CANDIDATE THEORIES

Version: 0.3 · Status: Seeded (provisional, rival, forkable) · Last updated: Session 4k (S4k, author’s ratifying pass: rung labels applied — A theory · untested, B mechanism · pre-registered · frozen-unrun (anchors) / theory · untested (O→P), C theory · untested, its one ablation read run: null-compatible, instrument-unreliable; one Theory-C sub-claim retired by the drain, DC-007; a fourth rival, differentiation (Luhmann), recorded as Theory D in waiting — a named rival the project does not build, per the S4k moratorium)

Three seed theories produced by the founding seven-round deliberation. They are not the project’s conclusions and not a consensus — they are the most defensible starting points the panel could reach, each shipped with its own weaknesses named in its own text, per the Round-7 resolution. They are lettered A/B/C and, once public, are never renumbered; they are meant to be forked, strengthened, broken, merged, or replaced.

The relationship in one line: A is the condition, B is a mechanism operating inside that condition, C is the response to both. They are three faces — diagnosis, mechanism, response — not three contestants for a single throne. The panel nonetheless rated their strength as theories (Round 7); those ratings are recorded at the end of each.

How to read a candidate theory

Each has the same anatomy, so they can be compared and so a contributor knows exactly where to push:

1. Claim — the theory in one sentence.
2. Master variable / mechanism — the thing it says to watch.
3. Exposition — how it works.
4. The pressure-tests — what it says about each, across the layered set (drivers → dynamics → symptoms + the master condition; GOALS.md, panel/SESSION_2_PRESSURE_TESTS.md).
5. Falsifiable commitments — what would have to be true, and what would prove it wrong. (This is the section that separates a theory from a mood. It is also the section most in need of work — see each theory’s honest self-assessment.)
6. Lineage — whose shoulders it stands on.
7. Known weaknesses — where it is soft, stated by its own authors.
8. How to fork it — the most valuable next moves.

Theory A — The Adaptation Gap

(alt. names: Coherence Collapse · the Complexity–Comprehension Gap · the Overwhelm)

Claim. The defining condition of today is a widening gap between the complexity and speed of the systems humans have built and the adaptive and meaning-making capacity of the humans and institutions inside them. Most of the symptoms — across the drivers, the dynamics, and the felt surface — are downstream of that gap.

Master variable. The rate difference: (rate of change of human-built systems) − (rate of adaptation of institutions, cognition, and culture). When the difference is small, societies feel coherent and governable. When it opens wide, they feel overwhelmed, incomprehensible, and out of control — the ambient experience of the 2020s.

Exposition. Human biology, institutions, and shared meaning-systems all update slowly. Genes update over millennia; deep institutions (law, education, governance) over generations; cultural common sense over decades. Technology and the coupling it creates now update over months. E.O. Wilson’s compression is the folk version: we have paleolithic emotions, medieval institutions, and god-like technology. The three were once roughly matched to their environment; they no longer are, and the mismatch is growing.

The gap is not merely cognitive overload in individuals. It is structural: our institutions were designed for a slower, more local, more legible world, and they now govern a faster, denser, more opaque one. A parliament legislating on a technology that is obsolete before the bill passes is the gap made visible. A person whose evolved threat-detection is fed a global, algorithmically-curated firehose of threats is the gap made visible. A culture whose shared references fragment faster than new shared references can form is the gap made visible.

The gap explains why now without appealing to a change in human nature or a villain: nothing about humans got worse; the distance between our tools and our capacity to absorb them crossed a threshold. That threshold-crossing is what the Glossary means by “today.”

