Context: This paper is intended for readers familiar with Humberto Maturana’s theory of autopoietic systems and with the still unresolved debate concerning the existence of non-biological autopoietic systems. Because the seminal work of the Chilean biologist has not yet been fully and correctly understood in other disciplines, I consider that it is necessary to offer a more generalized concept of the autopoietic system, derived by implication from Maturana’s grounding definition. Problem: The above-mentioned debate is rooted in a deficient application of some rigorous distinctions, definitions, and epistemological considerations introduced by Maturana when he coined the term “autopoiesis.” Some researchers think that social or economic organizations could be considered as autopoietic systems of a higher order because they appear to behave autonomously and be self-organized and self-producing. However, in practice some precise distinctions would need to be verified through observation in order to claim properly their autopoietic nature. These distinctions were defined by Varela, Maturana and Uribe in 1974 as a set of six decisional rules (“MV&U rules”) whereby an observer may possibly justify this stand. My aim is to pinpoint clearly the basic cognitive tasks that an observer should perform in order to ascertain such a claim. Method: I accomplish this with a thorough analysis of the entailments derived from each rule when applied to the most general case – when the observational domain where the system manifests itself is not specified. A bottom-up approach is used to avoid referring to “apparent autopoietic behavior” as a starting point distinction (top-down approach): the aim is to distinguish “autopoietic behavior” as an outcome of more basic distinctions, not as a premise for these. These may be used as abstract tools to facilitate a rigorous description of observations and lead to precise explanations of the emergence of complex self-generated dynamic systems. In theory, this conceptual frame is not limited to the macro-molecular domain and may therefore be applicable to non-biological systems. Results: According to MV&U rules, the most important distinctions are those that refer to intra-boundary phenomenology: this focus is necessary to explain how the key processes involved in the emergence of autopoiesis actually manifest themselves. These explanations are crucial to validating a claim about the “autopoietic nature” of an observed system. Implications: This work could help multidisciplinary researchers to apply properly the theory of autopoietic systems beyond the realm of biology and to settle ongoing debates. It could also help investigations related to the specifications of software simulation processes for modeling a minimal artificial autopoietic system. However, the rigorous focus on the role of intra-boundary phenomenology and self-production of components reveals that our chances of detecting “natural” meta-molecular autopoietic systems are scarce.

Key words: autopoiesis, self-production, autonomy, causation structure, artificial life, organizational knowledge

Context: In this paper I expand aspects of the generalized bottom-up explanatory approach devised in Part I to expound the natural emergence of composite self-organized dynamic systems endowed with self-produced embodied boundaries and with observed degrees of autonomous behavior. In Part I, the focus was on the rules defined by Varela, Maturana & Uribe (VM&U rules), viewed as a validation test to assess if an observed system is autopoietic. This was accomplished by referring to Maturana’s ontological-epistemological frame and by defining distinctions, concepts, and abstractions necessary to describe dynamic systems in any observational domain. This approach concentrates on pure causation flow rather than on domain-specific interaction mechanisms. Problem: It is essential to analyze the requirements imposed by the VM&U rules on the “intra-boundaries” phenomenology for compliance with the self-production capabilities expected from an autopoietic system. Beyond what is merely implied by the compact wording of the VM&U rules, a key point needs to be addressed explicitly: how to describe some “peculiar” capabilities that the components should be endowed with to participate in new component production (as macro-molecules do in the biological domain) so that system’s self-production can be assessed. Method: Using this approach, I first describe the process of constituting self-organized dynamic structures provided with embodied boundaries. Then I explain how a capability of self-organization emerges and how this results in ephemeral configurations that may evolve into self-regulated long-lasting dynamic system stability within a continuous causation flow inside the boundaries, up to the emergence of some “specialized” subsets of components. This explication allows us to distinguish the medium, the boundaries, and the core of a self-organized dynamic system and to focus attention on the “intra-boundaries” phenomenology that should be at the heart of self-production capabilities, as prescribed by the 5th and 6th VM&U rules. Results: I propose an abstract, domain-free description of the “peculiar” composition and decomposition transformation capabilities that components should possess while subject to state transitions triggered within the “intra-boundaries” causation flow. This is combined with a discussion concerning the “intra-boundaries” causation structure’s possible topological layouts that could be compliant with the 6th rule. Implications: The above-mentioned results allow us to improve our analytic criteria when observing dynamic systems existing in non-biological domains in order to assess their autopoietic nature. They also reveal that the task of consistently identifying possible non-biological autopoietic systems is harder than merely identifying self-organized dynamic systems provided with boundaries and some observable autonomous behavioral capabilities in a given observational domain. More implications will be discussed further in Part III.

