The use of the concept autopoiesis in the theory of viable systems
Brocklesby J. & Mingers J. (2005) The use of the concept autopoiesis in the theory of viable systems. Systems Research and Behavioral Science 22(1): 3–9. Available at http://cepa.info/2808
Table of Contents
Introduction: Autopoiesis and viable systems
This paper examines the application and usage of the idea of autopoiesis – a theory of living systems – within the context of viable systems theory. In recent years the term autopoiesis has extended beyond the domain of cellular biology where it originated and is now used extensively across a range of different disciplines, fields of enquiry, and professional practice. The paper seeks to provide clarification of a distinction between viable and autopoietic systems that appears to have become somewhat clouded as a result of the transfer of ideas and terminology from one domain to another.
Key words: Viable systems, autopoiesis, living systems, autonomous systems, cybernetics.
Introduction: Autopoiesis and viable systems
This paper examines how the idea of autopoiesis is deployed within the context of viable systems theory and contrasts this with its application in the field of biology where it originated.
At the outset it is important to acknowledge that there is a close connection between viable systems theory (Beer, 1979, 1981) and the concept of autopoiesis (Maturana and Varela, 1980), and there are some fairly obvious surface similarities. Since much of the inspiration for the development of the VSM arises out of Beer’s interpretation of the operation of the human nervous system, and since the latter is a living system that we know a good deal about, this intersection between the two sets of ideas should not come as a major surprise.
Beer’s choice of terminology also speaks of overlapping concerns. The term viability itself has biological connotations, and the vernacular use of the term survival which is used extensively in the VSM literature has to do with maintaining life. The central characteristic of viability – maintaining an independent existence – is similar to Maturana and Varela’s insistence that living systems are separate from and exhibit autonomy from the medium in which they exist. Hence all three authors speak of their respective systems as being ‘subject to their own laws’ (Beer, 1981: 103; Maturana and Varela, 1987: 48).
In addition to these natural intersections it is also important to acknowledge that Beer cites autopoiesis as a useful set of ideas for the VSM, and for practical applications such as the well- known Cybersyn project. He endorses Maturana and Varela’s work in the preface to the seminal Autopoiesis and Cognition (1980), and further claims that Maturana provides: ‘the quintessence of this (i.e. viable systems) cybernetic thinking’ (Beer, 1979: 405).
More widely, the term autopoiesis itself crops up from time to time in descriptions and applications of the viable systems model (Beer, 1979: 405-408; Flood and Jackson, 1991: 96; Yolles and Guo, 2003: 188). Finally, VSM researchers such as Espejo and Harnden (1989, for example) have articulated an intervention epistemology for the use of the viable systems model that derives from the phenomenological (as distinct from biological) aspects of autopoietic theory (see Maturana, 1988).
In extending the basic resonances between his thinking and that of Maturana, Beer remarks:
Naturally I had very closely compared the conditions for life as expounded by the theory of autopoiesis with the conditions for a viable system … to me they were complementary and mutually enriching... (Beer, 1981: 338)
Cementing the link between the two theories further, he says that we may argue that ‘a viable system is autopoietic (and) the autopoietic faculty ‘… is embodied in the totality and in its systems one …’ (Beer, 1981: 338).
In the light, then, of Beer’s conviction that a viable system is an autopoietic system, juxtaposing the two sets of theories involves assessing the grounds on which Beer makes this claim. Inevitably, since he claims that complex systems such as enterprises and social systems are also viable systems, this leads into territory that has been well traversed by others (see, for example, Zeleny and Hufford, 1992; Hejl, 1984; Mingers, 1992).
It is possible to use Maturana’s distinction between the ‘specification’ of a system, the basic form (or ‘organization’) that the specification implies, and the specific systemic forms (‘structure’) through which the basic form might be realized, to show that although autopoietic and viable systems have some features in common, they are nonetheless fundamentally different entities.
In his major phenomenological work, Maturana (1988) claims that objects do not have an independent existence; rather they are constituted through acts of distinction made by the observer. On this basis we can surmise that when an observer specifies some real-world entity as a ‘viable system’, he or she does this using the various concepts and terminology of cybernetics and the viable system model, i.e. Systems 1-5, meta-system, variety amplification/attenuation etc. If one were not to use such distinctions then it would be anathema to claim that the object ‘brought forth’ was indeed a viable system.
