Autopoiesis and social systems – 1
Geyer F. (1992) Autopoiesis and social systems – 1. International Journal of General Systems 21(2): 175–183. Available at http://cepa.info/2812
Table of Contents
Defining the social system: Cooperative unit or hierarchical command system?
The applicability of Varela’s criteria to the family as a social system
Key Point #1
Key Point #2
Key Point #3
Key Point #4
Key Point #5
Key Point #6
The applicability of the theory of autopoiesis to non-living social systems
In this rejoinder to Zelený and Hufford’s paper, The application of autopoiesis in systems analysis: Are autopoietic systems also social systems?”, it is argued that applying autopoiesis, a biological concept, directly and literally to social systems in general, and to the family in particular, encounters practical difficulties of interpretation and adds no conceptual clarification or explanatory value to research in those fields. Additionally, and contrary to what Zelený and Hufford imply, social systems are not limited to cooperative units but include hierarchical command systems as well.Discussion follows Luhmann’s efforts to widen the concept of autopoiesis beyond its original biological connotations, thus making it applicable to the social sciences: social and psychic systems are not living systems, but meaning-using systems, based respectively on communication and consciousness as modes of meaning-based production rather than on individuals or even actions. Two concepts (self- observation and self-reference) from second-order cybernetics are then presented and are illustrated with recent research examples. These two concepts are held to be important explanatory variables for describing the behavior of social systems.
Key words: Autopoiesis, communication. complexity reduction, family, free choice, self-observation, self-organization, self-production, self-reference, sociocybernetics.
This contribution, a reaction from a sociologist to the focal article, will largely concentrate on what Zelený and Hufford say about their third example (the human family) and about social systems in general. First, the example itself. With a bit of definitional twisting and a bit of fuzziness added [see the reservations regarding the applicability of Varela’s six criteria outlined below] the family might indeed be viewed as an autopoietic system. The point is, however, and we shall return to it later in more detail: what, if anything, does that prove? What added conceptual clarification or explanatory value does that bring to the field of family research? What percentage of the enormous variance of family behavior does it help to explain? Of course, being systems theorists, we all suffer from the same professional deformation and love to get that orgiastic “eureka”-feeling when realizing that completely different phenomena in completely different disciplines are subject to the same laws, can be explained by the same or similar theories, or can be caught in the same type of differential equation. Yet by concentrating on the common denominator, as in the present case, of eukaryotic cells, osmotic growths, and human families, one tends to overlook the fact that what is specific to each of them may be more important to the field in question than what is common to all three.
Defining the social system: Cooperative unit or hierarchical command system?
Sonic clarity is needed, first of all, regarding the definition of social systems. Granting the fact that especially the higher mammals can have social systems, the social sciences usually conceive of a social system anascopically (i.e., from the bottom up) as a group of individuals interacting according to some more or less agreed-upon rules with respect to some goals – which do not have to be shared by all participants, although they often are. Hierarchical command systems, including the concentration camps and armies at war mentioned by Zelený and Hufford, can therefore certainly be viewed in a value-neutral sense as social systems. Abstracting from individuals, a social system can alternatively be viewed katascopically (i.e., from the group level down) as an institution composed of a number of roles, filled by (a succession of) individuals. But in this conceptualization, too, the roles do not have to be voluntarily chosen.
When Zelený and Hufford speak about social systems they clearly have an additional condition in mind, beyond Varela’s criteria, and also beyond the usual ones employed in social science: that the “components” of the “unity” partake in that unity out of their own free will, i.e., that they have a realistic choice whether or not to participate, and do so because they expect that a freely embarked-upon venture with others will further their perceived interests to a larger degree than if they would be acting alone. In other words: Zelený and Hufford limit social systems to cooperatively interacting groups. There is nothing against such a narrowed definition, as long as it is clearly stated, although this condition of free choice yields additional complexity.
For a free choice to be meaningful, a number of other conditions have to be fulfilled, the main one being the existence of a reasonable degree of determinism in the environment. As Plate has argued, free choice is impossible without such determinism: it is only when action A results in environmental reaction A´ rather than B´, and action B in reaction B´ rather than A´, that there is a rational basis for choosing at all. Free choice also assumes at least the satisfaction of other implicit preconditions: that the individuals in question have a reasonably adequate representation of their environment, that they test it out regularly, cumulatively updating it in interaction with that environment, and that they have a (flexible) value hierarchy (also produced in interaction with the environment, and therefore usually subject to change) as well as procedural rules indicating what action to perform on the basis of the interactive combination of one’s value hierarchy (i.e., primary goals at the moment) and one’s insight in the outside world (environment mapping or representation). This free choice is obviously especially relevant in cooperative social systems, but also in hierarchical command systems, even including concentration camps. In these situations of near-total powerlessness the array of choices is extremely reduced, but making alert use of the options that do exist can mean the difference between survival and death, as any holocaust survivor can testify.
