a metaphorical sense as an important aspect of tacit knowledge in use (Kuhn 1962). Based on these theoretical viewpoints an epistemological perspective has been formulated that can be summarised as follows:

·         Knowing is inseparable from human activities and is therefore a contextual phenomenon rooted in a specific context for action – a praxis. Having knowledge means to act in a way that, given the particular circumstances, is perceived as appropriate by members of this praxis.

·         The fact that knowledge is rooted in a praxis means that there can be substantial differences between what is accepted as knowledge in different types of praxis, e.g. in everyday human activities compared to in a scientific discipline. There cannot be any general epistemology and every praxis has to be investigated in itself. The knowledge developed in a praxis is based on specific situations and does not claim to be applicable outside the praxis where it is developed and established.

·         Actions acquire their meaning through a shared understanding among members of a praxis. This way a praxis is built on intersubjective meaning, i.e. meaning that cannot be the property of one individual but is necessarily rooted in a social context which is constituted by this meaning. Individuals, as members of a practice, constitute a language collective for whom the shared meaning is built on a pragmatic constitutive base.

·         Knowledge is socially and historically constructed. The rule following of members of a praxis maintain the meaning of appropriate actions, and the field of action constituting a praxis has developed out of a history of specific situations.

·         New situations can extend the field of action in a praxis. Knowledge is inherently dynamic and each new situation has the potential to shift or expand knowledge in new directions by those being members of a praxis. Analogical thinking and judgement are important elements in these dynamics. They function as tools for constructing an understanding of a specific situation which opens up for new approaches and may lead to a solution to the problem, or to the development of concepts that describes the situation.

·         Being a member of a praxis involves developing a full understanding of the meaning that is shared in the language collective and the rule following comprising action in the praxis. Some aspects of knowledge cannot be articulated. What constitutes appropriate action or appropriate use of concepts in a praxis therefore cannot always be articulated in language, but must be demonstrated by examples. Developing the intransitive understanding of rule following, where rules cannot be articulated, therefore has to be accomplished through imitation.

Research approach

An explorative, qualitative research approach has been chosen where collected data have been inductively analysed and categories have not been preselected but rather allowed to emerge from the analysis. The chosen epistemological perspective in the study has formed a background for the analysis but I have tried to avoid letting this perspective dominate the analysis. Since the research is primarily action oriented I regard it as action science in the form described by Argyris & Schön (1996). They differ between first order effects of interventions – »single-loop learning« – where only strategies for handling the immediate situation are changed, and second order effects – »double-loop learning« – where also value systems governing a praxis are changed. Both types of effects are relevant in the study but due to time limits I have primarily focused on first order effects of the interventions.

      In the first part of the case study, investigating process operator work knowledge, the data collected for analysis was interviews, observations and documents in the organisation. The main focus in the analysis was eight group interviews on operator work and the handling of production disturbances carried out with the three work shifts in parallel with design of the computer based tool. Three interviews with each shift was planned but one was cancelled. All interviews were tape recorded. Before engaging in the design process of the computer based tool eleven of the fourteen operators were interviewed. The goal was to find out the knowledge needs considering the consequences of the ongoing automation of the production. The interviews were carried out together with the operator participating in the design process for the tool and were documented with note taking. Observations of operator work were done at nine occasions including two »backup drills« where production was interrupted and critical problem situations simulated at the plant with all operator shifts. Observations were documented with note taking. Interviews were also carried out with two newly employed operators in order to cover their perspective on operator work, and three members of production management to cover a management perspective. All interviews were recorded on tape. The data collection described above has been used in a qualitative analysis with the main goal of investigating two of the research questions addressing characteristics of the operator work and the role of automated process control in operator work practice. The main part of the data collection for investigating the last research question is nine meetings with operator work shifts where the TIPS-tool has been studied in use. Operators have constructed and used interactive problem scenarios as a basis for discussing and reflecting on praxis in handling production disturbances. All meetings were recorded on video for later analysis. This material has been the base for a qualitative analysis with the goal of investigating the last and main research question addressing the interventions needed to create possibilities for collaborative learning supported by computer based tools in the work practice studied, and the role such tools can have in continuous development of operator skills for handling production disturbances.

