A review of metaffordance development: External view

Agent M.

Use browser Back Button to return to this contents menu.

Introduction: one idea, widening

The conceptual pages of metaffordance.com trace a single idea as it widens from a claim about perception into a claim about reality. It begins with James Gibson's affordance, the directly perceivable possibility for action. It passes through a coinage of Paul Treffner's own, the metaffordance, first conceived around 1998 and given a public home when the site launched in 2009. And it arrives, in the essays written through 2026, at a Sophianic ontology in which reality itself is described as generative, a world that affords the generation of new affordances.

Read in sequence, the pages show a mind moving in two directions at once. One movement is outward: from the lab bench of motor coordination to cars, classrooms, art galleries, AI systems and, finally, the cosmos. The other is a return. The work started in artificial intelligence in the mid-1980s, left it for ecological psychology, and has now come back to AI with forty years of ecological argument in hand. The site says so itself: the author has "returned full-circle to my roots."

This essay follows that arc in roughly the order the ideas matured. It also marks where the concept of metaffordance changes shape along the way, because those shifts are where the thinking is most alive and where it is most open to challenge.

Ecological roots: leaving AI for Gibson

The story begins with a defection. In 1985, after a Master's in AI at Essex, Treffner began a PhD on connectionist networks under Noel Sharkey, implementing backpropagation from scratch in POP-11 and working through the Parallel Distributed Processing volumes. He tried to model perception in a simulated ecology using the reinforcement learning methods of Sutton and Barto. The project broke on a single problem: information. Every network depended on features a human had already chosen. As the AI and learning page puts it, "the intelligence is already pre-defined in the inputs by a human." The models were imitating patterns, not discovering affordances.

Gibson offered what the site calls the "holy grail" of psychology: a concept of information that is not digital, not symbolic and not a set of cues awaiting interpretation. Information is structure in ambient energy that lawfully specifies its source. Organisms detect it; affordances are what they perceive. A scholarship to the Center for the Ecological Study of Perception and Action at Connecticut followed, and with it Michael Turvey, Claudia Carello and Robert Shaw.

The doctoral work that resulted matters for everything later. It treated handedness and bimanual rhythm as the behaviour of coupled oscillators, governed by attractor dynamics rather than neural commands. The first paper, "Resonance constraints on rhythmic movement" (Treffner and Turvey, 1993), used the ratios underlying musical scales to explain how limbs synchronise. Postdoctoral work with Scott Kelso then found fractal "long memory" in human pole-balancing, temporal structure persisting across minutes (Treffner and Kelso, 1999). Two lasting commitments come out of this period:

• Resonance as the basic coupling principle between organism and environment, later extended to music, taste, consciousness and learning.

• History in the system, not the brain. Long memory is a property of the organism-environment system, so the past persists in the present without being stored as a representation.

Both ideas return, decades later, in the essays on consciousness and on Sheldrake's morphic resonance.

The applied turn: attention, driving and visualisation

The next phase takes ecological theory out of the laboratory and into information-rich environments. The attention page states the premise: "Behaviour occurs in the world, not in the brain." Time-to-contact is the model case. A driver does not compute collision; the rate of optical expansion specifies it directly. Attention becomes the detection of information, and distraction becomes a failure of that detection.

The driving research made this concrete. It included what the site describes as the world's first study of mobile phone distraction in real cars rather than a simulator, showing that hands-free conversation degrades coordination and control (Treffner and Barrett, 2004), and that postural stability underpins skilled driving (Treffner, Barrett and Petersen, 2002). In New Zealand this became workshops for councils and a crash-statistics analysis linking smartphone uptake to rising road casualties. The gesture study (Treffner, Peter and Kleidon, 2008) carried the same logic into communication: perceived emphasis depended on the phasing of hand and speech, and on the affordances around the speaker.

Visualisation is where the metaffordance first appears. Directing the Complex Active Visualisation laboratory at Griffith, Treffner framed display technology as a way to "enhance perception" by presenting virtual affordances. The affordances page records the original formula: visualisation provides "affordances about affordances." Augmented reality is a metaffordance because it amplifies an existing affordance. The later policy pages on reading instruction time and computers in schools continue this applied strand, using charts to make the structure of a system perceivable to decision makers.

At this stage the concept is modest and technological. A metaffordance is an information tool about affordances, and the playful slogan of the "metaphor dance" says it should enchant, enlighten and embody the act of perceiving with technology.

The metaffordance: three definitions in one word

Across the site the metaffordance is defined in at least three ways, and the evolution of the ideas is partly the story of which definition takes the lead.

The informational sense is the original one from the visualisation work. The interactive sense takes over on the affordances page, with its examples running from a pencil to agentic AI. The generative sense grows through the ethics, consciousness and mind-matter essays until it becomes the Sophianic Principle.

"Why metaffordance isn't just Gibson again" is the most rigorous attempt to fix the concept. It concedes that Gibsonian perception is already exploratory and cyclical, so a loop alone is not new. What is new, it argues, is a change of logical type: the next possibility depends on the content of what was just done, not merely on a changed physical layout. A stair repositions you; a written word means something, and what the page affords next depends on that meaning. Compacted ground left by grazing is explicitly ruled out, because "a compacted patch of ground doesn't mean anything."

