Dash, D. P. (1999). Vocabulary of agency: Development and assessment of a generic conceptual framework to guide action-oriented research in multiple domains [PhD Thesis]. Lincoln, UK: University of Lincoln. Available online: http://www.ximb.ac.in/~dpdash/PhD.htm

Chapter 4

Family of Management Systems Literature

 4.1 Introduction to the Review

The family of action research literature reviewed in Chapter 3 does not exhaust the entire range of literature on how a research-orientation and an action-orientation might be combined in a systematic way. The present chapter reviews the family of management systems literature where the conceptualisation of action-oriented research seems to have taken a unique direction characterised by the variously interpreted notion of a system.

The label management systems refers to a growing body of academic literature using systems ideas (e.g., ideas of system—open and closed, boundary, elements and relations, complexity, communication, control, self-regulation, hierarchy, structure, emergent property, evolution, etc.) to study and address the problems of management in organisations and society. This literature is usually identified within broader category labels such as ‘systems approach’, ‘systems thinking’, ‘systems movement’, ‘systems research’, ‘systems practice’, ‘systems theory’, ‘systems and cybernetics’, etc. Some commentators have used the label ‘systems movement’ to refer to the broad intellectual landscape within which this literature has flourished (e.g., Checkland, 1981; Flood and Jackson, 1991a; Jackson, 1991; Klir, 1991; see Cybernetics and Systems Theory pages at http://pespmc1.vub.ac.be/CYBSYSTH.html).

The review will focus on the conceptual elements discussed in the literature which might be used later to build a general conceptual framework to guide action-oriented research. It is not the aim to describe a number of management systems approaches in detail covering the associated tools and techniques, their strengths and weaknesses, etc., but to focus on how an approach reconciles an action-orientation within research (or even a research-orientation within action). Special attention will be paid to ideas which address the difficulties of action research discussed in the previous chapter.

 4.2 Systems Movement

There seem to be many different interpretations of how and why the systems movement has emerged. It has been linked with an awareness about the need to study the interconnectedness of phenomena and perhaps also of the mechanisms that generate the phenomena (e.g., Ackoff, 1981; Capra, 1982; Churchman, 1979; Flood and Jackson, 1991a). It has been linked with the need to study complex integrated wholes having ‘emergent properties’—properties not present in the constituent parts but which emerge from their interaction (e.g., Beer, 1979; Capra, 1997; Checkland, 1981). It has been interpreted as a ‘will to systems’ within a process of transformation of ‘epochal ontological orders’ (Fuenmayor, 1997). It has also been interpreted as involving a generalisation on the conventional notion of scientific ‘object’, in terms of ‘system’, amounting to a possible extension of the scope and methods of science (de Zeeuw, 1995).

The movement is generally traced back to the 1940s (although it is known that the Russian thinker A. Bogdanov had pre-empted some of the systems ideas between 1912 and 1917) (Capra, 1997, Chapter 3). Early work on systems was done in two somewhat distinguishable traditions of inquiry, namely general systems and cybernetics. One of the core notions of the general systems tradition was that of an open system (von Bertalanffy, 1950; also see Systems Theory at http://www.ccens.com/systhe.htm). Similarly, one of the core notions of the cybernetics tradition was that of control (Ashby, 1964; Wiener, 1961; also see Homepage of W. Ross Ashby at http://www.gwu.edu/~asc/biographies/ashby/ashby.html; Wiener: Ideas at http://www.well.com/user/mmcadams/wiener.html). The notions of homeostasis and self-regulation were discussed in both the traditions, as the notions were found to be of potentially more general (and perhaps, inter-disciplinary) interest than commonly recognised among researchers at that time.

Both the traditions seemed to evince an interest in studying a general set of problems of research obtaining in a wide range of disciplinary fields. The group of scientists who had formed the International Society for General Systems Research in 1954 (see International Society for the System Sciences pages at http://www.isss.org/) utilised the open system notion in studying a wide range of phenomena, including biological, organisational, and social phenomena (Emery and Trist, 1965; Katz and Kahn, 1978). The defining characteristics of open systems (namely importation of energy, throughput, output, cycles of events, negative entropy, information input, steady state or dynamic homeostasis, differentiation, integration and co-ordination, and equifinality) were instantiated in separate areas of inquiry. It seemed as if the language of general systems was capable of unifying knowledge from different disciplines and fostering transfer of knowledge among them. It seemed to provide a way of abstracting domain-independent knowledge from a set of domain-specific knowledges. A major product of this type of approach was the conceptualisation of ‘living systems’ (Miller, 1978). This attempted to abstract the defining characteristics of living systems from seven different domains (namely the domains of cell, organ, organism, group, organisation, society, and supra-national system). This type of activity within the systems movement has sustained an interest in studying general systemic principles manifest in different phenomenal domains, and applying these principles in finding and rectifying dysfunctions that might reduce a system's effectiveness in achieving its purposes (see, e.g., Applications of Living Systems Theory at http://www.newciv.org/ISSS_Primer/asem05jm.html). It was generally expected that this type of strategy should result in the development of a corpus of knowledge (e.g., about living systems) and an increased ability to diagnose and improve, or design, systems in different domains.

The research perspective involved here might be compared with the more traditional research perspective. For example, the characteristics of living systems might be seen as part of the natural order of things; once discovered these can be used for various practical purposes. These characteristics do not seem to depend upon the values (or the actions) of either the researcher or the user of the result. Such a system, even if artificially designed, might thereafter behave as if it were a part of the natural order of things, although it might require some kind of maintenance effort.

That some systems could pursue their own values (goals or purposes) was a crucial issue within the tradition of cybernetics (Ashby, 1964; Wiener, 1961). Early cyberneticians had shown that apparent goal-seeking behaviour might arise out of a feedback structure, e.g., that of an ‘error-controlled regulator’. Such a structure would involve a ‘black box’ which transforms inputs into outputs, a ‘controller’ that compares the outputs with a ‘target’ parameter and computes (using a model of the ‘black box’) how the inputs should be altered in order to reduce the error between the output and the target, and supplies that input to the ‘black box’. Ashby and Wiener had highlighted the generality of the principles of cybernetics in studying goal-seeking behaviour in a wide range of phenomena—‘in the animal and the machine’.

Later on, the language of cybernetics was used to study patterns of regulatory behaviour in many domains including organisational and social domains (e.g., Bateson, 1979; Beer, 1979; 1985). Over the years, there has been a spectacular expansion of the field of cybernetics and applications have been found in a wide range of areas, e.g., robotics, computing, bio-medical applications, environmental studies, management, social planning and administration, economics, international relations, etc., accounts of which have been reported in various systems and cybernetics journals (see Cybernetics and Systems Journals page at http://pespmc1.vub.ac.be/JOURNALS.html). The interest of cybernetics in studying the various implications of self-regulation have been manifest in these applications.

A cybernetically self-regulating system might or might not be a part of the natural order of things. It might have to be designed, for example. But, once a self-regulating system is designed (or brought forth into the world) it might thereafter be seen as part of the natural order of things, and therefore, could be observed as such by an outside observer. Of course, such a system might also require some kind of maintenance effort. Applications and explorations of cybernetic thinking in certain domains (e.g., human knowing, co-ordination systems, organisational design, social administration, etc.) have indicated the need for conceptualising alternative forms of interaction between an observer and a system. In one such form of interaction the system might be capable of observing itself and/or observing the outside observer, i.e., it might be an observing system (von Foerster, 1984). The contemporary discussion on second-order cybernetics pays attention to this type of system and interaction among such systems (Cariani, 1993; von Foerster, 1984; Glanville, 1993; Reyes, 1995, Chapter 3; Umpleby, 1990; 1991; 1994; van de Vijver, 1992; de Zeeuw, 1993a; also see Cybernetics & Human Knowing: A Journal of Second Order Cybernetics & Cyber-Semiotics at http://www.db.dk/dbaa/sbr/cyber.htm). The issue of whether anything like a part of the natural order of things emerges from such an interaction has been discussed in this area.

In both the traditions of the systems movement (i.e., general systems and cybernetics) discussed above, there appears to be an interest in either discovering or postulating (and designing) a system with stable characteristics such that it displays (or acquires) a status equivalent to the natural order of things. This could be interpreted as their ‘scientific’ orientation. That might be the reason why the systems approaches needed no great justification to be applied to numerous practical situations requiring some kind of improvement. There have emerged organisations specialising in providing consultancy and training in applying management systems approaches with a view to improve organisations and communities (see, e.g., Phrontis Limited at http://www.phrontis.com/). However, applied systems work has brought in issues which have necessitated further elaborations of systems thinking. That is what one notices in the contemporary literature of management systems. This literature has revealed certain issues in using systems notions to achieve local outcomes (valued by some client). Whether and how such applied work might produce some kind of research outcome as well will be the focus in the present review. Some of these approaches that present a more or less contemporary picture of management systems thinking will be taken up for review in the next section. The main aim will be to explore further how applied systems work inter-relates action-oriented and research-oriented interests together.

 4.3 Management Systems Approaches

4.3.1 Operational Research/Management Science

Operational Research/Management Science (OR/MS or OR) consists of an ever growing set of methods and tools to effect improvements in the conduct and co-ordination of operations, typically within organisations. This interest is borne out by the military roots of OR, the interests of the professional societies promoting OR (starting with the OR Society in the UK, founded in succession to the Operational Research Club which was set up in 1948, see http://www.orsoc.org.uk/ors/index_f.html), and applications of OR in private and public sector organisations in many countries of the world (see OR/MS Societies and Regional Groupings at http://www.maths.mu.oz.au/~worms/soc/soc.html).

