A Structural Methodology for Information System Domains: Four Layers, Cross-Domain Patterns, and the Methodology’s Range as a Domain in Its Own Right

1. Introduction

This paper describes a methodology for structural analysis of information system domains. The methodology is the contribution. What follows is an account of what the methodology is, how it operates, and what falls out when it is applied — to other domains and to itself.

The methodology was developed during the design of a distributed information system — the entity system of The Entity System and the companion protocol (see The Entity Core Protocol) — where the question was practical: which abstractions actually carry the system, and which are convenient restatements of others. The procedure that answered that question turned out to apply more generally. It is a discipline rather than a recipe: a sequence of analyst-authored steps with explicit convergence tests, dependency constraints, and cross-domain checks that keep the result honest.

Two strands fed into the methodology before it became its own analytical surface. The first was a proto-methodology applied to two design spaces within the entity protocol — type description and authorization — as a dimensional analysis (seven type dimensions, seven capability dimensions, scored across a sixteen-system comparison). That work, recorded as Dimensional Completeness, gave the analytical posture: extract irreducible dimensions, position systems against them, identify gaps. The second was a combinatorial reframing of the entity system’s six primitives, which stopped being a list and became combinatorial potential dimensions across three resolutions (presence, partial-level, internal). That reframing brought dependency constraints, coherence, attractor positions, and the distinction between substrate primitives and derived spaces (extensions, peer architecture) explicitly into view — the vocabulary that Layer 1 of this paper later adopts in domain-general form. The first structural pass through the combinatorial space used a pair-relationship lens: the fifteen pairs among six primitives, classified by structural load, with load-bearing triangles surfacing as a side effect. Subsequent application to biology, cognition, and abstract information substrates forced the move beyond pairs to multi-arity compositions at the arity each domain required, which is the position Layer 1 occupies today. When the matured methodology was later run back over type systems and capability systems as domains in their own right, it re-derived primitive sets that converge with that original dimensional analysis — a cross-check developed in §Cross-Domain Structural Patterns, not re-argued here.

1.1. The four layers

The methodology operates at four layers.

Layer 1 analyzes a single domain through a twelve-step procedure: extract irreducible primitives via three convergence tests; decompose each primitive into partial levels; specify the dependency DAG; enumerate and classify pairwise interactions; construct the dependency-filtered coherent sub-lattice; identify load-bearing compositions (core triads, hub primitives, anchor pairs); predict emergent properties; and validate against the instances surveyed before primitive extraction.

Layer 2 connects analyzed domains through a typed graph of inter-domain edges — realization, role-identification, configuration, enrichment, decomposition, feedback, selection, coupling — and treats substrate gaps as first-class domains in their own right.

Layer 3 extracts patterns across the populated graph: structural shapes that recur across independent domains, candidate Layer-3 abstractions promoted only when the cross-domain check stabilizes.

Layer 4 applies the methodology to concrete situations through a scope ladder from Sc=0 (universal) to Sc=4 (specific instantiated event). Different scopes admit different kinds of question and produce differently-bounded outputs; the scope dial is itself part of the methodology.

1.2. Reflexive applications

The methodology supports two distinct reflexive applications. Applied to itself as a domain, it produces a four-layer self-decomposition (twenty-four primitives across Layers 1–4) in the same structural vocabulary it uses for any other domain. Applied to its own range — the meta-domain of analyzable domains — it produces a primitive set, a dependency filter, a core triad, and a set of empirical attractors that bound where the methodology produces high-value output, where it produces partial output, and where it does not apply. Claims throughout the paper should be read with the attractor a given application occupies in mind; the methodology is not equally informative everywhere.

1.3. Applications and the posture of this paper

The methodology has been applied across biology, cognition, the entity system itself (see The Entity System; Convergent Evolution), physics, and mathematics. Two extended applications are recorded as separate companion papers: the structural decomposition of abiogenesis (see Abiogenesis as Progressive Hardening), and an exploratory application to physics treated as an information-substrate domain (see The Structural Methodology Applied to Physics). Both are presented as methodology demonstrations rather than as substitutes for domain-specific theories.

The applications are exploratory: we report what the procedure produces and let the reader judge whether the structural readings cohere. We do not adjudicate among competing domain-specific theories. Empirical cartography (trajectory regimes across multiple arrangements, cross-arrangement coupling at fine scope, a calibration architecture that attaches wall-time anchors to dimensionless rate models, a cross-corpus instrumentation track) is documented in the body because it is what we have done with the methodology, not because it is what the paper argues for.