In contrast, specifying the entity ‘living system’ implies a basic form – that is, constituted through a specific dynamic or, using Maturana’s words, a specific ‘manner of relating’ of molecules. This involves a circular self-producing process in which molecules produce networks of molecules that, in turn, produce more molecules. Through such means, the class identity ‘autopoiesis’ specifies the single abstract entity ‘living system’ which, in the concrete, can be realized in a myriad of different (‘structural’) ways.
In this very basic sense, autopoietic and viable systems are fundamentally different entities. However, in fairness to Beer, it is easy to see where the confusion might have arisen. In explaining this it helps to begin by drawing attention to Varela’s (1981: 103) claim that potentially there are many examples of systems that are ‘capable of specifying their own laws’, and therefore can be regarded as exhibiting autonomy. Under such a conception, autopoiesis and viable systems do indeed arise as autonomous systems, i.e. as specialized subsets of a more general class entity. Yet, as we shall now aim to demonstrate, autopoietic and viable systems are clearly not autonomous in the same way.
Maturana’s interest in the autonomy of living systems developed out of experimental work that he and others carried out in the 1960s. Subsequently he sought an explanation to support a conviction that the purpose of a living system is internal to it, and to counter an idea – apparently prevalent at the time – that human beings are relatively more independent from their environments than are other living systems. In this context, autopoiesis – components producing networks of components which produce more components – arises as the central mechanism that provides all living systems with a degree of autonomy and independence.
This, of course, is not to say that the environment is unimportant. The canonical example of the individually autonomous living system – the single biological cell – is not completely autonomous because it is embedded in a multicellular organism which in turn exists in, and is affected by, an environment. The cell’s autonomy has to do with its bounded self-defined nature, not its complete insulation from the systems in which it is embedded.
In contrast, as one might expect for someone with a cybernetics background, when Beer speaks of autonomy he means that the system is responsible for its own regulation. In complex systems, this means maintaining a stable internal environment – ‘homeostasis’ – in the face of environmental complexity. The latter is measured using the cybernetic concept ‘variety’ (Ashby, 1952). Under conditions of high external variety a requisite amount of internal variety is necessary; hence there is a ceding of autonomy to the various subsystems that ‘connect’ with the ‘outside world’. The human body, for example, has an autonomous nervous system because control mechanisms cannot be controlled consciously by the brain. In enterprises, the task facing management is to find the optimum level of operational autonomy that allows the system to be both externally responsive and internally cohesive.
It seems to us that Beer presents autonomy in functional terms, i.e. as something that is necessary for survival or task accomplishment. Viable systems are autonomous because their environments are complex, and the amount of autonomy ceded to subsystems varies in relation to complexity in the environment. The VSM can then be seen as a complex working out, in organizational or structural terms, of how – depending on circumstances – varying degrees of autonomy might be ceded and then managed so as to ensure viability.
For Maturana it is not a case of autonomy being necessary. Rather, autonomy is a constituent condition; it is more fundamental. Whereas Beer comes up with a list of characteristics of viable systems, of which autonomy is one, for Maturana autonomy is the fundamental condition that constitutes a living being. Living systems are autonomous because they are autopoietic, i.e. because the components are the result of a self-producing dynamic.
On this view, autonomy is a more fundamental condition. And it is autonomy that provides living systems with a particular identity that can be maintained in the face of significant and enduring structural change. In the theory of autopoiesis, autonomy is not a variable; living systems are living because they are autonomous in the manner in which their molecules relate. If molecules stop producing networks to produce more molecules then the system dies.
This is not to say that Maturana discounts the importance of Beer’s point that systems must respond to prevailing environmental circumstances. It is just that on Maturana’s view autonomy is not the mechanism through which this occurs. For Beer, autonomy is a mechanism for adaptation; for Maturana, the parallel mechanism is ‘structural coupling’ which is something different. The point is that while autopoietic and viable systems are both autonomous, the nature of the autonomy and the manner in which it is realized in the two cases are different.