There were many precursors to OR/MS in the arena of systematic support to the activity of co-ordinating organised work. The double entry system of book-keeping developed by the Italian mathematician and friar Luca Pacioli in 1494; Charles Babbage’s work on industrial productivity in the UK during the 1830s, the scientific management approach of Frederick Taylor in the US since the 1890s, and time and motion study (especially the work carried out for the cotton textile industry in the UK by the Shirley Research Institute from the 1920s onwards) could be mentioned among them (Duckworth, 1962, p.12; Keys, 1991; Kirby and Capey, 1998; see http://www.arthurandersen.com/ for the history of accounting).

A significant amount of work has been (and is being) done to review the origins of OR and its developments before, through, and after the period of the second world war (e.g., Keys, 1991; Cook and Shutler, 1991; Ranyard, 1995; Kirby and Capey, 1998). These as well as text books and guide books on OR (e.g., Duckworth, 1962; Ackoff and Sasieni, 1968; Budnick, et al., 1988; Hillier and Lieberman, 1990) have characterised OR by its emphasis on modelling the real situation in order to identify optimal states/levels of an operation.

It is the purpose here to consider how OR adapts the task of science/research to make itself useful in an operational situation. OR workers have tried to give various interpretations to the term ‘research’ in OR. Daellenbach, et al. (1983) have stated: "It is not scientific research, with its connotation of advancing fundamental knowledge in some science. … The objective is to improve the effectiveness of the system as a whole" (p-1). Hillier and Lieberman (1990) have argued that OR involves creative scientific research into the fundamental properties of operations. They have used the formulation of Ackoff (1956) to argue that the implementation of an OR solution might itself be a test of the hypothesis that the conclusions (solutions) obtained from the model are also valid for the real problem.

OR texts have emphasised the importance of carefully observing the real problem situation (e.g., Hillier and Lieberman, 1990). There does not appear to be much debate within the early OR literature regarding what constitutes data in an OR study although disquieting concerns have been expressed (e.g., Churchman, 1970). Statements obtained from primary and secondary sources within the client organisation (Graham, 1984), and perhaps also from outside, seem to constitute data for OR modelling. One of the characteristic features of OR is the way these data are used. The data are used for the purpose of modelling. An OR model is meant to be a ‘model of the system’ (e.g., Duckworth, 1962, p.19); it is supposed to ‘represent the system’ (e.g., Ackoff, 1956); and it is usually a ‘formal model’ (Pidd, 1991). The information created out of the model is typically called a ‘solution’ in the idiom of OR. The ‘solution’ of an OR model refers to a specific state of the system, or a specific set of outputs of the model given a specific set of inputs. It is that state (or that set of inputs and outputs) which is supposed to be of special interest to the client organisation—the state that the client organisation might not have identified without the contribution of the OR study.

It is sometimes expected that the activity of OR will result in the generation of knowledge regarding ‘the fundamental properties of operations’ (the expression is from Hillier and Lieberman, 1990, p-5). There is something like a ‘body of knowledge’ in OR which is transferred in professional training, and which continually grows by encompassing new formal models of operations, modelling and solution strategies, as well as consulting skills. OR has strongly emerged as a profession in the UK and elsewhere, with all its ramifications, namely membership, training, manifesto, institutional structure, ‘house journal’, etc. (Churchman, 1970; Kirby and Capey, 1998; Ranyard, 1995).

Churchman (1970; 1979) has highlighted what he labels as the ‘irrational’ side of the profession, mentioning the problems of ‘observing a system’, ‘peculiar mixture of attitudes about data’, problem of identifying the ‘whole relevant system’, ‘deeper problem of morality’, opposition to ‘political power structures’, and inflexibility associated with institutionalisation of the profession. Ackoff (1979) has highlighted various inadequacies in the overall scheme of OR, pointing out the difficulties of ensuring the reliability of data, simplifications involved in modelling, short life span of optimal solutions, omission of aesthetics, OR’s ‘pretension’ of interdisciplinarity, and its failure to take all stakeholders into account. That everything might not be going well with the profession of OR, can be gleaned from some of the recent studies concerning the state of the profession (e.g., Fildes and Ranyard, 1997; Ranyard, et al., 1997; Ranyard and Fildes, 1998). Discussions about the characteristics of OR have appeared consistently in the literature for over more than two decades now (e.g., Dando, et al., 1977; Jackson, et al., 1989; Jackson and Keys, 1987; Keys, 1991; 1997; Rosenhead, 1989; White and Taket, 1996.) One of the outcomes of these discussions about the nature of OR has been the emergence of what is now labelled as ‘Soft OR’ (Rosenhead, 1989), which retains the professional model of OR but introduces new features into the methodology.

The following issues might be noted here from the point of view of this review. OR is interested in action, in the sense that it is client-oriented. The objectives of the client determine (though not entirely) what system is to be modelled, what data are to be collected for the purpose, and finally what state of the system is to be identified as the most desirable given the client’s objectives. OR is also interested in research, at least in two different senses. Firstly, at the level of the local, OR is interested in identifying the particular state of the chosen system that would be the most satisfactory in fulfilling the client’s objectives. Secondly, OR is also interested in strengthening its body of knowledge, i.e., its library of formal models, solution procedures, and its general vocabulary, in an ongoing way. OR/MS certainly appears to offer an approach for action-oriented research—one that highlights the role of formal models in achieving client-oriented results as well as results of wider professional interest.

4.3.2 Soft Operational Research

As mentioned before, Soft Operational Research (Soft OR) emerged as a response to deal with certain problems with the OR/MS approach (Rosenhead, 1989). It was noticed during the 1970s that the traditional OR approach seemed to work well under certain conditions; e.g., it worked where,

  • the client organisation was structured in a tight hierarchy;
  • few of its members were analytically sophisticated;
  • the organisation performed a well-defined task, thus generating reliable data; and
  • there was general consensus on priorities (Rosenhead, 1996).

All these point to the difficulties associated with OR practice. There is the problem of identifying what constitute legitimate data in certain situations (e.g., involving ‘ill-structured problems’, ‘wicked problems’, or ‘messes’) (Simon, 1973). Similarly, Soft OR casts doubts on the possibility of a single uncontested representation of a problem situation, thus putting the classical OR notion of modelling in jeopardy. Traditionally OR had served clients like defence organisations, private and public sector organisations, etc., which were formal organisations with rather well-defined structures of roles and relationships. When OR was taken to non-traditional clients such as ghettos, youth clubs, housing co-operatives, community service organisations, etc., there was a difficulty in applying the methods of OR (Rosenhead and White, 1996). Even within formal organisations, sometimes ‘ill-structured’ problem situations (e.g., strategic problems) seem to emerge that make it difficult to apply traditional OR techniques (Rosenhead, 1996).

Soft OR seeks to resolve this difficulty by introducing something else into the situation that might make the situation again like the traditional, whereupon traditional OR might be applied. What is typically introduced is problem structuring. It has been argued that sometimes problem structuring (or ‘issue structuring’) might itself be an adequate support to the client, implying that it might not be followed up with more conventional OR modelling.

Different kinds of ‘problem structuring methods’ (PSMs) have been proposed in Soft OR (Rosenhead 1989; 1996). Some of the methods presented in the literature are: Hypergame Analysis, Interactive Planning, Metagame Analysis, Robustness Analysis, Soft Systems Methodology (SSM), Strategic Assumption Surfacing and Testing (SAST), Strategic Choice Approach (SCA), and Strategic Options Development and Analysis (SODA). From the variety of these methods, it would seem that Soft OR might be a broad and general approach to tackle ill-structured problem situations. Therefore, it might be more appropriate for this review to discuss Soft OR in general terms rather than focusing on the specific methods. (One of the methods, i.e., SSM, is reviewed separately, in Sub-section 4.3.7, as it seems to have been given special attention in the literature, independent of its association with Soft OR.)

A study of these ‘problem structuring methods’ indicates some of their general characteristics (Rosenhead 1989; 1996). These methods visualise a situation in which there is usually a ‘facilitator’ (or a team of ‘facilitators’) and a client group—typically involving multiple actors, having multiple or unclear priorities, multiple perceptions and uncertainties about the nature of the problems to be addressed, and multiple viewpoints and uncertainties about what would amount to a satisfactory solution. Soft OR methods seem to intervene in this situation with an aim to improve interaction and co-operation among actors, reduce differences in terms of perceptions, preferences, priorities, etc., guide constructive debate among the actors, develop images of desirable future for all the actors involved, orient the actors towards some desirable direction for action, etc. In this effort, different methods seem to contribute differently. Hypergame Analysis focuses on a better appreciation of the differences among the actors. Interactive Planning helps in developing images of the ideal future. Metagame Analysis structures the interaction among the actors such that there is some coherence in terms of preferences. SSM also structures the interaction among participating actors in order to generate and guide constructive debate especially about feasible and desirable changes in the situation. SAST provides a method for guiding constructive debate about strategies and assumptions with the aim of reducing the differences by mutual adjustments. SCA helps participants identify priority areas by exploring the interconnectedness among decisions and enables them to develop action plans. SODA uses the device of cognitive mapping to help participants externalise their perceptions about the problem situation and develop constructive debate about actions to be taken. Robustness Analysis helps the actors maintain useful flexibility while planning for the future despite uncertainties and changing environments.

The approach of Soft OR has been compared with the ‘process consultancy’ approach used within the Organisational Development (OD) literature (Rosenhead, 1989, p.348). Both Soft OR and process consultancy recognise and emphasise the existing experience and knowledge of the participants as a resource in transforming the group, situation, or organisation. The general strategy used by both approaches is to promote interaction among actors, socialise and purposefully organise the knowledge previously fragmented among the participants, promote thinking and dialogue, and generate commitment towards action. In order to highlight the distinctiveness of Soft OR methods in comparison with OD, Rosenhead has argued the following:

Each member of the PSM family incorporates as a core element the explicit modelling of cause-effect relationships. This gives PSMs their unambiguous operational research identity. It distinguishes them, for example, from non-OR modes of group working, such as organisational development (Rosenhead, 1996, p.120).