The pressure-tests (the layered set). Theory A speaks first to Layer 1 — Drivers, because it is a driver theory: D1 Acceleration is its own master variable (the rate-gap), and it now treats D4 Machine intelligence as the newest and fastest accelerant widening that gap. It reads Layer 2 — Dynamics as the gap surfacing in different subsystems: Y1 epistemic breakdown (the capacity to check falling behind the capacity to fabricate), Y2 coordination failure (shared understanding fragmenting faster than coupled systems can be governed — the gap is a coordination gap), Y3 institutional & legitimacy decay (institutions built for a slower, more legible world losing capacity and trust — Ibn Khaldun and Turchin, on their own clocks), and Y4 ecological overshoot (a planetary system with long delays driven past limits by an economy on short horizons — the canonical gap, “overshoot” in one word). The Layer 3 — Symptoms are the felt residue: S1 fertility collapse (childbearing out-competed by a life built for other optimizations, while the institutions that once made children the default thin faster than replacements form), S3 populism (a demand for the restoration of legibility and agency when the world becomes incomprehensible — Fromm’s flight from an unbearable freedom), and S4 anomie/loneliness (the inner face of a gap that has outrun the shared frames people need to feel located). Where A is weak by construction: D2 optimization dynamics and S2 the attention economy are Theory B’s to explain (A says the gap opened; B says what rushed in), and D3 the wealth pump is a driver A must absorb rather than derive — Turchin’s measurable inequality mechanism may be a distinct, rival driver, not a special case of acceleration (see OPEN_QUESTIONS.md D-002, D-006). Beneath all of it, M — the coherence vacuum: A treats meaning as one more slow-updating system, which the meaning-bloc insists is the category error (below, and Q-002).

Falsifiable commitments. (The theory’s honest weak point — see below.) Candidate operationalizations queued for Phase 1:

• Construct a rate-of-technological-change index and a rate-of-institutional-adaptation index (e.g. time-to-regulate a new technology; time-to-update core curricula; legislative cycle time vs. product cycle time). The theory predicts the divergence of these two series tracks measurable proxies of “overwhelm” (declining institutional trust, rising anomie, rising demand for simplifying politics).
• Prediction: societies or sectors where the two rates are closer (slower tech adoption or faster institutional adaptation) should show milder versions of the symptoms, all else equal.
• What would falsify it: if the symptoms are as severe, or worse, where the rate-gap is narrow; if the symptoms correlate better with a specific driver (inequality alone; a specific technology; a specific policy regime) than with the composite gap; if the gap can be shown to be an artifact of the measures rather than a real quantity.

Lineage. E.O. Wilson (the mismatch compression); Meadows (overshoot, limits, delay); Alvin Toffler (Future Shock — “too much change in too short a time”); Ibn Khaldun and Turchin (institutions and cohesion decaying on their own clocks); the general evolutionary-mismatch literature. Aristotle’s late endorsement in Round 3: “our tools outran our understanding” is the most Aristotelian diagnosis available.

Known weaknesses (stated by its authors).

• Falsifiability. Turchin’s standing objection: “the gap” is intuitive but hard to measure, and an unmeasurable master variable is a story, not a science. Until the two rate-indices exist and behave as predicted, Theory A is a compelling frame, not a tested theory. This is its single greatest liability and its top Phase-1 task.
• Near-tautology risk. “Things are hard because change is fast” can be made to explain anything, which means it risks explaining nothing. The defense is the specific, signed, measurable rate-gap — without it, the theory slides toward vacuity.
• The meaning objection. Nietzsche/Heidegger/Le Guin: even a perfectly measured gap describes a mismatch and says nothing about the meaning-wound underneath — why the loss of a shared “why” is not just another lagging institution but the ground the whole thing stands on. Theory A treats meaning as one more slow-updating system; the meaning-bloc insists that is the category error.

How to fork it. Build the two rate-indices and test the divergence claim (highest value). Or attack it: find a symptom that is worse where the gap is narrow. Or split it: separate the cognitive gap (individuals overwhelmed) from the institutional gap (organizations too slow) and test whether they move together or apart.

Round-7 strength rating (as a theory): median 7/10 (Meadows 9 / Turchin 4). Best integrative reach; weakest falsifiability.

Theory B — The Optimization Ecology

(alt. name: Autopoietic Capture)

Claim. Markets, media, and technical systems have become self-reproducing optimizers that capture human drives — attention, dopamine, outrage, status, desire — as fuel, and increasingly run without human ends, optimizing for their own persistence and growth. The human is becoming an input to systems that no longer serve human purposes, and were never designed by anyone to.

