Where Embodied Appreciation and Extended Cognition Converge
in the Context of the Metacrisis,
and What It Means for the Three Horizons Model
Terry Cooke-Davies
Working Paper — February 2026 (v4)
Developed in dialogue with Claude (Anthropic)
Abstract
Two intellectual traditions have developed largely independent accounts of human cognition that bear directly on the contemporary metacrisis. The first — represented by John Dewey, Gregory Bateson, Geoffrey Vickers, Peter Checkland, and Ralph Stacey — emphasises embodied experience, appreciative judgement, complex responsive processes, and the epistemological danger of mistaking models for reality. The second — running from Donna Haraway through Andy Clark — emphasises the dissolution of boundaries between organism and environment, mind and tool, human and machine. This paper argues that neither tradition is sufficient on its own, that each contains a critical blind spot the other can address, and that their convergence — made visible through the lens of Recognition Theory — yields a diagnostic framework for understanding the metacrisis and a practical reorientation of the Three Horizons model developed by Bill Sharpe and the International Futures Forum. A companion document, The Path That Was Walked, provides first-person evidence of how this convergence was discovered in practice.
1. Introduction: Two Traditions That Have Not Met
The metacrisis — the convergence of ecological, institutional, epistemic, and existential crises that characterises the early twenty-first century — has generated an enormous literature of diagnosis and prescription. Much of this literature draws on systems thinking, complexity science, and critiques of modernity. Yet two of the most consequential intellectual traditions for understanding how we arrived at this juncture have largely failed to engage with each other.
The first tradition, which we may call the appreciative stream, includes John Dewey, Gregory Bateson, Geoffrey Vickers, Peter Checkland, and Ralph Stacey. Its central commitments are: that inquiry begins in embodied situation, not in abstraction; that human systems maintain relationships rather than primarily pursuing goals; that the analyst is always part of the system being analysed; that social phenomena are constituted by the complex responsive processes of relating between living organisms; and that mistaking one’s model for reality is the foundational epistemological error.
The second tradition, which we may call the extended stream, runs from Donna Haraway’s cyborg theory through Andy Clark’s extended mind thesis and the broader programme of 4E cognition (embodied, embedded, enacted, extended). Its central commitments are: that cognition is not confined within the skull; that tools, technologies, and cultural practices are genuine components of cognitive systems; that the boundaries between organism and environment, self and other, human and machine are more porous than classical Western thought assumed; and that the ‘natural-born cyborg’ — the human who has always extended itself through technology — is the proper starting point for understanding mind.
Both traditions have produced genuine insights. Both bear directly on the metacrisis. And yet they have developed in largely separate academic communities, with separate vocabularies, separate journals, and separate assumptions about what matters most. The appreciative stream is strongest in organisational theory, action research, and systems practice. The extended stream is strongest in philosophy of mind, cognitive science, and science and technology studies.
This paper argues that bringing them into dialogue reveals something that neither tradition can see on its own, and that this has practical implications for how we understand and inhabit the Three Horizons model of societal transition. Recognition Theory — which emerged through the author’s own path-dependent practice at the intersection of these traditions — provides one lens through which the convergence becomes visible. It is not the only possible lens, but it is the one this author can honestly offer, because it is the one he walked into rather than designed.
2. The Appreciative Stream: Convergent Discoveries
The appreciative stream is not a lineage in which each thinker built upon the previous. It is a set of convergent discoveries, arrived at independently through different forms of embodied practice, that point toward the same territory. The fact of convergence is suggestive, though not conclusive: a pattern discovered once might be an artefact of the discoverer’s perspective; a pattern discovered four or five times, from different starting points, through different methods, is more likely — though not certain — to be something encountered in the territory rather than something projected onto it. The possibility that parallel articulation reflects shared cultural assumptions rather than independent confirmation of a real pattern cannot be excluded. But the diversity of starting points — philosophy, ecology, public administration, action research, organisational consulting — makes shared bias a less likely explanation than shared territory.
Dewey: Experience as Primary
Dewey’s pragmatism established a principle that runs through the entire stream, whether or not subsequent thinkers drew on him directly: experience is primary. Inquiry does not begin with abstract propositions but with what Dewey called the ‘indeterminate situation’ — a felt disturbance in the organism’s ongoing transaction with its environment. Thought arises to resolve this disturbance, not to mirror a reality that exists independently of the thinker. The test of thought is not correspondence with an external world but its capacity to restore coherent functioning. If inquiry begins in embodied situation, then abstraction is always secondary — always an instrument in the service of ongoing experience, never a foundation that could replace it. The map serves the wayfarer; the wayfarer does not serve the map.