To keep description and finding separate, the body uses an explicit cartography-versus-licensed-claim distinction: cartography is what the methodology produces from its analyst-authored inputs; a licensed claim is what survives a non-circular external check. We mark the distinction at each section that crosses it.

1.4. Organisation

The body develops Layer 1 across three chapters — the twelve-step procedure, the product lattice it produces, and the Bayesian-network interpretation of the dependency-filtered sub-lattice — and Layer 4 in a single chapter (scope ladder, unified manifestations, context domains, lifecycle patterns). Layers 2 and 3 are presented through their primary applications rather than as standalone chapters: the Convergence Domain chapter develops one confirmed Layer-3 abstraction; the Realization Chain chapter develops the chain of substrate domains connected by Layer-2 bridges; the Cross-Domain Structural Patterns chapter near the close of the paper gathers the patterns recovered across the populated graph (including the SSA topology as the second confirmed Layer-3 abstraction).

A methodological-discipline chapter introduces the cartography versus licensed-claim distinction. An empirical-cartography chapter keeps one illustrative slice of the applications — the trajectory regimes — and points to a companion note for the wider record (cross-arrangement coupling, the calibration architecture, the cross-corpus instrumentation), which is exploration recorded for transparency rather than material the argument rests on. The two reflexive applications follow in the abstract’s order — the methodology applied to its own four-layer structure, then the methodology applied reflexively to its own range as a meta-domain (its full nine-attractor map relocated to a companion note). The paper closes with cross-domain structural patterns, a short computational-implementation and reproducibility chapter (with the component detail in a companion note), related work, and an invitation to disproof.

Three findings would refute the methodology itself, not its applications: a domain where the convergence tests fail to stabilize, a domain where the dependency filter falls outside the predicted range, or a trajectory the cross-domain patterns fail to classify. None has been identified across the roughly twenty domains analysed so far.

2. The 12-Step Domain Analysis

Layer 1 of the methodology is a twelve-step procedure for analyzing one domain. The steps are sequential in their ordering but routinely recurse during application: identifying a primitive often forces a revision to partial levels, which forces a revision to dependencies, which surfaces new pair classifications. The procedure is therefore a discipline rather than a recipe. We describe each step briefly, note the iteration loop that ties them together, and list the refinements (R1–R13) that have accumulated through application across roughly twenty domains.

2.1. The cycle’s character

The twelve steps run as an alternating construct-and-reduce cycle — a modern, iterated instance of the classical analysis/synthesis method (Pappus, Descartes, Newton; see the methodological-context discussion in §Related Work). Each pass builds the analytical structure forward (construct: posit primitives, lay out dependencies, predict properties) and then reduces (analyze: remove what can be absorbed, collapse what is redundant, retract what cannot be derived).

The cycle has a dialectical character. Each reductive pass exposes a contradiction or redundancy in the current build — a primitive that turns out to be expressible from the others; a property that doesn’t derive; a dependency that proves spurious — and the resolution is a higher unified structure that both negates the prior distinction and preserves what it was tracking, expressed at a higher level. The pattern is the Hegelian thesis/antithesis/synthesis applied to engineering design and analytical method rather than to consciousness or history.

Convergence is a bilateral fixed-point criterion: the cycle stops when (a) no candidate primitive can be removed without losing a class of design moves the domain requires, AND (b) no candidate primitive can be added that is not recoverable from the existing primitives via the derivation discipline (Step 10b below). Both directions must reach the fixed point. Reduction stopping alone is not enough; addition stopping alone is not enough; both must. This is what makes “the primitive set” a structural claim rather than a stopping preference.

2.2. Steps 1–2: Information gathering and landscape analysis

Before naming primitives, read what exists. Survey the instances in the domain — working systems, documented designs, prior analyses where they exist — and note the design moves that recur. The landscape orients the analysis around shapes that the domain actually exhibits rather than around primitives the analyst would otherwise invent. The orientation is non-negotiable: skipping it produces elegant-looking primitive sets that fail step 11 (cross-domain pattern extraction) when the surveyed instances refuse to fit the primitives.

2.3. Step 1c: Level-of-description declaration

Before primitive extraction begins, declare the level of description at which primitives will be extracted. Primitives are level-relative in a way the classical analysis tradition leaves implicit: chemistry’s elements are irreducible at chemistry’s level and reducible at particle physics’s level. The declaration includes (i) what is accepted as primitive at this level without further reduction (the analytical floor), (ii) what is accepted as background without explicit modeling (the unmodeled context), and (iii) what assumptions hold about adjacent levels above and below.