However, the situation becomes more confused since Maturana (1991) does acknowledge that autopoiesis can exist in the conceptual as well as in the molecular/physical domains. Thus arises the possibility that the VSM might be a conceptual autopoietic system, if not a physical/ molecular one.
This widening of the applicability of autopoiesis beyond the physical/molecular domain is an inevitable consequence of how the term is defined. Maturana, for example, acknowledges that because autopoiesis describes a relational phenomenon that exists in the dynamic space, clearly it does not depend on any particular set of components, and is thus potentially applicable to other non-physical/molecular systems. He remarks:
When I first spoke about living systems as autopoietic systems, I was speaking of molecular systems. Later … I realized that it was necessary to make the molecularity of living systems explicit in order to avoid confusion. A computer model of an autopoietic system does not take place in a molecular space … and this is why we did not claim to have a living system in the computer. (Maturana, 1991: 376. Also cited in Mingers, 1995: 45)
In other words it is molecularity and autopoiesis that specify living, i.e. ‘a living system is an autopoietic system in physical space’ (Maturana 1981: 22-23). Or more precisely: ‘any such autopoietic system realized in physical space is a living system’ (Whitaker, 1996: 6).
It seems clear then that, when ‘living’ is defined in these terms, a viable system is not a living system. But is it a conceptual autopoietic system? If our interpretation of what Maturana and Varela mean by the term self-production is correct, then the answer to this question is also no.
In Beer’s preface to Maturana and Varela (1980), he latches on to the terms organization and structure. In true cybernetic style, he focuses on Maturana and Varela’s remark that an autopoietic system is a homeostat, i.e. a device for holding a critical systemic variable constant within certain limits. Beer then goes on to say, in the case of autopoietic homeostasis, regarding Maturana and Varela’s ‘definitive point’, that:
the critical variable is the system’s own organization. It does not matter, it seems whether every measurable property of that organizational structure changes utterly in the system’s process of continuing adaptation. It survives. (Beer 1980: 70)
In the same piece, he then goes on to say: ‘their (i.e. Maturana and Varela’s) ‘it’ (i.e. autopoiesis) is notified precisely by its survival in a real world’.
Elsewhere, Beer (1979: 405) remarks:
The enterprise(s) … staff may come and go, its departments may be closed down or opened up … and still it has and retains its identity. In cybernetic terminology, this enterprise is called autopoietic.
In trying to make sense of this, it is necessary to recall that in the case of autopoiesis, the term organisation draws attention to the circular self producing manner in which system components relate to one another. However the term is not tied to autopoiesis; it refers to the basic class entity of any system.
Now while we would agree with Beer that this distinction between organization and structure is very important, in tying it to autopoiesis we submit that he misrepresents Maturana and Varela. He interprets them as saying that autopoiesis is the continuation and preservation of a system’s identity in the face of change. Thus:
any cohesive social institution is an autopoietic system because it survives … and because it may well change its entire appearance and its apparent purpose in the process. (Beer, cited in Mingers 1995: 119)
Moreover, Beer remarks:
Look at any great institution: a hospital, a university, a multinational company, a social service, a country. All these things change, in so far as their elements are replaced; all these things change, insofar as some features disappear while others are invented. But Guy’s Hospital, Oxford University, the steel industry, education, and Britain itself, are recognizably themselves. There is, as in our bodies, every kind of change. But there is no alteration … In the concept of autopoiesis we have the final testimonial to viability. The viable system is directed towards its own production. (Beer, 1979: 405)
All of this suggests that in latching on to Maturana and Varela’s distinction between organization and structure Beer has substantially modified the meaning of the term autopoiesis. Autopoiesis is not, as he would have us believe, the continuation and preservation of identity in the face of change; it is a very specific type of organization that constitutes the class identity ‘living’. Molecular systems such as motor vehicles and items of footwear and clothing change constantly through use, and they maintain a specific identity; there is structural change with conservation of organization.
But this does not make them autopoietic systems. The same is true of hospitals, universities (to use Beer’s examples) and other human social systems. Beer is right in saying that these change constantly but retain a sense of (observer- defined) identity. But autopoiesis defines its own organization; it is not open to observer- dependent assessments of what it is that is conserved. Autopoietic systems define their own identity; they do not require an observer to do this for them.