It might be worthwhile to notice that the type of ‘explicit modelling of cause-effect relationships’ in the more conventional OR is expected to produce models that ‘represent the system’, but Soft OR seems to produce a different type of model. For example, the advocates of SODA have clarified the nature of the models they are concerned with:

… we were concerned with finding means of understanding and representing explicitly the way individuals perceive their world, that might be helpful to them in permitting a self-conscious reflective dialogue with their own thinking (Eden, et al., 1979, p.104).

Interest in Soft OR has revived an earlier interest in exploring the relationships between OR and the social sciences (see Journal of the Operational Research Society, 44, 6, Special Issue on the Interface between OR and the Social Sciences, 1993). One of the aims of this exploration is to identify the conditions for successful interventions, so as to sustain ‘systematic improvement’ in the field of operational research (Eden, et al., 1993, p.532). Among the various areas of the social sciences, the study of ‘facilitated work groups’ appears to be of particular relevance to Soft OR (Phillips and Phillips, 1993). Following the insight in this area of research, the facilitator seems to have a more active role than a conventional researcher. The facilitator becomes a participant in the group that s/he is facilitating and plays an important role in influencing the process so as to make the group work effectively.

From this, it would seem that Soft OR might be action-oriented in two different senses. Firstly, Soft OR is client-oriented, as is traditional OR. The outcome of a Soft OR exercise is expected to be useful locally, to the satisfaction of the client body. But, there might be a second sense in which Soft OR is also action-oriented. The Soft OR facilitator is not limited to the somewhat passive role of collecting data and producing representational models. The facilitator has to actively influence the group process in order to make the group become more like a collective capable of acting in solidarity. In every instance of a Soft OR application, there is the expectation of a local outcome that would be valued by the clients, e.g., a structured issue, a decision, a plan, a strategy, etc. Additionally, there is also the potential of an expansion in the tools and techniques for group facilitation, a developing understanding of the conditions of success of the Soft OR approach, and a progressive elaboration of the general vocabulary used in Soft OR type of work, in an ongoing way. As a possible approach to action-oriented research, Soft OR seems to highlight the importance of the researcher’s action in actively creating a collective capable of acting in solidarity.

4.3.3 Systems Analysis and Systems Engineering

The term ‘Systems Engineering’ is said to have become current in the 1950s (Checkland, 1981; 1989), although quite complex technological systems had been built by people over the ages (e.g., ships, irrigation systems, urban settlements, etc., in ancient civilisations, and later, railway systems, power generation and transmission systems, telephone networks, etc.) Systems Engineering strives to conceptualise in a systematic fashion the general pattern of conceiving, designing, and implementing systems of this type. The process generally starts with the specification of the purpose to be served by the object or system. Systems Engineering then works back from these specifications to design a system that would achieve these specifications using minimum resources, in an efficient and effective manner, without too many negative side effects.

Jenkins (1969) has described the Systems Engineering approach in detail. The notion of ‘system’ seems important in his account. In this scheme of things, a system could be designed to achieve an overall objective. Systems could have multiple objectives which might be mutually conflicting in some cases, necessitating certain compromises. A system would typically form part of a hierarchy of systems, in which there might be interaction between systems at the same level and at different levels. The functioning of a system would depend on these interactions.

The notion of ‘engineering’, in Systems Engineering, refers to designing, constructing, ‘knitting together’, and operating ‘works of public utility’. That makes Systems Engineering the activity of conceiving, designing, constructing, ‘knitting together’, checking, and operating individual systems so that jointly they perform efficiently and effectively as an overall system in pursuit of the prescribed objective.

The role of a systems engineer has been compared with that of a general medical practitioner who is supposed to look after the general health of patients, but might seek specialist medical knowledge from time to time (Jenkins, 1969).

The distinctiveness of Systems Engineering, over engineering per se, seems to lie in the broader view it takes in dealing with questions of operational efficiency and effectiveness. It seems to be focused on optimising overall criteria, associated with the functioning of a system. Another distinctive feature of Systems Engineering is that the approach might be extended to apply over a range of different ‘systems’, from technological to organisational and societal systems.

The process of Systems Engineering can be visualised in terms of four stages, although these might not necessarily be sequential (Jenkins, 1969):

(i) Systems Analysis: This stage involves a definition of the system to be engineered. It involves identification of various sub-systems and their interactions, and the wider system which contains the system being engineered. The objectives of the wider system clarifies the role and responsibility of the system being engineered. This helps clarify the objectives of the system being engineered as well as performance criteria for the system. If objectives cannot be clearly defined in the beginning, analysis can start with a working definition which might be reconsidered later if possible.

(ii) Systems Design: This stage involves building a model of the system that is being engineered. The model is built for the purpose of designing actions that will optimise the system. The model is used to compute the values of the performance criteria under different modes of operation, and an optimum solution is sought. A control system is designed to maintain optimal operation of the system under a varied range of conditions, so that the overall reliability of the system’s operation would remain at a desirable level.

(iii) Implementation: Construction of the designed system is undertaken at this stage. It is necessary to ensure that system builders understand the design in order to deal with uncertainties that might arise during implementation. Various kinds of project scheduling support might be provided to the system builders at this stage.

(iv) Operation: At this stage, various kinds of support are provided to the users of the system. Documentation about the system, user manuals, training, etc., are usually provided. Performance of the system is assessed from time to time. Re-optimisation might be needed after a period of time. Continued improvement of the system might be needed in a changing environment.

The sort of thinking embodied by Systems Engineering was also incorporated by a whole group of approaches for systematically achieving prescribed aims in an efficient and effective manner. Many of these approaches were developed during the 1950s and 1960s and were used in areas of engineering, planning and management, public administration, etc. ‘Systems Analysis’ refers to a class of these approaches said to have been initiated by the RAND Corporation although later developed by many different groups (Checkland, 1981; 1989; Miser, 1995; Miser and Quade, 1985; 1988; see Classic RAND Research page at http://www.rgs.edu/Classic.html). Quite like Systems Engineering, Systems Analysis also starts with the statement of an objective or a set of objectives. The steps followed typically involve the identification of alternative systems for achieving the objectives, estimation of the resources required by each of these alternative systems, use of models to assess the efficiency and effectiveness of these alternatives, and choice of one alternative on the basis of some preference criteria.

Both Systems Analysis and Systems Engineering provide a general way of thinking about practical situations requiring the creation of a system that achieves some pre-defined purpose. This highlights their action-orien]\=tation. The research-orientation might be seen in their use of systems and cybernetics ideas (e.g., system, sub-system, environment, self-regulation, control, etc.). Research results might be expected with respect to the general characteristics of this particular class of systems, the methods of their optimisation, the conditions of their smooth operation, etc. One of the research results to be taken up later in this review (in Sub-section 4.3.7) pertains to the need for extending the research vocabulary associated with Systems Analysis and Systems Engineering to make it more suitable for applications dealing with managerial systems in organisational and social situations (Checkland, 1981, Chapter 5; 1989). As a potential model for action-oriented research, Systems Analysis and Systems Engineering highlight the importance of following a particular image of systems (i.e., those which can be designed, built, and operated to accomplish specified objectives) in conducting practical projects.

4.3.4 Organisational Cybernetics

Organisational Cybernetics has emerged as a management systems approach by adopting cybernetic notions for the purpose of organisational and social management, especially by focusing on the Viable System Model (VSM) postulated by S. Beer (Beer, 1979; 1985; Espejo and Harnden, 1989; Espejo, et al., 1996; Flood and Jackson, 1991a; Jackson, 1991). Besides the VSM, there seems to be a wider discourse on social and organisational cybernetics, around the notions of sociocybernetics, autonomy, autopoiesis or self-production, self-organisation, second-order cybernetics, observing systems, etc. (Achterbergh, et al., 1997; Geyer and van der Zouwen, 1986; Jantsch, 1980; Willke, 1990; see The Challenge of Sociocybernetics at http://pespmc1.vub.ac.be/Einmag_Abstr/FGeyer.html and http://www.unizar.es/sociocybernetics/indice.html; Autopoiesis Related Web Sites at http://server.snni.com/~palmer/autopoiesis/websites.htm). However, from the point of view of management systems, the VSM still seems to remain an important instrument in the operationalisation of the organisational cybernetics approach and has produced a substantial body of management systems literature (the journal Systems Practice devoted a Special Issue to this topic in 1990, Volume 3, Number 3). Besides, the current literature related to the VSM seems to reflect some of the ideas of the wider cybernetics discourse (e.g., Harnden, 1990). This review will primarily focus on the VSM-related literature in organisational cybernetics.