The declaration matters because primitive-extraction debates often turn out to be level-of-description disagreements rather than disagreements about the primitives themselves. Once the level is declared, the disagreement either resolves or sharpens into a clear question about which level is most useful for the analytical question. The declaration also clarifies what the recursive partial-level decomposition (Step 3b and the sub-level structure discussed in §Product Lattice) is doing: it is re-entering primitive extraction at a finer level, with a new analytical floor and a new background.

2.4. Step 3: Primitive extraction

A primitive is irreducible at the chosen resolution. We use three tests, applied jointly:

• Structural minimality. Removing the primitive forfeits a class of design moves the surveyed instances actually make. A candidate that can be safely deleted is not a primitive.
• Compositional productivity. Combining the primitive with others yields new capabilities not present in any subset.
• Empirical recurrence. The primitive shapes design decisions across instances drawn from independent traditions (not all in the same lineage).

The tests are analyst-judgment-heavy. Step 11 disciplines them externally: a primitive that survives the three tests but fails to recur across domains is a candidate, not a confirmed primitive.

2.5. Step 3b: Partial-level decomposition

Each primitive decomposes into a gradient of partial levels — from absent to fully elaborated. Typical decompositions have four to six levels. Partial levels are not measurements; they are descriptive gradations the analyst names to capture where instances actually sit.

The 3/3b iteration loop is the methodology’s core reliability mechanism. Partial-level analysis routinely surfaces problems with the primitive set itself: one “primitive” turns out to be two features bundled together (the partial levels would need to track each independently); two “primitives” turn out to collapse at coarser resolution (their partial levels covary across all instances). Steps 3 and 3b iterate until the primitive set stabilizes against partial-level decomposition.

Partial levels themselves admit sub-level analysis when finer resolution is useful. The biology arrangement’s R0 $\to$ R2 transition is the canonical example: at one resolution the transition is a single edge in the chain; at sub-resolution it decomposes into eight sub-levels with their own primitive interactions, autocatalytic spirals, and crystallization events. The vocabulary is scale-invariant.

2.6. Step 3c: Evaluator identification (information-processing domains)

For domains that process information, one primitive typically plays the evaluator role: it translates encoding into function. The evaluator’s determinism level (typically labeled Kd, with partial levels from interpretive Kd1 to fully deterministic Kd4) is a critical structural variable; it tends to determine which other primitives admit which partial levels. Not all domains require this step — physical and abstract domains often have no evaluator primitive.

2.7. Step 4: Dependency specification

Primitives have partial-order dependencies: $B$ requires $A$ at some threshold before $B$ can advance beyond a corresponding threshold. We specify the dependency DAG explicitly. Most dependencies are conditional partial-level: $\mathrm{Dep}(A \geq x, B \geq y)$ — $B$ cannot reach level $y$ until $A$ reaches level $x$. The DAG filters the lattice (next chapter); the filter’s stringency is a measurable domain characteristic.

Step 4 also includes a domain-type declaration (R11): substrate, surface, ecosystem, context, or bridge. The domain type changes what the filter stringency means (substrates filter tightly; ecosystems filter loosely; bridges fall in between).

2.8. Steps 5–6: Pair enumeration and load classification

There are $\binom{n}{2}$ pairs of primitives. Each pair is classified qualitatively as heavy, medium, light, or negligible by four criteria: dependency strength, structural co-engagement, emergent-property contribution, and cross-instance recurrence. Hub primitives are those participating in many heavy pairs; anchor pairs are heavy pairs whose joint presence enables a load-bearing structural move.

2.9. Step 7: Coherent sub-lattice construction

The coherent sub-lattice is the set of lattice positions that satisfy the dependency DAG. Two resolutions:

• Coarse (primitive presence/absence): $2^n$ positions, filtered by dependency presence.
• Fine (partial-level): $\prod_i k_i$ positions, filtered by conditional partial-level dependencies. The fine lattice is much tighter than the coarse lattice because partial-level dependencies cut harder than presence-only dependencies.

The coherent fraction (filter stringency) varies by domain type and is one of the methodology’s most stable cross-domain observables.

2.10. Step 8: Hasse diagram walks

A walk is a monotone path through the coherent sub-lattice from the empty position to a fully populated position. Each walk is a build-up narrative: the order in which primitives can be added without violating dependencies. Walks at coarse resolution decompose into nested walks at fine resolution: a single coarse step (adding a primitive) typically expands into a multi-step fine walk through that primitive’s partial-level cascade.

