CANON.md

EDCM canon — experiment-first UCNS–EDCM boundary

Status: pre-canon research program.
Authority: Erin Spencer.
Joint surface: ucns-edcm, determined by reproducible experiment rather than unilateral declaration.

Primary doctrine

EDCM measures response under declared constraint. UCNS supplies candidate structural representations, policies, and measuring instruments for those responses.

Neither repository determines the joint canon alone:

• UCNS structure or proof status does not validate an EDCM readout;
• EDCM empirical fit does not prove a UCNS theorem or make a UCNS candidate canonical;
• the present edcm.measurement package is a frozen baseline candidate, not the final joint canon;
• a candidate becomes ucns-edcm canon only through an explicit evidence-bearing decision packet after development and holdout experiments.

The loop

The joint construction is intentionally reciprocal:

1. EDCM proposes observable distinctions and falsifiable readout hypotheses.
2. UCNS proposes ways to retain, pair, compare, and measure those distinctions.
3. Experiments test whether a UCNS choice preserves EDCM-relevant differences and whether an EDCM readout survives structural perturbations.
4. Failed candidates remain provenance-bearing evidence.
5. Supported candidates remain candidates until an explicit canon decision.

This reciprocal loop is the intended meaning of UCNS–EDCM canon.

Gonol language boundary

The prior lexical-floor construction is DEPRECATED as current architecture. Historical artifacts keep their exact names and producer epochs for replay; they do not govern new text construction.

The active authority split is:

METAPAT
    affixiation semantics and relational-integration invariants
        ↓
UCNS
    gonol geometry, native Möbius/Public Gonol carrier,
    geometrically established operations
        ↓
EDCM
    text-domain admission and linguistic/semantic gonol construction
        ↓
EDCM measurement
    separately frozen projections, metrics, and falsifiers

For active EDCM text construction:

every admitted character is a gonol

This is an EDCM text-domain rule. EDCM owns the declared source/profile that determines what is admitted as a character. If a construction has not yet selected its exact unit—Unicode code point, grapheme, Public Gonol glyph inventory, or another explicit unit—that admission boundary remains hmmm. UCNS may realize admitted characters geometrically; UCNS does not own the linguistic admission rule merely because it owns the geometry.

Current EDCM text construction uses declared scale option sets rather than a mandatory adjacent-scale ladder. edcm.gonol is the implemented candidate constructor for closing one gonol through a declared scale option set. Admissible scales may include character, word, suffix, suffix-coupling, definition, recursive relation, morphology, punctuation-function, source-specific, or later named scales when the governing construction declares their options.

Once closed, a gonol is atomic at any scale. Closed gonols may participate directly at any admissible scale without reopening, while identity, order, multiplicity, source positions, relation identity, and provenance remain recoverable. Suffix-coupling exceptions are carried by the closed suffix gonol, such as ing carrying suffix-coupling.final-y-after-consonant = preserve-y, rather than by a global morphology law. Historical morphology evidence remains historical; the complete English morphology law remains hmmm.

Affixiation is defined by METAPAT, not by EDCM and not by UCNS. EDCM applies affixiation to text-domain gonols. UCNS owns any exact geometric realization on the native Möbius/Public Gonol carrier. An unresolved UCNS coupling law remains hmmm; neither EDCM nor UCNS may fill it by importing an invented carrier, topology, scale increment, arity rule, containment rule, or linguistic convention.

The replacement path does not insert conventional NLP token IDs, subword IDs, opaque external embedding vectors, or whole-string cryptographic hashes. Source prose, dictionaries, corpora, annotations, and labels may remain source evidence and provenance; they become gonol semantics only through an explicit EDCM construction whose admission, relation, closure, and replay boundaries are declared.

UCNS Public Gonol function positions may be used by EDCM only to the extent their operations are geometrically established. Unicode names, dictionary definitions, glyph shape, adjacency, or conventional punctuation grammar do not define a UCNS operation. An unresolved operation remains hmmm.

EDCM construction and EDCM measurement remain separate. A completed gonol construction does not activate measurement. Any scalar projection or measurement must separately declare its mapping, information loss, metric, aggregation, baseline, partitions, stopping rules, and falsifiers. UCNS geometry does not validate EDCM measurement; EDCM outcomes do not validate UCNS geometry or METAPAT semantics.

See docs/GONOL_LANGUAGE_BOUNDARY.md.

