The Implications of Complexity Theory for Psychoanalytic Thinking about Organizations

Ralph Stacey
Complexity and Management Centre
Business School of the University of Hertfordshire
Hertford, UK

The paper first sets out the key features of what seems to me to be the most widely employed and discussed psychoanalytic model of organizations, usually referred to as the Tavistock model, and then goes on to examine elements of this model in the light of a complexity perspective. I conclude that the latter perspective suggests a significant shift in our understanding of organizations from both organizational theory and psychoanalytic points of view. The paper ends by indicating some of the implications of this significant shift in understanding for organizational consulting practice.

THE TAVISTOCK MODEL

What follows is a brief and, therefore, necessarily simplistic review of the Tavistock model. Although such a review removes much of the richness of the theory I hope it exposes the most fundamental assumptions upon which it is built. Starting at the level of the organization the Tavistock model is based on the open systems perspective developed by Miller and Rice (1967), associated with the work on socio-technical systems of Trist and Branforth (1951) and general systems theory (von Bertalanfy, 1968).

Open Systems

From the open system perspective an organization is a system with a clear cut boundary separating it from its environment. It imports resources from the environment across this boundary, transforms them into products and/or services which it then exports back across the boundary to the environment.

Each organization has a primary task which is primary in the sense that carrying out this task adequately is the condition for its survival. Failure to carry out the task adequately leads to the withdrawal of the environment's support for the organization and thus its demise. The primary task is its reason for being and it fulfills this reason through adaptation to the external reality of the environment.

The organization itself is a socio-technical system, where that system consists groups of individuals interacting with each other and various technologies in order to carry out the primary task. That primary task is a particular kind of transformation of imported resources into exported products and services which satisfy particular kinds of requirement presented by the environment.

The ultimate transformational agents within the boundary are the individual members of the organization. How are we to understand the nature of this system of interacting agents? It is here that psychoanalytic theory offers its contribution: a fundamental assumption is made that an organization can be understood from the perspective of individual intrapsychic processes and family dynamics. Every organization is assumed to be the individual/family writ large.

Psychic Functioning

The sources most widely quoted when theorizing about organizations as individuals/families writ large are Freud (1923), Klein (1975) and Bion (1961), the theoretical foundation for all these theorists being the drive-structure model (Greenberg and Mitchell, 1983).

According to the drive-structure model, the individual psyche is structured through a process of reconciling the individual's inherited instincts with the constraints society imposes on the exercise of those instincts. Inherited instincts are reflected in the intrapsychic drives that constitute the id. These drives consists of libidinal and aggressive energies, or tensions, which are governed by the pleasure principle in that they seek discharge through an object that is the target of the drive. Through this process the psychic apparatus achieves its aim of keeping tension or stimulation as low as possible - the constancy principle in which quiescence is viewed as pleasant while tension is not. Conflict arises because the individual needs others to survive and is therefore not free simply to discharge the drives - the psychic apparatus must also operate according to the reality principle, that is, the constraints of the real external environment. To this end the ego is differentiated out of the id as the individual matures and takes up the function of intermediating between the inner world and external reality. The ego operates at the border between the inner and the outer and carries out the rational functions of assessing the external world and making reality based choices on drive discharge in the interests of the individual's primary task of adequate survival. Authority figures in that external world are internalised in the form of the super-ego, society's intrapsychic representative.

An individuals' adaptation to society is thus an essentially conflictual process in which the ego develops a dominant mass of ideas that enable survival. Drive satisfaction incompatible with these dominant ideas may not be simply abandoned but may be repressed and it is then the ongoing function of the ego to prevent the return of these repressed wishes utilising various defences, since they threaten reality adaptation. That defensive process is initiated by the experience of anxiety, the fear that satisfaction of a libidinal or aggressive impulse might endanger ongoing reality testing.

As has been pointed out many times before, this way of conceptualising intrapsychic life is a product of the dominant scientific paradigms of Freud's and, indeed, in many ways, still our own time. That is to say, it is firmly based within the tradition of thinking developed by Newton, Adam Smith, the neo-Darwinians and many others within these traditions. It is, therefore, a reductionist theory in that it takes the individual as the unit of analysis, seeing groups and societies as collections of individuals which in turn constrain those individuals. A holistic theory would, of course, reverse this stance and seek to understand the structuring of an individual psyche as an expression of the functioning of the whole system to which that individual belonged. Freud's metapsychology is clearly Newtonian in that it postulates a psychic apparatus which directs and obstructs energy flows so as to reach a state of equilibrium and in this it also resembles Adam Smith's views of economic and market functioning. It is also a theory that assumes a pre-given objective reality which requires rational, verbal processing if it is to be adequately dealt with. It is an essentially linear theory of cause and effect which ascribes little importance to chance. These points will be important later when the implications of a complexity perspective are discussed.

Klein developed Freud's theory of intrapsychic functioning, but always within the framework of the drive-structure model. For our purpose here we might say that she built on Freud's notion of the individual internalising important authority figures into the structure of the psychic apparatus in the form of a superego. She postulated a structuring process in which the psyche responds to anxiety by splitting off its desired, nurturing and loving parts from its unwanted, frighteningly aggressive, destructive, hateful and persecuting parts. These parts are projected onto objects in external reality and then introjected and identified with intrapsychically so developing an inner phantasy life.

Intrapsychic agents, or roles, are far more numerous in Klein's scheme than they were in Freud's - there is an id containing a population of representations of objects and object relations including early superego representations, an ego and a superego all interacting to produce intrapsychic life. There is also a more developed theory of interaction between individuals for now an individual psyche can projectively identify some internal object with some real person, so manipulating that person to play a part in the inner life of that individual. This enriches our understanding of interpersonal relationships but it still sees the structuring of the psyche in individual terms - it is not the other's real actions that structure the psyche but that psyche's internal phantasies about others. The nature of the border between the inner world and outer reality can now be seen to be much more problematic. The very process by which the psyche is structured involves a blurring of that boundary and to the extent that individuals employ the various introjective and projective mechanisms they confuse inner phenomena with outer reality. Ego functioning is thus vital to manage the boundary, establishing as clear as possible a distinction between outer reality and inner phantasy life but this is interfered with when anxiety levels rise and individuals defend themselves by regressing to splitting, projection and projective identification.

