SELF-ORGANISING SYSTEMS

One of the most recent perspectives on studying organisations is self-organising sys-tems. Recent years have witnessed an increasing interest in the study of chaos⁴ and the science of complexity and their applications in various areas including business and organisational sciences. When such insights are applied to real-world systems, e.g. to busi-ness organisations, they shed light on dynamic processes of adaptation and survival. Seen through the lenses of chaos and complexity theories, business organisations come to be regarded as dynamic open complex adaptive systems (DOCAS), composed of inter-related parts, perhaps attaining a sustainable advantage, and in this whole process generating certain emergent phenomena.⁵ Concepts such as self-organising phase transitions, co-evolutionary adaptation, and movements toward an optimal location at a performance peak on a changing fi tness landscape (co-adaptation) spring to mind as ways of characterizing business phenomena like synergy in mergers and acquisitions or successful product innovations. (Holbrook, 2003)

The theory of autopoiesis – self-production – offers one perspective on changing and transforming organisational life. The theory was originally introduced by Matu-rana and Varela (1980, 1988). Living systems are systems that maintain themselves by producing components and organisation, that is, the system itself (Maturana &

Varela, 1980, p. 79, Di Paolo, 2005, p. 434). Structural couplings are “encounters with the environment resulting in perturbations to the autopoietic dynamics without loss of organization”. The concept of structural coupling is the basis of all that autopoietic theory has to say about cognition. By nature structural couplings are conservative, not improving processes. (Di Paolo, 2005, pp. 436-437, Maturana, 1975 in Di Paolo, 2005, Beer, 2004) In this study, this biological view of self-organising; autopoiesis, or self-regulation, self-correction is not built upon. In my ontological understanding human consciousness and purpose-orientation need to be taken into account when considering communities consisting of human beings. They cannot be reduced to the mere automatic search for adaptation, equilibrium and the survival of an organ-ism.⁶ Cognition requires “a natural centre of activity in the world as well as a natural 4. Chaos refers to the phenomenon wherein systems composed of inter-related parts or interdependent agents – each of which follows simple, regular rules of behaviour – generate outcomes that refl ect these interactions and feedback effects in ways that are nonlinear and unpredictable. Because of nonlinearities that refl ect interactions and feedback effects, tiny changes in inputs can make enormous differences in outputs. This concerns the sensitive dependence on initial conditions or the so-called butterfl y effect. (Holbrook, 2003, p. i)

5. See Schneider & Somers (2006) for a summary of complexity theory origins.

6. The theory of Maturana & Varela (1980, 1988) was originally defi ned for the contexts of biological systems. In the literature on autopoietic systems, the question of the nature of social systems has received considerable attention and also created a great deal of confusion. Originally Maturana’s and Varela’s questions were about the nature of autonomous living units such as cells and multicellular organisms. In their view living systems are systems that maintain themselves by producing components and organisation, that is, the system itself (see Maturana & Varela, 1980, p. 79, Di Paolo, 2005, p. 434). Extending the theory of autopoiesis to social systems is problematic because social systems are not biological organisms (see Maula, 1999 for one application of autopoiesis to organisations). Tuomi (1999, p. 192) claims that there are basically two ways to apply autopoietic theory to social systems. Firstly, it is possible to argue that societies, properly conceptualized, are autopoietic systems. This is the view selected by Luhmann (1989), who deals with autopoietic social systems. Secondly, some fundamental aspects of societies may be understood in the light of the theory

perspective to it”. Structural coupling alone cannot ground the concept of activity and autopoiesis alone cannot ground the concept of perspective (Di Paolo, 2005, p.

443). Thus, I will use the terms self-organising or self-steering instead of self-regula-tion, self-correction or autopoiesis. (See e.g. Järvinen, 2004, pp. 132-134 about the classifi cation of causal systems⁷)

In my view, some of the fundamental aspects of successful boundaryless work can be understood in light of self-organising. Conversely, successful boundaryless work requires certain features of self-organising. The boundaries in the organisation need to be fl exible enough to allow self-organising to take place. Thus, the perspective of self-organising systems brings into the picture organisations that are well tuned to the outside world, and are thus better at uncovering and exploring potential opportuni-ties and threats. In boundaryless contexts, the principle of self-organising brings the aspect of freedom into the picture. When the traditional hierarchies have eroded, it is the people in self-organising systems who take charge of the agile coordination in the context of boundaryless work. The vocabulary, and thus the mental models evoked and used in the studies arising from chaos and complexity theories are very different from the traditional linear understanding that has for so long been in active use and also the basis for organising work. Hatch (1998, p. 557) describes 21st century organisations using words like fl exible, adaptable, and responsive to the environment, loose boundaries and minimal hierarchy. Mirvis (1998) describes the dialectical emergence in organisa-tions in which “the new is latent in the old and seems to emerge naturally through the creative resolution of contrary forces”.