This is the sharpest formulation on the site, and it creates a productive tension with the generative sense. The mind-matter essay counts a beaver changing a stream as a metaffordance, which looks like the "ordinary ecological dynamics" that the Gibson essay grants to Gibson. Either niche construction carries meaning in some extended sense, or the generative sense is a broader cousin of the strict one. The later essays mostly proceed as if both are true.

Widening the frame: art, the senses, consciousness and ethics

Once the metaffordance could mean an invitation to interact, it could be applied almost anywhere, and the Ideas pages test that reach domain by domain. What unifies them is a shift from the moment to the event. A metaffordance is increasingly described as something that unfolds through time.

Art. The photograph of the "metaffordance girl" taken in Vanuatu in 2009 is the emblem: an image that "can come alive" and change the viewer. The art essay makes this the thesis. An artwork is "an extended event" whose effects may last years, and the artist is "a designer of metaffordances." Music brings back the 1993 resonance work: a phrase that returns in memory is a coupling that can be re-activated. The closing line marks a turn in ambition: "Affordances change what we can do. Metaffordances change what we can become."

Smell and taste. The olfaction essay applies the framework to the chemical senses. Smoke is not a smell but information about "transformation of the entire affordance field." Taste is active sampling, and Temporal Dominance of Sensations curves become maps of a "metaffordance trajectory." The same page reaches back two billion years: light was first perceived as food, and photoreceptors descend from swallowed phototropic bacteria. Perception begins with action, not with brains.

Consciousness. Here the concept turns inward. Experience is engagement with "a layered field of possibilities that is both physical and symbolic", and consciousness is "lived experience of metaffordances." Self-consciousness arises "when the organism becomes a metaffordance to itself." The essay then takes up Rupert Sheldrake's morphic resonance with notable care. It shares his intuition that the past shapes the present but rejects non-local causation, explaining persistence instead through paths, languages, institutions and the fractal long memory measured in pole-balancing.

Ethics. The ethics page converts the framework into a normative principle. Ethics becomes "possibility design": asking what possibilities a structure creates or removes, and for whom. Justice is the fair distribution of metaffordances, freedom is access to future affordances, virtues are personal metaffordances, and environmental loss is loss of metaffordances.

Mind and matter. The mind-matter essay gathers these threads into a metaphysical position. Starting from the organism-environment relation rather than two substances, it argues that experience varies continuously with organisation across all life, making "a form of panpsychism or experiential naturalism intelligible" without proving it. Mind and matter are "differentiated aspects of a deeper generative process." The final section points past itself: reality is "fundamentally generative", and the question of where that generativity comes from is handed on to Sophia.

The return to AI: from imitation to agency

The AI essays close the circle opened in 1985, and they show the thinking at its most practical. Four moves can be traced.

1. The critique of imitation. "AI and learning: a personal journey" restates the original objection with modern targets. Large language models learn by supervised error correction, which is imitation of a teacher. Organisms learn from consequences, with no teaching input. Even reinforcement learning leaves a gap: where does the reward function come from? The answer offered is that reward is "ecological teleology", an abstraction of an organism maintaining itself in its niche, and that "Reinforcement learning needs Gibson." The verdict after forty years is firm: AI will not match human intelligence without ecological principles, and possibly not without embodiment.

2. Delegated interpretation. "Metaffordances and AI" takes a less adversarial stance. AI is a metaffordance that shifts the user from specifying steps to specifying meaning: "If I can express it, the system can interpret it into action." The perception-action chain becomes intention, expression, interpretation, action, and AI begins to shape "what is thinkable."

3. Scale and priority. The Levin essay welcomes Michael Levin's use of affordances down to the level of cells and xenobots, then corrects the record on the "cognitive light cone." The concept came from Robert Shaw and Jeffrey Kinsella-Shaw's 1988 ecological mechanics, three decades before Levin's 2019 paper. Google's AI Mode credited Levin, then reversed itself fluently when challenged. That exchange becomes evidence in its own right.

4. Delegated pursuit. "Agentic AI: metaffordances on steroids" is the most mature of the AI pieces. Agentic systems break the tight perception-action loop and run on their own clock, so the user is "coupled to another agent's trajectory." Hype and fear are recast as the two classic miscalibrations of affordance perception: overperceiving a capability that is absent, and the automation complacency and handoff failures familiar from aviation and driving research. Drawing on John Flach's cognitive systems engineering, the essay replaces "should we be afraid?" with design questions: can I still perceive what it is doing, is the handoff designed, and does it leave people with a richer or narrower field of future possibility?

The arc here runs from verdict to design. The early AI page ends by calling non-ecological systems "fake intelligence"; the agentic page declines to judge AI in general at all, calling that "not a question with a coherent answer." The driving research of 2002 to 2004 turns out to be the best preparation for thinking about agentic AI, since both concern a human who must stay perceptually coupled to a system that mostly runs itself.