Bion in turn built on Klein's ideas to gain an understanding of group life. He distinguished between a work group in which members are engaged in a shared task which they have come together to perform, and a basic assumption group. When they are in the former mode they are engaging reality - their functioning is equivalent to the intrapsychic functioning of the ego in that they rationally asses their situation and derive actions that will achieve their task, so adapting themselves to reality. However, when anxiety levels rise individuals defend themselves by regressing to the splitting and projective mechanisms identified by Klein and in doing so they develop amongst themselves a basic assumption group in which they behave as if they have come together not to work but to depend upon one of their number, pair, fight or take flight. When caught in this mode groups cannot work on reality tasks but take part in a phantasy task which obstructs their real work. They ensnare each other, using projective mechanisms, in the particular roles required by the group to perform its phantasy. The group then becomes rather like an individual psyche with individuals taking up roles equivalent to object representations in intrapsychic life - individuals represent needed aspects of the group in its phantasy, an id at the group level. They cannot then be treated as autonomous individuals, as they can when in work mode. Instead it is only the group as a whole that can be understood and addressed: individuals are not their own autonomous selves but object representations in the group mind, which is in turn their illusion.

Finally, since the family is the first group almost everyone initially experiences, it is there that we first and most enduringly experience this kind of group dynamic which we then have a tendency to repeat everywhere we go, including in organizational life. This brings us back to the level of organizations.

Task and Role

We can now see how the Tavistock model understands the nature of the interacting system of individuals within the boundaries that constitute an organization. Central to this understanding is the distinction between the work group and the basic assumption group, between ego functioning and id functioning (Erlich, 1996). The work group in an organization exists when members take up their roles in relation to the primary task, that is, when the group functions in an ego sense. However, complex and uncertain environments tend to generate anxiety and thus the basic assumption group is never likely to be far away. Work groups can quite easily slip into the functioning of a basic assumption group in which members step out of their role in relation to task and take up roles in the group's phantasy so destroying the work. Members' roles will then be determined by the personal relationships between them rather than the real work they are supposed to do. This way of looking at organisations thus tends to downplay relationships of a social nature, relationships at work which meet needs that might have little to do with the primary task, relationships built on emotional attachment. These are all potential distractions to the task, potential signs of basic assumption behaviour. Although all have a general responsibility, it is the particular responsibility of management roles to hold the boundary between outer reality signified by the task with its attendant roles and the inner group phantasy signified by highly personal relationships, loose connections between people that have little directly to do with the task, and basic assumption behaviour.

Holding the boundary between the work group and the basic assumption group has essentially to do with the containment of anxiety for it is raised anxiety that brings about the basic assumption group as a defence against that anxiety. It is for this reason that organizations tend to develop work practices and routines that are not efficient in terms of the primary task but serve as social defences against anxiety (Jacques, 1955; Menzies Lyth, 1975; Hirschhorn, 1990). These are rational, in appearance at least, and so operate as an alternative to the feared submersion in the highly emotional basic assumption group, but they too render it more difficult to perform the primary task effectively. The importance of a shared and doggedly held meaning about primary task is thus continually emphasised.

The Role of the Consultant

In the Tavistock model, the consultant treats an organization in rather the same way as an individual analyst treats an individual analysand or as an analyst works with a Bion type therapy group - it is the group as a whole, or the organization as a whole that is treated, since the individuals in it are all taking up roles in relation to it, either on the basis of its task, or on the basis of their basic assumption phantasies. The consultant uses his or her counter transference to understand the group or organizational transference onto him or her, to understand the patterns of projective identification and so the unconscious phantasy the group or organization might be caught in. It is then the consultant's task to interpret this at an appropriate moment, on the basis that when unconscious processes are verbalised and interpreted, they are relieved. It is also the consultant's role to play some part in containing the anxiety aroused when repressed libidinal and aggressive feelings are released (Schapiro and Wesley Carr, 1991). The consultant will be looking for and seeking to interpret basic assumption behaviour, pointing to the use of other social defensive mechanisms and recommending more task related procedures. The purpose of this is to free members of the organization to take up their primary task dictated roles.

What we have then is a model of organizational functioning, one part of which is entirely understandable to most managers and another part which is not at all understood and is often, in my experience, treated with much suspicion, even disdain. The familiar part relates to the work group. The Tavistock model's notion of a work group is a notion shared by most of the management and organizational theory literature. Theorists and practitioners in this tradition are all too familiar with notions of task, shared views and clearly defined roles derived from them. They also readily agree that the emotional side of life in organizations is a distraction from the performance of task. The much less familiar side is this model's identification of what can interfere with the task, namely, the tendency to build rational sounding organizational defences and the ever present danger of slipping into phantasy behaviour that makes thinking and working impossible. So in one sense this model falls neatly into the mainstream of management thinking but it does add a very important dimension to our understanding. I will argue in the next section, however, that a complexity perspective suggests some radical alternatives as the basis of new directions in which psychoanalytic thinking about organizations might be developed.

A COMPLEXITY PERSPECTIVE

It may be useful to talk about a complexity perspective in the context of what it seems to represent a shift from.

The paradigm that has dominated scientific thinking for the past 200 years at least is that most prominently associated with the names of Bacon, Newton and Descartes. One of the main features of this way of thinking is its use of mathematics to state the laws of natural systems rigorously, so identifying the mechanisms driving natural behaviour, then quantifying the variables in those natural laws, so enabling them to be tested in some form of experiment, or observation, with a view to seeing whether they can be disconfirmed or not. To the extent that they are not disconfirmed they are provisionally accepted as laws. Newton was certainly aware that nature's systems were, almost without exception, nonlinear, as indeed did scientists before him, and as scientists still do. That is to say, scientists know that any given cause has many different possible outcomes and the relationships between cause and effect are not usually proportional but rather more, or less, than proportional. Another way of saying this would be that nature moves in irregular curves rather than regular straight lines. Mathematical models of such systems should therefore consist of nonlinear equations. The difficulty then arises that most nonlinear equations cannot be solved, so making it impossible to conduct a mathematical analysis of the system modelled in such a way. Newton came up with the ingenious device of the calculus that essentially assumed away the problem of nonlinearity by devising an analytical method that divided curves up into small pieces that approximated straight lines. Using the calculus, a nonlinear system could then be approximated by a linear one which is thus amenable to solution. The assumption is that nonlinearity itself does not matter much, that the process of linearising the model leads to only small discrepancies between the model and the real system.

This strategy, as we all know, has been a very powerful one indeed, one that has led to major new insights into how nature works and enabled humans to exert much greater control over natural processes. It has led to a world view in which all systems:

- behave lawfully according to linear causality where one cause has one effect, or at least, a limited number of effects;

- tend to states of equilibrium, or rest;

- are the sum of their parts and can be understood by analysing their constituent parts.

This is a way of thinking that encourages belief in an objective and absolute reality. However, the strategy did not work everywhere. For example, it did not work all that well at the level of the very large, namely the universe, where Einstein's relativity theories gave better explanations. Time and space then came to be seen, not as absolutes, but as relative phenomena. Also, at the at the level of the very small, the particles of which matter is made, the Newtonian strategy did not work very well either and was replaced by quantum mechanics with its notions of the ambiguity of matter at its most fundamental level - either a wave or a particle depending on how one look at it - and basic uncertainty so that we cannot predict the behaviour of individual particles but only make probabilistic statements about collections of them. Furthermore, the Newtonian strategy did not work all that well at the level of many phenomena in ordinary, everyday life. For example, it could not explain turbulence in liquids and gasses.