Beer has defined cybernetics as ‘the science of effective organization’ (Beer, 1975, p.425). In Beer’s version of organisational cybernetics, the role of scientific object is played by the notion of ‘viable system’, the postulated archetype of effective organisation. A system is said to be viable if it is able to adapt effectively to environmental changes although the changes are unforeseen. The background research and thinking behind such a notion has been presented in the literature (Beer, 1979; 1984). The VSM might be taken as an elaborate and formal description of the necessary components of a viable system, starting with the notion of an operational element (said to be a basic potentially viable system) consisting of an operation (O), in which is embedded a management unit (M), an environment (E) in which both are embedded, and a process of variety engineering between O and M, and O and E (Beer, 1979, Chapter 4). Using a notion of recursion, the model visualises that a viable system might include a collection of operational elements—together referred to in the model as System 1. A management unit is visualised for the entire System 1 (i.e., the collection), containing Systems 2, 3, 4, and 5, each having specified roles and orientations. This management unit, containing Systems 2, 3, 4, and 5, would look after the co-ordination of operational units (System 2), control and regulation of the affairs of the System 1 giving it a here-and-now orientation (System 3), interaction with the outside environment of the System 1 giving it a future orientation (System 4), and dealing with issues not dealt with by any of these, e.g., steering a balance between the here-and-now and the future (System 5). These systems and their interactions are elaborated within the VSM, including the necessary communication channels operating between them (Beer, 1979; 1985).

Many applications of the model have been reported (e.g., Espejo and Harnden, 1989). The most common type of application of the model seems to be in diagnosis (e.g., to assess whether a given organisation contains all the elements of the VSM or not). The model has also been applied in designing information systems (e.g., Murthy, 1994) and designing a management structure for complex computer networks (Latin, 1991). The applications have also included studies in which the role, mission, and structure of entire organisations have been clarified using the model (e.g., Flood and Zambuni, 1990; Leonard, 1989). Viable systems thinking has also yielded useful insights for organisational theory, e.g., in reinterpreting the notion of organisational effectiveness (Schwaninger, 1990).

In the recent literature of organisational cybernetics, there seems to be an exploration of alternative uses of the model. There seems to be a new emphasis on the use of the VSM vocabulary in a conversational process within an organisation (Harnden, 1990) and its use as a ‘tool for self-constructing organizations’ (Espejo, 1996). This exploration is expected to indicate alternative ways of translating the insights of VSM into organisational contexts, rather than imposing the model on the organisational members.

This area of organisational cybernetics seems to offer many insights for a richer comprehension of action-oriented research. Its reliance on the field of cybernetics is one of its strengths. The manner in which the basic cybernetic notion of a ‘self-regulating system’ is extended and generalised in terms of a ‘viable system’ is noteworthy. This does highlight the possibility of deriving a formal model (i.e., the VSM) based on a general notion (i.e., ‘viable system’) and the possibility of instantiating the formal model in material terms in order to realise local results for specific clients. Obviously, the process might not be as straightforward as it sounds. There could be many uncertainties, speculations, and unknowns involved in it. The literature also indicates the possibility of the general notion becoming something like an imposition in a local situation (Jackson, 1989). The response to this danger in terms of ‘conversation’ and ‘self-construction’ seems to be a potentially valuable line of thinking for action-oriented research. Although the literature of action research has consistently emphasised ‘participation’, recent organisational cybernetics literature seems to highlight the need for participation towards self-constructing something that might be considered as an instance of a general notion—e.g., a ‘viable system’. The literature of organisational cybernetics proffers a model for action-oriented research which involves the notion of a formal model, the process of instantiating it in material terms, and an interest in a systematic improvement of both.

4.3.5 Socio-Technical Systems Thinking

Herbst (1974) traces the development of Socio-Technical Systems thinking to the coal mining studies carried out in the 1950s in the UK. These studies found that mechanical mining had resulted in the break-up of the small self-regulating work teams in the mines. The new mechanical system compelled the miner to concentrate on a single task. The individual miners’ activities were integrated by the mechanical system. Although the performance of each miner was dependent on that of the whole shift, the performance-related rewards were based on piece rates. The studies concluded that the new organisation of work was leading to psycho-somatic disorders, interpersonal and inter-group conflicts, often accompanied by mutual scapegoating, absenteeism and, overall, a low performance.

Trist and Bamforth (see Herbst, 1974, Chapter 1), who had reported the first coal mining study in 1951, had interpreted the situation through a model that introduced the basic notion of a socio-technical system. According to them, the quality and the quantity of output in any work organisation depended on two distinguishable aspects of the organisation: the technological system and the social structure of the work system. They viewed the psycho-somatic disorders, and interpersonal and inter-group conflicts observed in the organisation as a consequence of the social structure of the work system. They hypothesised that the social structure of the work system was partly influenced by the technological system. This led them to formulate the notion of joint optimisation, i.e., optimising the performance of the work organisation by modifying both the technical and social systems in a suitable way. The notion of organisational choice began to emerge after a second series of coal mining studies carried out by Trist and his collaborators. The notion highlighted the degree of choice available to an organisation with respect to the type of form it could take, for a given production technology. Later in the 1960s, especially after the socio-technical thinkers became involved with the Industrial Democracy Project of Norway (Thorsrud, 1977), the notion of technological choice began to be discussed.

The notion of technological choice suggested that the production technology need not be taken as a given for any organisation; it could be chosen from the point of view of fulfilling certain objectives other than purely technological. This involved the notion of the ‘social ecology of industry’. The basic search seemed to be for a way of jointly optimising the technological and the social systems of work organisations such that, locally, the work organisation performs satisfactorily without creating human costs in the form of psycho-somatic disorders, and interpersonal and inter-group conflicts, etc., and globally, it leads to a ‘social ecology of industry’ that supports such jointly optimal work organisations.

This type of focus led the socio-technical thinkers to realise that organisations might be viewed as open systems which had to survive in different types of environment, some of which might not be passive or undifferentiated (Emery and Trist, 1965). They have observed cases which illustrated ‘what is meant by the environment becoming organised at the social level’. Describing the plight of a vegetable canning company that had invested a very high amount of money in a new factory only to find that its market share was dropping quickly while the factory was being built, Emery and Trist have argued:

They [the managers] failed entirely to appreciate that a number of outside events were becoming connected with each other in a way that was leading up to irreversible general change (Emery and Trist, 1965, p.24).

Following this line of thinking, Emery and Trist have proposed four ideal types of organisational environment, namely placid randomised, placid clustered, disturbed reactive, and turbulent environments. They have argued that the survival of an organisation would depend upon the compatibility between its strategy and the type of environment it confronts. In the latter three types of environment, the survival of an organisation would depend upon what other organisations in its environment do. In the placid clustered situation, survival becomes critically linked with what an organisation knows about its environment. But, this strategy does not work in a disturbed reactive situation, where what one knows can also be potentially known by the others in the environment. Therefore, the struggle for survival would include calculating reactions of the others, and planning counter-actions, in a game-like affair. In turbulent environments, ‘the "ground" is in motion’, where organisational actions and interactions might lead off in unpredictable ways, and where actions and their results might not remain linked in clear patterns. Emery and Trist have given an interesting proposal to deal with the vagaries of turbulent environments. According to them, a possible solution to this is the emergence of values that have overriding significance for all members of the field. In this they have used the Lewinian notion of ‘power fields’. It is the emergence of such fields which would make the environment ‘simplified and relatively static’ again.

The development of thought on Socio-Technical Systems suggests some fundamental insights concerning action-oriented research. Different types of action-orientation are involved here. Like the other management systems approaches, there is a client-orientation. Positive local outcomes are expected from proper organisational or technological choice. It should, for instance, improve performance, reduce human costs, etc. But, insights in this literature indicate how and when such local improvements cannot be sustained, e.g., in reactive and turbulent environments. Socio-Technical Systems thinking suggests that research should contribute not merely through local improvements, but also by indicating how to deal with the situation where attempts at local improvements do not seem to have the desired effect. One of the proposals is to identify methods of interaction among organisations such that something like a ‘power field’ develops which makes the situation again conducive to the pursuit of local improvements. This appears crucially significant in thinking about action-oriented research especially because it highlights that a systematic pursuit of local improvement might sometimes require interventions outside the local.

4.3.6 System Dynamics

J. Forrester pioneered the System Dynamics (SD) approach around 1958 (Dash, 1994; Flood and Jackson, 1991a; Forrester, 1958; Keys, 1990; see Jay W. Forrester page at http://sysdyn.mit.edu/people/jay-forrester.html). The early formulation of the SD approach involved the use of the concepts and ideas of Control Theory and Control Engineering in management and decision making. The approach was aimed at rectifying what Forrester perceived as the then prevailing tendency to view management issues as isolated problems at isolated points in time. Subsequently, Forrester elaborated on the potential application of SD for decision making and policy analysis in industrial organisations (Forrester, 1961), in urban planning (Forrester, 1969), and in world-level policy issues (Forrester, 1971, Meadows, et al., 1992).

SD is primarily orientated towards identifying (defining), modelling, and simulating a system (e.g., a factory, a city, an economy, or even an entire ecological system) in order to explicate how the system might behave under different policy regimes. The SD modelling approach makes use of notions like levels, rates, constants (parameters), physical flows, information flows, discrepancies, delays, loops, non-linearities, control strategies, etc. Typically, models are developed by first developing an influence diagram (also called a causal loop diagram), then elaborating it by means of a flow diagram, and finally encrypting it in some appropriate computer simulation language. Many computer applications have been developed to build SD models and simulate model behaviour under various policy regimes. Some of the applications used currently are Stella®, iThink® (see both in High Performance System pages at http://www.hps-inc.com/), and Vensim® (see Vensim pages at http://www.std.com/vensim/).

The SD approach deals with ‘dynamic systems’, i.e., aggregates of physical and abstract entities which are distinguishable from their surrounding environment as purposive wholes and which exhibit dynamic behaviour. The approach recognises that certain dynamic systems might behave in a counter-intuitive manner, i.e., an unexpected manner, especially due to the way such systems might be constituted. By trying to model dynamic systems and simulating them on a computer, the SD approach seeks to predict their behaviour pattern.

