Autopoiesis and Death (Semiology of Autopoiesis IX)

Thus, if I kick a dog or a snake, the dog and the snake will react according to their dog-ness or snake-ness, from their internal circuit—from within.

Pier Luigi Luisi, Essays on Life Sciences with Related Science Fiction Stories, Newcastle upon Tyne, Cambridge Scholars Publishing, 2020, p 51

This claim, by the chemist Pier Luigi Luisi, encapsulates the problems of autopoiesis as a theory of life and death. How we die doesn’t come from within. It is a long process of transformation bringing the outside into us. A mistreated dog doesn’t suddenly die when its living systems stop. It doesn’t merely react to a kick. It has been dying over a long period, given a particular kind of perishing from repeated cruelty.

Violent blows don’t leave the dog’s internal circuits and systems untouched. They alter them and connect them to a history of faithfulness and mistreatment, in loops where the distinction between inside and outside does not hold and is misleading. Ideas of ‘dog-ness’ and ‘snake-ness’ are abstractions from the interconnected and extended lives of particular dogs and particular snakes embedded in multiple environments.

These loops and their elision of boundaries between inside and outside do not mean we cannot identify a dog, or any given life. They imply that the identification is not final and clear; it is a temporary selection within many diffuse connections. Varied and always uncertain interconnections explain how any living thing is an individual. We are individual in our changing relations, not according to some fixed identity.

Each life is misidentified as a member of a category such as ‘dog-ness’ or ‘humanness’ if this identification is deemed sufficient. Every dog is more than its dog-ness, in life and in death. A life is also misidentified if its identity within a category is deemed necessary. Dog-ness tells us something about our knowledge of dogs at a particular time. This categorical knowledge is not necessary for identifying or interacting with any particular dog.

These initial criticisms of autopoiesis as a theory of life and death raise two objections. Do my points on knowledge and identity lead to a position that is anti-science and hence to claims inconsistent with the benefits of science in practical fields such as medicine? Is the denial of an identity with clear boundaries a denial of death itself, since without a well-defined life to end there is nothing that could be judged to have died?

I am not denying the importance or practical value of knowledge and science. My critical point is that general categories are neither sufficient nor necessary for defining individual lives. This means that there are aspects of a life beyond the categories it is judged to belong to. It means that a life can be determined without referring to general categories.

Scepticism about the necessity and sufficiency of categories does not imply that knowledge and sciences do not have important roles to play in theoretical and practical work on individuals. Where individuals are concerned, it implies that belonging to general groups is not necessary for defining them or acting upon them.

The answer to the objection ‘Don’t you have to know it’s a dog to treat it?’ is ‘We should treat symptoms, illnesses and injuries in individuals. It helps to know what category the individual belongs to, as a short-cut to a list of symptoms and treatments.’ A category is a useful device, for a given state of science, knowledge and practice, but not an accurate match for the lives it applies to.

We are only members of categories in our own ways. These individual paths depend on our connections to others and to a history of shared relations. Many interactions let us benefit from care, but they also make us vulnerable to destructive acts. Cruelty twists the limbs and reactions of particular dogs differently, from bent resignation to bristling rage. A dog situated within an environment is an individual. It develops the abstract idea of dog-ness in its own way.

A sad dying dog isn’t a closed autopoietic system, coupled to loving or cruel humans that feed, play with and exercise it. It is an extended network of living and perishing, sometimes characterised by sadness (no doubt an approximate human projection) but open to other and better flourishing; for instance, when a mistreated dog is rescued.

I have added perishing to death, following A N Whitehead, to reflect the way death happens over time at different speeds, through relations and environments. The passing of time, across multiple processes, brings many different types of flourishing and perishing together.

This togetherness is another reason why closure, independence, identity and self-sufficiency lead to a mistaken view of death. Those ideas entail death as a halt to them – an interruption – whereas perishing (dying in zones, patterns, networks and depths) has been happening all along and to many, not only one.

My remark about closure can be put more widely. Our weaknesses and strengths as defined by membership to a general category only reach their full significance when individuals are taken as embedded in environments, not as autopoietic systems, but as extended and multiple networks of processes. Death happens with an environment, not in it.

An injury, illness or hereditary disease is very different when we are alone or within a social support system. We could explain this difference according to a model where a self-sufficient organism is merely aided by the wider system, but that explanation misses the joint development of individuals and systems, and hence their much greater co-dependence.