Indeed, when looking at the overall organisation through the lenses of chaos and complexity theories, business organisations come to be regarded as dynamic, open, complex, adaptive systems that are composed of inter-related parts interacting with its environment, subject to resulting feedback effects, evolving over time adaptively to fi t the pressures imposed on it. In DOCAS (dynamic, open, complex, adaptive system) there are hierarchically nested levels. Figure 11 features an example of hierarchically nested levels of a dynamic open complex adaptive system. The systems, networks, or holistic combinations can occur at any level of hierarchical organisation within an overall nested structure. At any given level of organisation, via the nesting of systems, a DOCAS will be composed of lower-level micro networks and will, in turn be

em-of autopoiesis. Maturana and Varela (1980, 1988) have opted for this view. Tuomi’s (1999, p. 413) own conclusion and actually one of the major results of his dissertation is that the basic thesis of the autopoietic theory can not be right: living systems can not be strictly autopoietic. This also leads him to a question whether social systems can really be autopoietic. To overcome this problem he developed the idea of “almost autopoietic system” and defi ned a social system as a “self-maintaining meaning processing system”. In my ontological understanding the concept of “almost autopoietic systems” cannot bridge the gap between the biological sense of fi nding equilibrium and human consciousness and pur-pose-orientation. Di Paolo (2005) in his article examines the mediacy between organism and environment, autopoiesis, adaptivity, sensemaking and cognition (see also Weber & Varela, 2002).

7. Systems with a full causal recursion can be divided into self-steering systems, self-regulating systems, and systems steerable from outside and disintegrating systems. Goal-oriented nature of thought process is typical of human consciousness. The causal process that takes place in human mind in an alert state is the self-steering process, whereas living organisms are self-regulating systems. (Järvinen, 2003, pp. 132-134)

bedded within a higher-level macrostructure. The problem is to relate these multiple nested and interwoven systems so that, via modifi cations in the lower-level micro networks (the parts), the DOCAS (the whole) evolves to achieve a better fi t with the higher-level macrostructure (the surrounding context) – which is in turn composed of other DOCAS-level networks giving rise to processes of co-evolutions. (Holbrook, 2003, p. 22) In Holbrook’s exposition of the nested levels, the collision of biology and understanding human being as a conscious and intelligent is again visible. The term “adaptive” is derived from biology, and thus not easily attached to human beings who are capable of improvising in order to move towards a certain target. Biological phenomena with their adaptive orientation towards equilibrium and survival are more predictable than human beings who, with their own consciousness, are more unpre-dictable. (The direction of the organisation or certain parts of it may be something other than equilibrium and survival if human beings so choose.) The nested levels in Holbrook’s picture depict well the interrelatedness and nested nature of various parts of the organisation (for Holbrook organism). However, even within this picture the biological concepts of atom, molecule, cell, organism, species, ecosystem, biosphere and cosmos are mixed with the concept of community that is a creation of human beings with the above described features of object-orientation. Thus, for me, human beings can choose how they behave and act; the behaviour and practices based on

Atom Molecule Cell Organism Species Community Ecosystem Biosphere

Cosmos

Figure 11. Hierarchically nested levels in DOCAS (dynamic, open, complex, adaptive system) (Holbrook, 2003, p. 23)

improvisation is not thus involuntary, automatic behaviour but rather a consciously selected way to behave and act.⁸

There are many claims concerning the superiority of self-organising systems when it comes to turbulent and chaotic business companies. However, as noted earlier, at-tempts to organise work based on these principles are still rare (cf. Dyer & Ericksen, 2005). For example, Foster & Kaplan (2001) claim that in turbulent and unpredict-able environments, self-organising systems are superior to top-down organisations at allocating resources to their most productive uses. They are supposedly also relatively unemotional and apolitical about abandoning legacy commitments that no longer make sense.