2.11. Step 9: Load-bearing composition identification

A composition is a subset of primitives. Composition load measures whether the composition’s semantic content is irreducible to its component primitives: do the parts interact to produce something none of them alone does. Core triads are 3-primitive subsets whose pairs are all heavy and whose triangle is load-bearing. Higher-arity load-bearing compositions (quads and beyond) exist but are less common; the load classification operates at any arity.

2.12. Step 10: Emergent property prediction

Mapping load-bearing compositions to observable properties of the domain produces structural predictions. A system with a particular composition at particular partial-level positions is predicted to exhibit a particular property. These predictions can be tested against the surveyed instances and the literature. Predictions check against the world: empirical validity.

2.13. Step 10b: Derivation discipline

For every claimed emergent property at a composition, write an explicit derivation: how does the property emerge from the primitives’ interactions, given the composition’s structure and the partial-level positions of each primitive? The derivation must use only the primitive set, the dependency structure, the pair-relationship classification, and the composition rules established in earlier steps.

Derivation outcomes are graded on a four-level spectrum rather than binary success/failure:

1. Clean — the derivation goes through mechanically using only the primitive set, dependencies, and composition rules. The property is structurally grounded; this is the unambiguous case.
2. Plausible — the derivation is a reasonable structural explanation but requires some interpretive judgment in mapping primitive interactions to the property. The property is grounded with caveats; the interpretive choices are flagged.
3. Ambiguous — the property is empirically observed but multiple alternative derivations exist, or the structural mechanism is unclear at the current resolution. The property is not removed — empirical observation overrides derivation incompleteness. Flag for sub-level decomposition (§Product Lattice), 3/3b iteration, or additional composition rules.
4. Failed — no derivation appears possible from the current primitive set. One of three things is true: the primitive set is incomplete (a missing primitive that the derivation implicitly requires — re-enter the 3/3b loop); the property is mis-attributed to the wrong composition (move it); or the property is a higher-level observation that depends on additional context (escalate to a higher Sc level or add a context-domain dependency).

The discipline is conservative on removal: a property at Ambiguous status is kept on the emergent map with a flag, not removed. Removal requires a Failed derivation that survives 3/3b iteration and sub-level decomposition. Empirical observation outweighs analytical derivation when the two disagree — the methodology’s job is to make the disagreement explicit, not to police the empirical record.

Step 10 (prediction) and Step 10b (derivation) play complementary roles. Predictions check against the world (empirical validity); derivations check the methodology’s internal coherence (structural validity). A claim that predicts correctly but cannot be derived is a successful empirical observation about the domain, not a structural finding about the primitive set. A claim that derives cleanly but predicts wrongly indicates a derivation that doesn’t match how the domain actually behaves, suggesting either a model error or a domain anomaly worth examining.

The derivation discipline is the methodology’s analogue of Wierzbicka’s Natural Semantic Metalanguage paraphrase test, where every concept must be paraphrasable using only ~65 cross-linguistically stable semantic primes; paraphrase failure indicates either a missing prime or that the concept lies outside the metalanguage’s scope. The methodology’s version operates on structural emergent properties at compositions rather than on natural-language concepts, but the discipline is the same — the procedure’s outputs must be recoverable from the procedure’s primitives, and recovery failure is the test. (The NSM parallel is developed in §Related Work.)

2.14. Step 11: Cross-domain pattern extraction

After analyzing several domains independently, observe which patterns replicate across them: primitive counts in a narrow range, filter stringencies clustered by domain type, core triads with similar functional roles, hub structures. Patterns that replicate are candidates for Layer-3 abstractions (the convergence domain, the SSA topology). Patterns that fail to replicate are domain-specific features, not structural invariants.

2.15. Step 12: Literature alignment

Check the analysis against the established literature of the domain. A 12-step result that contradicts settled empirical knowledge needs revisiting; a result that aligns with settled knowledge demonstrates that the methodology recovers what is already known and may surface new structural framings of it. The point of step 12 is calibration against the domain, not endorsement by it.

