Wolfgang Wopperer Notes Essays

Fabric foundations: collective agency, attractors and power

Draft · Version from · 14 min read

Fabric foundations: collective agency, attractors and power

When a System can actively self-maintain its persistence, it is an Agent – a bearer of Agency in a minimal sense: it upholds a boundary, it carries a self-model, it takes action to preserve its own coherence.

This is a Scale-free Abstraction: the same pattern recurs from cells to organisms to Social Systems, so a movement or network can be a collective agent in a precise, non-metaphorical sense.

Importantly, in this view agency comes first: All systems have agency in this minimal sense, just to different degrees. There is no magic threshold of complexity above which agency appears

Agency is primary – and it is constituted in relationships. In other words, relationships are equally primary – they are the other half of agency. This is why agency in relationships is at the heart of our thinking.

Three descriptions of one thing

The terms at the top of our mountain occupy three registers that are easy to conflate. “Third attractor” is a dynamical description – a region of the civilisational State Space toward which the system tends, with a basin and a pull. “Collective agent in service of life” is an agential description – a system that maintains its own Information Closure: a self-model, goals, a boundary held from within. “Collective power” is a relational description – that agent’s capacity to influence other systems by acting as a Constraint on them. We want to keep the three distinct, and then show that they describe one phenomenon.

On the surface, the identity “Third Attractor = collective agency in service to life” is a category error. An attractor is a set of states a system tends toward; an agent is a system that maintains its own closure. A region of state space and a self-modelling entity are not the same kind of thing.

The agency-first frame dissolves this: in agential systems the Lyapunov function that (kind of) describes its attractor landscape is an outcome of goal-directed behaviour, the agent actively seeking out its Attractor states. An attractor defines a stable system – a system is a pattern in its state space. Put together: an agent is an attractor maintained from within.

A pendulum’s pattern persists because nothing disturbs it; an agent’s pattern persists because the system itself works to restore it. The attractor-view and the agent-view are the dynamics-eye and agent-eye descriptions of one and the same thing – the ant’s-eye/eagle’s-eye reconciliation Stu offered for protocols vs living systems, raised one level.

So “two names for one telos” is not loose synonymy. It is a substantive claim that the agency-first framework licenses: attractors as expressions of the agents that maintain them.

The third attractor is not yet an attractor

Collapse and control are attractors of the current world-system – they have basins, they exert pull now. The third attractor does not; the entire “conditions for emergence” apparatus exists precisely because it lacks pull. Strictly, it is at present an imagined attractor – and that is an alarm, because forcing systems into imaginary attractors doesn’t end well. Instead we need to better equip the system for the during- and post-collapse liminal space and ensuing search process rather than to specify and force a future state.

Fabric’s theory of change answers the concerns about the imposition of a future almost point for point. “Model, don’t centralise”; the self-model cannot be imposed from outside; cultivate the constraints from which a movement self-organises; membership as opt-in adoption. This is landscape engineering rather than attractor-forcing – Bakunin’s critique of the “chief engineer of the world revolution” taken seriously. The conditions-for-emergence activities are the competences-for-the-liminal-space move. Fabric is not “supporting a third attractor” that already exists; it is working to make an imagined attractor real by changing the landscape’s constraints – a different and more exciting claim.

Control, on this reading, is the institutionalised version of the same revolutionary error: it treats society as something that can be held at a designed point. It differs from the revolutionary attempt in content, not in form. The third attractor differs in form – it does not specify a configuration at all, it specifies a way of moving.

Collective agency is scale-free, and our theory of change depends on it

Our material uses “collective agency” at two scales: generically – Fabric’s own collaborations, a pilot pair of organisations, Lifehouse Stroud – and as the singular planetary telos, humanity and Gaia as one coherent agent. This is not equivocation; collective agency is a Scale-free Abstraction: the same pattern – information closure, a self-model constructed from within, mutual commitment constituting a normative space – recurs at every scale, and the theory of change works only because it does.

The generative cycle (commitment → interdependence → trust → deeper commitment) builds small collective agents; small agents become the components of larger ones; the planetary agent is the pattern at the top of the recursion.

Fabric’s “model and spread” strategy is therefore a bet on the scale-free character of agency: if the pattern were not scale-free, modelling it at Fabric’s scale would tell us nothing about its emergence at civilisational scale.

The four criteria as closure conditions

Read through the agent-eye lens, Schmachtenberger’s four criteria stop being a list and become closure conditions at the civilisational scale – what it takes for a planetary-scale Living System to hold a self-model and act on its own behalf:

The criteria can thus be got out of the concept of a living collective agent, rather than imported on authority.

Collective power is the relational face of agency

Power is a system’s ability to influence the behaviour of other systems, exercised by acting as a Constraint – and a constraint restricts trajectories from reaching some parts of the state space. So:

The geographic/local aspect (togetherness, resilience, power) is agency-building. Repair-and-renewal is power exercised constructively – reshaping media, economy, reparations. The global immune response is power in its purest constraint form: “making war and extractive and colonial projects undesirable or impossible” is, word for word, restricting trajectories from reaching parts of the state space.