For instance, modern dentistry isn’t simply an intervention on particular cases of bad teeth. As an ongoing practice, including prevention and intervention, it has accompanied individuals fortunate enough to grow up with it such that their bodies have been shaped and supported by a scientific, educational, social, political, alimentary and economic system. Take that away and you don’t merely have ‘humanness’ with bad teeth as some sort of external accessory. You have different kinds of individuals, suffering and thriving deep into their bodies, in different ways, according to different systems, such as one dependent on high sugar use and colonial exploitation.

Is this insistence on extended and interconnected networks a denial of death as the end of well-defined identities? Yes it is. But it is not the denial of death itself. Instead, death applies inseparably to networks and to their subsections: to dogs, companions, carers, limbs, organs, genes, animals, and to their shared histories, futures, dates and environments.

According to the process philosophy I am contrasting with autopoiesis, death occurs when a set of processes cease in part within a network. This ending has effects throughout the network and death is therefore not a full halt, even for the processes that have ceased.

There is always an ending, but only relative to wider continuations, to the point where death can only be understood as an event across a limitless system, rather than a complete coming to a stop for a separate identity. Even if it appears to be in the smallest ways, every death is cosmic rather than individual, if by individual we mean a limited identity.

My use of ‘in part’ raises a deep problem for process philosophy. How can we refer to a subsection of multiple and interconnected processes accurately if the full operation and sense of that section require its relations to the whole? The solution is to refer to parts, but with the proviso that this must be understood as incomplete and that further investigation of the costs of severing the part from the whole is required.

This strategy is legitimate because it is necessary. It isn’t possible to take in the whole as such. There is too much to encompass and it is in constant movement. Instead, starting from a contingent part and working back and forth to the whole is the only approach that can account for the whole and its interaction with parts.

Added to this, we must critically evaluate the implications of each contingent choice and partial view since, though they are unavoidable, they are always flawed in relation to other starting points and selections. Life and death will always be a matter of perspectives on a complex multiplicity.

The solution can appear paradoxical for the concept of death. When something dies, does it make sense to say that it is only ‘in part’? For a process philosophy of continuous networks there is no paradox. Death is concentrated on part of the network with further connections to the whole. It is therefore more accurate to say that death is an event occurring in a zone with a wider environment. There is no death without that environment and its relations to the zone.

For autopoiesis, the part-whole and zone-environment response remains a contradiction, because autopoiesis prioritises identities, process closure and boundaries. In this case, death in part makes no sense and should be thought of as a mere change or surmountable malfunction, but not death, so long as essential organising, reproduction and transforming processes are still viable.

The concept of viability draws attention to a further opposition. More than a disagreement about death, the difference between the two process philosophies is about definitions of life:

• For autopoietic theory, life is defined as autopoiesis. Autopoietic organisation is what ‘to live’ means. Any set of processes that renew, develop, create and maintain themselves independently within a boundary is living. Those processes that fail to satisfy these conditions aren’t life. Life is therefore to be capable of autonomous self-reproduction and development;
• For continuity theory, life is creative flourishing and perishing over the whole of the network. To live is to belong within an environment as its processes renew, develop, create, maintain, flourish and die in part. An independent entity, or one defined as independent, is not fully living. Life is always participating in the whole living multiplicity.

Both theories have unusual and surprising implications. For autopoiesis, self-replicating and self-creating computer simulations might have to be considered to be living. The same holds for autopoietic social organisations, such as a design studio, legal system or medical practice. Whereas for continuity theory, no animal or plant lives independently and parts usually thought to be inanimate, such as a dental implant, a ruined building or a dried carcass, participate in overall concept of life.

The opposition between the two process accounts of death can be understood through two models. The first is borrowed from Deleuze and Guattari. When part of a rhizome dies, such as the shoots or roots of a Japanese Knotweed, those parts may be cut away and burnt, but the full rhizome continues and thrives again.

The second model is taken from autopoietic theory. The biological cell is the overriding example of autopoiesis. When processes essentiel for the maintenance of a cell within its membrane cease, then the cell dies. The fact that other cells around it are still living is irrelevant to that death.

The cell model can be used as a counter to the rhizome model. Accordingly, the rhizome model has simply ignored the idea of essential processes within a boundary. If we define these as the overall maintenance of the full rhizome, then we can say that a rhizome can die within a boundary. The reply to this counter is that the rhizome model isn’t about specific and limited rhizomes, but about the implication of every process in a limitless rhizome.