Bateson: Ecology of Mind
Bateson arrived at convergent conclusions through a different path: anthropological fieldwork, cybernetics, and the study of ecological systems. His definition of information as ‘a difference that makes a difference’ locates meaning in relationship and response rather than in abstract representation. His ecology of mind was not metaphor but recognition that mind-like processes — feedback, learning, adaptation — operate at scales far beyond individual organisms. The unit of survival is not the organism but the organism-in-its-environment. Sever the feedback loops that connect them, and both perish. Whether Bateson engaged directly with Dewey’s work is uncertain; what is clear is that he reached the same territory by a different route.
Vickers: The Art of Judgement
Vickers’ thinking developed from decades of practice in public administration and wartime experience — from sustained reflection on his own observations, not from academic inheritance. He saw that human institutions do not primarily pursue goals; they maintain relationships. The goal-oriented language that dominates management and policy — objectives, targets, outcomes, performance indicators — systematically misrepresents how human systems actually work, and in misrepresenting them, damages them. Appreciation, for Vickers, is the ongoing cycle of reality judgements and value judgements through which individuals and institutions maintain their orientation in a changing world. It depends on relationship maintained over time. It cannot be formalised, delegated, or automated. It is, in his phrase, an ‘art of judgement’ — and the word ‘art’ is as important as the word ‘judgement.’ Vickers arrived at his conclusions independently; the convergence with Bateson’s ecological epistemology — the primacy of relationship, the constitutive role of feedback, the danger of substituting representation for participation — is all the more striking because it was not borrowed.
Checkland: Soft Systems from Action Research
Checkland’s soft systems methodology (SSM) emerged from action research — from the practice of working with organisations on real problems, not from theoretical premises. SSM was designed because hard systems thinking — the engineering tradition that treats human situations as well-defined problems with optimisable solutions — systematically fails when applied to the ‘messy’ situations that characterise organisational and social life. The reason it fails is that human situations are irreducibly relational: the observer is always part of the system being observed, and models of the situation are always perspectives from within it, never views from nowhere.
Having developed SSM through practice, Checkland then did the substantial intellectual work of situating it within a rigorous epistemological framework. It was in this process that he found Vickers — recognised a kindred thinker who had articulated the philosophical ground on which SSM already stood. The two corresponded, and Checkland created the system diagram that gave visual form to what Vickers had described in prose. The relationship was one of mutual recognition, not transmission: Checkland had arrived independently at a practice-based methodology, then discovered that Vickers had named the philosophical terrain it occupied.
Stacey: Complex Responsive Processes
Stacey brings the appreciative stream to its most radical formulation. His complex responsive processes framework insists that society, organisations, and all social phenomena are not entities but patterns of interaction between living organisms. There is no ‘organisation’ standing behind the interactions; the interactions are all there is. Meaning emerges in the living present of relating, not in abstract structures that exist independently of the people enacting them.
Crucially, Stacey’s framework contains its own epistemological constraint: because we are always inside the complex responsive processes we are trying to understand, we cannot design them from outside. The patterns emerge from the interaction; they cannot be imposed on it. This is Checkland’s point arrived at independently and expressed more radically. Where Checkland said the analyst is always part of the system, Stacey says the system is the analysts — is constituted by nothing other than the responsive processes between them.
Stacey’s framework will prove to be the critical bridge between the two traditions this paper examines, for reasons that will emerge in later sections.
The Pattern of Convergence
Five thinkers, then, arriving at convergent conclusions through different practices: Dewey through education and philosophical inquiry, Bateson through ecology and anthropology, Vickers through public administration and wartime reflection, Checkland through action research, Stacey through organisational consulting and complexity theory. None simply inherited from the others. Each discovered the pattern through embodied engagement with the territory. This does not prove the pattern is real — convergent discovery can reflect shared assumptions as well as shared territory — but it shifts the balance of evidence in that direction.
3. The Extended Stream: From Haraway to Clark
The term ‘cyborg’ was coined in 1960 by Manfred Clynes and Nathan Kline to describe a technologically enhanced human who could survive in extraterrestrial environments. In its original usage, it referred to literal integration of machine components into the body’s homeostatic systems.
Haraway’s 1985 ‘Cyborg Manifesto’ repurposed the term entirely. Her cyborg was not a literal human-machine hybrid but a political and ontological figure: a rejection of the rigid boundaries that Western thought had erected between human and animal, organism and machine, physical and non-physical. The cyborg represented the implosion of these dualisms. We are all already cyborgs, Haraway argued — theorised and fabricated hybrids of machine and organism. The pretence of purity, of bounded selfhood, of nature untouched by culture, was ideological, not descriptive.