Clearly, these additional preconditions for the type of cooperative social system described by Zelený and Hufford are not fulfilled by the elements of the eukaryotic cell or the osmotic growth, and go beyond the minimal criteria necessary to determine the existence of autopoiesis. When atoms A and B meet, their freedom is limited, and they have few options for interaction; when persons or institutions A and B meet, their options for interaction are so manifold that social science is not able to predict them with any degree of certainty. This is caused among other things by the emergent properties at this level of complexity: by the relatively high degree of “internal” complexity of A and B, by the complexity of their mutual (but never completely shared) environment, and by the fact that this environmental complexity gives rise to feed-forwarding phenomena like self-reference in (groups of) individuals in their effort to reduce environmental complexity to manageable proportions.
The applicability of Varela’s criteria to the family as a social system
Now, admitting that the family might be viewed as an autopoietic system to some degree (if one does not take Varela’s criteria literally), it may nevertheless be useful to mention a few reservations regarding their applicability to this case and to similar cases of other social systems.
Key Point #1
The family boundary is not as clearly defined as it seems to be at first sight. Zelený and Hufford mention the possible use of fuzzy set theory to give persons outside the nuclear or extended family a certain membership grade. However, the same should likewise be done for the core family members themselves: the membership of the set is not the family, but a part of each family member; varying per person and through time. Unlike the component parts of the eukaryotic cell, no family member belongs totally to the set: each of the family members – with the exception of the newborn – is also part of many other sets, and these other membership sets may take up even more of one’s time and involvement than the family does.
Key Point #2
The individuals in the family (at least to the extent their “family part” is concerned) can certainly be viewed as components, although the definition of what these components are is not unequivocally clear and is subject to a negotiated consensus, which is not the case for the components of the first two examples Zelený and Hufford present. Such a definition is observer-dependent and problem-dependent, as the authors implicitly indicate themselves: one can have a kinship viewpoint (mother, father), an economic one (wage-earner), and a moralistic one (“black sheep”).
Key Point #3
Only on a very specific level of aggregation and under very specific conditions can the unity of the family be considered a mechanistic system, like the other. two examples. While family members certainly have system-derived properties (aspirations, goals, role sets, norms) as a result of genetics and especially socialization, the result of their interactions cannot be forecast like the result of the interactions of the parts of a reliable and well-functioning machine. “The interactions and transformations of these components” are definitely not determined, although admittedly some interactions are more likely than others. There are many reasons for this, including differences in genetic, endowment, socialization (e.g., first-born vs. later-born), edu‑cation, knowledge base, and life history. Moreover, each family member has a different degree and different set of extra-family interactions which co-determine his/her world view, character, goals, modus operandi, etc..
Key Point #4
It is a moot point whether the boundary of the family is maintained and defined more by the family members or by the persons as well as the structures in their environment. Just look at the differences in divorce rate between Californian yuppies and strictly religious groups such as the Amish, where the environment’s strict social control tends to prevent divorce, even if one or both partners are extremely unhappy. In a highly mobile, modern, multi-group society like California’s, with its implicit relativity of norms, its high degree of geographic mobility, and its opportunities for women to be economically independent, there is both less social control and more opportunity to escape it when it is there. It is therefore precisely in the less free societies where, with a little help from one’s “friends,” one tends to remain married – although this is not to deny that often family members do indeed fiercely defend the family boundary.
Key Point #5
At issue in Varela’s criterion are “the components of the boundary of the unity,” i.e., of the family boundary, not within the family boundary or within the family. As argued in Key Point 1 above, the boundary of the family goes “through” rather than “around” the family members. It results not only from intra-family interaction (i.e., is not only “produced by the interaction of the components of the unity”), but also from interaction between the family (or family members) and its/their environment. This latter interaction often produces an environmental pressure in the form of social control, even the threat of ostracism, to keep the family from disbanding as a unity.
