Cartography.

Context Signals and the retired tensor claim

The Context Tensor's tensorial claim and 1,707-dimension headline were retired in the KTP specification revision. Current vocabulary distinguishes Context Signals from Risk Factors. The earlier formulation remains part of the research history; a collection of measurements alone does not establish the transformation law a tensor requires.

Next test or repair: Make the replacement visible wherever a historical entry is opened, while preserving its original wording and source history.

What a cross-domain mapping must preserve

Shared vocabulary and similar diagrams can suggest a comparison. A structural claim must also name the states and operations being mapped, what stays unchanged under the mapping, and where the material rules differ. A resemblance between networks does not by itself establish the same dynamics.

Next test or repair: Specify an intervention and a boundary case for each proposed mapping, and trace evidence to its source before counting independent support.

A ≤ E: principle, authorization rule, measurement

The environmental-capacity principle guides the rule that permitted action should fit a declared safe operating range. Testing it as an empirical claim requires demand and capacity measures defined independently of the outcome. A useful authorization rule does not, on its own, establish a universal physical law.

Next test or repair: Choose a bounded application, define the two measures in advance, and compare its failure predictions with a competing account.

The Trust Force Equation remains a developing model

The current model proposes how relational conditions shape trust accumulation over time. A condition that stops new accumulation does not, by itself, erase trust already accumulated. Expectations about future interaction and processes of forgetting are distinct. The model still needs independently scored observations and comparison with rival accumulation rules.

Next test or repair: Fix the measurement scale and distinguish changes in accumulation from losses of existing trust before running a discriminating test. The full instrumented form remains in private research.

Cost can support honesty; it is not information itself

Signaling cost can help sustain honesty under specified incentives and constraints. It is not identical to information content: signaling models can have honest equilibria with zero realized signal cost. Applying this result requires identifying what is signaled, the available misleading alternatives, and who bears their costs.

Next test or repair: Test the proposed mechanism in its setting. Neither a cheap signal nor an expensive one establishes trustworthiness by price alone.

Network collapse and the possibility of repair

A sharp transition in a simulation does not establish a universal transition class or prove that recovery is impossible. When potential links remain and the rules permit renewed cooperation, low trust can recover. Removing links without a reconnection rule can instead leave a network disconnected. These are different model conditions.

Next test or repair: Separate retained-link recovery from link removal, state exactly what counts as activity, and test reactivation under declared rules. A finite run without recovery is not proof that recovery is impossible.

The Trust Force Equation

A proposed model of how relational conditions shape trust over time, with tests written to probe its boundaries. This is a dense region of the graph, whose empirical claims depend on independent measurement and the case-recoding gate. The current assessment above distinguishes new accumulation from trust already held.

Ostrom × Percolation

Ostrom's eight design principles are read as maintenance functions on a trust network — a proposed structural mapping in Paper 1. The candidate ninth principle names temporal edge persistence. Its measurement and the simulation's recovery conditions remain open work; a dense cluster does not establish either the mapping or a universal transition class.

KTP Specification

The versioned protocol stack, organized by layer. The graph preserves specification history alongside the structural relations between layers. The retired Tensors RFC and Context Tensor remain labeled as historical entries here; current specification vocabulary separates Context Signals from Risk Factors.

Cultivated Priors

Priors are perceptual organs — the lens through which experience is constructed, not a belief held alongside it. The adversarial default installs the watcher: the operator's optimum is a target who watches themselves on the operator's behalf, indefinitely, through priors carried forward by the cognitive propagation chain. The framework's defensive transmissions arrive from opposite cultural registers and converge on the same architectural commitment — organs cultivated by craft, not installed by adversary. Brand trains pessimism for failure-mode detection (the maintainer's posture); Haidt transmits Mary Oliver's three-stage discipline for attention and wonder (the cultivated witness). DCTI is the framework's working method; the Flourishing course at NYU Stern is its 12-year empirical anchor.

Forgiveness Mathematics

Forgiveness ratio, metabolic rate of repair, the asymmetry between betrayal and reconciliation. Independent convergence from primatology (de Waal) and lived experience (Denali Wilson), both arriving at the Ninth Principle as the missing operator the original eight assume but never write down. Formalized as operations on the trust integral.

The framework rests on cross-domain synthesis. The Trust Force Equation has forty-nine author-scored cases (the recoding gate is open). The Ostrom-percolation paper has simulations, not field data. Temporal edge persistence is named as a candidate principle without measured renewal functions in real systems. No prediction has been registered, run, and either survived or been falsified outside the program's own writing.

The model's relational conditions need measurements that independent observers can score consistently. Expected future interaction is a prospective belief; forgetting is a separate process that would need its own specification. Units, observation windows, and a common outcome scale must be fixed before competing models can be distinguished. These measurement questions remain open.

The framework-ignores-power steelman is held open. The trust force equation has explicit power-weighting terms, but Foucault's deeper claim — that power *constitutes* systems rather than *distorting* them — is only partially answered. The framework celebrates structural constraint as a virtue; from the constitution-of-power angle, structural constraint is the most effective form of capture, the kind so embedded the governed cannot perceive it as governance.

