[HOME PAGE] [CURRICULUM VITAE] [DRAFTS AND SKETCHES]
Autopoiesis-Dialognet Saturday, 15 May 2004
There is a substantial literature on autopoiesis and social systems. Searching Google for [autopoiesis "social systems"] gives more than 3,000 hits. One that I have found to be a good starting point is Randall Whitaker's (1995) Autopoietic Theory and Social Systems: Theory and Practice.
Maturana and Varela's concept of autopoiesis (Maturana, 1970, 1974, 1978, 1978a, 1988; Maturana and Varela, 1973, 1980, 1987; Varela 1979, 1994; Varela et al.., 1974) is fundamentally biological - i.e., autopoietic individuals are genuine living organisms, and in their conception, information cannot exist independently from the organism responding to it.
Nicklas Luhmann (1986, 1990, 1995) has started a significant school of social systems theory with his variant concept of autopoiesis based on systems of interaction on the level of disembodied information flows (Google gives more than 5,700 hits on [autopoiesis luhmann]). Assuming that I understand Luhmann's ideas correctly (I don't read German and am just coming to grips with the large literature), Luhmann still traced his ideas back to Maturana and Varela but took them in directions that they (and I) would not agree model reality.
My approach to social ("third order") autopoiesis is based on understanding of different levels of biological organization ("hierarchy theory" - Pattee 1973; Rosen 1991; Salthe 1985, 1993, 2001) able to support organizationally closed cybernetic systems. Biologists have no problems with recognizing cells (first order) and multicellular organisms (second order) as living in their own rights (i.e., being autopoietic).
Social systems of living organisms (e.g., insect colonies, biological species, firms and other human organizations) represent a third order of biological organization. Steven Gould (2002; Gould and Lloyd 1999) discusses the existence of evolutionary entities at several levels of biological organization. Whether such evolutionary entities can be considered to be autopoietic is an interesting problem from several points of view. I think at least some organizations are clearly autopoietic What follows is some more information relating to the arguments sketched in Hall 2003.
An important criterion for being able to discriminate an entity as being autopoietic is that it is complex and identifiably bounded. From our human points of view as multicellular autopoietic individuals, first order autopoietic entities are clearly bounded by cell membranes, and second order autopoietic entities are clearly bounded by skin or other kinds of epidermal structures (e.g., plants have epidermal or cortical coverings). On the other hand we can't see unambiguous boundaries around the organizations that we as people, have allegiance to from time to time. For example, there is an interesting literature on the concept of a "boundaryless career" (Arthur 1994; Arthur et al. 1999 - more than 2,000 Google hits) highlighting the fact that people can belong to several organizations at the same time and over the lengths of a career, showing that organizations don't totally own or control those working for them.
The boundary problem is a result of what I would call the problem of "situated perception", that Gould (2002; Gould and Lloyd 1999; ) also discussed in relationship to our difficulties in perceiving higher orders of evolutionary entities. Basically, we have great perceptual difficulties in seeing structural entities as individuals that have vastly different scales in space and time than ourselves. Our failure to see them, doesn't mean they don't exist, but it does mean that we have to construct our concept of a boundary logically rather than simply seeing or touching it through native perception. [Note that this idea is elaborated further in my subsequent sketch, A Biological Theory of Knowledge.]
I suspect that Luhmann's failure to situate his concepts of autopoiesis in the direct physical and economic interactions of people was due to his lack of a solution for the problem of properly situating his perception in the complex hierarchical structure of the social system (Salthe 1985, 1993, 2001). [Note: At the time this sketch was first written I was not aware of Howard Pattee's work on the epistemology of complex systems. The applicability of his ideas of semantic closure and the "epistemic cut" are described and their applicability to organizations is briefly discussed in my subsequent sketch, A Biological Theory of Knowledge.]
The logical resolution of the boundary problem that enables an autopoietic organization to discriminate itself from the environment and other organizations is that people accept "tags" that identify them as members of particular organizations at particular times and for particular purposes. Tags can include such social artifacts as membership cards, contracts, uniforms, oaths of allegiance, verbal agreements, badges, etc... Acceptance of the tag means that the person is a member of the particular organization for some time for the purposes of some defined/agreed subset of all the things the person does.