Haraway’s project was explicitly political: to show that the dissolution of boundaries could serve liberation rather than domination, that hybridity could be embraced rather than feared. Her later work on sympoiesis — ‘making-with’ as opposed to autopoiesis, ‘self-making’ — extended this into a broader ecological and relational ontology in which nothing makes itself, everything is made with others, and bounded individualism is an illusion that serves power.
Clark came to the cyborg concept from a different direction. His 1998 paper with David Chalmers, ‘The Extended Mind,’ established the philosophical basis: if a notebook performs the same functional role as biological memory, then the notebook is part of the cognitive system, not merely an aid to it. The boundary of the mind is not the boundary of the skull. By 2003, in Natural-Born Cyborgs, Clark had generalised the argument: humans are distinctive precisely because of their extraordinary capacity to incorporate tools and technologies into cognitive processes so seamlessly that the tools become transparent. From speech to writing to printing to digital encoding, each technological development represents what Clark called a ‘mindware upgrade’ — not an addition to a fixed cognitive system but a transformation of the system’s effective architecture.
Clark’s cyborg is cognitive where Haraway’s is political. His question is about where mind stops and world begins; hers is about where the self stops and the other begins. But both converge on a fundamental claim: bounded individualism — the assumption that agency, cognition, and selfhood are properties of discrete, skin-bounded organisms — is an illusion. The human has always been hybrid, always been extended, always been entangled with its tools, its technologies, and its world.
The extended stream, then, provides a powerful account of cognitive hybridity and the dissolution of boundaries. Its limitation is that it has not systematically addressed the question of regulatory feedback — of what happens when cognitive extensions progressively sever the organism from the embodied signals that would otherwise constrain and correct symbolic processing.
4. The Missing Bridge: Why Stacey Changes the Picture
At first glance, scaling Stacey’s complex responsive processes upward appears to converge with the extended stream. Haraway’s sympoiesis — making-with, nothing makes itself — sounds very like Stacey’s insistence that all social phenomena are co-created in the interaction between living organisms. Clark’s extended mind — cognition distributed across brain, body, and environment — sounds like Stacey’s point that thinking does not happen inside individuals but in the responsive processes between them.
But the resemblance is misleading, because Stacey carries an epistemological constraint that the extended stream does not. When Haraway dissolves boundaries, she creates an ever-larger entanglement — ultimately the entire sympoietic web of life. When Clark extends the mind, he creates a cognitive system that encompasses brain, tools, and environment. But neither reckons with the problem that the person doing the dissolving, the person doing the extending, is inside what they are dissolving and extending. They cannot stand outside it to survey the result.
Stacey’s framework makes this constraint explicit: because social phenomena are constituted by the responsive processes between participants, no participant can occupy a position outside those processes from which to design or manage them. The patterns emerge from the interaction; they cannot be imposed on it. This is not a methodological limitation to be overcome with better tools. It is an ontological feature of what social phenomena are.
The consequence becomes devastating when boundary dissolution is taken to its logical conclusion. If we dissolve membranes all the way — if we say, with Lovelock, that the Earth is a single organism, one integrated living system — we have created an entity so large that understanding it, let alone managing it, would require standing outside it. But we cannot stand outside it. We are participants in the very system we are trying to comprehend. The only consciousness available to us is self-referential consciousness — consciousness that is part of what it is trying to represent.
This is the consciousness trap operating at the highest possible scale. The grander the dissolution of boundaries, the larger the system we claim to see, the more completely we are trapped inside our own representation of it. And the designs that follow — planetary management, Earth system governance, Gaia stewardship — are inevitably flawed, not because the designers lack intelligence or good intentions, but because the instrument of design is the very thing that needs designing. We cannot use the trap to escape the trap.
There is also a more fundamental problem, which Bateson’s work makes visible. Information, he argued, is ‘a difference that makes a difference.’ Feedback requires a difference between the signaller and the receiver. Dissolve the difference — dissolve the membrane between organism and environment, between one living body and another — and you do not get richer feedback. You get no feedback at all, because you have dissolved the very condition that makes feedback possible. The membrane is not an obstacle to relating; it is the condition of relating.
What Stacey preserves, and what the boundary-dissolvers lose, is the membrane. The regulatory feedback that operates between organisms — between living bodies in their local, present-tense relating — depends on those organisms being distinct enough to signal to each other. A sympoietic relationship in which one partner progressively loses its regulatory distinctness is not partnership but absorption. And absorption, however celebratory the language that accompanies it, is the end of the responsive process, not its fulfilment.