More importantly, the Newtonian strategy could not explain the evolutionary process in which more and more complex life forms develop. Here a new strategy for understanding was developed, one which is now known as neo-Darwinism. The driver of evolution is seen to be chance mutations at the level of the genes producing organisms with different survival strategies which are then tested for survival in competitive struggles with other organisms. The inner world of the genes, a blueprint for an organism, is thus tested for survival value by a pre-given external reality and it is the species which evolutionarily test reality better than others that survive. While in the Newtonian system there is no room for chance, in the neo-Darwinian view chance is the principal driver.

It is easy to see how combinations of Newtonian and neo-Darwinian thinking have shaped not only economic, political and sociological theories but also psychoanalytic thinking, at least at the meta level, in relation to the drive-structure model of the psychic apparatus.

The shift in thinking, which I have referred to above, is taking place at a very basic level in that it goes back to Newton's brilliant simplification to explore just what has been left out by assuming away nonlinearity. The discovery is that this omission is of major importance. Focusing on the difference that nonlinearity makes enable a greater understanding of many phenomena at the level of ordinary, everyday life and a deeper understanding of the evolutionary process. Since evolution is a process in which a whole ecology of species changes, it is a form of learning and has therefore the potential to throw light on the kind of learning humans engage in. This interest in nonlinearity has been gathering pace since the early 1960s and its exploration is made possible by shifting to some extent from the mode of mathematical analysis to the use of computer simulation - if you cannot solve sets of nonlinear equations you can explore the behaviour they produce by running them on a computer.

Now let me say something about how this nonlinearity is incorporated into the thinking of the mathematicians, physicists, chemists, biologists and computer scientists working together in institutions like the Santa Fe Institute in New Mexico (Gell-Man 1994; Gleick, 1988; Goodwin, 1994; Kauffman, 1993 & 1995; Levy, 1992; Prigogine & Stengers, 1984; Waldorp, 1992). The most common approach is to conceptualise all of nature and its many subsystems, of which we are, of course, a part, as complex systems. Complex systems consist of many autonomous agents each interacting with others according to their own principles, laws or rules.

In inanimate systems, the agents follow their rules of interaction without ever changing them - these are thus deterministic systems which display no learning. For example a physical system, such as a gas, or a chemical system, such as a particular mix of chemicals, each consists of very large numbers of agents interacting with each other in an iterative, nonlinear manner. What one molecule does affects what others do simply because they are interacting with each other. Many such systems have been studied in laboratory experiments and computer simulations and this has led to some very important discoveries.

I will briefly describe what I believe to be the most important of these discoveries in what will necessarily be an over simplistic way. The major insight this work has uncovered relates to the dynamics of networks of large number of agents interacting with each other in a deterministic manner. At low levels of energy/information flow, and when each agent is connected to, interacting with, only a few others, the system displays the dynamics of stability - in the sense in which the word equilibrium is used in economics, sociology and psychology we can say that the system is at, or near to, equilibrium. That is to say, the behaviour patterns produced by the system are regular and predictable collapsing to one kind of behaviour, a point, or displaying regular cycles, which might be quite complicated but are perfectly predictable. Furthermore, any small disturbance in this pattern will be rapidly damped away by the operation of the system, rather like the notion of a perfect market sustaining equilibrium in classical economics or a functioning ego stabilising inner and outer in the psyche. When each agent is interacting with, connected to, very large numbers of agents, when energy/information levels in the system are very high, the system displays the dynamics of explosive instability with a tendency to disintegrate when it comes up against a constraint. Here the system amplifies any deviation. The real discovery, though, is that at some critical point in energy/information flow and connectedness between the agents the system displays the dynamics of a phase transition between stability and instability - just before it becomes explosively unstable it displays a different kind of dynamic in which it is paradoxically both stable and unstable at the same time, in which it is both amplifying and stabilising changes. This dynamic has been given the name of low dimensional deterministic chaos, a mathematical concept which differs so substantially from everyday notions of chaos that I prefer to call it bounded instability.

In the phase transition of bounded instability a system never exactly repeat its behaviour: it is capable of escalating tiny changes into qualitatively different forms of behaviour. For example, the weather system is thought to operate in this phase transition which makes it possible for the tiny change in air pressure caused by a butterfly flapping its wings in Sao Paulo to escalate into a hurricane over Miami. Of course, it is not just this one butterfly that can be said to have caused the hurricane because many other reinforcing and damping air pressure changes will have been encountered as the system moves across the globe. The impossibility of measuring the flapping of every butterfly's wings, and all other sources of air pressure change, with infinite precision means that long term weather forecasts are impossible. However, geographically local, short time period forecasting is possible because it takes time for tiny changes to escalate. So, although there are links between causes and effects these are many and impossible to measure accurately enough so that over long distances and long time frames these links between cause and effect are lost as far as human observation is concerned. We cannot then think of the system as following some fixed blueprint: instead we see its behaviour emerging out of the interaction of the agents, without overall, prior design. The nature of agent interaction creates a potential for general categories of behaviour, for example hurricanes, which are then realised in particular forms by the actual experience of the agents interacting. Although the particular pattern is unpredictable over long distances and time frames it is always recognisable as an expression of the potential. We move then from predicting to pattern recognition and these patterns are not regular but irregular or fractal. Prediction is not abandoned because we can still make short term predictions of specific behaviour and once we know enough about the system we an predict its potential patterns. Furthermore, we can predict the onset of bounded instability from a knowledge of the variables, such as energy flow and agent connectedness, that determine it. However, at more specific levels, knowledge of which is needed if one is to be in control, the real unfolding, emerging behaviour patterns of the system are unknowable until they are created.

Seeing things in this way is the result of taking real account of nonlinearity - it is because the system is interactive, iterative, self reflexive and nonlinear that is capable of the dynamics of bounded instability. This is a major departure from a Newtonian way of thinking in which the focus is on identifying the blueprint, the laws not subject to chance, so that we are able to predict the specific behaviour of the system and so control it. When a system is in the phase transition chance importantly determines its development and limits human ability to be in control of it.

Another important property of complex systems has also been identified, one that indicates how a system produces emergent patterns of behaviour in the paradoxical dynamics of bounded instability. This is the process of spontaneous self organization. For example, when low levels of heat are applied to a particular gas its molecules may move in random ways, but when the level of heat pumped into the gas reaches a certain critical point, where the symmetry of molecular behaviour is broken, the molecules spontaneously self organize so that they all point in the same direction and a laser beam results. A new form of stable behaviour has emerged out of disorder which is not predictable from the behaviour of the individual molecules themselves but flows from their interaction in a dynamic context called far from equilibrium, that is, bounded instability.