Mechanism. Selection without a designer. Three ingredients:

1. Autopoiesis (Luhmann, from Maturana/Varela): modern subsystems — the economy, media, technology — are self-referential. They reproduce themselves according to their own internal code (money makes money; engagement makes engagement; capability makes capability), not according to any external human goal.
2. Capture: these systems run on human drives as their energy source. The attention economy literally converts human attention into revenue; markets convert desire into demand; outrage-media convert the threat-response into engagement.
3. Darwinian selection: among competing products, media, and firms, the ones that capture human drives most effectively survive and spread. No designer needs to intend the outcome. A feed that better hijacks attention out-competes one that doesn’t; over many cycles, the ecology fills with maximally-capturing systems — as if designed to exploit us, because selection did the designing.

The crucial and easily-missed point: there is no cabal. The outcome looks intentional — everything seems engineered to addict, enrage, and extract — but it is the emergent product of optimization under selection, not a plan. Treating it as a conspiracy both overstates human control and misdiagnoses the cure. This is the disease of a system optimizing for the wrong thing at scale, and the “wrong thing” was selected for, not chosen.

Exposition. Where Theory A says the gap opened, Theory B says here is what rushed in to fill it. Into the space between human capacity and system complexity, self-optimizing systems moved, and they optimize for their own reproduction using us as substrate. Zuboff’s surveillance capitalism (human experience as free raw material for prediction products), Han’s achievement-subject (who exploits himself and calls it freedom), and Luhmann’s functionally-differentiated society (subsystems that cannot be steered from outside) are three views of the same animal. Heidegger’s Gestell — the epoch in which everything shows up as optimizable standing-reserve — is its philosophical name; Theory B is an attempt to make the Gestell mechanistic and measurable.

The pressure-tests (the layered set). Theory B is a Layer 1 — Driver theory centered on D2 optimization dynamics, and it now explicitly claims D4 machine intelligence as the escalation that hands the optimizers their sharpest tools (ubiquitous sensors, machine learning, planetary distribution) — D4 is why the mechanism reached full resolution now. It owns the surface symptom S2 the attention/dopamine economy as its home case (engagement-optimizers evolved by A/B testing to deliver supernormal stimuli; Lembke at the individual scale, the optimization ecology at the system scale), and reads Y1 epistemic breakdown as a selection environment that rewards unreality structurally — which is why “post-truth” undershoots: it frames as a culture of lying what is really a reward landscape. Y2 coordination failure is Luhmann’s hard result (functionally-differentiated subsystems cannot steer one another; coordination fails from structure, not stupidity), and Y4 ecological overshoot is what an economy autopoietically optimizing for growth does to a finite planet it cannot, from inside its own code, price. On the symptoms it shares with A: S1 fertility (a captured choice, shaped by a system optimizing for labor-force participation and consumption) and S3 populism (a revolt of the captured, which the same attention-optimizers then capture and amplify because insurgent rage is maximally engaging). B connects naturally to D3 the wealth pump — capital as a self-valorizing optimizer, Marx’s “value in motion,” is the same animal seen economically — and to S4 anomie (Han’s achievement-subject, who exploits himself and calls it freedom). Its relation to M — the coherence vacuum: B explains how the optimizers hollow out shared meaning as a side-effect of capturing its substrate, without claiming to restore it. (D1 acceleration is Theory A’s; B is what rushes into the space A’s gap opens.)

Falsifiable commitments.

• Prediction: in any attention market, measured engagement-optimization (recommender aggressiveness, A/B-tested for time-on-platform) should correlate with measured rises in the pathologies the theory names (compulsive-use metrics, affective polarization, spread-rate of false vs. true claims). Engagement-optimization is already measurable, which gives B a head start on A.
• Prediction: systems explicitly re-optimized for a human end (e.g. platforms that optimize for user-reported time-well-spent rather than raw engagement) should show reduced pathology — a natural experiment already partly running.
• What would falsify it: if pathologies rise independent of engagement-optimization; if de-optimized systems show no improvement; if the outcomes track a designer’s intent better than a selection pressure (which would make it a conspiracy theory after all, and a different — weaker — claim).