Key Point #6
Barring daddy’s artificial limb, mom’s dental work, and possibly daughter’s IUD (sub-“components of the unity” definitely not “produced by components of the unity”), I have no objections re the applicability of Point 6 to the family. This is intended not in a merely jocular way, but to raise an important issue: defining a component is problem-dependent. In what context, with what problem in mind, does one want to study the family? It should be stressed that a component does not necessarily coincide with an individual. In systems-oriented family therapy, for example, if two family members gang up against two others, with the fifth one as a mediator, it may be useful to distinguish only three components instead of five. Sometimes non-family members have to be included as well: see for example Bateson’s analysis of the alcoholic’s game which requires four supportive roles, not necessarily family members. And if daddy’s artificial limb is included within his boundary, then why not other sources of support one can avail oneself of when needed, such as knowledge, personal networks, etc.?
Summarizing: the family might perhaps be viewed as an autopoietic system when Varela’s six criteria are reinterpreted to some degree – when a generous dose of fuzziness or interpretative leeway is added. However, what conceptual clarification, explanatory value, or reduction of variance does this bring to family sociology? Perhaps it is precisely because autopoiesis is indeed something in common among units of widely differing degrees of complexity such as eukaryotic cells and human groups that the conclusion, “The family is an autopoietic system,” does not sound terribly surprising. Not that it is untrue or uninteresting as such, but it explains only such a small part of the variance. There is so much more, and even the social sciences are not terribly good as yet at explaining that “more” – although there are interesting systems-oriented approaches to it within the rapidly-developing field of sociocybernetics. Possibly the social sciences need the additional boost from what has been called “the emerging science of complexity to do that.
The applicability of the theory of autopoiesis to non-living social systems
Zelený and Hufford implicitly point to an interesting direction for further research when they state (p. 156) that “some autopoietic systems are non-biological, i.e., self-producing in inorganic milieus.” They refer to the low-complexity end of the evolutionary scale, but the above probably goes for the high-complexity end as well. A very successful attempt at an interesting theory transfer, in this respect, is Luhmann’s reconceptualization of the concept of autopoiesis to make it applicable to the social sciences. We will paraphrase and quote here the main points of his rather complicated line of argument, a novel and revolutionary approach not only for general systems theory but also for the social sciences themselves.
Luhmann argues that the term autopoiesis, originally coined to define living systems, has been extended on the wrong premises to other fields, and unsuccessfully at that. While it seems tempting to consider psychic and social systems as living systems – since they presuppose (biological) life after all – one immediately gets into trouble trying to define “precisely what the ‘components’ of psychic and social systems are whose reproduction by the same components of the same systems recursively defines the autopoietic unity of the system.” The term ‘closure’ is equally hard to define. Consequently, Luhmann considers it desirable to “abstract from life and define autopoiesis as a general form of system building using self-referential closure.” A “truly general theory of autopoiesis” would thus avoid “references which hold true only for living systems.” The point then obviously becomes which attributes of autopoiesis would “have to be dropped on behalf of their connection with life.”
Social systems and even psychic systems are not viewed as living systems; “the concept of autopoietic closure” forces us to view “meaning and life as different forms of autopoietic organization.” Meaning-using systems, i.e., social and psychic systems, are based upon a type of autopoietic organization other than living systems: namely on communication and consciousness, respectively, as modes of meaning-based reproduction.
Luhmann thus defends the quite novel thesis that, while social systems are self- organizing and self-producing systems, they do not consist of individuals or roles or even actions, as commonly conceptualized, but of communications. Communication can be viewed as “a synthesis of three selections: namely, information, utterance, and understanding,” a synthesis which is “produced by the network of communi‑cation, not by some inherent power of consciousness, or by the inherent quality of the information.” It should be stressed in this respect that not only was the concept of autopoiesis developed while studying living systems, but that also much of present-day social science still considers human individuals as its basic unit of analysis. Apart from the extension of autopoietic theory to osmotic growths (low-complexity non-life or “pre-life”), Luhmann thus extends the applicability of the concept to highly-complex but equally non-life situations, although they presuppose life.
While communications rather than actions are thus viewed as the elementary unit of social systems, the concept of action is admittedly necessary to ascribe certain actions to certain actors. The chain of communications can be viewed as a chain of actions, enabling social systems to communicate about their own communications and to choose their new communications, i.e., to be active in an autopoietic way. Such a general theory of autopoiesis has important consequences for the epistemology of the social sciences: it acknowledges “that observing systems are themselves autopoietic systems,” subject to the same conditions of autopoietic self-production as the systems they are studying. Thus the general theory draws a clear distinction between autopoiesis and observation, without which “the system could not accomplish the self-simplification necessary for self-observation.” As Luhmann states:
“Autopoiesis and observation, communication and attribution of action are not the same and can never fuse. Nevertheless, self-observation in this specific sense of describing itself as a chain of clear-cut and responsible actions is a prerequisite of autopoiesis as such. Without this technique of using a simplified model of itself, the system could not communicate about communication and could not select its basic elements in view of their capacity to adapt themselves to the requirements of autopoiesis. This particular constellation may not be universally valid for all autopoietic systems. In view of the special case of social systems, however, the general theory has to formulate the distinction of autopoiesis and observation in a way which does not exclude cases in which self-observation is a necessary requirement of autopoiesis as such.”