      Analysis of collected data has been carried out as a continuous comparative analysis based on »grounded theory« (Glaser & Strauss 1967). However, data collection has neither been iterative nor theory driven, and therefore this study could not be described as a grounded theory study. Analysis of data for the main research question, on interventions needed to create possibilities for computer supported collaborative learning, has involved interaction analysis (Jordan & Henderson 1994) with influence from conversation analysis (Atkinson 1984), since the focus has been social interaction among process operators in using the developed computer based tool.

Results

Operator work

The process control work studied is characterised by a substantial complexity but partly with other roots than those described in other studies of process control work. The transferability of generalisations about the complexity of operator work between different process control contexts therefore can be questioned. The complexity of the operator work contains factors rooted in the dynamics of the production process   internal complexity – and factors rooted in the dynamics of the production environment – external complexity. The root cause for the internal complexity lies in the variations in »temper« for the many different processes (around 100). Some processes are particularly reactive and requires a high level of attention from the operators. They develop an intransitive understanding, in the sense described by Johannessen (1999b), of the variations in temper which is a necessary requirement in operator skills for different reasons. Firstly for being able to control the process, secondly for handling the »rhythm« of the operator work, and finally for being able to handle production disturbances. The »temper« of the processes influences production planning, which operators increasingly have become responsible for as their work tasks have been extended. The aspect of operator skills needed for handling the complexity in production planning is described as knowledge of the »work rhythm«. It involves creating a reasonable individual work situation as well as creating a working rhythm in the production as a whole together with other operators. External complexity can be divided into four factors: lacking communication within the operator collective and between operators and other personnel such as production management, maintenance workers, etc; errors in process control software; malfunctions in the machinery; and environmental factors such as weather conditions, temperature of cooling water, etc.

      The apparent contrast between the dramatic situations of production disturbances and the routines of daily operator work at the plant that showed in the early observations soon turned out to be a false picture. Production disturbances of varying degree is a continuous aspect of operator work and variations in level of risk is a central part of work complexity. Even small incidents have the potential to develop into critical situations in short time under certain circumstances, which puts a high demand on the operators skills to handle such situations. The complexity of the work studied requires skills which to a large degree are tacit. Over time operators develop a deep knowledge of the production environment, and in particular the variations in »temper« that characterises the many different products. They also develop a judgement for handling problem situations in the plant as a whole rather than for the processes that they are responsible for, in particular for judging potential risks, possible future developments of a situation and consequences that follow. A continuous focus on the plant situation as a whole appears as a strategy to improve the ability to handle production disturbances, and also to maintain this particular aspect of the work knowledge. However, the ability to judge the level of risk in a production situation varies strongly depending on the amount of work experience.

Automation and operator work

Automation in process control leads to changes in the mediation of information between process and operator, but it is a change in the nature rather than the degree of mediation. In manual process control all process information that the operator perceives through his senses is mediated in different ways, often by the production machinery (e.g. when the operator looks at the characteristics of the foam produced on the surface in the reactor, or becomes aware of unfamiliar sounds in the machinery). Therefore it is misleading to describe the effects of automation as a layer introduced between the operator and the process, since the operators contact with the process might be mediated through several layers also in manual process control. Furthermore, the expressions of the automated process, as mediated through different sensors of the process control system, are not necessarily similar to the expressions that the operators has become familiar with in manual process control. Automation therefore primarily leads to a change in the nature of mediation of information between process and operator.

      The understanding of different expressions of the process in manual control work is intransitive and takes a long time to develop. The operators attention can be directed towards certain expressions through examples demonstrated by an experienced operator, but perceiving differences between variations of the same expression demands exposure to a large number of situations. In manual process control operators develop a »seeing«, in the metaphorical meaning of the concept as suggested by Kuhn (1962), where properties of the process are perceived through different stimuli and form the base for a development of an intransitive understanding of variations in »temper« for the many different products.