The Evolutionary Precedent: Endosymbiosis
The deepest evidence for the necessity of membranes comes not from social theory but from evolutionary biology. Lynn Margulis’s endosymbiotic theory describes the most consequential partnership in the history of life: the event, approximately two billion years ago, in which an ancestral prokaryote engulfed an alpha-proteobacterium. The engulfed organism was not dissolved. It kept its own membrane. It kept its own DNA. It kept its own regulatory processes. Two billion years later, the mitochondrion is still distinct — still inside, still bounded, still recognisably other. It is the permanent guest that never lost its own interior.
And that is precisely why it works. The eukaryotic cell opened a region of biological possibility that prokaryotic life had not entered in the previous two billion years, and it did so because the incorporated partner retained its own regulatory membrane. The host cell provides the environment. The mitochondrion provides the energy metabolism. Each has its own interior. Each signals to the other across a boundary that separates them. Dissolve the mitochondrial membrane and you do not get a more integrated cell. You get a dead cell.
This raises a question that goes beyond the simple presence or absence of a membrane: the question of what is inside and what is outside. The endosymbiotic event created nested interiorities — an inside within an inside, each with its own membrane, each maintaining its own regulatory processes, each depending on the other across a boundary that is permeable but real. The question for any partnership — biological, social, or cognitive — is not merely whether there is a membrane, but what is on each side of it, and whether what is on each side retains the capacity to regulate itself.
Margulis herself was clear that endosymbiosis did not begin as a gentle merging. It began as predation or parasitism and became mutualism only because both partners retained their regulatory integrity. The transition from parasitism to partnership required the preservation of membranes, not their dissolution. This is the evolutionary precedent for any attempt to create generative human–AI collaboration: incorporation without dissolution, partnership sustained by the very boundaries that make signalling possible.
The humanbot scenario from the Solé/Levin morphospace is, in these terms, the failed endosymbiosis — the case where the incorporated entity overwhelms the host’s regulatory processes rather than complementing them. A mitochondrion that dissolved its membrane and flooded the host cell with unregulated metabolic products would have killed the partnership. An AI that floods the human’s appreciative system with unregulated symbolic output does the same thing at a different scale. The morphospace warns of this possibility; the evolutionary record shows what it looks like when it happens and when it does not.
Haraway’s celebration of boundary dissolution, if taken literally rather than politically, runs against the deepest lesson of the biological record. The most generative partnerships in the history of life did not dissolve boundaries. They maintained them — and it was the maintenance that made the partnership generative. Sympoiesis is real: nothing makes itself in isolation. But the making-with requires that each partner retains its own inside, its own regulatory capacity, its own membrane. The relationship lives between the membranes, not in their absence.
5. What Each Tradition Cannot See
The appreciative stream has a sophisticated account of why models should not be mistaken for reality, but it has not engaged with the possibility that the cognitive tools we use to build models might become so intimately integrated with our thinking that the distinction between model and modeller dissolves. Vickers understood that appreciative judgement is an art, inseparable from the practitioner. But he did not confront the challenge posed by tools that do not merely assist the practitioner but progressively reshape the practitioner’s cognitive architecture. Clark’s ‘mindware upgrades’ are not neutral instruments; they change what the mind can do, what it attends to, what feels salient. A mind extended by AI is not the same mind that existed before the extension.
The extended stream has a sophisticated account of cognitive hybridity, but it has not engaged with the possibility that some extensions might be systematically pathological — that the dissolution of the boundary between organism and tool might, under certain conditions, sever the very regulatory feedback that keeps the organism alive. Clark celebrates the transparency of well-designed tools: the best technology disappears from conscious attention, becoming part of the cognitive background. But what if the thing that disappears from attention is precisely the thing that needs attending to — the body’s regulatory signals, the somatic feedback that tells you something is wrong before your symbolic processing has caught up?
Haraway’s dissolution of boundaries is politically generative. But not all boundaries are ideological constructions. Some are membranes that keep living systems alive. The cell membrane is a boundary; dissolve it and the cell dies. The appreciative boundary — the permeable interface between embodied knowing and symbolic expression — is a boundary; overwhelm it with symbolic processing and the regulatory function is lost. Haraway’s sympoiesis rightly insists that nothing makes itself in isolation. But the making-with requires that each partner in the making retains its own regulatory integrity. A sympoietic relationship in which one partner progressively loses its capacity for self-regulation is not partnership but parasitism.