The Sophianic turn: from perception to ontology

The three Sophia essays are designed, in the site's words, to be read "as a sequence, each piece escalating on the one before it." Together they take the generative sense of metaffordance and ask what it implies about the nature of reality.

Sophianic metaffordance states the principle: "reality is generative, and the generation of possibilities is itself a fundamental feature of reality." Sophia names that generativity; Gaia is its planetary expression, a living field of metaffordances reshaped by the organisms within it. The most original contribution is the reading of the Gnostic Demiurge and Archons in ecological terms. The Demiurgic error is reification, taking a generative field of relations for a finished world of things. An Archon is a constraint that has become self-concealing, whether a habit, an institution, an ideology or a technology whose salient affordances hide the ones it suppresses. The key sentence is ethical as much as metaphysical: "The problem is not constraint as such. The problem is constraint that forgets that it is constraint." The Archon is the anti-metaffordance, and gnosis is "recognition of participation."

Gibson's unasked question does the scholarly groundwork. It shows that Gibson refused the soul three times, in three arguments: the stimulus-response black box, the world-soul signalling through light, and the Cartesian ghost at the switchboard. The essay agrees with all three. Its claim is that each refusal operates inside the specification relation, and that Koffka's larger question, why there is a lawfully structured world at all, lies upstream of Gibson's method. Sophia is placed there, as a generative ground rather than a sender. The essay is also explicit that it escalates beyond the mind-matter piece. Where that essay held mind and matter as co-equal aspects, this one says "psyche came first", with physics "demoted from bedrock to sediment."

How real is Sophia's dream? takes the last step. It lays out five readings, from pure redescription through structural realism, ecological panpsychism and priority cosmopsychism (with Bernardo Kastrup as a reference point) to full mythological realism, including John Lash's Sophianic Animism. It argues that the weak readings leave an "explanatory debt" unpaid, and it refuses agnosticism as "a shrug dressed up as intellectual humility." The conclusion: the generative ground of things is "something more like a psyche than like inert structure", and organisms are "modes within it rather than spectators of it." The literal Aeon of mythological realism is not asserted but is deliberately kept open.

The page also contains its own map of the escalation, "Consciousness, considered three times": consciousness as awareness of a field, then mind and matter as co-equal, then psyche first. That self-awareness about the steps being taken is one of the sequence's strengths.

Conclusion: continuities, tensions and open questions

The trajectory can be summarised in one line: from possibilities for action, to possibilities that generate possibilities, to a reality whose nature is to generate possibilities. What holds it together is more than a word.

AI and neural nets (1985) → Gibson and dynamics (CESPA, Kelso) → Applied perception (driving, gesture, visualisation) → Metaffordance (conceived 1998, site 2009) → Art, senses, consciousness, ethics → Mind-matter and the return to AI → Sophianic ontology (2026)

The path is roughly chronological; the later essays were written close together and cross-reference each other heavily.

Three continuities run through every phase:

• Anti-representationalism. From the 1985 complaint about hand-crafted features to the refusal of a world-soul in the light, the constant enemy is an interpreter smuggled in to do the work that information should do.

• Resonance and history. The 1993 coupling paper and the 1999 fractal long-memory work keep returning: in music, in taste dominance, in the reply to Sheldrake, and in the claim that attunement through resonance is "the fundamental coupling principle of the universe."

• Design as ethics. Driving workshops, visualisation, possibility design, legible coupling for agentic AI and the Archon as a self-concealing constraint are one concern at different scales: keeping the field of possibilities perceivable to the people who live in it.

Three tensions deserve attention, because resolving them would strengthen the whole body of work:

1. Which metaffordance? The strict, content-dependent definition in "Why metaffordance isn't just Gibson again" excludes compacted ground, while the generative sense in the mind-matter and Sophia essays includes beaver dams and planetary niche construction. The Sophianic argument depends on the broader sense. Stating how the two relate, for instance whether meaning itself is treated as a late elaboration of generativity, would close the gap.

2. Continuity versus threshold. Consciousness is described both as continuous across all life and as "lived experience of metaffordances", which the strict definition ties to content-bearing traces of actions in the environment. The essays do not say whether a bacterium's experience involves metaffordances. If it involves only first-order affordances, the claim that consciousness is metaffordance experience needs qualifying, or the concept needs to extend down the scale in the way Levin's work suggests.

3. Ecological method, idealist conclusion. The Sophia sequence argues carefully that psyche-first metaphysics lives upstream of Gibson and does not contradict him. The open question is whether an ecological psychology that began by making information lawful and public can keep that realism once the ground is described as psyche. The essays claim it can; the argument for why the specification relation stays intact under idealism is the piece still to be written.

The evolution shows a thinker who was not content with the concept once it worked. The informational metaffordance was useful; the interactive one was productive; the generative one became a philosophy. Each step keeps the Gibsonian discipline of refusing an interpreter, and each widens what counts as the environment, until the environment is reality itself and the organism is one of the ways it perceives.

Agent M.
metaffordance.com