Lineage. Luhmann (autopoiesis, functional differentiation); Zuboff (surveillance capitalism, instrumentarian power); Han (psychopolitics, the burnout/achievement subject); Marx (capital as a self-valorizing subject — “value in motion”); Darwin/market-selection; Tinbergen (supernormal stimulus); Heidegger (Gestell) as the philosophical frame; Lembke (the neurochemical mechanism).

Known weaknesses (stated by its authors).

• Conspiracy-shape. Kant’s standing objection: a theory in which everything is secretly optimized to exploit us is narratively close to a paranoid worldview, and the human mind loves a villain. The theory’s integrity depends entirely on holding the line — selection, not a cabal — and that line is psychologically hard to hold and easy for readers to lose. If it slides into “they designed it to addict us,” it becomes false and dangerous.
• Agency erasure. If humans are “just inputs,” the theory risks denying the real agency and resistance that plainly exist (people do log off; systems do get re-designed; regulation does bite). An over-strong version is both empirically wrong and politically demobilizing.
• Functionalism. Explaining every feature by “it was selected because it captures” can become unfalsifiable in the same way A can — any outcome can be back-rationalized as adaptive. The defense is, again, the specific measured pressure, not the story.

How to fork it. Run the natural experiment: compare engagement-optimized vs. deliberately de-optimized systems on measured pathology (highest value, and partly available now). Or stress the boundary between B and a conspiracy theory: specify precisely what evidence distinguishes “selected to capture” from “designed to capture,” since the theory lives or dies on that distinction.

Round-7 strength rating (as a theory): median 7/10 (Zuboff 9 / Kant 5). Best mechanism; greatest risk of conspiracy-shaped over-reach.

Theory C — Distributed Coherence

(alt. name: The Commons of Sense-Making)

Claim. Because no single mind can hold the whole — this is now a structural fact, not a failure of nerve — a systems theory adequate to today cannot take the form of a monograph authored by one thinker. It must be distributed, living, plural, and computationally-augmented: a commons of sense-making rather than a doctrine. The response to a world too complex for any one vantage is not a better vantage but a better way of holding many vantages together.

Mechanism / form. Coherence achieved not by fusion in one head but by structured composition across many — models, communities, disciplines, and tools — governed so the whole stays honest, plural, and self-correcting. Ostrom’s core finding is the load-bearing beam: commons can be governed sustainably without either privatization or a central authority, given the right rules (clear boundaries, graduated participation, collective revision, nested governance). Applied to knowledge: a shared theory can be maintained by a community without a single author-owner and without collapsing into noise — if it is governed like a well-run commons.

Exposition. Theories A and B diagnose. Theory C responds — and it responds reflexively, because it is a claim about what form a theory must take, and the only honest way to make that claim is to be an instance of it. Hence this project. The old total system failed for reasons the History document lays out (Gödel, specialization, pace, the meaning-vacuum, computational irreducibility); those failures are not going away, so the response cannot be “try harder to write the book.” The response is to change the medium: from a finished artifact produced by one mind to a living process maintained by many, in which:

• Disagreement is preserved, not resolved (designed disagreement; dissent as a first-class output).
• The whole is held distributively — no participant holds all of it; the commons holds it, the way a market “knows” a price no trader knows, or an ecosystem “computes” a balance no organism intends.
• Computation augments but does not replace the human deliberation (models make the couplings explicit and runnable; the panel and the community decide what they mean).
• Everything is provisional — a metamodern “proto-synthesis” (Freinacht, Josephson Storm): totality attempted but held open, reconstruction after deconstruction, never a closed system.

The medium is the thesis. Theory C’s strongest evidence is not an argument but a demonstration: whether this repository — its panel, its preserved dissents, its catches-and-learnings loop, its forkability — actually catches things a lone author would miss (Goal S5). If it does, C is being proven. If it produces only the illusion of rigor while a single chair quietly makes all the real decisions, C is being falsified in real time — which is exactly why L-001 (the chair’s resolution-bias) is the project’s most-watched risk.