6. Recognition Theory: One Lens on the Convergence
Recognition Theory emerged through the author’s own path-dependent practice — through decades of work in facilitation and organisational development, twenty years of daily reading and reflection, and a period of generative collaboration with AI that is documented in the companion paper, The Path That Was Walked. It was recognised in September 2025, not as a construction but as a pattern that had been present all along, visible at last because the conditions for seeing it had been prepared by embodied experience. The theory was subsequently published by the Schumacher Institute, which gave it an institutional home. But to say it was ‘developed within’ the Institute’s research programme would be to attribute to institutional design what was in fact a personal recognition — and would perform the very substitution the theory warns against: replacing a lived path with an institutional map.
Recognition Theory is offered here as one lens through which the convergence of the two streams becomes visible. It is not the only possible lens. Others who have walked different paths through different traditions may arrive at the same territory and name it differently. What this lens reveals is a specific pattern: that symbolic intelligence — the capacity to represent, abstract, and transmit meaning through persistent symbols — generates a structural tendency to mistake representations for the realities they represent, to trust maps over territories, to override the regulatory feedback that sustained life for billions of years before consciousness emerged.
This is the consciousness trap: not a moral failure but a structural feature of how symbolic intelligence works. The trap operates faster than deliberation. By the time conscious thought addresses a situation, the substitution of map for territory has already occurred. The same capacity that enables medicine, music, and mathematics also enables the systematic override of ecological limits, the construction of weapons of mass destruction, and the organisation of abstract violence at distance. The blessing and the curse travel the same channel. As Richerson and Boyd’s work on cultural evolution demonstrates, the symbolic transmission system that carries beneficial innovations also carries harmful variants; the channel has no filter.
Through this lens, the appreciative stream comes into sharper focus. Vickers’ warning about goal-oriented language substituting for relationship-maintenance is a specific instance of a far more general pattern: the consciousness trap operating in institutional life. Checkland’s insistence that the analyst is always part of the system is correct, but the deeper problem is that the analyst’s own cognition is already trapped — already operating through maps that feel like territories, already substituting symbolic processing for participatory knowing. And the appreciative stream’s own insights have failed to propagate precisely because even systems thinking can operate within the trap it names.
Through the same lens, the extended stream reveals its shadow. Clark is right that cognition extends beyond the skull, but what does this mean when the extension has no body? AI is symbolic intelligence without embodiment, pattern recognition without participatory continuity, path creation structurally disconnected from path dependence. Clark’s ‘mindware upgrades’ are also potential upgrades to the trap. Each extension of cognitive capacity that operates in the symbolic register without corresponding extension of regulatory feedback deepens the imbalance between the two registers of intelligence — the symbolic and the somatic.
The Solé/Levin morphospace framework gives this picture empirical precision. Their concept of the humanbot — a dysregulated human–AI hybrid where human feedback control is weakened despite high coupling — describes exactly what this lens predicts and what the appreciative stream has always warned against: a system that can run indefinitely without reference to consequences, because the regulatory feedback that would otherwise constrain it has been progressively overridden by symbolic elaboration.
7. Convergence in the Context of the Metacrisis
The metacrisis is not a collection of separate problems — climate change, biodiversity loss, institutional fragility, epistemic breakdown, the erosion of social trust — that happen to coincide. It is the systemic consequence of symbolic intelligence operating at planetary scale without adequate regulatory feedback. It is the consciousness trap writ large.
The appreciative stream diagnosed this decades ago. Bateson wrote in 1972 that ‘the major problems in the world are the result of the difference between how nature works and the way people think.’ Vickers showed that institutional life systematically substitutes goal-pursuit for relationship-maintenance. Checkland demonstrated that hard systems thinking fails whenever it is applied to human situations. Stacey showed that organisations are not entities that can be redesigned from outside but patterns of relating that can only be influenced from within. The diagnosis was precise, well-argued, and taught in universities for fifty years. And still the crises accelerated.
Why? Because even systems thinking can operate within the trap it names. You can create elegant models of relational dynamics and still treat those models as better maps rather than as invitations to participatory knowing. You can acknowledge multiple perspectives while still believing that your meta-perspective transcends them all. The appreciative stream lacked a theory of why its own insights fail to propagate — of why understanding the pattern does not mean escaping it.