The framework proposes that relational conditions shape trust accumulation over time. How those conditions can be observed when one party is an AI agent remains unsettled. A zero input can stop the model's contribution to new accumulation; it does not automatically erase existing trust. Signaling theory also does not establish that zero realized cost means zero information. The boundary needs operational definitions and competing predictions.

The acequia works at twenty-five families. The internet commons (1969–1993) worked at ten thousand nodes. The framework's answer to the scaling steelman (that the acequia is romanticized) is *federation through nested enterprise*. But federation as a working mechanism — what makes federated trust networks composable, where they break, what bridges them — is mostly hypothesis. The graph holds the simulation results; it does not hold a working theory of federated repair at civilizational scale.

The KTP red lines are theoretically held but operationally untested. No production deployment exists. The framework's strongest commitment — that KTP scores actions, not persons — has not been pressure-tested by an enterprise that wants a person score and is willing to pay for one. The Specification-Existence Gap names this honestly; closing it requires a reference implementation in the wild.

Reconciliation Dynamics — forgiveness ratio, metabolic rate of repair, the asymmetry between betrayal and reconciliation — arrives in the framework as independent convergence (Denali, de Waal). The convergence is striking. The mathematics is not yet a theory: there is no specified function for how much repair metabolizes how much betrayal, no clean derivation of the asymmetry, and no cross-substrate confirmation that the same dynamics operate in different repair domains.

The program ran the Reduction Attack on its own vocabulary: fourteen constructs, each forced against the strongest fielded equivalent in established computer science, by two adversarial reviews run blind to each other — one by ChatGPT under explicit kill instructions, one by Zia, the framework author's Claude-based assistant — with both tables committed before comparison. Eight constructs died twice, independently. E_base is credit-risk composition (FICO's stress-weighted history, backing, thin-file caps). The Trust Score, E_trust = E_base × (1 − R), is multiplicative risk discounting — among the oldest moves in decision theory. Deautomation is adaptive automation (Sheridan's levels; Airbus control-law degradation; Simplex runtime assurance). Sponsorship is the surety bond plus proof-of-stake slashing. The Kinetic Envelope is seccomp's no_new_privs plus Macaroon-style attenuation. Blue Zones are MILS separation plus federation assurance mapping (NIST SP 800-63C-4). The Context Tensor is multi-criteria decision analysis with one hard constraint — CVSS ships base, temporal, and environmental scoring today, and nothing about the object transforms like a tensor. The Flight Recorder is event sourcing plus PROV-O provenance; the name concedes the reduction. This page applies a level hierarchy to other people's parallels; run against the framework's own vocabulary, several constructs were Level 1–2 wearing Level 4 clothes.

The A ≤ E cascade

The founding claim, traced through its operational consequences. Each step tightens a screw the prior step left loose, ending in the explanation of why straightening environments produces weak agents — a result the framework predicted before it was named.

Trust Force → Forgiveness

From the formal equation to the maintenance work that keeps a trust network alive. Percolation theory provides the network substrate; the ninth principle names what the original eight assume; the forgiveness ratio sets the metabolism.

The Straightening Watershed

A proposed pathology pathway: removing opportunities for consequential interaction may slow the formation and renewal of trust. Whether existing trust then declines depends on the specified processes of loss and repair. This path names a hypothesis to test; it does not imply that removing friction instantly erases a relationship.

The Panel names the evidence dilemma — govern now without evidence, or wait and it may be too late — and then instantiates it: the body producing consensus evidence runs an order of magnitude slower than the field it measures. The framework's reading is that you do not speed the slow pole, you gate the fast one; but the Panel has no gating authority by design (policy-relevant, not policy-prescriptive), so the tension is structural to its mandate. Held open; watched via the thematic-brief cadence and the Global Dialogue.

An ideologically charged tension surfaced by the 12 Levels source material. The framework holds the catastrophic reading as default but cannot dismiss the succession-as-good argument without engaging it on its own terms. Quotes are verbatim; framework voice is the holding ground, not either side.

If extinction is not the worst outcome, the failure-mode taxonomy needs an additional axis — value lock-in versus continuity. The graph currently encodes continuity as the dominant axis; this tension argues for a second.

The political-economy critique. Steelmanned across Foucault, Michels, the literature on capture. Held open because the framework's response — that the equation already encodes power — is partial; the deeper claim about constitution, not distortion, has not been answered to satisfaction.

A live empirical question the framework treats as open. The gate at /research/methods is the public commitment to report whatever the recoding finds, including findings that move the equation off its current form.

Held under deep analysis. The measurement-independence objection is not yet defeated. The program's response is a future empirical study, not a current rebuttal — and the cartography page exists partly to keep that gap visible until the study lands.

The framework treats this as named and addressed rather than open. Listing it here keeps the question visible: federation is the answer the framework owes, and the federation work is still partly hypothesis.

A ≤ E's unification claim survives both reviews as an untested hypothesis — exactly that. The audit sharpens the unfalsifiability charge from prose to protocol. The criterion on the wall: a successful experiment the framework predicted is evidence for the model; an experiment that could only have succeeded if the framework were true is evidence for the ontology. The program is aimed at the second.