The theory of autopoiesis thus belongs to the class of global theories, i.e., theories that point to a collection of objects to which they themselves belong. Classical logic cannot really deal with this problem, and it will therefore be the task of a new systems-oriented epistemology “to pay attention to at least two fundamental distinctions: … between autopoiesis and observation on the one hand, and … between external and internal observation (self-observation) on the other.” Traditional epistemology searches for the conditions under which external observers arrive at the same results and does not deal with self-observation. Consequently, societies cannot be viewed, in this perspective, as either observing or observable.
We can only agree with Luhmann that sociology sorely needs an epistemology in which the concept of self-observation plays a central role. Whether this is also the case for biology as well as psychology remains to be seen and will depend upon the possibility of applying the concept of self-observation not only to social systems and conscious systems, but also to living systems. “Within a society, all observations of the society are self-observations,” and this fact has unexpected consequences. One of the main characteristics of social systems, and of the individuals composing them, is their advanced potential for self-reference. This means that the knowledge accumulated by the system about itself in turn affects the structure and operation of that system.
In this respect, social systems are different from many other systems, including biological ones. There is a clearly two-sided relationship between knowledge about the system on the one hand, and the behavior and structure of that system on the other hand. Biological systems, like social systems, admittedly do show goal-ori‑ented behavior of actors, self-organization, self-production, adaptation, and learning. But it is only social systems that arrive systematically, by means of experiment and reflection, at knowledge about their own structure and operating procedures with the obvious aim to improve these.
In social self-referential systems, feedback loops exist between parts of reality on the one hand and models and theories about these parts of reality on the other hand. Concretely, social scientists systematically accumulate new knowledge about the structure and functions of their society or about subgroups within that society. This may backfire, however: the paradox often is that the accumulation Of knowledge leads to a utilization of that knowledge, both by the social scientists and the objects of their research, which may change the validity of that knowledge. For instance, when the new knowledge is subsequently made known through publication in the mass media, it can be invalidated because the research subjects may react in such a way that the analyses or forecasts made by the social scientists are falsified.
To illustrate this more concretely, several research examples can be given:
1) Henshel, for example, has extended the notion of self-fulfilling prophecies to serial self-fulfilling prophecies, where the accuracy of earlier predictions, themselves influenced by the self-fulfilling mechanism, has impact upon the accuracy of subsequent predictions. He distinguishes credibility loops, where source credibility (i.e., the credibility of the forecaster) becomes significant, and confidence loops, where continuity across predictive iterations in the prediction itself is at issue.
2) Van der Zouwen, engaged in non-experimental methods research, investigated the consequences of self-reference for research methodology in the social sciences, and poses the question, “Given the fact that social systems are self-referential, to what extent is the accumulation of valid knowledge about them even possible, for researchers who – either as individuals or as a group – are themselves self-referential systems?” He stresses the enigma that it is precisely that methods research which hampers its own development, and in two different ways: by an increasing standardization of research practice and by anticipatory behavior of survey researchers. Standardization reduces the variance in the data collection procedures, which results in unreliable estimates of the effects of the methods on research outcomes. And when there are differences with respect to the methods used, these are largely caused by the anticipation of the researchers on the effects of their methodological decisions. As these anticipations become more frequent and more “reliable,” the relations between the characteristics of the methods used and the data obtained increasingly become artefacts of these anticipations – i.e., the data become more unreliable.
3) Anderson uses frame methodology derived from developments in artificial intelligence to study political systems, arriving at similar conclusions as Van der Zouwen. Luhmann has stressed that a political system can recognize only those problems it is “programmed” to recognize;’ problems sometimes become important only because the means for their solution exist. The self-reference of the political system comes out clearly in the fact that a successful solution of a high-priority problem or the failure to solve such a problem will strengthen or weaken the relevant part of the political system. The success as well as the failure of previous efforts to solve specific political problems feed back on present-day efforts and tend to produce a standardization of the solutions that are deemed possible in certain cases, while political decision-making obviously thrives on anticipatory behavior of politicians. If a politician has learned his lessons well, he knows what manipulative stimuli to give in order to elicit specific reactions from the public, not unlike Van der Zouwen’s methodologist who more or less determines the answer distribution of his respondents by using certain methods.