Many applications have been reported in the literature, especially in the journal System Dynamics Review. Some of the applications have been in complex forecasting, e.g., in forecasting the year in which world petroleum oil production will reach its peak (see World Oil Forecasting Program page at http://www.halcyon.com/duncanrc/index.html), in improving the management of the commercial software development process (see Software Process Modeling with System Dynamics pages at http://www-rcf.usc.edu/~madachy/sd/), modelling ecological systems for the purpose of planning (see, e.g., Landcare System Dynamics pages at http://www.agfor.unimelb.edu.au/LSD/links.html), designing information and control systems in a colliery (Wolstenholme, 1990, Chapter 7), evaluating management information systems (Wolstenholme, et al., 1993), studying the interactions of local structures and decision-making policies in the US economy (see MIT System Dynamics Group pages at http://sysdyn.mit.edu/sd-group/home.html), improving educational processes by using SD models (see System Dynamics in Education pages at http://sysdyn.mit.edu/home.html), etc.

One of the success stories of SD-type feedback modelling was that of FOSSIL2, an integrated model of energy supply and demand in the United States of America, which has been used to prepare projections for energy policy analysis in the US Department of Energy (Nail, 1992). The model has shown, inter alia, that the expected reduction in oil imports under the National Energy Plans might not be achievable.

The literature in this area seems quite diverse. One particular strand among those distinguishable in the literature pertains to organisational learning (Senge, 1990; Senge and Sterman, 1992). Although some of the basic notions of the SD approach are still used in this strand of literature, the guiding notion does not seem to be a ‘dynamic system’ any more. The focus seems to have shifted to the use of the ‘dynamic systems’ language for the purpose of creating ‘learning laboratories’:

Our research attempts to develop learning processes aimed at (1) improving managers’ shared mental models so that they become more systemic and more dynamic, and (2) developing managers’ abilities to view new situations systemically and dynamically (Senge and Sterman, 1992, p.139; also see Learning Laboratory Projects pages at http://learning.mit.edu/pra/pro/learnlab.html).

The SD approach seems to have a number of suggestions for action-oriented research. In its earlier form, the SD approach sought to create value for its clients by providing them with superior policy insights. The process was also expected to contribute to the general fund of knowledge about ‘dynamic systems’, e.g., in terms of ‘generic structures’, control strategies, etc. In the learning-oriented applications of SD, clients are expected to benefit from participating in learning laboratories, e.g., by developing dynamics-related insights, shared mental models, etc. Even these applications are not entirely limited to local improvements. In such applications, the power of the ‘dynamic systems’ language is demonstrated, and the language itself might be expected to develop as a result, e.g., in terms of its symbols and the range of interactions it makes possible. The ‘dynamic systems’ vocabulary seems to reinforce the SD approach, whether the approach is used for policy analysis or, more actively, for creating ‘learning laboratories’.

4.3.7 Soft Systems Methodology

Soft Systems Methodology (SSM) is said to be a development from Systems Engineering (Checkland, 1981; 1989; Checkland and Scholes, 1990; see SSM Internet Resource Page at http://www.bf.rmit.edu.au:81/~andrewf/ssm-res.html ). One of the basic starting points of Systems Engineering, i.e., the specification of the purposes to be served by the system to be built, was found difficult to secure in a range of situations which nevertheless required the securing of some type of improvement. Instances of such situations were obtained in ‘messy, changing, ill-defined problem situations with which managers have to cope in their day-to-day professional lives’ (Checkland, 1989). In such situations, SSM seeks to introduce a process typically described as a never-ending cycle of learning, which might be implemented at the level of an individual or a group representing one or more organisations, professions, communities, etc. (Checkland, 1995).

In SSM, a situation is viewed as a product of history of which no unique account is usually possible. A situation is expected to contain some ‘would-be improvers’ of it, who are likely to have various perceptions of the situation, of the purposes being pursued in the situation (‘tasks’), and of the various things about which there are disagreements (‘issues’) (Checkland and Scholes, 1990, p.28). These tasks and issues are typically used in SSM to develop ‘root definitions’, i.e., definitions of a relevant system that is consistent with a particular view of the situation. Each definition is then elaborated in the form of a ‘conceptual model’, which spells out the necessary activities required to fulfil the task identified in the root definition or to deal with the issue. It is clear from the ever-growing literature on SSM applications that the conceptual models are normally based on the generic notion of a ‘self-regulating system’. (See, for example, the characteristic ‘monitor’ and ‘take control action’ activities that constitute the feedback in the model in Figure 2.11; Checkland and Scholes, 1990, p.40).

The use of a conceptual model in SSM is quite different from that in Systems Engineering. Unlike in Systems Engineering, a conceptual model designed here is not meant to be implemented and operated. It is used, alongside a number of other conceptual models (arising out of different views in the situation) to orchestrate a type of debate among the ‘would-be improvers’ who participate in the SSM process. The debate requires participants to compare the conceptual models with what they perceive the actual situation to be and arrive at an ‘accommodation’ on what changes could be brought about in order to improve the situation.

Many applications of SSM have been reported (Checkland 1981; Checkland and Scholes, 1990; see past issues of the Journal of Applied Systems Analysis). Applications include redesigning of the role, structure, and activities of a department in an organisation (Checkland and Scholes, 1990; Chapter 3), developing a more comprehensive way of evaluating the performance of a particular specialist service within a wider health service system (ibid., Chapter 4), redesigning the role and organisation of a public sector agency (ibid., Chapter 5), reorienting the focus and activities of a division or function within a business and clarifying how a decision support system (DSS) might be designed for the division (ibid., Chapters 6, 9), using SSM to define, plan, and manage major organisational change programmes (ibid., Chapters 7 and 8), etc.

Reflecting about the role the SSM approach plays in a project, Checkland and Scholes have commented that:

… the role of the approach is akin to that of the cavalry in nineteenth-century war: it can add a certain tone to what might otherwise be a vulgar brawl (Checkland and Scholes, 1990, p.302).

Development of SSM is said to be predicated on the difficulties of using the image of a self-regulating system in order to address problematic organisational situations so as to secure some improvement. The SSM literature indicates the potential of alternative images, e.g., that of an appreciative system (Checkland, 1995; Checkland and Casar, 1986), a learning system (Checkland, 1985), or a system of inquiry (Checkland and Scholes, 1990, p.18).

SSM can lead not only to purposeful action to improve a problem situation; more generally, it also helps to orchestrate the process of ‘appreciation’ in Vickers’s sense, sharpening views and making choices of action more explicit (Checkland and Scholes, 1990, p. 147).

Whereas systems engineering methodology is a system concerned with achieving objectives, SSM is a learning system (Checkland, 1989, p.78). (The ‘is’ appearing twice in this statement might be interpreted as ‘articulates and helps to bring about’.)

What was found to be needed was a broad approach to examining problem situations in a way which would lead to decisions on action at the level of both ‘what’ and ‘how’. The solution … was a system of inquiry (Checkland and Scholes, 1990, p. 18, emphasis in the original).

To the extent that an ‘appreciative system’ or a ‘learning system’ or a ‘system of inquiry’ is viewed as a more general type of notion than a ‘self-regulating system’, SSM thinking might be regarded as a creative enlargement of systems and cybernetic thinking (this point has been emphasised in Checkland and Holwell, 1998a, p.158). It seems that SSM has a lot to offer as a potential model for action-oriented research. Unlike in the Operational Research (OR) or Systems Engineering approaches, in SSM the client is not always a unitary authority who could prescribe what objectives must be pursued. Therefore, in order to be client-oriented, SSM has devised a set of action-like steps (akin to problem structuring of Soft OR) in order to ‘add a certain tone’, or to create a client, as it were! But unlike Soft OR, SSM emphasises a repetition of the procedure. The repetition aims to achieve continuous learning (or ‘appreciation’) and maintenance of the ‘accommodation’ against disturbances. An explicit ‘framework of ideas’ and a ‘recoverable’ research process are regarded by the developers of SSM as central to any research (Checkland, 1991; 1995; Checkland and Holwell, 1998b). (There will be occasion to review the research thinking of SSM in more detail in Chapter 5.)

4.3.8 Critical Systems Heuristics

Critical Systems Heuristics (CSH) (also labelled Critical Heuristics of Social Systems Design or Critical Heuristics in short) has been proposed by Ulrich (1983; 1987). Discussions on CSH by other writers are also available (e.g., Flood and Jackson, 1991a; Jackson, 1991; Schecter, 1991). CSH provides a method to be used by planners (as well as affected citizens) to practise practical reason, i.e., ‘to lay open, and reflect on, the normative implications of systems designs, problem definitions, or evaluations of social programmes’ (Ulrich, 1987). The notion of normative implication refers to the values about the social consequences and side effects that enter the process of planning and designing social programmes. This hints at the difficulties of dealing with values in conducting research and applying research results in the social domain. CSH might be interpreted as constituting a special type of response to these difficulties. The response involves a kind of safeguarding against the potential harms of the value components that invariably enter the process of creation and utilisation of knowledge.

Like SSM, CSH also appears to view the givenness of the purposes in Operational Research and Systems Engineering as problematic (Jackson, 1991, p.188; Ulrich, 1988). The aim of CSH is to help people discuss this givenness in practical situations of planning and deliberate on what ought to be done. CSH does not provide a way to identify ‘the right’ purpose, but does provide a method by which purposes (or presuppositions) and their inevitable partiality might be kept constantly under review.