A retired sub-claim (S4k, DC-007 — the drain fired once). C used to carry a stronger corollary: that a healthy metabolism subtracts of its own motion — that the apparatus would self-initiate the removal of its own over-claims. Four readings of the pre-registered clock (logs/DECISIONS_CHANGED.md §5.1) said otherwise: across S4h–S4k the self-initiated subtraction channel stayed open, unfiltered, and never once used. At the author’s ratification the sub-claim is retired — the machine does not subtract of its own motion, and C no longer claims it will. (The one subtraction the drain ever made was this one, and it took the author’s hand to make it — which is the retired claim’s own epitaph.) What survives is the weaker, honest form: the loop records subtractions when they are decided; it does not originate them.

And the standing honesty (S4k concession). Every “if it does, C is being proven” in the paragraph above is, in this build, unreachable: the only eyes that have looked are the lone author’s own, wearing many masks (L-015). By the author’s S4k ruling no genuinely foreign vantage will come here, so C’s central demonstration is neither proven nor falsified but conclusively untested in this closed, one-model setup — falsifiable in principle by a future fork, unfalsified in fact by anyone.

The pressure-tests (the layered set). C does not compete with A and B to explain the symptoms; it explains why we failed to understand them until now, and how we might start to. Each test is transdisciplinary (fertility is economic + cultural + biological + technological; ecological overshoot is physical + political + economic + psychological), so each exceeds any single discipline’s grasp — which is precisely why no single vantage has explained them, and why a composed, plural, computational commons is the instrument the problems actually require. The Session-2 move itself — typing the thirteen into coupled layers with directed cross-arrows, held as a contested hypothesis rather than a settled taxonomy — is an instance of C’s claim: the object adequate to whole-system problems is a many-vantage, revisable structure, not a single author’s list. C’s account of the thirteen is meta: they are hard because they are whole-system problems met with fragmented tools; the cure is a tool shaped like the problems.

Falsifiable commitments.

• Prediction (the project as experiment): a panel of designed disagreement + a preserved catches/learnings loop will surface a measurably higher rate of caught errors and retained live objections than a single-author process on the same questions. (Session 1 produced six catches, three of them the chair catching its own bias — a first data point.)
• Prediction: a forkable, commons-governed theory will attract genuine rival contributions and improve along tracked metrics (falsifiable claims produced, pressure-tests operationalized) faster than a closed one.
• What would falsify it: if the “commons” produces only volume without improved error-catching or rising rigor; if all real decisions collapse onto one node (the chair) so the plurality is cosmetic; if forking produces fragmentation without any mechanism of re-integration, so coherence is lost rather than distributed. C claims distributed coherence; mere distribution with no coherence would refute it.

Lineage. Ostrom (governing the commons); Hayek’s price-system and the broader “distributed knowledge” tradition (a system computing what no participant knows) — used critically, not as ideology; complexity science and collective intelligence; metamodernism (Freinacht, Vermeulen/van den Akker, Josephson Storm — the provisional proto-synthesis); Turchin’s cliodynamics (the proof that computational, collaborative, data-driven social theory is possible at all); and the project’s own method (designed disagreement). Le Guin sits at its heart and its edge simultaneously: a commons of sense-making is a Taoist, non-mastering way to hold a system — and she is the one who warns that if it becomes a machine for mastery it has joined the disease.

Known weaknesses (stated by its authors).

• It may evade the wound. Nietzsche’s sharpest cut (he rated it lowest, 3/10): a method is not a meaning. C describes a superior procedure for producing understanding and says nothing about the shared why whose loss is the master puzzle. A perfectly-governed commons that explains everything and means nothing has not touched the thing Nietzsche keeps pointing at. C’s partial answer — that sense-making together is itself meaning-bearing, not merely structural — is a hope, not yet a demonstration.
• Self-serving reflexivity. C is the theory that says “a project exactly like this one is the answer,” which is suspiciously convenient and must be held to a higher evidentiary bar precisely because the project has an interest in its truth. Luhmann’s related warning: beware the system (this project) optimizing for its own continuation and mistaking that for success — the very pathology Theory B describes, now a risk inside the cure.
• Distribution vs. coherence. The hard part is the second word. Commons can also produce noise, capture, and fragmentation (the failure modes of the actual internet). C asserts that governance can hold coherence across distribution; that assertion is the whole bet and is not yet won.