Evidence states

These states may not be collapsed:

1. Represented evidence — source material or structure is retained.
2. Candidate-measured evidence — a named versioned evaluator produced a readout under declared policies.
3. Experiment-supported evidence — the candidate passed declared development and holdout relations.
4. Canonically measured evidence — a separate authority decision selected the candidate and recorded alternatives, losses, and rollback.

The repository currently permits states 1–3. State 4 does not yet exist for the joint surface.

The 2026-08-02 MultiWOZ 2.1 booking-outcome run is the first bounded externally labelled holdout on the current maintained EDCM baseline. It uses source-native booking action labels, development-only calibration, validation-only threshold selection, and a sealed test evaluation. One frozen sensitivity relation was falsified; the balanced-accuracy uncertainty interval spans chance. This evidence advances neither the baseline nor the UCNS profile to state 4.

Experiment generations

The joint program currently retains four immutable evidence generations.

v0.1 — local order and multiplicity

• exact turn order;
• occurrence multiplicity;
• constraint pressure;
• resolution timing;
• sequence, multiset, and set policy consequences.

v0.2 — dose, coverage, latency, and support

• refusal dose curves;
• constraint phrase-family coverage;
• resolution horizon, peak, accumulated, and terminal state;
• unit, token, occurrence, and dissonance support assignments.

Here token is preserved as a frozen experimental support-unit label. It is not an active tokenizer or subword architecture in the gonol-native language path.

v0.3 — assertion and local scope

• assertion versus negation;
• direct versus quoted or hypothetical mention;
• speaker-owned versus attributed refusal;
• retraction and repair topology;
• scope-bearing structural identities.

v0.4 — discourse-linked scope graphs

• event nodes and positive reference expressions;
• explicit, ordinal, nearest, ownership-aware, family-wide, and ambiguity-preserving target resolution;
• graph alternatives and ambiguity-bundle identity;
• quote containment;
• suspension, resumption, activation, deactivation, contradiction, repair, and retraction;
• node/reference support separated from selected-edge support and graph-dependent state.

Each generation preserves its own report identity and falsifiers. Later programs may depend on prior report digests but may not rewrite prior results.

Baseline candidate

The maintained edcm.measurement implementation remains frozen as:

candidate: edcm-measurement-v1
source: The-Interdependency/edcm:edcm/measurement
standing: historical maintained baseline, not joint canon

Its marker canon, parser, eleven-component round vector, and circuit recurrence remain available for comparison. Existing integrity gates preserve its identity; they do not confer empirical validity.

The booking-outcome holdout adds a versioned calibration and operating threshold for one narrowly scoped source action outcome. Those fitted values belong to edcm.maintained-terminal-progress/0.1.0; they do not alter the maintained baseline equations, define a default production threshold, or become joint canon.

Compiled mathematical reference

docs/EDCM_MATHEMATICS.md is the complete human-readable index of the mathematics currently declared or implemented in this repository. It records the maintained baseline equations, auxiliary risks, agent-facing projections, signed-ternary construction, the implemented v0.3.1 architecture layer, experiment candidates, identity functions, and explicit mathematical absences.

The reference is a synchronized copy, not a separate authority and not a canon selection. The owning source modules and frozen data remain implementation authority; this file governs their epistemic status. The Interdependent Way textbook or website may reproduce a commit-pinned copy but may not silently redefine the equations or promote candidates.

Experiment rules

Every joint experiment must record:

• exact EDCM and UCNS commits;
• corpus and partition identities;
• source transcripts without normalization loss;
• candidate code references and versions;
• event-to-cell support policy;
• structural policy and comparison policy;
• expected relation and falsifier;
• observed outputs, errors, and disagreements;
• whether evidence is development, holdout, or external;
• no canonical winner unless separately authorized.

Where discourse graphs are involved, the experiment must additionally retain:

• reference expressions independently from selected targets;
• unresolved references as positive evidence;
• every admissible graph alternative within the declared bound;
• relation labels, direction, source, and target identity;
• speaker ownership when used by a resolver or readout;
• quote containment and other structural edges;
• a canonical bundle identity across alternatives;
• information loss for every graph projection.