In Beer’s examples it is a particular observer- defined organization (Maturana’s term) of the system that is conserved. When Beer says that Guy’s Hospital and Oxford University have existed for hundreds of years, he is saying that something (tangible or intangible) has been conserved. It could be the way people interact, it could be the manner of production, or it may just the name. The point is that something has been conserved, and it this that gives the system its identity to some observer.
That all of this is the case is not, as Beer seems to think, autopoiesis. All systems exist as long as there is conservation of whatever it is that defines them. All systems arise this way. But we cannot say, as Beer does, that because systems exist or are viable over time, then they are autopoietic. It is only when a particular organization is conserved – when there is a conservation of a specific manner of relating of the components – that we can speak of autopoiesis.
Specifically we cannot see anything in the theory of viable systems that implies self- production of components. In specifying the entity ‘viable system’ an observer does not have to specify that the components of System 1 must be produced by the system itself. All that is specified is that they directly produce whatever it is that an observer decides the system produces. Such is the case with viable systems such as enterprises. If an observer were to attribute the purpose ‘knowledge production’ to a university then the various knowledge producers within the university and its component organizational units would fulfill System 1 functions. But there is nothing to say that such researchers must have been produced within the system, even though some may have been, nor that they are engaged in producing systems to produce more knowledge producers.
Again it is possible to see what might have led Beer to make the claims that he does. He is correct when he says that people can enter and leave an organization and that departments may be closed down or opened up – yet it still has and retains its identity. Ostensibly this resonates with Maturana’s claim that molecules enter into a living system, participate in the autopoietic dynamic and then leave. In both cases there is a flow of matter through the system. But it is only when the molecules participate in a specific dynamic – producing networks that produce more molecules – that we can speak of autopoiesis. Indeed it is only when the molecules participate in this dynamic that we can legitimately regard them as components. Participating in the production of any other specific identity is something different entirely. Autopoiesis is not just the conservation of any identity, it is the conservation of a particular identity – the class identity of living systems – self-production of components.
In debating the distinction between viable systems and autopoiesis, the teleological nature of the former is significant. Whereas notions of purpose, function and goals are pivotal to the theory of viable systems, these concepts are irrelevant to autopoiesis. No matter how you look at viable systems it is difficult to get away from the idea of purpose. Observer-dependent judgements on the identity of the system will often have recourse to notions about ‘the basic purpose of the system’. In turn these rely on assessments being made about the relationship between the viable system and the wider environment: for example: ‘the purpose of a university is to prepare students for the world of work’. This, of course, is possible for an observer who can observe both the system and the environment and explain the former in relation to the latter.
Internally the various components of the VSM are defined functionally (or, ‘allopoietically’) in terms of the purposive role that they play in supporting the overall purpose of the system. For example, the purpose of System 1 is to do what the system does; the purpose of System 2 is to coordinate the activities of operational units; the purpose of System 5 is to formulate policy, and so on.
Inevitably, then, when the VSM is applied to some sort of social system, it is brought forth in functional terms. In other words it is constituted as something that fulfils some external purpose and its internal structure is defined in functional terms.
All of which is very different from autopoiesis. Maturana acknowledges that in our culture we often have a need to speak about the function or purpose of systems. However, he argues that these distinctions are misleading, because they obscure how the system operates, which is as a result of how they are made or built, their components and relations. It is not because of a purpose. The distinction purpose belongs in the domain of the observer; it is a tool that people use to explain their experiences, and it arises as an explanation when the observer can observe the interaction between the system and its environment.
Purposes are not to be found in nature. The basic function of a living system is to produce itself and no more (see also Capra, 1996: 106). On this view the notion that internal components have purposes in relation to the system of which they are a part is equally misleading. The components of biological systems have no conception of an ‘outside’.
Having said all of this, it would be wrong to suggest that Beer believes that the purposes of organizations are pre-given. Attributing one or more purposes to a viable system is one of the first tasks that an agent must carry out in a VSM intervention. And for Beer, the purpose of a viable system is what it does, as this is defined by an observer. Thus, purpose is a mental construct imported by the observer to explain what is really an equilibriul phenomenon of polystable stable systems’ (Beer, in Maturana and Varela, 1980: 67).