This settles a terminological oscillation in our source documents, which alternate between Fabric “supporting collective agency in the service of life” and its purpose being “building collective power in the service of life”. These are not interchangeable, and the order matters: agency is prior, power is derived. Collective power without collective agency-from-within is not a degenerate case – it is the control attractor.

Control and collapse are entangled

It is tempting to map the three attractors onto three dynamical regimes – chaotic, point, strange attractors. Bu this is wrong at the level of actual system behaviour. A more accurate picture is this:

Agents always move on strange attractors – they steer a course between Scylla and Charybdis, between too much order and too much chaos. Collapse is Charybdis – drift past the chaotic edge, loss of coherence. Control is Scylla – sub-critical over-order, sustained only as metastability. The third attractor is the near-critical course between them, which is why it is the only one of the three a living collective agent can actually inhabit.

Keeping an agent away from this course is work: the continuous expenditure of energy on repression, surveillance and propaganda. Franco’s clerical-fascist Spain expended lots of energy to stay stable, and in the end did not: such regimes are metastable states that rigidify too much, lose resilience, and then collapse. Dynamically, a control regime is a metastable basin within a strange attractor, held open by constraint work, not a simple attractor the landscape provides for free.

The point attractor therefore does not live in the dynamics. It lives in the strategic model of the controlling part: the configuration it is trying to hold the system at (uniform stability) and the relation it wants to occupy (control). Point-attractor is the aim, metastability is the achievement, and collapse is the fate. The mechanism of that fate is a variety mismatch: a controller can never match the requisite variety of a complex society, so it must instead reduce the society’s variety – legibility, suppression of fluctuation, elimination of dissent. That is drift away from criticality toward the over-ordered edge; the regime progressively blinds and embrittles itself, loses the information-processing capacity that keeps a living system adaptive, and collapses. Control is not an alternative to collapse but a slower, temporarily stabilised route into it.

The third attractor is a way of moving, not an end state

The third attractor cannot be a settled state. Living Systems dynamically avoid attractors, staying near-critical, away from any stable attractor, because that is what maximises their information-processing capacity. If the collective agent in service of life is a living system – and it must be – then the third attractor is necessarily a strange attractor: states always similar, never the same; an ongoing way of moving, not a destination.

This gives “realistic-yet-utopian” a technical meaning: not a utopia the system arrives at, but a Non-equilibrium Steady State it continually re-achieves. It is also a guard-rail on the vision language – any rendering of the third attractor as a settled future world quietly re-imports the rigid-attractor picture, i.e. the control attractor’s form with nicer content.

Sarri’s critique is structural, not ethical

“We’re not trying to just make coordination go better” – Sarri Bater’s insistence on distributing power rather than merely coordinating better – becomes structural in this vocabulary rather than a caveat. Coordination without distributed agency is movement toward the control attractor. A network that improves coordination while closure remains imposed from a centre is building the second attractor with better vibes. This is why the critique-and-contemplation layer (the reflection layer, renamed from “Compass”) is non-optional: it is the mechanism that detects drift between the two configurations, which are locally indistinguishable – both look like “collaboration going well”.

”In service to life”: being in right relations

Planetary-scale coherence alone does not distinguish the third attractor from the second: the control attractor is also humanity behaving coherently. “In service to life” does the discriminating work – and its content comes into focus through the distinction between the dynamical and the agent perspective, applied to life as a whole.

Dynamically, life is not a set of separate organisms but one vast four-dimensional system – the entire web of relationships and nested systems, extended through time, that living things participate in. (The individual, on this view, is a “time-like” pattern, a worldline correlated with others; cf. Individuals are patterns in larger systems.) To be part of life is simply to be embedded in this web, enmeshed in its relationships and its larger systems.

“In service to life” is the agent-perspective correlate of that embeddedness: being in right relations within the web. From the agent perspective the primary reality is relationships; from those relationships the dynamics of the overall system emerge; and when those dynamics stabilise into a coherent higher-order pattern, that pattern can itself become an Agent – one that then imposes constraints on its constituents in order to stay stable (top-down Causation), while the constituents must “do their part” to keep the whole going. Right relations are what let a higher-order collective agent form at all, and what keep it coherent once formed.

Michael Levin‘s cancer research is a sharp illustration. A cell becomes cancerous when it is cut off from the feedback of the cells around it: disconnected, it “no longer knows” it is part of a wider organism with a role to play, and reverts to the small goal of its own proliferation. In the vocabulary of Agent, its cognitive light cone shrinks back to the single-cell scale; it defects from the higher-order agent and, unchecked, destroys it – and itself. Cancer is not extra malice but lost interconnection: a part whose agency has detached from the whole.

Being “in service to life” therefore means maintaining the interconnection and feedback that keep one in right relations with the larger living systems one is part of – the opposite of the cancerous, separative mode. The experience of separation is the civilisational analogue of the cell cutting itself off, and much of modernity’s pathology is separation at scale (cf. Machado de Oliveira). A collective agent in service to life must therefore itself be built on a self-model of interconnectedness, or it will reproduce the very separation it opposes.

This also gives the control attractor a sharper name: control is cancer at civilisational scale – a part that detaches from the whole’s feedback and subordinates it to its own persistence, killing its host by the metastable route traced above. Testing for that cancerous mode, on Fabric itself, is exactly the job of the critique-and-contemplation layer.

References