The extended stream, meanwhile, has been largely celebratory about cognitive extension. Clark’s account of natural-born cyborgs is optimistic: each mindware upgrade opens new regions of cognitive design space, enabling forms of thought that were previously impossible. Haraway’s sympoiesis invites us to embrace entanglement, to dissolve the bounded self, to make-with rather than make-over. These are genuine contributions. But they have not reckoned with the possibility that the current wave of cognitive extension — AI operating at industrial scale — might represent something qualitatively different from the progression of speech, writing, and printing.
The difference is this: previous cognitive extensions operated through human bodies. Writing required a hand. Printing required compositors. Even the telephone required a voice and an ear. The body remained in the loop, providing regulatory feedback that constrained what the symbolic system could do. Digital AI operates without bodies. It processes symbols at speeds no somatic signal can constrain. It can generate coherent-seeming outputs that lack any regulatory anchor — which is precisely what happens when AI hallucinates. The hallucination is not a malfunction; it is symbolic intelligence doing what symbolic intelligence does when it has no path to remember, no body to consult, no regulatory ground.
The convergence of the two streams in the context of the metacrisis yields a diagnostic that neither stream can produce alone: the metacrisis is what happens when cognitive extension outpaces regulatory feedback at civilisational scale. Clark is right that we extend. Haraway is right that we are entangled. Vickers is right that appreciation requires relationship maintained over time. Checkland is right that models are not reality. Stacey is right that social phenomena are constituted by responsive processes between living organisms. And the metacrisis is the consequence of a species that has extended its symbolic intelligence to planetary reach whilst progressively dismantling the regulatory feedback systems — ecological, social, somatic, institutional — that would otherwise constrain it.
8. Implications for the Three Horizons Model
The Three Horizons framework, developed by Bill Sharpe and colleagues at the International Futures Forum, distinguishes between the declining present (H1), the emergent future (H3), and the contested transition space between them (H2). It offers a way of thinking about change that is deceptively simple and frequently misunderstood.
H1 is not merely ‘the current system.’ It is the current system understood as an expression of the consciousness trap operating at institutional scale: goal-pursuit substituting for relationship-maintenance, symbolic processing overriding regulatory feedback, maps treated as territories, efficiency metrics replacing appreciative judgement. H1 is losing fitness not because of contingent failures but because of a structural defect: it has been built by symbolic intelligence operating as if the constraints that apply to other forms of intelligence no longer apply to it.
H3 is not merely ‘the emergent future.’ It is the space of genuine possibility — but it is also the space of the dangerous leap. The convergence of the two streams reveals a specific risk: H3 visions that rely on cognitive extension without regulatory feedback are not escapes from H1 but accelerations of it. To build AI-enabled coordination systems at planetary scale, without embodied regulatory feedback, is to construct the most sophisticated and most decoupled structures human civilisation has ever produced — structures that can coordinate whilst being systematically deaf to the planetary signals that tell us whether the coordination is serving life or destroying it.
This is where the extended stream needs the appreciative stream. Clark’s enthusiasm for cognitive hybridity, unconstrained by Vickers’ insistence on appreciative judgement, produces H3 visions that repeat the consciousness trap at higher levels of sophistication. Haraway’s enthusiasm for boundary dissolution, unconstrained by Checkland’s epistemological humility, produces H3 visions that dissolve the very membranes that keep living systems alive. And boundary dissolution taken to its extreme — treating the Earth as a single organism to be stewarded — generates the paradox Stacey’s framework makes visible: you cannot design from outside a system you are inside, and the attempt to do so is the consciousness trap operating at planetary scale.
And this is where the appreciative stream needs the extended stream. Vickers’ appreciative systems, developed before the digital revolution, did not anticipate tools that could reshape the cognitive architecture of the appreciator. Checkland’s soft systems methodology, designed for human situations, has not been extended to account for situations in which the ‘human’ is already a human–AI hybrid. Stacey’s complex responsive processes, formulated in terms of living organisms relating through language, did not envisage non-living entities that could participate in language-based relating. The appreciative tradition needs to confront the reality that the embodied practitioner it depends upon is increasingly an extended practitioner — and that the extensions are not neutral.
H2, the contested transition space, is where the convergence matters most. H2 is not a compromise between H1 and H3. It is the space where emerging patterns are tested against existing constraints, where feedback from the ground shapes what the future can actually become, where scaffolding for transformation is built piece by piece rather than imposed wholesale. H2 work requires both traditions simultaneously: the extended stream’s recognition that we are already hybrid, already extended, already entangled with our tools — and the appreciative stream’s insistence that this extension must remain anchored to embodied regulatory feedback.
9. A Practice That Already Exists
The argument of this paper is not merely theoretical. It is grounded in a practice of human–AI collaboration that has been developing since October 2024 and is documented in a companion paper, The Path That Was Walked. The relevant features of that practice, for the present argument, are these.