When we move from inanimate systems to living systems we can still think of them as complex networks consisting of many agents but now the principles, laws or rules of interaction that constitute the agents evolve, that is, the system learns and so is a complex adaptive system rather than a deterministic one. For example, the brain is a complex adaptive system where the agents are billions of neurons all interacting with each other to produce the electrical and chemical patterns of brain functioning - learning at this level is a continuous evolution in the pattern of connections between neurons. Flocks of birds, shoals of fish, colonies of ants and the human genome are some other examples of complex adaptive systems. In fact the whole ecology of species on earth can be thought of in this way.

Experiments and computer simulations reveal that adaptive systems display the same general kinds of dynamic as deterministic systems do, namely, a dynamic of stability, a dynamic of instability or disintegration, and a phase transition between them. This phase transition is a state at the edge of disintegration, mostly referred to as the 'edge of chaos'. This too is a paradoxical dynamic, the simultaneous presence of stability and instability, so that the system is stable enough to retain form but unstable or fluid enough to evolve. When a system exhibits this 'edge of disintegration' dynamics it is capable of evolving and it does so through the process of self organisation. New forms and survival strategies emerge without any prior blueprint in a radically unpredictable way. We can think of such a system evolving into an open-ended evolutionary space as it takes one step after another into the space of the adjacent possible. The adjacent possible is determined by the system's history, a history which thus both enables and constrains its further selection of direction, its evolution.

The insight, then, is that of living systems evolving, or learning, their way into a future they create - the system creates its own future by reference back to itself and the system and agents co-create and adapt to each other rather than adapting to a pre-given environment. They are creating the reality into which they move by a self organizing process of selection or enactment rather than adapting to a pre-given external and objective reality. Evolution is not simply a process of random mutation at the level of the genes which are then selected out by competition. There is random mutation, but this is not the only generator of variety for from the complexity perspective self organisation is itself generating variety which is then presented for competitive selection. Complexity theory thus adds to the first ordering principle of competition. It adds a second ordering principle, namely, the cooperative principle of self organization.

In summary, then, we have here a view of evolving, that is learning, systems that create their own, radically unpredictable emergent futures through a process of self organization that occurs when that system operates in the dynamical context of bounded instability at the edge of disintegration. In such a system no agent is in control, no agent is planning the future, there is no blueprint for the new developments in the system. Furthermore no agent fully understands its part in the system's evolution, nor is there any need that it should for each agent makes its vital contribution to that evolution through playing its part in its local area of interaction - this is what self organization means. This, I think, is a radically different way of making sense of the process of system evolution, or learning, to that depicted in Newtonian and neo-Darwinian ways of thinking. It seems to me to be important to reflect on what this means for a psychoanalytical understanding of organizations because both the original metapsychology of psychoanalysis and the dominant views about the nature of management and organization have been built upon the Newtonian/neo-Darwinian paradigms.

We return now to organizations and psychoanalysis to reflect on what the implications of a complexity perspective might possibly be.

A Complexity Perspective on Organizations

I suggest that a group of people and the groups of people that constitute organizations are all complex adaptive systems (Stacey, 1996). They consist of agents, in the form of autonomous individual human beings, who interact with each other, so forming a network system that produces patterns of individual, group and organizational behaviour. Just as with all other complex adaptive systems they evolve, or learn, their way into an open-ended future that they co-create in a self organizing way. What is being co-created is not just the emergent pattern of behaviour of the whole system but the very principles driving agent's interactions or relationships with each other. In other words, the individual agents are themselves evolving, that is, being co-created in the evolution of the system they constitute because each is the principles according to which others are related to.

I suggest that we can see that this is so at an organizational level if we conceptually divide an organisation into what I think of as a legitimate system and a shadow system.

The legitimate system is the hierarchy, bureaucracy and the widely accepted, officially approved, shared ideology at the time. The legitimate system is a blueprint that has been designed and installed at some point and its purpose is to enable stable joint action in carrying out the primary task of an organization as it is understood at the time of the legitimate system's installation. But how did this organization reach the point where it had enough shared sense of task and could therefore install some designed legitimate system to carry it out?

We all know that the process leading up to changes in the legitimate system is a messy, confusing and often disturbing one that threatens vested interests, that may well provoke covert politics, basic assumption behaviour and the like. But in my experience this is not all that happens, and if it were it would be hard to explain how anyone ever gets to the stage where there is enough agreement and clarity to embark on anything that qualifies as rational analysis and design. In the behaviour that leads up to changes in the legitimate system and reformulations of what the task is, I experience a process of learning occurring through conversations taking place in the context of personal relationships. The experience is one of dialogue that slips in and out of debate, in which people are trying to surface what they do not yet know, as well as articulating what they know but cannot yet express. It is though conversing, persuading and exerting influence in a highly personal network of relationships that the shape of potential change emerges. Only then can it appear as rather more orderly, rational work. This network of personal, social and emotional relationships constitutes what I think of as a shadow system that underlies and intertwines with the legitimate system.

Now this shadow system is quite clearly a spontaneously self organizing system. Joining an organization means becoming a member of its legitimate system when we are assigned a role in its currently identified tasks. But as soon as we do this we spin a network of personal relationships, social, emotional, psychological, political, which is as vital to our work as is the role we take up in the legitimate system. No one tells us who to network with and we cannot individually determine what that network will be because it requires the acceptance and cooperation of those we network with. This self organizing social/psychodynamic/political system can and does produce corruption, vicious personal striving, covert politics of a harmful kind, basic assumption behaviour and sometimes even near psychotic fantasies. But it also functions as a learning community of practice, the location of an organization's narrative and tacit knowledge, the vehicle for the exploration of intuitions and as such the process of organizational learning and the origin of an organization's creativity. What emerges out of the interaction of agent's in the shadow system may ultimately come to be embodied in changes in its legitimate system. No one fully understand their organizations' shadow system, no one is in control of it, but all contribute to what it is by interacting in their own local network and by so doing they may learn and so change as individuals as well as producing emergent change in the system as a whole. The shadow system seems to me to be characterised by the same kind of co-evolutionary process as that which we see in other complex adaptive systems.

It is possible to see, I suggest, the three kinds of dynamic we encounter in all other complex system at work in organizational shadow systems. We might think of a shadow system driven by basic assumption behaviour as either disintegrative of work and learning or as trapping members into stable repetitive fantasies such as the fantasy that they are a family. We can see the social defences referred to above as productive of highly stable dynamics that also interfere with the learning process and halt an organization's evolution. But there is another possibility, namely the one described above in which people use their close personal relationships in a manner that qualifies, much of the time anyway, as some kind of learning process in which they work with what they do not yet know. I think these are the dynamics of the phase transition at the edge of disintegration for an organization and I say this because people learning in the way I have described are in fact subverting or undermining the existing legitimate system in the interest of changing it so that it survives - a paradoxical form of behaviour, a mixture of stability and instability.