So to the extent that Beer does give credence to the role of the observer, he and Maturana share common ground. However, this does not suggest that there is some fundamental point of similarity between a viable system and autopoiesis. Among the other points of contrast that we have discussed, the teleological nature of the viable system and the non-teleological nature of autopoiesis identifies them as fundamentally different kinds of entity. This leads us to conclude that while there can be abstract autopoietic entities, Beer’s viable system is clearly not one of these.
The main contention of this paper is that Stafford Beer’s assertion that a viable system is an autopoietic system rests upon a conception of autopoiesis that is at variance with the original intended meaning of the term and which stands in contrast to the manner in which it is used in the domain of biology and elsewhere. Notwithstanding this, some of the concepts and propositions that show up in Maturana and Varela’s explanation of autopoiesis, but which are quite independent of it, are potentially highly relevant to those who have an interest in viable systems and who use the VSM in a diagnostic and system improvement manner. This opens up the possibility that distinctive concepts such as ‘organization’, ‘structure’ and – in particular – structural coupling might be employed in harness with the VSM to gain a better understanding of how (viable) systems such as social organizations and enterprises operate, and how they might be better managed. This proposition is pursued in a follow-up paper.
Ashby W. R. (1952) Design for a Brain. Chapman & Hall: London.
Beer S. (1979) Heart of Enterprise. Wiley: Chichester.
Beer S. (1980) Autopoiesis: the organization of the living. In Autopoiesis and Cognition: The Realization of the Living, Maturana H. & Varela F. (eds). Riedel: Dordrecht; 63-72.
Beer S. (1981) Brain of the Firm. Wiley: Chichester.
Capra F. (1996) The Web of Life: A. New Synthesis of Mind and Matter. Harper Collins: London.
Espejo R. & Hamden R. (1989) The VSM: an on-going conversation. In The VSM: Interpretations and Applications of S. Beer’s VSM, Espejo R. & Hamden R. (eds). Wiley: Chichester; 77-100.
Flood R. L. & Jackson M. C. (1991) Creative Problem Solving. Wiley: Chichester.
Hejl P. (1984) Towards a theory of social systems: self- organization and self-maintenance, self- reference and syn-reference. In Self-Organization and Management of Social Systems: Promises, Doubts and Questions, Ulrich H. & Probst B. (eds). SpringerVerlag: Berlin.
Maturana H. (1981) Autopoiesis. In Autopoiesis: A. Theory of the Living Organization, Zeleny M. (ed.). Elsevier: New York; 25-32.
Maturana H. (1988) Reality: the search for objectivity or the quest for a compelling argument. Irish Journal of Psychology 9: 25-82.
Maturana H. (1991) Response to Jim Birch. Journal of Family Therapy 13: 375-393.
Maturana H. & Varela F. (1980) Autopoiesis and Cognition: The Realization of the Living. Reidel: Dordrecht.
Maturana H. & Varela F. (1987) The Tree of Knowledge: The Biological Roots of Human Understanding. Shambhala: Boston, MA.
Mingers J. (1992) The problems of social autopoiesis. International Journal of General Systems 21(2): 229-236.
Mingers J. (1995) Self-Producing Systems: Implications and Applications of Autopoiesis. Plenum Press: New York.
Varela F. (1981) Describing the logic of the living: the adequacy and limitations of the idea autopoiesis. In Autopoiesis: A. Theory of the Living Organization, Zeleny M. (ed.). Elsevier: New York; 75-92.
Whitaker R. (1996) The Observer Web: the internet nexus for autopoiesis and enaction. http: //www. informatik.umu.se/˜rwhit/AT.html
Yolles M. & Guo K. (2003) Paradigmatic metamorphosis and organizational development. Systems Research and Behavioral Science 20: 177-199.
Zeleny M. & Hufford C. (1992) The application of autopoiesis in systems analysis: are social systems also autopoietic systems? International Journal of General Systems 21(2): 145-160.
Found a mistake? Contact corrections/at/cepa.infoDownloaded from http://cepa.info/2808 on 2016-10-29 · Publication curated by Alexander Riegler