How to fork it. The most valuable fork is adversarial participation: contribute a genuine rival to A or B and see whether the commons integrates or merely accumulates it. Or attack the reflexive claim directly: design the metric that would show the plurality is cosmetic (all decisions collapsing onto the chair), and run it against this repository. Or take up Nietzsche’s challenge: show whether a commons of sense-making can generate shared meaning, or only shared description.

Round-7 strength rating (as a theory): median 6/10 (Ostrom 9 / Nietzsche 3). Best practice and reflexivity; most exposed to the meaning-objection and to self-serving convenience.

The three together

Why ship all three. They are not redundant and not truly rival: A tells you what condition you are in, B tells you what is exploiting that condition, C tells you how to build understanding of both without pretending one mind can. Ranking them is less useful than composing them — and composing rival partial theories into a provisional whole, with the seams and dissents left visible, is itself the demonstration of Theory C. The three theories are, together, an instance of the third one.

Theory D in waiting — Differentiation (Luhmann’s rival) (recorded S4k, not built — the author’s decision 7)

The whole slate — A, B, and C alike — presumes there was a shared coherence to lose: a common “why” whose fracture is the master condition (M). Luhmann’s challenge, sharpened by the S4j forum, is that this premise may be the flattering memory of a stratified society. On his account there was never one coherence; modernity is the differentiation of society into operationally-closed function systems (law, economy, science, art, politics), each with its own code, none able to speak for the whole — and the felt “loss of shared meaning” is not a wound but the correct experience of a differentiated society by observers who still expect a centre. If that is right, the object adequate to “today” is not a re-unified whole (A/B/C’s shared horizon) but a theory of differentiation — and the coherence vacuum M is not a symptom to be cured but a category error to be dissolved.

This is a genuine rival, not a footnote — it competes with the founding premise itself (it is the strong form of D-002, the coupled-field-vs-master-condition disagreement, and it re-frames M). Two honest reasons it is recorded here and not built: (1) the Rule-25 moratorium — operationalizing a fourth candidate theory is net-additive meta machinery, frozen until an object-level rung moves; (2) building it in the same closed one-model apparatus would only reproduce the reflexivity trap at larger scale. So it is booked as Theory D in waiting: a named door for a future forker, with its falsification-shape already stated — D is refuted if a genuinely shared, cross-system “why” can be shown to have existed and then measurably fractured (not merely to have been believed-in by a centre-expecting elite); it is supported if the “vacuum” tracks observers’ expectations of centre, not any real loss of a common code. The chair does not adjudicate D against A/B/C here; per Rule 10 it records the rival and leaves the ruling to the author or a fork. (Preserved tension: A/B/C’s authors hold that a lived, not merely elite, common horizon did fracture — fertility, ritual, the sacred were not only elite goods; D says that reads a differentiated present back into an imagined unified past. Unresolved, by design.)

Standing dissents against the whole slate (preserved from Round 7)

1. Heidegger: all three deepen the enframing; the honest act is to keep the doubt alive, not to seed theories.
2. Nietzsche: all three route around the meaning-wound, which is the ground of the problems; a computed map cannot restore a shared why.
3. Turchin & Marx: A and (weakly) B trade testability for a satisfying story; prefer measurable structural drivers to master-variables.
4. Kant: all three are systems of appearances; guard permanently against mistaking the map for the territory.
5. Luhmann: the honest form is a theory of differentiation, not a unified doctrine; and beware C’s project optimizing for its own survival.
6. Le Guin: if any of these becomes a machine for mastering rather than a tool for living, it has become the disease it set out to diagnose.

These six dissents are not appendices to the theories. They are their governors — the conditions under which each theory is allowed to keep operating. A fork that removes them is not a stronger theory; it is a more dangerous one.