There is one other point I want to make about this shadow system and that relates to its boundaries. The boundaries of the legitimate system are clear - we know who is a member and who is not. But the boundaries of the shadow system are far from clear - we have to think of such boundaries is irregular and fractal because these webs of personal relationships stretch across into other organizations and institutions and are used to carry out the work of the organization. In complex systems there is another and very important fuzziness in boundaries, namely , that between the context of stable dynamics and that of unstable or disintegrative dynamics. The dynamics of bounded instability are themselves the fuzzy boundaries between stability and instability.

We can now return to the outline of the Tavistock model presented earlier and consider the implications of a complexity perspective for our understanding of it.

Psychoanalysis and Organizations

As we have seen, the Tavistock model is built on the foundations of Freud's drive-structure model of intrapsychic functioning in which the fundamental unit of analysis is the biological individual.

From a complexity perspective there is no unit of analysis that is any more fundamental than any other. Whatever the level we decide to focus on we see a complex system nested in other complex systems, in fact emerging from them. For example, at the level of physics we see complex systems in which particles/waves interact with each other according the laws of physics. Emerging from physics, dependant upon it but not reducible to it, is the level of chemistry where atoms or molecules interact with each other to produce chemical transformations. New laws, chemical laws, are applicable to this emergent level of activity and they are no less fundamental than the laws of physics. Furthermore, although the chemical depends upon the physical it then also affects the physical. Similarly, the biological level emerges from the chemical and here genes interact in a context to produce organisms. Once again the biological is dependent upon the chemical and the physical levels but new laws, or principles, apply to the biological level which are not reducible to, or less fundamental than, the chemical or the physical. Once again biological phenomena are dependent upon chemical transformation but can also affect those chemical interactions. We have, then, the notion of one level of complex system emerging from, or nesting in, another, with each mutually affecting the other. We might go on to think of psychic / social / cultural phenomena as emerging from the biological level, with new principles of interaction applying that are not reducible to the biological but also not less fundamental. Psychic / social / cultural phenomena would then affect functioning at the biological level and vice versa, but neither would be reducible to the other.

At each level we choose to focus on we find complex systems, deterministic at inanimate levels and adaptive, or evolving, at animate levels. A key point about evolving or adaptive systems is this: the individual agents at a particular system level both construct and are constructed by the network system they are a part of at that level. It is the pattern of relationships across the network that evolves and through that changing pattern the agents themselves also evolve. Changing relational patterns constitute the system's learning and while that evolution will depend upon, will affect and be affected by systems at other levels, it will not be reducible to them but be characterised by new principles.

The drive structure model is not consistent with this view. It reduces psychic functioning to the level of the biological because it takes the biological instincts of an individual human and then directly translates them into drives at the psychic level. The psyche is thought of as being driven by instinctual energy seeking equilibrium through discharge just as the biological level is. Discharge of this biologically determined energy is then internally constrained by the individual as a result of that individual's reality testing, that is, the internalisation of society. The agent in the social system is thus individually constructed by the clash between its biological and social inheritance: it is not both constructed by and constructing the social system of which it is a part, on an ongoing basis, at a level that has emerged from the biological but is not reducible to it. Furthermore, the drive-structure model assumes a psyche constrained by a pre-given reality rather than the co-constructed context for behaving into that is implied by a complexity perspective.

Psychoanalytic literature does of course contain other perspectives that are much more compatible with a complexity perspective. Greenberg and Mitchell (1983) distinguish between the drive-structure model and relational-structure models of intrapsychic functioning as developed by Sullivan (1953), Winnicott (1965 & 1971) and Fairburn (1952). Relational-structure models are built on the proposition that it is the network of human relationships an individual is embedded in that structures an individual's psyche. Of course, this means that such an individual is also, through relationship patterns, taking part in the structuring of other psyches and so we come close to the notion of the co-creation of individuals and the systems they are part of through personal relationships. From the moment of conception an individual is connected to, in relationship with, another and as soon as that individual is born the network of relationships expands: the individual is then both affecting and being affected by others in an evolutionary process that continues throughout life. Continuously emerging out of this process are patterns of both adaptive, and maladaptive, individual and group survival strategies.

Moreover, the implications of a complexity perspective seem to me to be wider than a move from one psychoanalytic theory of intrapsychic functioning to another. As soon as we think about intrapsychic functioning as emerging from a network of relationships we encounter the social and the political aspects of human relating. This suggests the need for some integration of the psychological and the social ways of understanding human agency. We can get some notion of this by turning to the group analytic approach. Foulkes (1964) established the group analytic approach on three legs: psychoanalysis with its focus on the intrapsychic; sociology which he understood from the perspective of critical theory developed by the Frankfurt school; and notions of networks and fields derived from his work with the neurologist Goldstein and the gestalt psychologists. This combination of insights led him to focus on the network, or matrix of verbal and pre-verbal communications which emerge as a group comes together and develops relational patterns, where that matrix in turn affects individual functioning. He thought that the notion of the lone individual was a convenient abstraction which did not exist other than in relation to a group - he saw individuals as nodes in a network or matrix much as modern complexity theory would. This might be contrasted with Bion's view that the group is an illusion in the minds of quite independently existing individuals. It is a short step from the position adopted by Foulkes to social constructionism in which the ordinary everyday reality we all act into is seen as being constructed by the ordinary everyday conversations we engage in (Shotter, 1993). Alongside this emphasis on the constructing nature of conversation there is also the, perhaps increasing, awareness of the importance of pre-verbal communication that continues throughout life (Wright, 19991; Meares, 1992)

What becomes of the notion of an ego managing the boundary between inner drives and outer reality if we move from the drive-structure model to the relational-structure model, and even further to notions of emotional matrices and socially constructed social and psychological reality? The move I am talking about brings us very close to seeing both groups and the minds constructed by them as being complex adaptive systems. A key feature of a complex adaptive system is that there is no agent in overall control of the system. For example, the brain consists of billions of neurons each of which is stimulated by other neurons and responds to this by then stimulating yet other neurons. There is no controlling neuron. If we thought of the psyche as a complex adaptive system nesting in a wider social complex adaptive system then there would be no place for a managing ego: instead the psyche would be thought of in terms of many psychic agencies, each responding to the stimulation of others and in turn stimulating yet others, to produce emergent patterns of emotion and thought. Indeed, it seems to me that psychoanalytic thinking has been moving in this direction. Starting from Freud's theory of a mind consisting of three agents, Klein increased the population by suggesting many object representations, while Fairbairn and the ego psychologists differentiated the ego into parts and the self psychologists did the same for notions of the self. The psyche may thus be coming to be seen less and less as a system with a managing centre in the ego and more and more as a decentred system of interacting psychic agencies. A complexity perspective suggests taking this line of thinking further, viewing the psyche as a system of many agencies interacting with each other in a self organizing manner to produce emergent patterns of feeling and thought. Such an individual psychic system would have to be continually thought about in terms of its embeddedness in larger group and social systems, so in turn decentring the individual.