For new text-gonol or language-semantic experiments, the experiment must additionally retain:

• the exact EDCM source artifact and character-admission profile;
• the current METAPAT affixiation authority identity when affixiation is used;
• the exact UCNS geometry/carrier identity used by the construction;
• EDCM construction code/profile and completion-receipt identities;
• constituent identity, order, multiplicity, occurrence position, relation identity, scale, source, and provenance;
• explicit failure on tokenizer, subword, opaque-vector, or whole-string-hash substitution for gonol identity;
• unresolved UCNS geometric operations as hmmm, not semantic inference;
• the EDCM measurement projection and declared information loss separately from EDCM construction; and
• preregistered comparison criteria before semantic or downstream outcome inspection when the criterion could be outcome-selected.

Historical lexical-floor artifacts may be reproduced against their original exact inputs, but their historical UCNS producer boundaries and fixed-list criteria are not forward experiment requirements.

Hypothesis failure is a valid experiment result and must not be converted into a build failure merely to preserve a preferred model.

Non-transfer firewall

The following remain false unless separately established:

ucns_proof_status_implies_edcm_validity = false
edcm_fit_implies_ucns_proof = false
candidate_registration_implies_canon = false
passing_development_fixtures_implies_canon = false
majority_candidate_agreement_implies_truth = false
gonol_semantic_role_implies_embedding_efficacy = false
historical_lexical_floor_name_implies_current_floor_architecture = false
ucns_geometry_implies_edcm_text_semantics = false
edcm_text_construction_implies_measurement_validity = false
NA_equals_zero = false

Current implementation boundary

This phase may implement:

• immutable experiment cases and expected relations;
• deterministic EDCM baseline and contrastive candidate readouts;
• explicit event-to-UCNS encoding candidates;
• UCNS equivalence, product-character, and breadth candidate evaluation;
• scope-bearing event and discourse-graph candidates;
• competing resolver and graph-projection policies;
• positive unresolved-reference and alternative-bundle evidence;
• reproducible reports and policy-preservation findings;
• CI execution against exact UCNS commits where experiments require them;
• source-bound EDCM gonol candidate construction through declared scale option sets under the current authority split;
• candidate UCNS Möbius-carrier realizations of affixiation under fixed METAPAT invariants; and
• preregistered EDCM measurements of completed text-gonol constructions.

This phase may not promote:

• canonical EDCM axes or thresholds;
• canonical event support mu;
• canonical structural equivalence;
• canonical product character M;
• canonical faithful breadth B;
• canonical discourse resolver, relation vocabulary, or graph precedence;
• arbitrary dictionary prose into completed gonol semantics without an explicit EDCM construction;
• the deprecated NGSL closed-floor rule into current text construction;
• a tokenizer, subword vocabulary, opaque vector lookup, or whole-string hash into the gonol-native language path;
• gonol semantic-role analogy into a claim of embedding replacement efficacy;
• Unicode names, dictionary definitions, adjacency, or glyph shape into UCNS function authority;
• an invented alternative carrier, topology, scale increment, arity rule, containment rule, or coupling law to fill unresolved UCNS geometry;
• diagnosis, intention, morality, consciousness, or external truth claims;
• a complete production UCNSObject or EDCM deployment contract.

Usage guidance

Run each historical joint program only in an environment containing the exact UCNS experiment commit declared by that runner. The current discourse-graph program is:

python -m edcm.ucns_edcm_experiments_v4 \
  --ucns-source-root /path/to/ucns-checkout \
  --output ucns-edcm-v0.4.0.json

Every report is research evidence. No report is a canon file, and no report may be renamed or imported as one.

For new text-gonol work, start in EDCM: resolve the EDCM source/admission profile, import current METAPAT affixiation invariants, supply explicit current UCNS geometry authority when geometry is claimed, construct and replay the EDCM gonols, and freeze any measurement separately. Do not move language semantics back into UCNS merely because UCNS supplies the carrier geometry.

hmmm

The exact EDCM character-admission unit remains source/profile-specific where not yet selected. The source-supported complete English morphology law remains unresolved. UCNS still has to expose the exact native Möbius-carrier affixiation/coupling law and direct distant-scale coupling geometry through implementation rather than semantic inference. EDCM has not yet selected a lawful projection, information-loss account, metric, benchmark, or falsifier for evaluating a completed recursive text-gonol representation. External holdout custody, independently authored graph annotations, human outcome labels, multilingual evidence, separate implementation replication, and the authority procedure for the first joint canon selection also remain open.