What makes an activity possible?

Back to the AE guide

A well-run meal depends on more than having food. People need equipment, skills, time and a way to work together. Preparing one meal can also change what they can do next time.

An ordinary example

Imagine volunteers organising a community meal. They agree roles, prepare food, serve people and clear up. In doing so, they may learn useful routines and discover who needs extra support. The next meal becomes easier to organise. They could also finish exhausted, with equipment damaged and nobody willing to return.

AE asks how the activity becomes possible and what carrying it out does to the possibility of doing it again. This is a constructed example, not a reported research finding.

Two words, explained

Operability means the capacity to do something particular under particular conditions. A kitchen can be usable for making tea while being unsuitable for preparing fifty meals. We need to say which activity we mean.

Autopoiesis means self-production. The word originally described living organisation: a living cell continually produces components that help maintain the organisation producing them. It does not mean making everything from nothing.

AE develops a broader philosophical proposal: what happens helps bring about the organisation that makes it possible. At the meal, volunteers do more than follow a plan. By agreeing roles and learning to work together, they also create arrangements they can use. This comparison does not establish that the volunteers or the kitchen form a living cell or one self-producing system.

Questions to try

What are people trying to become able to do?
Which arrangements make that possible now?
After doing it, what is available for next time?
Which conditions are renewed elsewhere, or simply used up?

The meal uses and changes the conditions for future meals. The return arrow asks what is renewed, changed or depleted; it does not claim that the meal is one self-producing system.

Optional: the research behind this idea

Keep three questions separate: how a capacity is constituted; whether its conditions are renewed, changed or depleted over time; and whether a specified unity qualifies as autonomously self-producing. A repeated activity does not answer all three. Luhmann already analyses recursive reproduction and material dependence. AE’s wider generalisation must earn its value through explanation, rather than treating all recurring activity as biological autopoiesis. See AE2.6, chapters 2–3, and Luhmann, Theory of Society, volume 1, pp. 50–57.

Read the AE2.6 manuscript (PDF). This is a developing manuscript. Examples in this guide are illustrations, not reports of research findings.

Read next: What is still open after a decision?.

Optional: extended discussion and scholarly background

Autopoiesis concerns the production of the organisation through which a system exists and operates. In AE’s generalised formulation, the question is how operations realise the organised capacity that makes those operations possible. “Operability” names that capacity to perform a determinate role. An event can acquire the force of a decision, for example, through procedures, records and relations of recognition. Its force needs explanation. (Watson and Brezovec, 2026, AE2.6)

Maturana and Varela developed autopoiesis to characterise living organisation. A network of processes produces components that realise that network and constitute a living unity. A cell is materially dependent on its surroundings; self-production does not require manufacturing all its own matter or energy. The distinctive claim concerns the organisation of production. (Maturana and Varela, 1980)

Luhmann reconstructed the concept for social systems, treating communication as the operation through which further communication becomes possible. This changes the domain and the meaning of production. It also creates a theoretical dispute: which features of biological autopoiesis survive the extension? Mingers’s critical examination is useful because it requires the proposed social unity and its self-production to be argued rather than assumed. (Luhmann, 1995) (Mingers, 2002)

What AE proposes

AE2.6 retains the proposition that all operations are autopoietic. Its current argument concerns constitutive reciprocity: an operative organisation is realised through the domain of operations it enables. This is broader than biological component production. The extension is therefore a philosophical proposal with additional burdens of explanation, not the discovery that every process satisfies the biological definition. (Watson and Brezovec, 2026, AE2.6)

Consider an assessment decision. A mark counts because assessment arrangements give it a particular role; those arrangements acquire practical reality through authorised acts of setting tasks, judging responses, recording outcomes and responding to challenges. The claim is about how this organisation is constituted. It does not establish that every decision improves or even preserves the arrangement.

An institution’s final decision could dissolve it. That decision still has institutional force, although it ends the institution’s continuation. Similarly, a technical operation can be fully effective on one occasion while consuming a resource needed later. These cases show why constitution and successful maintenance over time must be examined separately. (Watson, 2026, Operability)

Three questions that should remain separate

Constitution: through which relations can this occurrence perform its identified role?
Renewal: which conditions of further activity are actually retained, replenished, transformed or exhausted over the period studied?
Autonomous unity: what organisation, production relations and boundary justify treating the proposed target as self-producing?