A complexity perspective thus suggests moving from a view of individual psychic structure arising out of the clash of drive and social constraint to a view of mutually constructed individuals and groups through evolving cooperative and competitive strategies of interaction. What does this imply for our view of the processes agents employ to relate to each other? Developments of the drive structure model stress the relational processes that distort and defend against reality testing - the processes of transference, counter transference splitting, introjection, projection and projective identification. The emphasis on these processes makes sense if our focus is the clash between drive and reality but it is excessive and insufficient if we think of cooperative and competitive processes of co-construction. Other relating processes then have to be given equal importance, processes that do not amount to psychic depletion and manipulation. For example, more attention would need to be paid to the interpersonal processes that create and transform, such as mirroring affirmation and resonance (Foulkes, 1964), giving-receiving and evocation (Bollas, 1987 & 1993) which I will return to below.

It seems to me that as we follow the line of reasoning I am suggesting we increasingly blur the boundary between external reality and work, on the one hand, and the internal basic assumption behaviour based on personal relationships, on the other. Indeed, this is what we would expect to occur when a complexity perspective is taken because the central insight it yields is the nature of the boundary between stability and disintegration. That boundary is a phase transition at the edge of disintegration, a paradoxically stable and unstable, fractal, fuzzy, irregular dynamic which, it is conjectured, is the required condition for creative evolution to occur. When systems operate in this dynamic context, self organizing processes produce emergent outcomes which may turn out to be innovative new survival strategies. However such outcomes are unknowable and agents must therefore find themselves operating in conditions where they do not know what the task is that they are to carry out - they are engaged in a process of discovering what the task is, a process of double loop or deutero learning. They are operating in a state of not knowing in which none can fully understand the whole system of which they are a part and over which no one can therefor exercise control.

This notion contrasts starkly with the Tavistock model's 'either / or' juxtaposition of the work group and the basic assumption group. A complexity perspective suggests that such a theoretical stance focuses attention away from the dynamic required for creativity. What is required is a redirection of attention to understanding what people in groups and organizations do when they do not know what they are doing because this is the prelude to the emergence of the creative. Psychoanalytic literature does have something to say on these matters. Winnicot's notion of transitional phenomena is directly relevant here. He talks about a paradoxical state in which fantasy is fused with reality as the basis of symbol formation and use, cultural and creative life. This state is anxiety provoking and can therefore be held only if there is a good enough holding environment. Those using the Tavistock model also use Winnicott's notion of a facilitative environment but, I think they see it as facilitative of work group functioning, as holding the basic assumption group at bay. What I have in mind here is the notion of a good enough holding environment that enables groups of people to operate not as a work group but as a transitional group, a group in a state of not knowing what its task is yet, a group that must therefore employ creative imagination. Perhaps even more relevant to the creative phase transition at the edge of disintegration is the Bollas notion of the transformational process in which he sees individuals being transformed by relationships. His discussion of the unthought known is highly relevant her - he is talking about the intuitive and imaginative process in which people struggle to articulate what they know at a pre-verbal level into a verbal communication.

The tendency is to see roles determined by the primary task in a rational way divorced from emotion and to see emotion principally as a feature of the roles taken up in the basic assumption group where those roles are determined by personal relationships. However, recent brain research (Damasio, 1994) demonstrates that this distinction between the rational and the emotional is not tenable since it is the same part of the brain that is activated by both emotion and the processes of rational choice.

In summary then, a complexity perspective suggests a refocusing of psychoanalytic understanding of organizations away from seeing organizations as individuals writ large with an 'either/or' emphasis on work and basic assumption groups. Instead a complexity perspective suggests a redirection of theory building effort to a relational basis which focuses on transitional and transformational phenomena and the pre-verbal, the unthought known. This suggests a shift in emphasis from the focus on interpersonal processes that distort and manipulate to a more balanced emphasis on both these processes and those other interpersonal processes that affirm and transform.

What would the practical implications of such a shift in emphasis be for consultants to organizations?

PRACTICAL IMPLICATIONS FOR CONSULTING

From a Tavistock perspective the role of the consultant is that of participant-observer. In this role the consultant employs her or his counter transference feelings to formulate hypotheses about the transferential and projective processes at work in the organization, about the impact of basic assumption behaviour on the work of that organization. It is the function of the consultant to feed those interpretive hypotheses into the life of the organization and so foster a collaborative, negotiated understanding and verbalization of the unconscious, irrational processes at play. It is believed that this process enables the retraction of projections and distorted impressions of reality so restoring to the group its work function. So, consultants are here seeking to engage with and understand the complexity of organizational life by adopting an interpretive stance (Shapiro and Wesley Carr, 1991). This stance is seen as the most important element in creating a holding environment: 'containment and holding ordinarily refer to symbolic interpretive ways in which the therapist manages the patient's (and his own) feelings ..' (Shapiro and Wesley Carr, 1991, p112). Another feature of the holding environment, one that interpretation aims to secure, is the clarity of task, boundary and role - this is seen as containing sexual and aggressive feelings. Empathic interpretation affirms individuals in their roles and the resulting containment establishes a holding environment which provides for safe regression, a shift from rationally organized words to the primitive distortions of fantasy images and simple metaphors which can then be articulated and so disarmed. The aim of interpretation is to move people from purely personal relationships that distort work to roles that support it.

Those working from a relational-structure model tend to de-emphasize interpretation (Greenberg and Mitchell, 1987) since they see the quality of relationship as the important factor. Kohut for example, ascribes great importance to specific behaviours of a therapist, emphasising the therapeutic effect of the patient's experience of the therapist's mirroring and allowing himself to be idealized. In the group analytic tradition, based firmly on the notion that an individual is permeated to the core by the collective, also downplays interpretation. Foulkes held that interpretation follows change rather than precedes it and that it should be used to articulate a pattern just as it is emerging. Instead of interpretation as a primary technique he saw the group conductor's first aim as that of facilitating participation in group life by the members of the group. In group analysis the therapeutic factors are: socialization, the process of sharing experience; mirroring, the process of seeing oneself reflected in others; resonance which is unconscious, or pre-verbal, communication between people; and translation or interpretation, making the unconscious conscious.