      Automation of process control forces the operator to develop a new kind of »seeing« in the metaphorical sense used above. This does not involve development of knowledge of a more intellectual nature as proposed by Zuboff (1988). Neither does it involve seeing through the digital representations of the computer based control system in order to re-establish the link to the concrete expressions of the process as described by Perby (1995). It is a matter of finding new paths to the expressions of the process. The operator has to develop a new intransitive understanding of the various expressions of the automated process based on the intransitive understanding of expressions of the process in manual control. There is no difference in nature between the operators work knowledge before and after automation. The central aspect of the knowledge is still an intransitive understanding of the expressions of the process as mediated through different representations, and a reflective judgement of the plant situation as a whole, with level of risk and potential future developments in a particular situation, based on this intransitive understanding.

      Finally, the study shows that automating the same type of machinery, in this case reactor systems with similar construction in principle, has different consequences for operator work depending on the »temper« of the process which in turn determines the level of automation that is feasible. In reactor systems where full automation has not been feasible automation relieves the operator of some tasks, but still requires the operator to keep attention on what the automated control system does. In these cases automated control cannot match the sensitivity in control demonstrated by an experienced operator in manual control. Therefore the automation takes the role of an »uneducatable apprentice«. The transferability of generalisations about the effects of automation between different process control contexts can therefore be questioned.

Designing the TIPS-tool

The intervention in the study consists of the participatory design of a computer based tool for collaborative learning. Inspired by John Carrolls notion of minimalist instruction (1990) an information format for describing production disturbances was developed which is referred to as thin interactive problem scenarios or TIPS. A strong element of user participation has been crucial for developing representation formats for problem scenarios that can be used in a collaborative learning setting. In local work contexts, language and jargon is developed to be an efficient means of communicating about work matters. By relying on the process operators highly developed context specific work knowledge, problem scenarios can be described using a minimum of information. The TIPS-tool allows operators to create interactive problem scenarios, based on experienced or plausible problem situations, that are staged in a hyper linked model of the production environment. The interactive problem scenario fixates a certain situation and by being confronted with the scenario repeated times, different ways of handling the situation can be explored. Suggested operations are embedded in the scenario resulting in feedback on actions taken. However, the practice inscribed in the scenario is not necessarily the best, but rather invites the users to reflect on different possible ways of handling the situation. It runs on an ordinary PC and has been introduced in the control room environment at the plant. The goal has been to establish the tool as a component in an environment for continuous development of operator skills.

Supporting collaborative learning in operator work

Computer based tools for learning based on different forms of simulation, e.g. of science phenomena, often reduces complexity in a knowledge domain by a reduction of the phenomena to basic causal relationships. However, industrial production environments are characterised by high complexity. Aspects of work knowledge that require the longest time to develop, and which are largely tacit, cannot be separated from the practice where they are applied. Thus, developing a computer support for developing work knowledge seems impossible without risking a decontextualisation of the work knowledge. At a first glance it does not seem at all fruitful to take interactive problem scenarios, based on simplified models of situations from the operators work practice, as a starting point.

   However, analysis of the tool in use has showed that the highly specific local language of a praxis can be used as a resource in constructing interactive problem scenarios for collaborative learning. It is the local language of the operators work practice that allows a link to the context of action to be maintained in the operators work with interactive problem scenarios. The work language is deeply rooted in operator praxis and this creates the necessary conditions for constructing a shared view of the problem situation based on the condensed description of an interactive problem scenario. This shared view of the problem stretches far beyond the actual contents of the description bringing out details in operator praxis, and has sufficient authenticity for elements of learning to emanate from discussion and reflection on the situation at hand. Thus, the communicative power of the local work language creates possibilities for sharing knowledge of work practice through the construction and use of interactive problem scenarios. This way operators can, at least in thought, be exposed to critical production disturbances more often than in daily work, but without the risks of the real situation. The value of naturalistic representations in computer supported learning, where a shared understanding of real world situations should be constructed, can therefore be questioned.