Self-production in the autopoietic social organization is achieved by induction processes that persuade individuals to join and accept the organizations tag and associated rules of interaction that make the joined individuals productive for organizational purposes while they are members of the organization. Note that there is nothing in this statement that precludes a particular individual from belonging to more than one organization at the same time.
Maturana and Varela stressed the idea that autopoietic systems were circularly closed (~autonomous) systems of dynamic interactions. In other words, to be autopoietic, organizations must be autonomous, having the requisites within their boundaries to maintain their self- control, at the same time that they are open to exchanges of matter and energy with the surrounding world.
Living organisms are dissipative systems, powered by and conforming to the laws of thermodynamics (Prigogine 1955, 1977, 1997, 1999, 2000; Prigogine and Antonious 2000; Morowitz 1968, 2002; Chaisson 2001). Processes within the organism are driven by fluxes of matter and energy from sources of high potential to sinks of lower potential through cyclical transport processes (i.e., metabolism: Morowitz 1968). In this sense, the organizational closure required by the concept of autopoiesis states that these processes must be cybernetically self-regulated within the autopoietic entity to homeostatically maintain the fluxes necessary to maintain the entity's state of organization.
In a social organization, fluxes of matter, energy and cash - which tokenises exchanges of matter and energy, maintain the self-regulative capacities of organizational processes in ways that are precisely analogous to the metabolism of cells and multicellular organisms. In fact, the purpose of the autopoietic social organization such as a firm is to provide its members with resources they need to support their own bodily metabolisms, in the same way that multicellular organisms exist to provide the circumstances to support the metabolic requirements of their individual cellular constituents.
The difficulties we have in seeing this kind of metabolism at work at the organizational level is again the problem of situated perception.
Some authors have argued that autopoiesis requires an "invariant organization" - such that an organization with a continually changing membership could not be autopoietic. I still need to backtrack and verify that this was what was actually written, but I think the issue is the result of misunderstanding what Maturana and Varela meant by organizational invariance.
My understanding of the concept is that the self-regulatory/self- productive properties of the organization that make it autopoietic must remain invariant - or that it is no longer autopoietic. Once the feedback capacity to self-regulate/self-produce is lost, dis-integration of the system/entity is inevitable as processes go out of control and become destructive. That is, life requires pre-existing life. The alternative is death.
Social organizations certainly exhibit this kind of invariance. They are able to persist as long as they can maintain adequate staff, financial reserves and social capital to maintain their existence against perturbations from the environment. When the self-regulatory capacity is lost (e.g., via bankruptcy), the organization disintegrates.
To me, given that I am an evolutionary biologist who has been working as a computer systems-based documentation and knowledge management analyst for the last 20+ years, the interesting problem is to understand how organizational cybernetics actually works. My understanding of Maturana and Varela's self-referential language equates "cognition" in the autopoietic entity with the sum of the cybernetic self-regulatory and self-productive processes that work to maintain the state of autopoiesis against external and internal perturbations. See Whitaker's (2001) summary of their thinking: http://www.enolagaia.com/EA.html#Cognition
As neurobiologists, the authors of autopoiesis theory were not specialists in genetics, population biology or evolution. Although autopoiesis is about dynamic processes, because they were seeking a minimal definition for the distinction between life and death, M&V explicitly excluded requirements for growth, reproduction, competition and evolution in the basic definition of autopoiesis (Varela et al. 1974). Thus, M&V, and many others who have adopted the concept of autopoiesis, have not seemed to understand or have neglected to think about is how autopoiesis arises in the first place, or how autopoietic entities can change, reproduce and evolve through time.
Harold Morowitz (1968) showed theoretically and Stuart Kauffman (1993) showed experimentally that fluxes of energy constrained to traverse a material system in order to be conducted from sources to sinks will drive the system transporting the energy to develop increasingly elaborate cyclic complexity. With sufficient complexity the transport system becomes autocatalytic (i.e., autopoietic).
In an environment where there is competition for limiting resources, systems able to successfully acquire needed resources will grow at the expense of systems less able to do so. However, in early evolutionary days when there are only small entropic differences between sets of building blocks that are sufficiently well organised and integrated to be autocatalytic/autopoietic, and collections of un-integrated catalytic components, an equilibrium will exist between sets that are actively autocatalysing more components fit to participate in autopoiesis, and products of previously autocatalytic sets that have for one reason or another dis-integrated.