If death is defined simply as an ending within a defined boundary (this dead plant or animal) then the roles and meaning of wider networks are detached from it (these other loves, this shared grief, those distant blows, that hidden affection, this slow bleeding, that wild run, these unasked for caresses, this goodbye, all these causes, all those effects, these precursors and these inheritors).

All those wider processes reverberate across the boundaries defined by autopoiesis and beyond any identified subsection of any network. Since treating an enclosed life in abstraction from its connections gives a misleading definition of life, it also gives a misleading account of death. Life might have ceased in part, but life was always more than this and death shouldn’t be seen as a final end, but rather as a disruption, often a wrenching one no doubt, yet this devastation is a sign of the continuities around any circumscribed ending and identity.

Autopoietic theory has two resources to counter the definition of death as continuous extension rather than as a final halt happening within a boundary when an identified process ceases.

First, the concept of coupling explains how independent autopoietic processes can nonetheless be linked. Two or more processes can work together, providing energy, information and catalysts for homeostatic processes for each other, yet retain their independence. Animals and humans can interact and thrive together but remain fully determined by ‘dog-ness’ or ‘humanness’.

Second, the dependence on a necessary internal autopoietic organisation does not mean that there can’t be powerful connections and interactions across boundaries, so long as that core organisation remains independent. There can be external effects that lead to death and a death can have external consequences, but death itself is determined by the ending of those core autopoietic processes.

Affects such as intense joy and extreme fear might be caused by the death of another living being and might lead to the death of the being experiencing them, but that does not mean that their identities as autopoietic processes have been affected. Even the overwhelming power of grief might not indicate an inner transformation in essential processes, since it could have been a latent and necessary process of autopoiesis, a feature of ‘humanness’ that is triggered but not made by an external cause.

Given these responses, is the extension of identities into continuous networks of processes a devaluation of death in favour of a false definition and, perhaps, a false consolation? My answer is to reverse this claim. The false view is the one where identities are the place of death and the false consolation is identity itself, in the longing for something fully self-owned and disconnected from the outside.

The point of my argument is more restricted than simply identity versus extension. It is about an opposition between two theories of process and death. For autopoiesis, death occurs when a process ceases within defined limits and according to an account of what the essential processes of self-creation and maintenance are within those limits. Whereas I support the idea of a multiplicity of processes with no requirement for defined limits or for essential processes for any given identity.

For the former, death is the end for an identity (it is dead). For the latter, death is an ongoing process with extended networks of processes that can be associated more closely with parts of them, but need not (death is occurring).

To explain and understand these different views of death, a semiology of autopoiesis traces the presuppositions and effects of the new autopoietic definition of death as system breakdown. It observes its oppositions to earlier religious, scientific, social and philosophical definitions of death. It follows the consequences of the novel definition: what it commits us to and how it shapes our thinking and feelings.

Luisi gives this definition of autopoietic death: ‘From all this it is clear that death, from a systemic standpoint, is a process, a stepwise course of action: first the entire organism, then the organs, then the cells.’ (33) Death occurs when the parts of an integrated system no longer interact with each other.

This autopoietic death is a process of breakdown in stages and we can select particular parts or stages to prioritise in practical decisions about death: ‘[…] from a systems point of view, it is reasonable to state that death occurs when we reach the situation [corresponding] both to pulmonary-circulation death, as well as to brain-death [as systems].’

The flaw in this definition is its understanding of system and its association of death with the breakdown of a closed circuit. If systems are defined as interdependent and forming a continuous network, then death cannot be limited to a circumscribed breakdown, limited in time to the final moments of a failing system.

When we take account of the extended networks of causes and effects round any death, the spatially limited and time-bound definition of death is implausible. A system doesn’t simply stop. It does so because of a history of causes. It does so with a legacy of effects. These extensions make each death individual, rather than determined according to essential features such as pulmonary circulation.

Death is painful not because it is a standard halt, but because it is the loss of an individual part that cannot be separated from the whole. Each death is shared, communal and cosmic, not as the excision of a separable identity from a group, but as a wounding and yet healing alteration to joint processes and patterns.

Sadness and mourning, commemoration and thankfulness, perishing and flourishing, loss and renewal are hard and yet sometimes joyful ways of marking death, not because we have lost something we can never have back, but because we have suffered a deep alteration in valued and long lasting loving. That’s why we struggle and succeed in finding new ways of continuing with them.