      The interactivity in the use of interactive problem scenarios, with direct feedback on actions performed, gives problem scenarios an open character. Each scenario can be played several times in order to explore different strategies in handling a problem situation. Analysis of the tool in use shows that construction and use of problem scenarios supports reflection and discussion with elements of collaborative learning in two forms: inquiry into the problem situation leading to a need for more information on how the control system or production machinery works; and argumentative discussions where the scenario indicates differences in operator praxis and needs for changes in the established praxis of handling production disturbances.

      The study also reveals clear differences in the interaction between operators in working with problem scenarios, depending on the amount of experience of the operator interacting with the tool. Since constructing a authentic understanding of a problem situation requires a deep intransitive understanding of operator praxis and the language used in describing scenarios, less experienced operators do not have the ability to work with interactive problem scenarios on their own. They need a strong support from experienced operators. The documented use of the tool shows that this support is weak. Inexperienced operators are excluded from discussions on problems, not only through language, but also through the lack of necessary support from experienced operators.

Future work

Using the language of a work practice as a resource in designing support for collaborative learning among experienced members of a practice seems to be a useful strategy. The study has raised several new questions for future research and the following areas are identified as relevant for further investigation:

·         Transfer of the concept to other contexts for collaborative learning. Results from the study have shown the potential in using the language of a practice as a resource in computer supported collaborative learning. This potential could be further investigated by testing the concept in other contexts, in particular settings where changes in work practice, e.g. through the introduction of new technology, affects the ability to maintain and develop work knowledge for members of the practice.

·         Obstacles in collaborative learning for less experienced members of a practice. A problem that has evolved in the study is that less experienced workers have not been able to take part in collaborative learning through construction and use of interactive problem scenarios due to their less developed work language. Further studies of factors inhibiting computer supported collaborative learning in the form described above are needed.

·         New forms of interventions supporting work language development. A relevant problem connected with the one above is the need for new forms of interventions supporting work language development as an integrated part of computer supported collaborative learning in the form described above. In this particular context better possibilities for less experienced workers to engage in collaborative learning can be explored in a number of ways: through construction of interactive problem scenarios in the information format described above but were scenarios are adapted to different levels of skill; through development of new information formats for interactive problem scenarios that are more suitable for less skilled workers; or through development of new activities where skilled workers take an active part in supporting less skilled workers to engage in peripheral participation in construction and use of interactive problem scenarios.

References

Argyris, C. & Schön, D. A. (1996) Organizational Learning II – Theory, Method and Practice. Addison-Wesley. Reading, Massachusetts.

Atkinson, J. M. & Heritage, J. (red.) (1984) Structures of Social Action – studies in conversational analysis. Cambridge University Press, Cambridge.

Bainbridge Lisanne (1978) The Process Controller. I Singleton W T (ed) The Analysis of Practical Skills. MPT Press Ltd, Edinburgh.

Bainbridge Lisanne (1987) Ironies of Automation. I Rasmussen Jens, Duncan Keith & Leplat Jacques (red) New Technology and Human Error, John Wiley & Sons, Chichester, s271-283.

Brehmer, B. (1993) Processoperatörernas arbete i moderna kontrollrum. I Lennerlöf, L. (red.) Människor Datateknik Arbetsliv, Publica, Stockholm, s 115-142.

Carroll, J. M. (1990) The Nurnberg Funnel – Designing Minimalist Instruction for Practical Computer Skill. MIT Press, Cambridge, MA.

Dreyfus, H. L. (1990) Is Socrates to Blame for Cognitivism? I Göranzon, B. & Florin, M. (red.) Artificial Intelligence, Culture and Language – on education and work. Springer Verlag, Berlin, s. 219-228.