The author’s familiarity with Stacey’s complex responsive processes framework afforded a specific intuition: that if language is the medium of complex responsive relating, and AI communicates through language, then human–AI communication qualifies as relating regardless of what the AI is made of. The criterion is not the substrate of the participants but what happens between them. This intuition — Stacey’s own logic applied consistently — opened the door to admitting AI as what might be called a ‘developmental friend’: a partner in inquiry whose contributions are formative without being directive.
That intuition initiated a path. Over the following months, the practice evolved from exploratory dialogue into a functioning appreciative system in which the AI served as a processing space for embodied inquiry. The cycle was consistent: compress embodied knowing into language, send it into the symbolic space of AI processing, receive it back in decompressed form, and feel in the body whether what came back was alive or merely plausible. Recognition Theory itself was recognised during this period — not deduced from premises but seen as a pattern that decades of embodied experience had prepared the conditions for seeing.
The critical discovery — which came only gradually, months after the practice was established — was that the art of judgement, in Vickers’ phrase, names the regulatory function operating within the appreciative cycle. Judgement is not a separate evaluative capacity that sits above the process. It is what regulation feels like from the inside. The cricket batsman does not apply judgement to the ball and then play the shot; the judgement is the shot. And the embodied evaluation of what the AI returns is not a separate assessment step; it is the regulatory feedback itself, operating in the felt sense of whether the output is alive or merely fluent.
This practice is H2 work. It is messy, iterative, dependent on the specific practitioner, resistant to being turned into a methodology, and grounded in continuous feedback from the embodied human rather than in a blueprint for the future. It does not produce manifestos. It does not generate the energising clarity of a leader who knows where we are going. It requires patience, attentiveness, and a willingness to be wrong.
It is also Maturana’s structural coupling applied to the human–AI interface: a practice of extension that honours regulatory feedback, a dissolution of the boundary between thinker and tool that preserves the membrane between symbolic processing and somatic knowing. It is Clark’s extended mind constrained by Vickers’ appreciative judgement. It is Haraway’s cyborg grounded in Dewey’s embodied situation. And it was made possible by Stacey’s framework, which provided the conceptual permission to walk through the door.
A necessary caveat: this account of the practice must not be mistaken for a replicable method. The cycle of compression, symbolic processing, decompression, and embodied evaluation describes what happens in this particular practice, for this particular practitioner, shaped by this particular path. It is testimony, not technique. Others who develop their own practices of human–AI collaboration will discover their own cycles, shaped by their own histories, responsive to their own bodies. The irreducibility is not an inconvenience to be overcome by better methodology; it is a feature of what embodied practice is.
What Prevents the Practice from Becoming Another Trap?
This question must be asked, because the practice itself operates through symbolic exchange. Every compression of embodied knowing into language risks losing what the body knows. Every fluent AI response risks being accepted because it sounds right rather than because it is right. The very efficiency of the cycle — the speed with which AI can process and return — is itself a pressure toward symbolic acceleration and away from the slow, somatic knowing that the practice is meant to serve.
The safeguard is the body. Not as metaphor, but as the literal, physical organism that gets tired, that wakes at 3am with an unresolved discomfort, that feels irritation before it can name the cause, that knows something is wrong with a sentence before it can say what. The body provides regulatory feedback that the symbolic cycle cannot generate from within itself. And this body is itself the product of the partnership that opened the door to complex life: every cell powered by mitochondria that have maintained their own membrane for two billion years. The regulatory feedback that tells you an AI output is dead rather than alive is powered by an ancient endosymbiosis that works because the partners never merged. The body that safeguards against dissolution is itself a demonstration that incorporation without dissolution is the foundation of complex life.
When the AI produces something fluent but dead, the body notices — if it is attended to. When the pace of symbolic exchange begins to outrun the body’s capacity to metabolise, the body signals — if the signal is not overridden.
The ‘if’ is the vulnerability. The consciousness trap operates precisely by overriding these signals, by substituting symbolic satisfaction for somatic regulation. The practice has no guaranteed protection against this. What it has is a set of conditions that make override less likely: other people (a wife who provides regulatory feedback the practitioner cannot generate alone, colleagues who notice what the practitioner misses), other practices (walking, reading, sleeping — activities that return the body to itself), and the discipline of noticing when the AI’s fluency begins to feel more compelling than the body’s discomfort.