4) Hornung describes the construction of knowledge-based systems for the analysis of development problems in health care planning. Health care systems are viewed as autopoietic (self-organizing and self-referential) sociotechnical systems in line with Luhmann, quite apart from the fact that the participating individuals are autopoietic systems in the biological sense. Self-reference enters at four levels here: a) the level of individual learning, exemplified by interaction of the modeller with his cognitive model; b) the level of generating group expertise about a problem, by interaction among the modeller, the model, and other participants in a modelling or planning group; c) the level of self-organization in the scientific subsystem, i.e., interaction between the modeller or modelling group and the scientific community; and d) the level of management and policy-making in the health subsystem, a national system, or even the international system, consisting of interaction between modellers and decision makers at the corresponding levels. A cooperative and participative planning process is necessary on all these levels, as Hornung stresses, and seen all the more so if one agrees with Maturana and Varela” that in any strict sense there is no flux of thought from one person to another and that denotative functions of messages lie only in the cognitive domains of the observer.
Many more examples could be given, but these will suffice to demonstrate that self-reference and self-observation play a crucial role in social systems, explain more of the variance in their behavior than autopoiesis does, and are among the important variables that make them different from biological systems.
Summarizing: viewing the family, or any other social system, as an autopoietic unity in the strict biological sense leads to the necessity to twist and fuzzify Varela’s six criteria and is not very helpful in analyzing the family. However, the family can indeed be viewed as an autopoietic system if one abstracts from the family members and if it is conceptualized as a system of communications, rather than of biological individuals. Self-reference and self-observation, then, well-known phenomena from second-order cybernetics, turn out to be important explanatory variables, in addition to the original six criteria for autopoiesis, for describing the behavior of social systems.
I want to thank Gerard de Zeeuw for his comments on a draft version of this text and have invited him to contribute an addendum following this contribution.
 M. Zelený and K. D. Hufford (1991) “The application of autopoiesis in systems analysis.” International Journal of General Systems, 21, pp. 145-160.
 J. R. Platt (1973) “The Skinnerian revolution.” In Beyond the Punitive Society, edited by H. Wheeler, W. H. Freeman & Co., San Francisco, pp. 22-56.
 G. Bateson (1972) Steps to an Ecology of Mind. Ballantine Books, New York.
 L. Langman (1986) “The family: A ‘sociocybemetic’ approach to theory and policy.” In Sociocybernetic Paradoxes – Observation, Control, and Evolution of Self-Steering Systems, edited by F. Geyer and J. Van der Zouwen, SAGE, London, pp. 27-43.
 F. Geyer and J. Van der Zouwen (eds.) (1986) Sociocybernetic Paradoxes – Observation, Control, and Evolution of Self-Steering Systems, edited by F. Geyer and J. Van der Zouwen, SAGE, London.
 I. Prigogine and I. Stengers (1984) Order Our of Chaos – Man’s New Dialogue With Nature. Heinemann, London.
 N. Luhmann (1986) “The autopoiesis of social systems.” In Sociocybernetic Paradoxes – Observation. Control, and Evolution of Self-Steering Systems, edited by F. Geyer and J. Van der Zouwen, SAGE, London, pp. 172-192.
 N. Luhmann (1986), p. 179.
 R. L. Henshel (1990) “Credibility and confidence loops in social prediction.” In Self-Referencing in Social Systems, edited by F. Geyer and J. Van der Zouwen, Intersystems Publications, Salinas, CA, pp. 31-58.
 J. van der Zouwen (1990) “The impact of self-referentiality of social systems on research methodology,” In Self-Referencing in Social Systems, edited by F. Geyer and J. Van der Zouwen, Intersystems Publications, Salinas, CA, pp. 59-68.
 I. B. Anderson (1990) “Frames and dynamic models of political systems.” In Self-Referencing in Social Systems, edited by F. Geyer and J. Van der Zouwen, Intersystems Publications, Salinas, CA, pp. 69-84.
 B. R. Hornung (1990) “The construction of knowledge-based systems for the analysis of development problems in health care systems.” In Self-Referencing in Social Systems, edited by F. Geyer and J. Van der Zouwen, Intersystems Publications, Salinas, CA, pp. 115-142.
 H. R. Maturana and F. J. Varela (eds.) (1980) Autopoiesis and Cognition. D. Reidel, Dordrecht/Boston.
Found a mistake? Contact corrections/at/cepa.infoDownloaded from http://cepa.info/2812 on 2017-01-04 · Publication curated by Alexander Riegler