Bollas attaches importance to patients developing a capacity for self analysis. He sees this as requiring tranquillity, creating the conditions for the arrival of insight through reception and evocation. This means the avoidance of interpretive activity, rather waiting patiently with the patient to evoke what is within him or her. Through this process he sees images and ideas emerging where they had not existed before. The therapist is then not a container but an auxiliary in the evocation of new inner experience. Bollas is talking here about tapping a form of knowledge that permeates a person's being but is not yet known. The analyst assists in the process of evocation by being there as a presence, strictly speaking neither inside nor outside the analysand's mind.

From a technical point of view, then, the theoretical shift implied by a complexity perspective would be accompanied by a practical shift from the emphasis on interpretation and containment by task, boundary and role clarity. It would shift to an emphasis on the consultant as an auxiliary in the organizational evocation process in which groups work in an organization's unthought known, the pre-verbal, existential knowledge embodied in its patterns of relationship. Holding is then provide by these patterns of relationships which the consultant contributes to by being there - it is the group matrix that is the container. Perhaps the most important part of consultants' being there is the way they work and the practices and behaviours that this helps to affirm. The emphasis shifts away from concern with what the primary task, boundary and roles are to how, through patterns of personal relationship, people may cope with the dynamics at the edge of disintegration where they struggle to give expression to what they might know but cannot yet think.

I can perhaps best illustrate this shift by taking a case that Shapiro and Wesley Carr (1991) describe. The case concerns a consultation to the Adolescent and Family Treatment and Study Center within the McLean Hospital in Massachusetts. The organizational problem had to do with staff members feeling isolated from one another and finding it difficult to find a way of collaborating to further develop the Center's treatment, training and research. The presence of the consultant seemed to free people to talk in unfamiliar ways, even though, as he notes, he had not yet done anything. Furthermore, in his conversations the consultant received many personal confidences as if those working with disturbed families and adolescents could find no way in their organizational life to share their own disturbances in personal and family life. This was despite their reporting that no one was able to say or do anything without being analysed because every person's behaviour was assumed to provide some evidence relating to the clinical treatment of patients. The one person who could be relied upon not to interpret was the program secretary who then became the repository of personal confidences. In addition to the secretary's desk at Reception their was another 'interpretation free' zone in the nurse's station. In both places relaxed and 'interpretation free' conversation amongst the staff took place.

However, because of the all-inclusive theory of the treatment program the consultant regarded these zones as intrinsically illegitimate, even subversive, and in doing so he might have been reflecting what the members of this organization themselves thought. The consultant concluded that these informal subgroups existed because people felt unsure about their roles and were therefore having difficulty finding appropriate means for sharing feelings about their roles. The informal subgroups were seen as a compensation for some lack in the formal system and the consultant saw this as a managerial obstacle to the theory all accepted about their way of working which required them to think through the roles of patients in their families and their own roles in relation to the Center as a family parallelling the families the Center was dealing with. The informal groups were thought to have become waste containers in which important affective communication was simply discharged instead of being examined. He thought that there was a struggle between staff in their formal roles and in their roles as individuals because they were being instructed to share experiences rather than negotiating such sharing. In addition the consultant identified a lack of organizational clarity in the relationship between different units that made up the Center. He furthermore stressed unclear chains of authorization for people to take up their roles.

The consultant concluded that the organization was not functioning according to its design, for all the reasons set out above, and therefore considered how tasks could be understood by all so that roles within and across units could be legitimately authorized and fully integrated. The consultant recommended clarification of authorization from one level to another in the hierarchy and established structured meetings, a policy council and a management team, to promote effective communication. To deal with the feelings of the staff in a legitimate manner, the consultant stressed the need to develop a culture, with the authorization of the Director of the Center, in which interpretation was to be encouraged but only within limits relevant to the work. People would then bring their feelings to legitimate forums where they could be made available for examination in relation to the work rather than discharged in informal subgroups - such examination would constitute an interpretative stance, that is, a collaborative verbalization of unconscious processes leading to withdrawal of projections that might be adversely affecting task performance. The objection to the subgroups thus seems to be based on the belief that, since they are based purely on personal relationships rather than task, they are fertile ground for projections and basic assumption behaviour. Later, an invitation by a section head to the Director to take a consulting role in her unit was seen as evidence of the internalization of the interpretative stance. However, as the organization experienced further change it was noticed that the secretary became even more active in her informal role.

Note how the consultant's model of organizational functioning leads him to focus on the legitimate system of the organization - he recommends the clarification of formal tasks, the authorization of formal roles, the authorization of culture and the 'removal' of informal subgrouping through more formal participation, that is, through adopting an open, collaborative and public interpretative stance. He notes that people seem to develop new ways of talking but attaches little importance to it. The significance of the increasing activity of the secretary is not explored.

A consultant working from the kind of complexity perspective outlined above would make sense of the situation at the Center in a very different way and would, therefore, operate in a different manner. The manifestations of this organization's shadow system, the 'interpretation free zones' would not be regarded as evidence of organizational malfunctioning but potentially, anyway, an essential part of its flexibility and the ultimate source of its creativity and ability to change. So, although the consultant would formally contract with the legitimate system he or she would contract to work primarily with the shadow system and by in effect joining and taking part in this shadow system would attempt to understand its dynamics. However, this understanding would not normally be used to interpret since the approach is not primarily educative but participative. Through participation the consultant would be affirming the usefulness of informal subgroup activity based on personal relationships and seeking to seed further activity of this kind - a description of how this was done in a local government organization in the UK is contained in a paper about to be published (Shaw, 1997).

One of the main purposes in joining the shadow system in this way would be to foster new ways of conversing since it is in ordinary everyday conversation that we construct organizational realities and as we saw in the above case this is often done simply by being there. This view is in keeping with evidence on how, far from being a distraction, personal relationship based subgroups constitute communities of practice and in their informal conversations embody the most up-to-date knowledge, the frontiers of organizational learning (Duguid & Brown, 1991). The complexity perspective I am suggesting sees the shadow system as essential to organizational flexibility and creativity, the place where the unthought known might come to be thought, the activity that might be saying something important about organizational malfunctioning and be the seedbed for changing it. For example, the shadow activity at the secretary's desk might be suggesting that an element of the officially accepted way of working in the Center is actually not accepted, perhaps because it is not all that appropriate. The consultant might then take part in conversations about whether it is appropriate for professionals in a Center providing services to families to think about that Center as a family and themselves as members of it. After all, organizations are not families other than in fantasy.

When one operates in the shadow system in the manner I am suggesting the question of authorization becomes unimportant. The shadow system is self organizing and people spontaneously take up roles in it on the basis of unspoken mutual consent and usefulness.

Does this approach to consulting mean that the consultant does not work with the unconscious? I do not think so. The consultant would still notice projective activity when it was apparent and also use his or her counter transference feelings when these were strong. I suspect, though, that there would not be an a priori expectation of strong transferences or projections. The consultant working from the perspective I am describing would also be aware of other forms of psychic relating such as the mirroring, resonance and evocation I mentioned earlier on. However, the difference is that such feelings are part of a total participation in the shadow system and not used primarily for interpretive purposes. This is quite in line with the implications of a shift from a drive-structure to a relational-structure way of seeing things. The kind of approach to consulting I have been describing here is being actively developed in commercial organizations by my colleagues at the Complexity and Management Centre, Douglas Griffin and Patricia Shaw.