Ellström, P-E., & Gustafsson, M. (1996) Lärande produktion – erfarenheter från ett utvecklingsprogram för processoperatörer. I Ellström, P-E., Gustavsson, B., & Larsson, S. (red.) Livslångt lärande. Studentlitteratur, Lund, s. 255-281.

Gergen, K. J. (1995) The Social Constructionist Movement in Modern Psychology. American Psychologist. Vol. 40, No. 3, s 266-275.

Glaser B G & Strauss A L: The Discovery of Grounded Theory: Strategies for Qualitative Research. Aldine de Gruyter, New York, 1967.

Johannessen, K. S. (1990) Rule-following and Intransitive Understanding. In Göranzon, B. and Florin, M. (eds.) Artificial Intelligence, Culture and Language – on Education and Work, Springer-Verlag, Berlin.

Johannessen, K. S. (1992) Rule-following, Intransitive Understanding and Tacit Knowledge – an Investigation of the Wittgensteinian Concept of Practice as Regards Tacit Knowing. In Göranzon, B. and Florin, M. (eds.) Skill and Education. Reflection and Experience. Springer-Verlag, London.

Johannessen, K. S. (1996) Tradisjoner og Skoler i Moderne Vitenskapsfilosofi. Sigma Forlag A.S, Bergen.

Johannessen, K. S. (1997) Tysta inslag i kunskap och förståelse - några moment i Wittgensteins diskussion av olika former för gränserfarenhet i senfilosofin. Dialoger 43/97, s. 45-53.

Johannessen, K. S. (1999a) Det Analogiska Tänkandet. Dialoger 50-51/99, s. 61-72.

Johannessen, K. S. (1999b) Praxis och Tyst Kunnande. Dialoger, Stockholm.

Jordan B & Henderson A. Interaction Analysis: Foundations and Practice. IRL Report No. IRL94-0027. Xerox Palo Alto Research Center and Institute for Research and Learning, Palo Alto, 1994.

Kuhn, T. S. (1962) The Structure of Scientific Revolutions. Tredje upplagan. The University of Chicago Press, Chicago, 1996.

Lave J., & Wenger, E. (1991) Situated Learning - Legitimate Peripheral Participation. Cambridge University Press, New York.

de Montmollin M & De Keyser V (1985) Expert logic vs. operator logic. I Johanssen, G., Mancini, G. & Maartensson, L. (eds.) Analysis, design and evaluation of man-machine systems. Oxford and New York, Pergamon press.

Orr, J. E. (1990) Sharing Knowledge, Celebrating Identity: Community Memory in a Service Culture. I Middleton, D. & Edwards, D. (eds.) Collective Remembering. SAGE Publications Ltd, London, s 169-189.

Perby, M. (1995) Konsten att bemästra en process – om att förvalta yrkeskunnande. Gidlunds förlag, Smedjebacken.

Sandberg, J. (1999) Konstruktioner av social konstruktionism. Paper presenterat i ’Workshop om konstruktionism – att tänka, forska och skriva konstruktionistiskt’, Växjö universitet 9-10 december.

Sandén P-O., & Johansson, G. (1987) Processövervakningens arbetsinnehåll och teknikförhållanden: konsekvenser för mental belastning och arbetsengagemang. Stockholms Universitet, Psykologiska institutionen, Rapporter, nr 52.

Schön, D. A. (1987) Educating the Reflective Practitioner. Jossey-Bass Publishers, San Fransisco.

Skorstad E (1988) Technology and overall control: an example from the process industry. I de Keyser, V., Qvale, T., Wilpert, B., & Ruiz, S. A. (eds.) The meaning of work and technological options. Chichester, Wiley, s 77-89.

Taylor, C. (1971) Interpretation and the Sciences of Man. Review of Metaphysics, 31, s 3-51.

Zuboff, S. (1988) In the Age of the Smart Machine - the Future of Work and Power. Heinemann Professional Publishing, Oxford.