CSH recognises that every systems design involves ‘boundary judgements’, e.g., judgements with respect to the ‘value basis’, ‘basis of power’, ‘basis of know-how’, and ‘basis of legitimation’ involved in the design (Jackson, 1991, p.191; Ulrich, 1987, p.279). The heuristic consists of interrogating any given design in order to make explicit its normative bases (i.e., the four types of basis mentioned above, typically linked with the following four categories: client, decision maker, planner, and witness). Three types of question are framed for each of the bases, thus resulting in 12 questions. The 12 questions are asked in two modes: ‘is’ and ‘ought’ modes. These questions (in the ‘ought’ mode) are reproduced below:

    1. Who ought to be the client (beneficiary) of the system S to be designed or improved?
    1. What ought to be the purpose of S, i.e., what goal states ought S be able to achieve so as to serve the client?
    2. What ought to be S’s measure of success (or improvement)?
    3. Who ought to be the decision taker?, that is, have the power to change S’s measure of improvement?
    4. What components (resources and constraints) of S ought to be controlled by the decision taker?
    5. What resources and conditions ought to be part of S’s environment, i.e., should not be controlled by S’s decision taker?
    6. Who ought to be involved as the designer of S?
    7. What kind of expertise ought to flow into the design of S, i.e., who ought to be considered an expert and what should be his role?
    8. Who ought to be the guarantor of S, i.e., where ought the designer seek the guarantee that his design will be implemented and will prove successful, judged by S’s measure of success (or improvement)?
    9. Who ought to belong to the witnesses representing the concerns of the citizens that will or might be affected by the design of S? That is to say, who among the affected ought to get involved?
    10. To what degree and in what way ought the affected be given the chance of emancipation from the premises and the promises of the involved?
    11. Upon what world-views of either the involved or the affected ought S’s design be based? (Ulrich, 1987, Table 1)

The answers to the questions are expected to inform the process of argumentation between ordinary citizens on one side and presumed ‘experts’ like planners, scientists, and decision takers on the other.

Three key concepts are said to inform CSH. These are: (i) Justification break-offs as boundary judgements, (ii) a priori concepts of practical reason, and (iii) polemical employment of boundary judgements. The 12 heuristic questions, when used in a practical situation, make explicit the justification break-offs which are usually hidden, using a priori concepts such as ‘value basis’, ‘basis of power’, ‘basis of know-how’, and ‘basis of legitimation’, and are potentially capable of being used in a polemical way—enabling ‘ordinary people to expose the dogmatic character of the expert’s objective necessities’ (Ulrich, 1987).

CSH is expected not only to create an advantage of argumentation for the affected people, but also to be a tool for ‘experts’ and ‘applied scientists’ to deal critically with their ‘justification break-offs’, i.e., the point in an argumentation where the premise cannot be justified any further. It ‘seeks to render them more self-reflective and democratically minded with respect to their quest for improvement’ (Ulrich, 1996b).

As long as he does not learn to make transparent to himself and to others the justification break-offs flowing into his designs, the applied scientist cannot claim to deal critically with the normative content of these designs (Ulrich, 1987, p.277).

Commentators on CSH have interpreted it as a potential contribution to the extension of the OR philosophy (Flood and Jackson, 1991a; Jackson, 1991). CSH seeks to counter any imposition of plans and designs and the associated rationality by one group of people over another group of people; it aims at orchestrating a systematic debate between opposing points of view; and it seeks to unravel the bases on which opinions are formed. Using the language of cybernetics, it might be said that while CSH does not deny the possibility of goal-seeking (or value-oriented) behaviour in social situations, it does highlight the partiality of goals and values. The remedy comes in the form of a process (that is argumentative in nature) that would keep these goals and values under perpetual review. The image that comes to mind is that of an ‘intelligent thermostat’ that would keep the target temperature under constant review, not on its own, but through a process of argumentation with the relevant others (see Sub-section 3.3.2 on Action Science). Ulrich has provided a similar image in envisioning a ‘future-responsive management’ or a ‘future-responsive systems design’ that requires the creation of appropriate institutional arrangements to promote critical debate among differing perspectives about the preferred designs for the future (Ulrich, 1994).

One such arrangement has been suggested and practically tested by Peter Dienel [references provided] in Germany, namely the planning cell (Plannungszelle). It has been successfully used for developing citizen reports on technological projects (Bürgergutachten). Using statistical rules for random sampling or other procedures that ensure a representation of different concerns, a government body invites citizens to serve on a committee that examines design proposals and comes up with suggestions for better designs (Ulrich, 1994, p.31).

Ulrich goes on to elaborate the difficulties of argumentation within institutional arrangements like the above where different perspectives about the ideal future encounter each other. He has suggested CSH as an appropriate form of communication that might serve to deal with these difficulties.

CSH seems to offer many lessons for action-oriented research. Ulrich, the main developer of CSH, has commented about the similarities and differences between CSH and action research (Ulrich, 1996b) and has proposed a way to use CSH within action research (Ulrich, 1996a). Ulrich has argued that CSH might bring some methodological rigour into action research. From the point of view of this review, the notion of ‘future-responsive systems design’ (or to slightly modify it, ‘future responsive system’) is recoverable as a potentially more general type of notion than either ‘self-regulating system’ (Ulrich, 1981; 1988) or even a ‘learning system’ (Ulrich, 1988), as the latter two types of system do not deal with the ‘conditions of imperfect rationality’. It seems that the sort of system Ulrich has envisaged would have the ability to learn not ‘monologically’ but in a process of communication (or argumentation) with the relevant others in the environment. However, for such a system to be instantiated (e.g., in the form of a planning cell) there would be the need for an appropriate form of communication within the system, and between the system and its environment.

This perspective does seem to offer a way, howsoever preliminary, to address the tension between the local and the global in action-oriented research. Local interventions can have effects outside the local. These effects might not always be positive for the various relevant others in the environment. CSH suggests that the difficulty might be handled by devising suitable forms of communication (perhaps, forms of argumentation) within the local and between the local and its environment. The effect of this process would be a discursively maintained boundary judgement that guides and informs the local improvement effort.

4.3.9 Total Systems Intervention

Total Systems Intervention (TSI) is said to be a method of operationalising the main principles of Critical Systems Thinking (Flood and Jackson, 1991b, Chapter 16). Critical Systems Thinking (CST) itself seems to be a result of visualising new directions in management science (Jackson and Keys, 1987; Jackson, 1991; Flood and Jackson, 1991a; 1991b). The first elaboration of TSI had appeared in 1991 (Flood and Jackson, 1991a). Since then a number of different characterisations of TSI and CST have emerged, some of which might not be fully compatible with each other (Flood, 1994; 1995a; 1995b; Flood and Romm, 1996b; Jackson, 1997). Therefore, for the purpose of this review, a somewhat broad picture is presented, especially from the point of view of what insights might be secured for action-oriented research. TSI was initially described in the following terms:

Total Systems Intervention (TSI) represents a new approach to planning, designing, "problem solving" and evaluation. The process employs a range of systems metaphors to encourage creative thinking about organisations and the difficult issues that managers have to confront. These metaphors are linked through a framework, the "system of systems methodologies", to various systems approaches, so that once informed agreement is reached about which metaphors most thoroughly expose an organisation’s concerns, an appropriate systems-based intervention methodology (or set of methodologies) can be employed. Choice of an appropriate systems methodology will guide "problem solving" in a way that ensures that it addresses what are found to be the main concerns of the particular organisation involved (Flood and Jackson, 1991a, p. 45.)

From this (and also from a detailed study of the TSI literature) it would seem that the TSI approach does not belong to the same category as Operational Research, Systems Engineering, Organisational Cybernetics, Soft Systems Methodology, Critical Systems Heuristics, etc. Each of these latter approaches seems to visualise a type of system, e.g., ‘self-regulating system’, ‘viable system’, ‘learning system’, ‘future-responsive system’, etc., or a type of systemic effect, e.g., self-regulation, self-maintenance, learning, ‘accommodation’, discursively maintained boundary judgement, etc., and describe the means (methods, methodologies, forms of communication, heuristics, etc.) of creating such systems in a given practical context to achieve local outcomes of value to some clients. TSI does not seem to operate with this type of a logic. On the other hand, TSI seems to put forward a way of using judiciously the various management systems approaches in practical situations, perhaps aiming to fill a gap in the literature in this area. To do this, it needs a framework of ideas that would guide the TSI way of using the systems approaches. Such a framework of ideas, or a theory for the use of systems approaches seems to be offered by CST, seeking for example to avoid ‘authoritarian usage [of systems approaches] by powerful decision makers’ (Jackson, 1991, p.201)

The theory for the use of systems approaches that CST seems to embrace involves two broad elements: (i) certain commitments (critical awareness, social awareness, dedication to human emancipation, complementarism at the theoretical level, and complementarism at the methodological level) and (ii) a scheme for classifying the systems approaches (on two dimensions: simple-complex and unitary-pluralist-conflictual, resulting in six ideal types) (Flood and Jackson, 1991a; Jackson 1991, Chapter 7).

As might be expected, when there is a theory for the use of any approach, it becomes possible to disagree with the theory by devising some alternative type of use (Cummings, 1994), or using the approaches ‘obliquely’—to use a term coined by Flood and Romm (1996b). In fact, most of the current discussions about TSI and CST in the management systems literature seems to be dealing with this difficulty: How to establish the academic and practical superiority of a ‘theory for the use of’? One type of answer seems to suggest that there might not be any global criteria at all (Cummings, 1994). However, certain criteria have been explored by others. The notion of ‘learning’ has been used in order to address this question. On this view, a version of CST is preferable because when its prescribed ‘theory for the use of’ is followed, it would generate the ‘greatest possible learning’ (Flood and Romm, 1996b) ¾ a claim dificult to establish. The notion of ‘flexibility’ has also been used to tackle this question. Accordingly, a version of CST would ‘ensure maximum flexibility in an intervention’ (Jackson, 1997, p.6) ¾ again a claim dificult to establish.