Barrett (1998) guides managers in constantly evolving complex adaptive systems to escape all limits of their own patterned routines, to embrace errors and turn them into opportunities, to establish minimal structures that permit maximum fl exibility, to achieve a state of dynamic synchronization, to combine materials into retrospective sense-making and fi nally to play both leading and supportive roles. Barrett’s guidance is for contexts where a “group of diverse specialists live in a chaotic turbulent environ-ment; making fast, irreversible decisions; highly interdependent on one another to interpret equivocal information”. They “fabricate and invent novel responses without a pre-scripted plan and without certainty of outcomes; discovering the future that their action creates as it unfolds”. Weick (1998) uses the word improvisation about “order and control that are breached extemporaneously while the new order is being created.”

His conjecture is that improvisation may indeed be part of the infrastructure present in all organisations. Crossan (1998) ties the role of improvisation directly to the nature of chaos: “The value of improvisation is in the potential it holds to enhance the quality of spontaneous action… Improvisation is one of the few concepts and tools we have to develop to be innovative in the moment – a key requirement of organizations in the twenty-fi rst century”. For Crossan (1998) improvisation is more than a metaphor. It is an orientation and a technique to enhance the strategic renewal of an organisation.

It is an extension of more traditional skills.

In self-organising systems “quick and novel out-of-the-box responses are more fruitful than slower and more conventional in-the box approaches” (Holbrook, 2003, p. 7) Based on Chelaniu et al. (2002, p. 146) improvisation is a more suitable response to rapidly changing environments than planning. “In uncertain, complex and rapidly changing environments detailed planning may be a waste of time and resources…dan-gerous… perilous.” (Chelaniu et al., 2002, p. 146) I agree with Chelaniu to a certain 8. See for activity theoretical explanation for the difference between the adaptation in biological sense and human activity in chapter 4.1. Animal activity has an adaptive nature; it does not have the capacity for making, utilizing and preserving tools systematically. The breakthrough to human cultural evolution and human form of activity happened when the

“emerging mediators” became “unifi ed determining factors”. At the same time what used to be ecological and natural became economic and historical. “What used to be adaptive activity is transformed into consumption and subordinate to the three dominant aspects of human activity – production, distribution and exchange (or communication).” (Engeström, 1987, pp. 74-80)

degree but also concede the importance of high-level plans and a certain level of clarity on the organizational goals. Plans are needed for coordination but there still needs to be a capability to change, adapt and renew plans based on the need and situation.

What is needed is the continuous alignment of effort. In Orlikowski’s (2002) case study, development work in distributed environment calls for effi cient and continu-ally ongoing coordination. In her case the alignment was done via consistent use of proprietary project management models and tools. Too much planning might discourage improvisation. Balancing between planning and improvisation is indeed about aligning effort.

Moorman & Miner (1998, pp. 1-5) defi ne improvisation in product development as a “convergence of composition and execution where planning and performance or design and implementation occur simultaneously. It is about a collective system of interac-tion that creates and enacts the scene simultaneously… The joint activities of individual people create a collective system of improvisational action. The occurrence and effective-ness of such improvisation are likely to refl ect the impact of environmental turbulence and timely information fl ows concerning internal or external surprises.” This view is related to the parallel activation of the whole network to an effi cient “concerted effort”

(cf. Docherty et al. 2002, p. 6).

The emergence of self-organising systems is often observed in environments characterized by crisis: in hospital emergency rooms, among electricity teams after a hurricane or ice storm, and in the military units cut off from their normal chains of command (Pascale et al., 2000, pp. 135-147).⁹ Dyer & Ericksen (2005) state that very similar crisis situation behaviours should take place in agile organisations. In such situations, the people involved are expected to take personal responsibility for de-termining and deciding what needs to be done and how. Their behaviours may include improvising solutions to remove unanticipated obstacles, obtaining necessary informa-tion and resources, focusing “furiously” on the task at hand, and disengaging when their contributions are no longer required. Further their behaviours may include spontaneous collaboration with other people who happen to be needed to complete the task or to fi nd some relevant information.¹⁰ This concurs with Orlikowski’s (2002) collective recreation of knowing in practice. It is equally compatible with Döös et al.’s (2005) results from an R & D company also described as a “loose state”, where the engineers were trying to “catch” the latest situation and to “collect bits and pieces of information and relations between them.” Even Blackler (1995), when introducing his typology of organisations based on various types of knowledge, maintained that “ad hocracy”

is a feature of communication-intensive organisations. Ad hocracy as a way of

work-9. More recently the U.S. army has started to use this approach as a routine way of operating in certain combat condi-tions. (Pascale et al., 2000, p. 135-147 ). See also Sonnenwald & Pierce (2000) for situational awareness in battle situations.