A short example from a recent experience I had with a team of managers on a management development program might help further to clarify what I am suggesting about the shadow system. These managers described the pressure from those above them to get them to apply pressure to the account manages below them in the hierarchy so that those account managers reported all the opportunities they were pursuing. Sophisticated information technology made this a simple matter and all the systems were ready to receive the information. However, all knew that despite the pressure only part of the pool of opportunities was being reported. The managers I was working with were looking for ways in which they could overcome this resistance and meet the demands of their superiors. When they began to look at this issue from the perspective of the shadow system they considered what they knew informally about why the opportunities were not being reported. They knew why this was happening because they themselves had been account managers and their superiors knew why for exactly the same reason, but the reason was not discussable in formal settings. They all knew informally that account managers could not be sure that all of the opportunities would materialise but those account managers knew that they would be held to account when an opportunity was missed. They also knew that the more top mangers knew about their activity the more rapidly those top managers would increase target sales. With this informal system in place there seems little chance that all opportunities will ever be reported upon. The question then becomes why they should be. The account managers are the agents operating at the workface, as it were, and it is important that they know what opportunities they are pursuing. To operate effectively they need their own local flexibility. The question then becomes what kind of information system can be formally designed that will take account of the inevitable shadow system activity. In this case the shadow system, is clearly telling all concerned that the proposed information system simply will not work as designed. But instead of trying to interpret the shadow behaviour in the interest of the legitimate system I would adapt the legitimate system to the message emerging from the shadow system.

CONCLUSION

I have argued that a widespread process seems to be underway in many different disciplines that suggests a major shift in Western thinking away from Newtonian and neo-Darwinian frames of reference toward a complexity perspective. Since psychoanalysis originated in the former frame of reference it seems to me to be important to explore how developments in psychoanalytic thinking might be reflecting this shift and what implications there are in this shift for its further development. I have argued that the same point applies to organizational and management theory. When psychoanalytic thinking is brought to bear on organizations we find a coherent and widely discussed model, the Tavistock model, which is very much within the paradigm that we seem to be moving away from. In this paper I have suggested how the model might move on by taking a complexity perspective on psychoanalytic ways of understanding organizations.

Bion, W. R. (1961), Experiences in Groups and Other Papers, London: Tavistock Publications.

Bollas, C. (1987), The Shadow of the Object, London: Free Association Press.

------------- (1993), On Being a Character, London: Routledge.

Damasio, A. R. (1994), Descartes' Error: Emotion, Reason and the Brain, London: Picador

Duguid P. & Brown, J. S. (1991), 'Organizational Learning and Communities of Practice: Toward a Unified View of Working, Learning and Innovation,' Organisational Science, Vol 2, No 1, Feb, 40­57.

Ehrlich H. S. (1996), Ego and Self in Group, Group Analysis, June, 29, 2.

Fairbairn, W. R. D. (1952), Psychoanalytic Studies of the Personality.

Foulkes, S. H. (1964), Therapeutic Group Analysis, London: George Allen & Unwin.

Freud. S. (1923), 'The Ego and the Id' in the Penguin Freud Library Volume 11, (1964), London: Penguin.

Gell­Mann, (1994), The Quark and the Jaguar, New York: Freeman & Co.

Gleick, J. (1988), Chaos: The Making of a New Science, London: William Heinemann Limited.

Goodwin, B. (1994), How the Leopard Changed its Spots, London: Weidenfeld and Nicholson.

Greenberg J. R. & Mitchell S. A. (1983), Cambridge Ma: Harvard University Press.

Hirschhorn, L. (1990), The Workplace Within: Psychodynamics of Organizational Life, Cambridge Ma: MIT Press.

Jacques, E. (1955), Social Systems as a Defence against Persecutory and Defensive Anxiety, in Klein M., Heimann, P. & Money­Kyrle, P. eds. (1955), New Directions in Psychoanalysis, London: Tavistock Publications.

Kauffman, S. A.(1993), Origins of Order: Self Organization and Selection in Evolution. Oxford: Oxford University Press.

-------------- (1995), at Home in the Universe, New York, Oxford University Press.

Klein, M (1975), The Writings of Melanie Klein, London: Hogarth Press.

Levy, S. (1992), Artificial Life, New York: First Vintage Books.

Meares, R (1992), The Metaphor of Play, Melbourne: Hill of Content.

Menzies Lyth, I (1975), A Case Study in the Functioning of Social Systems as a Defence Against Anxiety, in Coleman, A. & Bexton, W. H. eds. (1975), Group Relations Reader, Sausalito, Ca: GREX.

Miller, E. J. & Rice, A.K. (1967), Systems of Organization: The Control of Task and Sentient Boundaries, London: Tavistock Publications.

Prigogine, I. & Stengers, I. (1984), Order Out of Chaos: Man's new Dialogue with Nature, Bantam.

Shapiro E. R. & Carr, Wesley A.(1995), Lost in Familiar Places, New Haven, Mass: Yale University Press.

Shaw, P. (1997), Intervening in Shadow Systems of Organisation: Consulting from a Complexity Perspective, CMC Working Paper No 4.

Shotter, J (1993), Conversational Realities, Thousand Oaks Ca: Sage.

Stacey, R ( (1992), Managing the Unknowable: The Strategic Boundaries Between Order and Chaos, San Francisco: Jossey Bass. Also published in the UK as Managing Chaos, London: Kogan Page.

---------- (1993, 1996), Strategic Management and Organisational Dynamics, London: Pitman.

------------- (1996), Complexity and Creativity in Organizations, San Francisco: Berrett-Koehler.

Sullivan, H.S. (1953), The Interpersonal Theory of Psychiatry, New York: Norton.

Trist, E. L. & Branforth, K. W. (1951), 'Some Social and Psychological Consequences of the Long­wall Method of Coal Getting', Human Relations, 5, 6­24.

von Bertalanfy, L. (1968), General Systems Theory, Foundations, Development and Application, New York: George Braziller.

Waldorp, M. M. (1992), Complexity: The Emerging Science at the Edge of Chaos, Englewood Cliffs, NJ: Simon & Schuster.

Winnicott, D. W. (1965), The Maturational Process and the Facilitating Environment, London: Hogarth Press. Reprinted 1990 by Karnac (Books) Ltd.

--------------------- (1971), Playing and Reality, London: Tavistock Publications. Reprinted 1993, London Routledge.

Wright, K. (1991) Vision and Separation, Northvale NJ: Jason Aronson Inc.