Autopoiesis is a conceptual framework that originated in cellular biology. Humberto Maturana and his student Francisco Varela proposed autopoiesis as a criterion for distinguishing living from non-living things, and as a unifying concept for biology, thus programmatically extending it to biology more generally and to developmental and evolutionary processes.
Etymologically, the word means self-making or self-creating.
In cellular biology, this ‘self-creating’ refers to homeostatic maintenance of the chemical components which must be present and in proximity to one another within a cell to enable the catalytic and other reactions which are involved in metabolism and in the activities of the cell as a living control system. (Terms like “living control system” are not ordinarily a part of autopoiesis discourses, but in the present post I am relating autopoiesis to Control Theory.) These chemical components are consumed (or ‘disintegrated’, as they say) in the course of these processes, and must be replenished. In addition, the very mechanisms of maintenance themselves must be maintained, the means by which the necessary concentrations of enzymes, catalysts, and other essential molecules are sustained at the levels that are requisite for the cellular structure as a whole to be maintained as a discrete system.
Note well that there are three levels of ‘maintenance’ here: concentrations of molecules, mechanisms that maintain those concentrations, and the cellular structure containing and sequencing those mechanisms and molecular concentrations. The chemical reactions that result from the proximity of the necessary molecules (their ‘concatenation’) are universal properties of chemistry and physics. The particular sequencing (‘concatenation’) of one chemical reaction and then another chemical reaction employing a resultant of the first is a structural property of ‘processes’ within the cell which may themselves be ‘concatenated’ (may have their outputs and inputs connected). More on this ambivalence of terms below.
Subsequently, by way of cybernetics, ‘second-order’ cybernetics, and systems theory, autopoiesis has been projected to psychology (to cognition, but not so much to behavior), sociology, and other fields. See e.g. Kauffman (2019). The publisher’s summary of Maturana and Varela (1984) expresses and promotes this:
“Knowing how we know” is the subject of this book. Its authors present a new view of cognition that has important social and ethical implications, for, they assert, the only world we humans can have is the one we create together through the actions of our coexistence.
Varela (2002) did not agree with these extensions, but rather held that the concept is only valid at the molecular level.
The molecular domain is the only domain of entities that through their interactions give rise to an open ended diversity of entities (with different dynamic architectures) of the same kind in a dynamic that can give rise to an open ended diversity of recursive processes that in their turn give rise to the composition of an open ended diversity of singular dynamic entities.
You may see that clarity and precision are not strong points in this literature. Perhaps we are spoiled by the excellent qualities of Bill Powers’ writing.
Multicellular organisms are considered second-order autopoietic systems. However, Maturana and Varela are not in agreement about this, and as noted the term has been generalized by way of second-order systems theory to encompass social systems as well.
There has indeed been much inconsistency and confusion about autopoiesis, especially in these wider extensions of the concept. Much of the confusion may be traced to ambiguities in its definition. Ambiguity is ubiquitous in language and is typically invisible to the language user because one interpretation is just plain obvious to the speaker, hearer, writer, or reader. The difficulty arises when what is obvious to one person differs from what is obvious to another, and these difficulties become pernicious ‘talking past each other’ when these assumptions are unstated. They are usually unrecognized and therefore inaccessible to be stated.
The original 1972 definition by Maturana and Varela was as follows (1973:69):
Una máquina autopoiética es una máquina organizada como un sistema de procesos de producción de componentes concatenados de tal manera que producen componentes que: i) generan los procesos (relaciones) de producción que los producen a través de sus continuas interacciones y transformaciones continuas, y ii) constituyen la máquina como una unidad en el espacio físico.
In English:
An autopoietic machine is a machine organized as a system of processes of production of components concatenated in such a way that they produce components which: i) generate the processes (relations) of production which produce them through their continual interactions and transformations, and ii) constitute the machine as a unit in physical space.
The definition appears to be recursive, though on at least one interpretation it is not, and for each recursion there is more than one possible (or candidate) re-entry point. The pronouns “they” and “them” (suffixes on Spanish words) have at least four possible referents. Abbreviating P = “processes” and C = “components”:
P1 produce C1.
The C1 are concatenated (brought into physical proximity within the cell) so that
“they” (P1? C1?) produce C2
C2 generate P1? P2?
which are relations of production
which produce “them” (P1? C1? P2? C2?)
this production is through “their” (P1? P2? C1? C2?) continual interactions and transformations
“they” (P1?2? C1? C2?) constitute the machine as a unit in physical space.
This has resulted in diverse interpretations, and has led some to conclude that the concept of autopoiesis is tautological, trivial, or vacuous.
Catalysis and other chemical reactions occur when the participating molecules are in proximity to each other. The need for proximity is consistent with a probabilistic model of the origin of life (Dyson 1982, Kauffman 2019).
In the ramifying analogical extensions of the concept from cellular biology to other fields, closure or boundedness is a somewhat fraught term. A cell is bounded by a membrane which contains the necessary molecules in proximity to one another, as is required for the metabolic and other processes within the cell. The boundaries of social systems and even of multicellular organisms are less easy to specify. (Bateson gives the example of a blind man walking with a cane.) Varela holds that the concept of autopoiesis is no more than a metaphor if there is no boundary delimiting the system.
Even for the cell the requirement of a boundary has been questioned.