10. One needs to add to Dyer & Ericksen’s reasoning that the people involved need to have attained a certain level of maturity to be able to act in a self-organising manner.

ing is close to the mental models evoked by the improvisation and spontaneity of self-organising systems.¹¹

I will next present Dyer’s and Ericksen’s (2005) framework and principles for

“managing” human resources in an organisation based on self-organising. Their idea is to present a model for human resource scalability that is based on the idea of self-organising system. (In the context of self-organising systems the verb “manage”

should be understood as creating conditions for the right kind of behaviour.) They base their reasoning on the effi ciency and fl exibility of self-organising systems. The basic idea is in creating favourable conditions that would allow discretionary work design where fl exible job roles and fl exible boundary crossings to new job roles are enabled. An optimal system seeks to promote both freedom and fl exibility and also enough discipline and order to keep the system viable. The goal is to create limitless opportunities for employee initiative, at the same time directing and restraining the chaos that can result from the pursuit of boundless opportunities. Dyer’s and Ericksen’s (2005) model (see Figure 12) is one of the rare attempts to apply the principles of self-organising to how to organise work in a fi rm. They themselves call their attempt

“modest, tentative and somewhat ephemeral” and call for more real-life experimenta-tion as well as research into self-organising systems. This model seems also to contain many features of a learning organisation (cf. Argyris, 1999, Argyris & Schön, 1978).

Kontoghiorghes & al.’s (2005) study suggests that organisational designs based on the principles of connectivity, redundancy, and self-organisation facilitate innovation and rapid change adaptation.

For Dyer & Ericksen (2005, p. 184) optimizing internal fl uidity is the main is-sue in the human resources management in self-organising systems. Internal fl uidity refers to the ease and speed with which the continuous self-allocation of existing talent and effort occurs. Self-organising systems “cannot be managed or directed, only nudged and disturbed”. This type of emergence happens best when the system is operating “at the edge of chaos”. This concept needs to be perceived as a state attained when forces favouring initiative, spontaneity, and improvisation are delicately but paradoxically bal-anced with forces favouring focus and direction. The task of management is to provide and continuously revise a bare minimum of guiding principles that both promote freedom and fl exibility and at the same time provide enough discipline and order to keep the system from spinning out of control. Thus, self-organising is not automatic, non-voluntary behaviour; there are always human beings with their consciousness behind the nudging, disturbing and improvising. Tharumarajah (2003) is aligned with Dyer & Ericksen’s thinking when saying that “if we are to design enterprises that are highly adaptable and self-organising, consideration should be given to the design that brings out such behaviours.” Doz & Kosonen (2008a) look at the same thing from the 11. Dusya & Crossan (2005) remind that improvisation does not always lead to positive performance. They also remind that improvisation is spontaneous but spontaneity tends to be overemphasized; improvisation relies on rules and routines pre-established in the organisation.

strategic and managerial perspective: “fast decisions in complex environments call for rapid resource deployment for their implementation. Since choices and commitments cannot be decided and planned well ahead of time, reactivity is needed: resource com-mitments need to be sudden and vigorous” (p. 29). They further state that resources like funding and investments are much easier to reallocate compared to for example competencies that actually are sticky both in location (cannot be moved easily) and in time (cannot be grown or redeployed fast) (pp. 31-32).¹²

Figure 12. Context for fostering human resource scalability in an organisation based on self-organising (Dyer & Ericksen, 2005, p. 185)

The fi rst of Dyer’s and Ericksen’s guiding principles is based on the idea that the fl uid organisation is defi ned according to what everyone in the organisation does rather than a place where they all do it. Hierarchies should be minimized and instead leadership should be forced to emerge when and where it is needed. The mental model should

The fi rst of Dyer’s and Ericksen’s guiding principles is based on the idea that the fl uid organisation is defi ned according to what everyone in the organisation does rather than a place where they all do it. Hierarchies should be minimized and instead leadership should be forced to emerge when and where it is needed. The mental model should