The questions can inform one another, but an answer to one cannot substitute for answers to the others. Repeated AI outputs demonstrate repeated production of outputs. Further evidence is needed to establish the self-production of the organisation producing them. Equally, a component can contribute substantially to an institution’s capacities while being maintained by other organisations. (Watson, 2026, Operability)

A worked example

In our constructed assessment case, an AI-generated suggestion becomes a premise for a teacher’s judgement. The teacher asks a follow-up question; the student’s response changes the available evidence; the revised judgement becomes a record. That record later conditions another assessment. Following these relations can explain how a capacity for assessment is constituted and altered.

A second inquiry asks whether the sequence sustains warranted assessment over a term. It might increase opportunities for explanation, or it might consume review time and leave inaccurate records in circulation. A third inquiry would ask whether any identified technical or institutional target meets a specified criterion of autonomous self-production. Calling the entire episode “autopoietic” without separating these inquiries would obscure the work still required.

Montévil and Mossio’s biological account specifies mutually dependent constraints at relevant timescales. It offers a useful comparison for accounts of organised dependence, while retaining a biological explanatory target. Its results do not independently establish AE’s universal philosophical thesis. (Montévil and Mossio, 2015)

Zönnchen, Dzhimova and Socher question technological extensions that insufficiently demonstrate self-production and re-entry. Their challenge is relevant to early AE-related arguments about technology. It does not make every computational contribution to meaning or organisation insignificant; it sharpens the need to specify what is being attributed to which target. (Zönnchen, Dzhimova and Socher, 2025)

AE’s proposal should be assessed against alternatives that explain capacities through dispositions, external organisation or domain-specific mechanisms. Its value depends on what reciprocal constitution makes intelligible. A successful example cannot prove universality, and a persuasive philosophical argument cannot establish the effectiveness of a particular intervention.

Rough closure examines the unsettled conditions of definite operations. Ecological hyperknots follows relations among different continuations. The history of AE places this generalisation in its scholarly context.

Optional: full references and further reading

Sources support the particular claims discussed; their inclusion does not imply endorsement of AE. Publication and manuscript status are identified below.

Watson and Brezovec, 2026, AE2.6

Watson, S., and Brezovec, E. (2026). Autopoietic Ecology: Rethinking Systems, Meaning, and Matter. AE2.6, revised manuscript dated 14 September 2026. Unpublished manuscript. Relevant locators: preface; chapters 3–6, 10–12 and 15–18; appendices A–C.

Read the AE2.6 manuscript (PDF).

Maturana and Varela, 1980

Maturana, H. R., and Varela, F. J. (1980). Autopoiesis and Cognition: The Realization of the Living. D. Reidel. Source.

Luhmann, 1995

Luhmann, N. (1995). Social Systems. Translated by J. Bednarz Jr. and D. Baecker. Stanford University Press. German original published 1984.

Mingers, 2002

Mingers, J. (2002). Can social systems be autopoietic? Assessing Luhmann’s social theory. The Sociological Review, 50(2), 278–299. Source.

Watson, 2026, Operability

Watson, S. (2026). Generative AI and the recursive constitution of operability: Rough closure, semantic transduction and ecological hyperknots. Cambridge EdgeLab working paper, version 1, 18 September. Manuscript consulted; not peer reviewed. Constructed case and prospective research propositions. EdgeLab working-paper page · Open PDF.

Montévil and Mossio, 2015

Montévil, M., and Mossio, M. (2015). Biological organisation as closure of constraints. Journal of Theoretical Biology, 372, 179–191. Source.

Zönnchen, Dzhimova and Socher, 2025

Zönnchen, B., Dzhimova, M., and Socher, G. (2025). From intelligence to autopoiesis: Rethinking artificial intelligence through systems theory. Frontiers in Communication, 10, 1585321. Source.

Guide revision: 15 October 2026.