The idea that the production of the components of a barrier is necessary for the maintenance of the system comes from the image of life as a cell with molecules in a liquid state dissolved in an aqueous solvent. However, the physical state of cellular cytoplasm is viscous (Schroder 2010) and under certain circumstances might maintain its unity (and therefore its border) by cohesion and surface tension (as does a drop of water surrounded by air). It is probable that in the origin of the first cells on Earth the production of a membrane was by the self-assembly of simple lipids synthesized in the environment, while the internal production of membrane components was a derived process evolved by natural selection (Hargreaves and Deamer 1978; Hanczyz et al. 2003; Budin and Szostak 2011). On the other hand, it is possible that in the origin of life the membrane was not a “limit” which separated the interior from the exterior of the living being, but rather the membrane itself was the living being (the membrane would be the interior of the living being …).
(Razeto-Barry (2012) 9th unnumbered page of the online publication, at fn. 7)
The cell membrane delimits an especially stabilized part of the environment of those molecules which come and go through the membrane. Their continual disintegration (by intracellular processes) and replenishment (by trans-membrane cellular processes) is at the heart of the original definition of autopoiesis which Maturana never abandoned.
Conceptually, it is said that an autopoietic system is closed in its organization, and this is equated with ‘circular causation’. As shown above, there is much confusion as to what causes what in this circular manner, because of the ambiguity in the definition of the concept that I analyzed above. Letelier refers to ‘metabolic closure’, and Varela talks of ‘operational closure’. An autopoietic system is closed in respect to the interconnections of homeostatic processes, but it is open in its structure because ‘components’ enter and leave. More generally, living things take in nutrients and excrete waste. Living control systems close their behavioral control loops through the environment.
It is relevant to note here an unusual characteristic of nerve cells (with analogs in mycelia and perhaps elsewhere). Their functioning extends the proximity required for molecular interactions between distal locations. Synapses bring cells in close proximity, but axons and dendrites extend between sometimes quite remote locations.
Now we can relate autopoiesis to control. Homeostatic mechanisms, such as those which maintain specified concentrations of essential molecules within the viscous medium of cytoplasm within the cell, are special cases of control systems with genetically defined reference values and genetically defined input functions and output functions in the cell membrane. (The location of these functions in the cell wall supports the notion, mentioned in the quotation above, that it is the membrane which is the autopoietic system, and that it controls inputs to its protected and stabilized interior environment and outputs to its exterior environment.) These homeostatic mechanisms maintain the control capacities of the cell as a living control system.
Generalizing, any living control system must have a meta-control system. In the extension of autopoiesis from cellular biology to multicellular organisms (and the more distant metaphorical extensions) there must be meta-system processes that repair and maintain the functional structures of control loops (comparator functions, perceptual and reference input functions, error output and effectors). Without an ‘immune system’ so to speak, a control system is defenseless against disturbances that would disrupt its functional structure, as distinct from disturbances to its controlled perceptual inputs. It would not survive outside the laboratory. The populations of autonomous control systems within a multicellular control system (gut and skin biomes, mitochondria and chloroplasts within cells) further clutters any requirement of closure and boundary.
In itself, a control system has no perception of itself as a whole, such that it might resist disturbances to this autoperception.
(Be careful here: what we call a self-perception in humans and perhaps some other animals is a system concept, and we do defend the story that we tell about ourselves. Even for a deeply practiced PCT theoretician, however, this system concept is not a perception of the control system that is controlling that perception among many other perceptions. It is a construct of words and concepts.)
It would be sufficient for the metacontrol system to sense only failure to control, or degradation of control. (Bill talked of quality of control as a perceived variable.) In other words, it must sense some biological consequences of error.
What parts of the biological implementation of a living control system can sense biological consequences of error? A cell senses chemical changes in its environment. In the immune system it is the cells which carry out the changes which constitute the immune response and the repair of systems. In the nervous system it is the cells which actually carry out the changes that constitute reorganization.
In sum, there are three aspects of the definition of living things:
- Replication. I haven’t touched on this, but reproduction is essential, and errors in replication enable evolution.
- Control. Maturana is concerned with a special case of control, homeostatic mechanisms within the cell, though he seems willing enough to indulge in metaphors and analogies based on this (e.g. Maturana & Varela 1984). Varela and the general systems folks do not have a good grasp of hierarchical control with variable, internally generated reference values.
- Meta-control: Reorganization, regeneration, immune response, healing.
============== References ================
Kauffman, Stuart A. (2019). A world beyond physics: the emergence and evolution of life. Oxford University Press.
Maturana, H.R. (2002). “Autopoiesis, structural coupling and cognition: A history of these and other notions in the biology of cognition”. Cybernetics and Human Knowing. 9: 5–34. http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.471.3749&rep=rep1&type=pdf
Maturana, H. R.; Varela, F. J. (1991). Autopoiesis and Cognition: The Realization of the Living. Springer Science & Business Media. [Tr. of De Maquinas y Seres Vivos (English: ‘On Machines and Living Beings’) (1973) Editorial Universitaria S.A.b (Chile). 2nd ed (1980) Dordrecht: Reidel; 3rd Ed. (1991) New York: Springer.] https://antropologiafractal.files.wordpress.com/2015/08/de-mc3a1quinas-y-seres-vivos-autopoiesis-la-organizacic3b3n-de-lo-vivo.pdf
Maturana, H. R.; Varela, F. J. (1984). The tree of knowledge: Biological basis of human understanding. [Rev. ed. (1992) The Tree of Knowledge: Biological Roots of Human Understanding. ISBN 978-0-87773-642-4
Razeto-Barry, Pablo (2012) Autopoiesis 40 years Later. A Review and a Reformulation
Origins of Life and Evolution of Biospheres 42(6). DOI:10.1007/s11084-012-9297-y