"Natural selection" in this kind of environment will cause components that more readily participate in autocatalytic combinations to produce more such components will cause those "favourable" components to increase in the environment at the expense of materials that to not readily participate in autocatalysis
The story goes on, but it is at this point where the concept of "knowledge" becomes useful.
My ultimate authority on the theory of knowledge is Sir Karl Popper's later thinking, published in his (1972) Objective Knowledge: An Evolutionary Approach. He continued writing for some time after, but in the area of interest here did not add significantly to what can be found in this book.
The book contains two major ideas, that of evolutionary epistemology (which provides the basis for critical rationality), and his "three worlds" (which in some ways bears a superficial resemblance to Aristotle's idealism).
In Popper's evolutionary epistemology, all claims to knowledge are fallible, in that it is logically impossible to prove that any particular claim to know reality is in fact actually true. However knowledge grows in accuracy and scope through iterated cycles of theory construction and the selective elimination of theories demonstrated to be false when tested against reality:
Pn is a problem (in this case, continuation of autopoiesis), TS is a tentative theory or solution to the problem (i.e.) a combination of catalytic elements, EE is an error elimination process that removes theories that give bad results (i.e., dis-integration of a set of catalysts that are not fully autocatalytic), and Pn+1 is a restatement of the problem based on observations of the first results (bad combinations don't catalyse the formation of elements that are not autocatalytic in combination), which in turn, allows a new tentative theory to be proposed based on learning what didn't work in the prior cycle (the next cycle will have a higher frequency of potentially autocatalytic elements in proportion to those that don't participate successfully in the formation of autocatalytic sets). Thus, the surviving autocatalytic entities encompass a form of knowledge about what kinds of elements successfully work together to be autocatalytic.
This is basically the cybernetic logic of the Boyd OODA loop (Observe Orient Decide and Act) as described Hall (2003)
Two web sites preserve and present Boyd's thinking (http://www.belisarius.com and http://www.d-n-i.net/), and another discusses the kinds of knowledge transformations that take place over many iterations of the OODA cycle as an autopoietic entity learns, adapts and develops strategic power in a competitive environment (http://web.archive.org/web/20010302160925/http://www.eclectic.au.com/speclty/im/iml.htm) Although both the above sets of ideas were developed in a military affairs environment, they have obvious applicability to all kinds of autopoietic systems. [Material from these sites is elaborated in a later sketch, A Biological Theory of Knowledge.]
Although so far as I have been able to determine no one has previously combined the two conceptual frameworks, I find that Popper's concept of three worlds of existence with Maturana and Varela's concept of autopoiesis to be mutually synergistic in ways that provide major help to clearly understanding and applying both concepts to the real worlds of biology and human affairs. At least to me, the joint concept makes the ideas of organizational cognition and organizational knowledge crystal clear.
Popper divides existence (all that is) into three worlds:
World 3 is the most difficult of the three worlds to grasp, but it also has the greatest explanatory value. Both world 2 and world 3 have their physical manifestations in world 1. World 2 represents ongoing processing as cybernetically regulated. World 3 represents experiential knowledge (built as described above in the P1 --> TT --> EE --> P2 cycle) encoded in persistent products. These persistent products are manifested as world 1 artefacts such as organizational structures, enacted workflows, DNA molecules, hand written documents or printed books, the pattern of magnetic domains on a magnetic disk or electrons in a silicon chip, the spoken word as captured by audio recordings, etc. However, Popper (1972) argues (convincingly to me) that the encoded knowledge has an objective existence that is quite independent from the physical artefact that encodes it; as is demonstrated by the ability of the content to influence what happens in world 1 as mediated by cognitive processes in world 2. Exactly the same content can be transformed from audio tape, to a printed page, to a computer memory, to a laser disk, to another computer, and eventually back into a human brain.
Various forms of world 3 knowledge are produced by all three orders of autopoietic entities.
Some of the knowledge organizations transport and process to govern their activities is tacitly held in peoples' heads, but substantial quantities are stored in world 3 as an organizational memory and retrieved, processed and applied transferred in organizational processes that may be enacted in electronic workflows or interpersonal processes. Knowledge may be associated with workflows in several forms, embedded in the construction of the process itself, transported as an object from one point in the organizational structure to another, placed in new contexts through the building of links and associations, transformed through the addition, modification or deletion of content, and so on.