Following G. Morgan, Jackson (1989; 1997) has called for reflective conversation among the management systems perspectives (or ‘paradigms’ in the TSI vocabulary). Indirectly, this seems to be a call for developing new languages (or other communicational devices) that would allow such conversation among systems languages contributing to their concurrent evolution. If the management systems approaches are interpreted as offering diverse alternative ‘frameworks of ideas’ and ‘methodologies’ for doing action-oriented research, then TSI/CST seems to pose the problem of dealing with this diversity.

The literature of TSI/CST seems to be quite instructive for thinking about action-oriented research. If the latter is going to involve ‘methodologies’ in the sense used in the management systems literature, it would come up against issues similar to those debated in the TSI/CST literature. Furthermore, if action-oriented research is going to involve general notions or vocabularies or multiple ‘frameworks of ideas’ to guide and develop research practice, it would have to deal with the problems of methodology choice, learning, flexibility, reflective conversation, development of languages, etc., in order to allow a concurrent evolution of these ‘framework of ideas’ and the associated models, languages, ‘methodologies’, etc.

 4.4 Management Systems and Action Research

The review set out to explore how the two types of orientation, i.e., research-orientation and action-orientation, intermesh within management systems thinking. The exploration seems to have yielded a number of interesting results. These results will be recapitulated in this section in terms of different types of action-orientation and research-orientation recovered from the literature of management systems. These will be used to develop the discussion about action research presented earlier in Section 3.4.

4.4.1 Action-orientation in Management Systems Thinking

Different types of action-orientation were found in the management systems literature. All the management systems approaches reviewed here are oriented towards bringing about some improvement in a practical situation according to some relevant local criteria. Therefore these approaches might be seen as forms of support given to some relevant set of actors in a practical context of action. However, based on the form of support, different types of action-orientation can be found in these approaches. Some of the approaches strive to produce the necessary information, knowledge, structure, system, etc., that might be used by the relevant actors (who are being supported) according to their own purposes. Some other approaches strive to orchestrate processes where new purposes, even new actors, might emerge and be supported.

In the Operational Research (OR) approach, the objectives of the client determine what system would be modelled. Once the objectives and the system are identified, the methodology of OR clarifies what data would be collected and finally what state of the system would be identified as the most desirable (or optimum). In this case, the objectives of the clients are taken as external to the form of support that is provided. However, the Soft OR approach seeks to ‘structure’ problems (or issues) such that the clients can come to a better understanding, an accommodation, a decision, a strategy, etc. The Soft OR facilitator actively influences the group process in order to make the clients become more like a collective, capable of acting in solidarity. This is akin to the creation of a new actor, in this case the collective actor.

Both Systems Analysis and Systems Engineering (SA/SE) strive to contribute in practical situations by creating a system that achieves some pre-defined purpose. The purpose is external to the SA/SE approach; it comes from the client. In that sense, it is similar to the OR approach. Soft Systems Methodology (SSM) in contrast strives to produce the purposes that might be pursued in the context. It also aims at some type of capacity building among the relevant actors so that they are able to appreciate their situation and intervene in it in an ongoing manner, as if to create and maintain their collective actorship.

In the applications of VSM, the client organisation improves its fitness to deal with its environment. This is expected to enhance the capability of the client organisation to pursue its purposes as identified from time to time. The VSM itself does not dictate what purposes the viable organisation ought to be pursuing. However, in the more recent adaptations of Organisational Cybernetics, there is an interest in producing the capabilities for new types of action, e.g., through the creation of ‘self-constructed organisations’.

The Socio-Technical Systems approach seeks to produce positive local outcomes by helping the client organisation make proper organisational or technological choice. The aim is to reduce human costs, improve performance, etc. This might be seen as a form of capacity building within the client organisation. This type of support might take either of the two forms discussed above. Consequently, the existing actors might be supported through better organisational/technological systems, or new actors might emerge through organisational change or inter-organisational co-operation.

The System Dynamics (SD) approach seeks to create value for its clients by providing them with dynamics-related insights about their situation, helping them develop shared mental models, etc. The SD-based ‘learning laboratories’ provide the participants with a space for experimentation, understanding, learning, etc. This seems to provide both the possibilities: supporting actors through better observation of the dynamics of a situation and creating new actors through the effect of learning, developing shared mental models, etc.

The CSH approach seeks to improve the ability of its users to participate in debates about plans and designs imposed on them. The aim is to produce new actors (e.g., a planning cell) who keep their actions under constant critical review.

Finally, the TSI/CST type of thinking recognises the potential contribution of the above management systems approaches in improving practical situations. It focuses on the difficulties of choosing the appropriate form of support in any given situation.

The two types of action-orientation can now be articulated:

  • In the first type of action-orientation, the purposes of some actors are taken as given. The action-orientation involves producing the relevant information, knowledge, system, etc., that would support the actors in pursuing their purposes. In this type of action-orientation, the actors’ purposes, intentions, and actions which are supported, are external to the form of support that is provided. In other words, the system analyst’s actions and the actors’ actions are by and large distinguishable.
  • In the second type of action-orientation, there is a process of constitution of a new purpose, or a new actor. The new purpose or new actor is expected to emerge through ‘issue structuring’, ‘accommodation’, ‘conversation’, ‘self-construction’, ‘critical debate’, etc. That which emerges might have certain qualities, e.g., it might be ‘viable’, ‘future-responsive’, etc. In this type of action-orientation, the system analyst’s actions and the client’s (and other actors’) actions are subsumed under the process of constitution of the new context. The client, the other relevant actors, as well as the analyst are more like participants in this process.

4.4.2 Research-orientation in Management Systems Thinking

Different types of research-orientation are found in the management systems literature. Although the management systems approaches are oriented towards creating some value in the context of application, none of the approaches reviewed here guarantee any form of success. Some of the approaches are explicit about the conditions under which the approach might not work effectively. For example, Socio-Technical Systems thinking indicates that organisational strategies might not work under reactive and turbulent environmental conditions. There seems to be a critical literature around most of the approaches, indicating where and how an approach might or might not work (e.g., Brocklesby, 1994; Dickover, 1994; Rosenhead, 1996). There seems to be an attempt to produce new versions of existing approaches in order to increase the chances of their success in application (e.g., Checkland and Scholes, 1990; Flood, 1995a; Harnden, 1990). A closer look at the literature does reveal a number of research-oriented concerns.

The OR approach seems to be research-oriented in at least two different senses. Firstly, at the level of the local, it is interested in identifying the particular state of the chosen system that would be the most satisfactory in fulfilling the client’s objectives. Secondly, the literature pertaining to the OR approach also points towards a systematic extension of its own library of formal models, solution procedures, and its general vocabulary, in an ongoing way. It has also maintained a critical debate about its methodology. Adaptations to the methodology of OR can be found within the discussion about Soft OR. This is found to be a feature of many of the management systems approaches reviewed here. The Soft OR approach also seems to combine an interest in producing locally relevant results with an interest in a systematic expansion (and improvement) in the tools and techniques for group interaction, group facilitation, etc., besides an improvement in its general vocabulary and its basic framework of thinking.

The research-orientation of SA/SE approaches might be seen in their use of systems and cybernetics ideas. These approaches strive to build instances of systems which need to be operated and maintained by people towards some practical end. Research results from such activity might be expected with respect to the general characteristics (and limitations) of such systems. Indeed, the application of SA/SE approaches to managerial problems and the associated critical debate has contributed to the research thinking in this area in a number of interesting ways. The development of the Soft Systems Methodology (SSM) and soft systems thinking in general is said to be a result of this. More specifically, the ideas of ‘appreciative system’ and ‘learning system’ as a basis for research in organisational and social situations have emerged from this.

The research-orientation in Organisational Cybernetics derives from the field of cybernetics. Organisational Cybernetics has extended the basic notion of a ‘self-regulating system’ by generalising it in terms of ‘viable system’. The more recent reformulations in organisational cybernetics appear to strive for a different type of research-orientation (see Sub-section 4.3.4). Properties like ‘viability’, ‘autonomy’, etc., are now sought to be attained through processes of participation, conversation, and ‘self-construction’. The new contexts these processes help create might be viewed as somewhat analogous to research results, although grounded locally. The literature in this area seeks to identify ways of systematically studying and improving these processes so as to increase the chance of their success in future applications.

Socio-Technical Systems thinking indicates how the pursuit of local improvements might be thwarted in reactive and turbulent environments. It suggests that research should contribute not merely through local improvements, but also by indicating how to deal with the situation where local improvements do not seem to have the desired effect. The proposal in this area is to identify methods of interaction among organisations such that something like a ‘power field’ develops which makes the situation again conducive to the pursuit of local improvements. This type of research-orientation might be compared with the recent reformulations in organisational cybernetics, where methods of interaction are being sought which would bring forth certain desired types of systemic effect.

The SD approach seeks to contribute to the general fund of knowledge about ‘dynamic systems’, e.g., in terms of ‘behaviour patterns’, ‘generic structures’, control strategies, etc. In the learning-oriented applications of SD, the power of the ‘dynamic systems’ language is demonstrated. The ‘dynamic systems’ vocabulary seems to reinforce the SD approach, whether the approach is used for policy analysis or, more actively, for creating ‘learning laboratories’ or ‘micro-worlds’.