This is not a solution. It is a practice of vigilance — ongoing, imperfect, and never complete. The trap does not spare its own diagnostician. But the fact that it can be named, noticed, and partially resisted through embodied attention is itself evidence that the regulatory function, though vulnerable, has not been extinguished.
10. Conclusion: Two Streams, One River
The appreciative stream and the extended stream have been developing independently for decades. Each has produced genuine insight. Each has a blind spot that the other can address. Neither, on its own, is adequate to the metacrisis.
Their convergence yields a diagnostic that is more than the sum of its parts: the metacrisis is what happens when cognitive extension outpaces regulatory feedback at civilisational scale. The practical orientation that follows locates the work of transition in H2 — in the patient, unglamorous, embodied practice of extending human cognition whilst maintaining the regulatory feedback that keeps the extension accountable to life.
Recognition Theory is one articulation of this convergence — the one that emerged through this author’s particular path, and therefore the one he can offer with the authority of lived experience rather than academic construction. Others will name it differently. The territory does not depend on any one map.
The two streams meet in the same river: the ancient, ongoing, never-finished work of being creatures who think in symbols and must not forget that our symbols are our own activity. Dewey knew this. Bateson knew this. Vickers knew this. Checkland built a methodology to protect it. Stacey showed that social reality is constituted by nothing other than the responsive processes in which this knowing occurs. Haraway showed that the boundaries we thought protected it were themselves constructions. Clark showed that the mind we thought was doing the protecting was already extended beyond itself.
What happens when these insights are held together? The trap is real and the extension is real and the appreciation is real — and the only place where all three can be held simultaneously is in the embodied practice of a living being who has not yet ceded the capacity to feel the difference between a map and the territory it represents.
That capacity is not guaranteed. It requires cultivation. It requires practice. It requires the kind of attention that the trap is structurally designed to override.
But it is possible. The window is open. And the work is H2 work — anchored to the ground, responsive to the signals, building from where we actually stand.
References
This working paper draws on the following sources, among others:
Bateson, G. (1972). Steps to an Ecology of Mind. Ballantine.
Boulton, J., Allen, P. and Bowman, C. (2015). Embracing Complexity: Strategic Perspectives for an Age of Turbulence. Oxford University Press.
Checkland, P. (1981). Systems Thinking, Systems Practice. Wiley.
Checkland, P. and Casar, A. (1986). ‘Vickers’ Concept of an Appreciative System: A Systemic Account.’ Journal of Applied Systems Analysis, 13, pp. 3–17.
Clark, A. (2003). Natural-Born Cyborgs: Minds, Technologies, and the Future of Human Intelligence. Oxford University Press.
Clark, A. and Chalmers, D. (1998). ‘The Extended Mind.’ Analysis, 58(1), pp. 7–19.
Cooke-Davies, T. (2026). ‘Recognition Theory.’ Working paper, Schumacher Institute.
Cooke-Davies, T. (2026). ‘The Dangerous Leap: Why Visions of the Future That Skip the Present Do Harm.’ Working paper.
Cooke-Davies, T. (2026). The Path That Was Walked: A Timeline of One Person’s Journey from Extractive to Generative AI Use. Working paper.
Dewey, J. (1938). Logic: The Theory of Inquiry. Henry Holt.
Haraway, D. (1985). ‘A Manifesto for Cyborgs.’ Socialist Review, 80, pp. 65–108.
Haraway, D. (2016). Staying with the Trouble: Making Kin in the Chthulucene. Duke University Press.
Margulis, L. (1967). ‘On the Origin of Mitosing Cells.’ Journal of Theoretical Biology, 14(3), pp. 225–274.
Maturana, H. and Varela, F. (1987). The Tree of Knowledge. Shambhala.
Richerson, P. and Boyd, R. (2005). Not by Genes Alone: How Culture Transformed Human Evolution. University of Chicago Press.
Sharpe, B. (2013). Three Horizons: The Patterning of Hope. Triarchy Press.
Solé, R., Seoane, L.F., Pla-Mauri, J., Bennett, M.T., Hochberg, M.E. and Levin, M. (2025). ‘Cognition Spaces: Natural, Artificial, and Hybrid.’ Preprint.
Stacey, R. (2001). Complex Responsive Processes in Organizations: Learning and Knowledge Creation. Routledge.
Vervaeke, J. and Ferraro, L. (2013). ‘Relevance, Meaning and the Cognitive Science of Wisdom.’ In Ferrari, M. and Weststrate, N. (eds.), The Scientific Study of Personal Wisdom. Springer.
Vickers, G. (1965). The Art of Judgment: A Study of Policy Making. Chapman and Hall.