Key words: autopoiesis, dynamic system, self-organization, self-production, boundary, autonomy, operational closure, causation, structure determined, meta-molecular, embodiment, interaction network, organizational invariance

Context: There is an ongoing debate about the possibility of identifying autopoietic systems in non-biological domains. In other words, whether autopoiesis can be conceived as a domain-free rather than domain-specific concept – regardless of Maturana’s and Varela’s opinions to the contrary. In previous parts my focus was, among other matters, on the rules defined by Varela, Maturana, and Uribe (“VM&U rules”). These rules were viewed as a validation test to assess if an observed system is autopoietic by referring to Maturana’s ontological-epistemological frame. I concluded that identifying possible non-biological autopoietic systems is harder than merely identifying self-organized dynamic systems that are provided with boundaries and some observable autonomous behavioral capabilities in a given observational domain. This is because no assessment could be valid without examining such systems’ “intra-boundaries” phenomenology and proving actual compliance with the VM&U component production rules. Problem: Any rigorous approach to investigating possible self-production capabilities within a given dynamic system needs to drill down on the composition and physical conditions of the system’s core dynamics. My aim now is to discuss the problem of choosing the adequate spatial and temporal scales to be applied when observing and describing dynamic systems in general. When trying to detect an autopoietic system in a given observational domain, the observer needs conceptual tools to apply rigorously the VM&U rules and decide on the matter. This is particularly useful when dealing with systems with spatially distributed components interacting through cause-effect couplings that are independent of the distance between them, as is the case of social systems. Results: For observing dynamic systems, the choice of appropriate spatial and temporal scales of description is not a trivial operation. The observer needs to distinguish between “instantaneous” phenomena and phenomena possessing extended “durations.” I argue that the observer can easily extend the notions discussed by Maturana and Varela to observational domains where the system’s components do not constitute an entity showing a topological “form” in physical space. Furthermore, I show that a diachronic perspective must be applied by observers to explain component production/destruction mechanisms as the outcomes of processes involving structure-determined coordination over relatively long time intervals. Finally, these considerations lead to establishing a link with Varela’s fundamental concept of autonomy. Implications: The adequate choice of spatial and temporal scales of observation and description are essential (a) to discuss the problem of a possible identification of social autopoietic systems, and (b) to analyze the possibility of designing virtual simulated autopoietic systems in software domains (“computational autopoiesis”).

Key words: Autopoiesis, complex dynamic system, self-organization, self-production, boundary, causation, structure-determined, meta-molecular, interaction network, niche, synchronic/diachronic, processes.

Upshot: Reflecting beyond the original intent of my paper, I respond to Luhmann-inspired commentaries by raising ontological-epistemological issues that stand before any attempt to build bridges between Maturana’s and Luhmann’s approaches to “autopoiesis.” I propose to look at the social from a vantage point from which human actors and their social doings (communications, among others) appear as equally relevant objects of knowledge in sociological theory-building.

Context: In previous papers, I suggested six rules proposed by Varela, Maturana and Uribe as a validation test to assess the autopoietic nature of a complex dynamic system. Identifying possible non-biological autopoietic systems is harder than merely assessing self-organization, existence of embodied boundaries and some observable autonomous behavioural capabilities: any rigorous assessment should include a close observation of the “intra-boundaries” phenomenology in terms of components’ self-production, their spatial distribution and the temporal occurrence of interaction events. Problem: Under which physical and components’ relational conditions can some social systems be properly considered as autopoietic unities compliant with the six rules? Results: Dynamic systems can be classified according to “degrees of autonomous behaviour” that they may acquire as a result of the emergence of organizational closure (i.e., autonomy. Also, the different “degrees of attainable systemic autonomy” depend on the “degrees of autonomy” shown by a system’s dynamic components. For human social systems, a necessary balance between individuals’ autonomy and the heteronomous behaviour brought about on people by social norms (laws, culture, tradition or coercion) sets limits to the “degree of systemic autonomy” that human organizations may acquire. Therefore social systems, defined as dynamic systems composed of physical agents, could not attain the high “levels of systemic autonomy” ascribable to autopoietic systems without constraining the autonomy of agents to “levels” that are incompatible with spontaneous human behaviour. Also, social organizations seen as composed of physical agents interacting in physical space cannot be construed as autopoietic systems. Alternatively, if seen as composed of “process-like” entities, where agents participate as actors within processes, some social systems could be described as autopoietic wholes existing in the abstract space in which we distinguish interactions between processes, provided that we can assess compliance with the rules for some specific cases. Implications: These conclusions contribute to the debate on the possible autopoietic nature of some human social systems and to grasping the opportunity to shift focus to the more interesting issue of the “degrees of systemic autonomy” that human organizations could acquire (if needed) without imposing unbearable limitations on the autonomy of human actors. Also, the conceptual framework of this explanatory approach could be used in practical terms to assist the development of new dynamic modelling languages capable of simulating social systems.

Key words: Autopoiesis, autonomy, heteronomy, self-organization, social system, organizational closure, causation, Maturana, Varela, Luhmann.

Open peer commentary on the article “The Autopoiesis of Social Systems and its Criticisms” by Hugo Cadenas & Marcelo Arnold. Upshot: Cadenas and Arnold contribute towards a better understanding of what is at stake in the long debate concerning the applicability of Maturana’s autopoiesis concept to social systems. However, their target article has two shortcomings: (1) it does not provide a deeper understanding of the reasons why Luhmann’s adoption of the autopoiesis concept has proved to be sterile after decades of debate; and (2) their historic account does not integrate recent developments that need to be highlighted to improve interdisciplinary research.

Open peer commentary on the article “In Maturana’s Wake: The Biology of Cognition’s Legacy and its Prospects” by Randall Whitaker. Abstract: With respect to pedagogic goals, Whitaker’s work stands for itself. First, I suggest valuing Maturana’s first-person approach to social interaction analysis and further exploring his late work and experience in applying biology of cognition to practical problems in social relations. Then I criticize Whitaker’s insufficient explanation of his distinction of the “abstract (non-physical)” nature of what he calls the “novel domain” engendered by the feature of operational closure of the nervous system. I end by discussing the difficulties encountered by the author when trying to “extract” a pedagogical methodology from biology of cognition.