SSM and the associated thinking has been regarded as a creative enlargement of systems and cybernetic thinking. It has produced a body of literature on the nature of research in the social domain and the possibility of using alternative ‘frameworks of ideas’ in conducting research for situation improvement. One of the main lessons from this literature seems to be the point that these ‘frameworks of ideas’ do not have to be veritable representations of how the world might be like, but the frameworks should only allow a researcher to carry out research even when different types of world are obtained. It emphasises the importance of the need to make the framework as explicit as possible, develop ‘methodologies’ to instantiate the framework, instantiate the framework in practical contexts, demonstrate its contribution in such practical contexts, improve the framework and the ‘methodologies’ in a programmatic way, etc. Besides these general insights, the SSM literature also hints about the potential of a research framework that might be built around the basic notion of an ‘appreciative system’ (or a ‘learning system’ or a ‘system of inquiry’). The power of this notion has been related to its greater generality compared to the notion of a system as used in the SA/SE approaches. In implementing the notion of ‘appreciative system’, SSM provides a way for the actors in a situation to function together (as a team, or collective, as if like a new actor) by ‘appreciating’ their context, generating ‘accommodation’ about the broader directions for action, and generating the appropriate actions themselves. The process is supposed to be never-ending. This implies that the emergent capacity to act together is not self-maintaining; it requires continuous effort (e.g., learning) in order to be maintained (and presumably developed).

The CSH literature seems to highlight an important difficulty associated with the frameworks of ideas used for research and action in worlds which do not conform to representational frameworks. Sometimes, merely instantiating such (non-representational) frameworks might amount to an imposition on some people. The traditional research thinking deals with such issues by trying to limit itself to representational frameworks alone. CSH seems to propose a way to deal with the situation where the representational frameworks are not (yet) found, but some form of systematic support to action is still felt to be necessary. The solution proposed is to establish a degree of collective control over the framework by designing a process of argumentation about the framework, such that it is maintained discursively. This is the so-called ‘critical solution’ to the ‘problem of practical reason’ in the vocabulary of CSH. It has been claimed in the literature that such discursively maintained frameworks can result in the creation of ‘future-responsive systems’.

The TSI/CST literature seems to acknowledge explicitly the possibility of many research (and action) frameworks to be used within the management systems area. It also acknowledges the various conceptual and practical difficulties arising out of this. On the one hand, the TSI/CST thinking might be viewed as a set of guidelines for using various management systems methodologies in order to magnify their research-orientation and action-orientation. For example, TSI/CST thinking seeks to clarify that the learnings from a System Dynamics (SD) modelling exercise is best expressed using the vocabulary of ‘dynamic systems’—one that is associated with the framework of ideas on which SD is based. On the other hand, referring to the more recent TSI/CST thinking, a different type of focus might be recovered. There is an emphasis on ‘reflective conversation’ among different frameworks of ideas, hinting at the need for new conversational devices which can accommodate different system notions, e.g., ‘self-regulating system’, ‘dynamic system’, ‘appreciative system’, ‘future-responsive system’, etc., and help produce a general body of thought for the management systems area.

An attempt is now made to articulate different types of research-orientation within the management systems literature based on what has been described in this chapter. It is to be noted that the literature focuses on the need for a systematic support to action within social and organisational situations and the need to maintain a research-like critical debate about it with the hope of improving the chances of the forms of support to be more successful in future applications. The general lesson in this literature about improving the chances of the forms of support to be more successful in future applications points either towards the innovation of alternative forms of support or towards the improvement (and enrichment) of the existing forms of support so as to make them more reliable, easier to use, more robust to changing circumstances, easier to be taught, etc.

The above general lesson captures some of the recurrent themes in the research literature pertaining to the management systems approaches. However, for the purpose of this review, it is also important to identify how some of the management systems approaches might be unique and different from some others in terms of their research-orientation. This has been done in the following paragraphs by trying to classify what the approaches strive to achieve and improve through research. It is to be noted that the classification presented below constitutes one way of discussing some of the similarities and differences among the management systems approaches. There can be alternative ways of discussing the same. However, the following way of discussing seems to point towards the specific contributions management systems thinking can make to the current debates in action research—a theme taken up later in the next sub-section.

  • One type of research focus seems to be on what might be labelled as ‘systems in the world’. The form of support involves the (re)organisation of a local environment so as to produce in the world an instance of a system having some observable properties, e.g., optimality, self-maintenance, etc. This is made possible by innovating methods of intervention and the means of ensuring that the results are of a desired quality and durability. The research-orientation of OR, SA/SE, and parts of Organisational Cybernetics might be interpreted as focussed on bringing forth such systems and maintaining a critical debate pertaining to such systems: i.e., their formal characteristics, their effect when brought forth, ways (methods) of bringing them forth, conditions of success of such methods, shortcomings of such systems and methods, etc.
  • A second type of research focus seems to be on what might be labelled as ‘systems to generate collective action’. The form of support involves the introduction of a specialised environment for interaction (e.g., ‘microworlds’, ‘learning laboratories’, ‘facilitated groups’, etc.) and a specialised form of interaction (e.g., language, rules, games, artefacts, etc.) among a set of actors such that they are able to initiate purposeful action as a collective. This is made possible by innovating suitable environments and forms of interaction to develop a shared understanding of the situation (as in the Soft OR approaches and in the learning-oriented applications of SD), ‘accommodation’ among viewpoints (as in SSM), etc. The literature in this area takes the first type of system notion (mentioned above) as a point of departure. The research-orientation associated with the ‘soft’ approaches might be interpreted as focussed on bringing forth such systems which generate collective action and maintaining a critical debate pertaining to such systems: i.e., the problem of describing such systems (worlds in which such systems might exist), the ways (methods) of bringing them forth in different contexts, the problem of describing the methods (or the rules that govern the methods), durability of such systems, the contribution of such systems in practical contexts, shortcomings of such systems and methods, etc.
  • A third type of research focus seems to be on what might be labelled as ‘systems to generate collective control over action’. The form of support involves the creation of suitable institutional arrangements and conversational devices to create conversations among a set of (individual or collective) actors such that they are able to keep each other’s actions under some type of collective control. This is made possible by identifying suitable institutional arrangements as well as innovating methods of conversation which will have the desired effect. An interest in this type of system can be noticed in the research literature pertaining to CSH and TSI/CST. The literature in this area appears to take the second type of system notion (mentioned above) as a point of departure. It highlights the possible side effects of bringing forth and maintaining a system to generate action without any form of collective control over it from the participating and non-participating actors. The literature indicates the need for designing suitable contexts where specific forms of ‘conversation’ or ‘argumentation’ can be used to keep the inevitable ‘boundary judgements’ or ‘paradigms’ under continuous critical scrutiny. The literature also refers to the difficulties of doing this as well as the difficulties of identifying whether or not any other system notion might offer a more appropriate form of support in any given situation.

4.4.3 Revisiting the Current Debates in Action Research

An attempt is now made to identify elements in the management systems literature which might be introduced within the current debates in action research (see Section 3.4) to enrich those debates. The debates pertaining to action research focus on the problems of ‘normal’ research (see Sub-section 3.4.1). The gist of the problems discussed in these debates might be expressed in terms of the following: The so-called ‘normal’ research does not seem to ensure that its results will always bring about improvements in how people behave and act in various practical settings. The general approach of ‘normal’ research to focus on observed patterns that can be expressed (in terms of laws), transferred in scientific communications, and progressively made more and more precise through the effect of repeated observations, is viewed as inappropriate for the types of aim action research strives to achieve. Having rejected this approach, the action research methods seek to orchestrate various types of interaction within the local contexts of action so as to effect an improvement in how people behave and act in those contexts. However, as soon as the focus shifts to orchestrating local interactions to achieve improvements in action, the issues of demarcation of research from other activities, the quality and durability of the results, effects of such activities outside the local context, the need for producing transferable results, the difficulty of conducting any critical scrutiny of the activities and the effects, etc., emerge to the fore. This is where the debate seems to get embroiled in the questions of ethics, epistemology, research paradigms, dilemmas, etc.

An appreciation of the development of the management systems literature and the current debates therein reveals that there might be alternative ways to formulate the core issues in the current debates in action research. To do that, it will be necessary to re-look at the general approach of ‘normal’ research which the action research debates seem to reject so emphatically. The so-called systems movement (see Section 4.2) also seems to have been based on an appreciation of the difficulties of using the normal methods of research in domains that display interconnectedness, emergence, openness of the object of study, etc. However, this seems to have been dealt with by introducing the notion of a ‘system’ that might not only refer to observed patterns, but also refer to entities that might be brought forth through purposeful action (including interaction, communication, etc.) such that the entities provide the basis (and support) for some other purposeful action. A review of the management systems literature reveals three general categories of such ‘systems’ being discussed there: systems in the world, systems to generate collective action, and systems to generate collective control over action (see Sub-section 4.4.2). The research focus in this area then shifts from the problems of identifying and refining observed pattern to the problems of specifying relevant types of system and identifying and refining the methods of bringing them forth in various practical contexts. This type of reformulation seems to overcome most of the difficulties described in the current debates in action research. The reformulation does not provide a direct answer to the question of how to improve action and behaviour in a practical situation; it simply reframes the question in ‘systems’ terms: What type of system could be brought forth in a practical context such that it proves to be a robust type of support to action and behaviour within that context? How to improve the methods of bringing forth such systems when the demand for such systems arise in future in similar or different environments? Whether it is possible to produce a systematic body of knowledge pertaining to various types of system and the methods of bringing them forth?

A summary of the key ideas recovered from this chapter will be presented in Chapter 5 (see Sub-section 5.5.2). The review of literature continues in Chapter 5 to explore the possible contributions of research to the improvement of action by delving into a wider body of literature on research as a form of support to action.

Acknowledgements

Contents

Preface

Synopsis

Chapter 1

Chapter 2 

Chapter 3

Chapter 4

Chapter 5 

Chapter 6 

Chapter 7

Chapter 8

Chapter 9

Chapter 10

Appendix A 
(Bibliography) 

Appendix B 
(Organisations)