2.16. Methodological refinements (R1–R13)

Thirteen refinements have accumulated through application. We list them briefly and note where each lives in the procedure:

• R1: Domain-kind declaration. Each analysis explicitly declares whether the domain is substrate / surface / ecosystem / context / bridge. Filter stringency expectations depend on the kind.
• R2: Filter-stringency reporting. The coherent-fraction percentage is reported per analysis. It is one of the cross-domain stable observables.
• R3: Firm vs borderline compositions. Compositions are marked as firm (multiple criteria) or borderline (single criterion). Borderline compositions are candidates for revision under further domain study.
• R4: Literature mapping. Every primitive set carries a mapping to vocabulary in the established literature of the domain, so cross-readers can locate the analysis against known terms.
• R5: Cross-domain mapping. Every primitive set carries a mapping to the corresponding Layer-3 abstraction (typically SSA roles or convergence-domain primitives) where one is identified.
• R6: Count-sensitivity honesty. When primitive counts could reasonably go either way (e.g., 5 vs 6 vs 7 depending on whether two candidates are merged), the analysis declares the alternative and notes the structural consequence.
• R7: Engagement with existing analysis. When prior analyses of the same domain disagree with this one, the disagreement is named and the structural source of it discussed.
• R8: Mode A / Mode B zoom. Mode A operates at a single resolution; Mode B zooms across resolutions. The mode is declared explicitly.
• R9: Candidate triage. Primitives that pass two of the three extraction tests but not the third are marked as candidates and carried through to step 11.
• R10: Architectural asymmetry acceptance. When the dependency DAG is genuinely asymmetric, the asymmetry is documented rather than smoothed out. Architectural asymmetry is often structurally informative.
• R11: Domain-type declaration. (Listed separately because it governs filter expectations.)
• R12: Level-of-description declaration (Step 1c). Declare what is accepted as primitive at the chosen level of description, what is background, and what assumptions hold about adjacent levels. Made explicit because several primitive-extraction debates in the corpus turned out to be level-of-description disagreements rather than disagreements about the primitives themselves.
• R13: Bilateral fixed-point criterion (§The cycle’s character). The cycle stops only when both reduction and addition reach fixed points. Reduction-side: no candidate primitive can be removed without losing a class of design moves. Addition-side: no candidate primitive can be added that is not recoverable from existing primitives via Step 10b. Made explicit when the derivation discipline was added; before that, the addition side was implicit and the stopping rule was under-specified.

2.17. Five domains compared

The methodology has been applied to roughly twenty domains. The five substrate-level domains run most frequently — entity system, biology, cognition, the convergence domain, and Layer 4 itself — are summarized in the table below. (The exact filter values are SageMath-computed from the dependency DAGs and are documented in the computational implementation chapter.)

The five analyses were authored independently from the literatures of their respective domains. The recurring shapes — primitive counts in a narrow band, filter percentages clustering by domain type, the presence of one or more core triads in every domain — are the patterns step 11 picks up.

3. The Product Lattice and Structural Properties

The 12-step procedure produces a primitive set, partial-level decompositions, and a dependency DAG. From these three inputs the methodology constructs a product lattice and reads several structural properties off it. The construction is mechanical; the properties it exposes are what the methodology uses to compare domains.

3.1. Lattice construction

Given $n$ primitives with $k_i$ partial levels each, the full product space has $\prod_i k_i$ positions. The dependency DAG, expressed as a set of conditional partial-level constraints $\mathrm{Dep}(A \geq x, B \geq y)$, restricts this to the coherent sub-lattice: positions satisfying every constraint simultaneously.

Two resolutions are useful in practice:

• Coarse ($2^n$): primitive presence/absence only. Each position is a subset of $\{p_1, \ldots, p_n\}$. Filtered by presence-only dependencies. This is the resolution at which Hasse walks are typically drawn.
• Fine ($\prod_i k_i$): full partial-level positions. Filtered by all conditional partial-level dependencies. This is the resolution at which exact filter stringency and walk counts are computed.

The fine resolution is much tighter than the coarse, because partial-level dependencies cut harder than presence-only dependencies. A primitive may be present at the coarse level but required to be at level $\geq 2$ before another primitive can advance beyond level $1$; the coarse lattice sees the presence, the fine lattice sees the threshold.

3.2. Filter stringency

The filter stringency is the fraction of the full product space that survives the dependency filter — the coherent sub-lattice size divided by the unfiltered lattice size. Across roughly twenty domains, filter stringencies cluster by domain type:

These ranges are empirical, not derived. They cluster because domains of similar kind have similar dependency densities; substrates sit at the bottom of realization chains and accumulate downward constraints (every layer above them must be compatible), so their filters tend to be tight. Ecosystems sit at the top and accumulate upward freedom, so their filters tend to be wider.

The empirical ranges are themselves a structural observation. Any new domain analysis whose filter stringency falls far outside its domain-type’s range is a candidate for re-examination of the primitive set or dependency DAG. We have once found a primitive in the entity arrangement that conflates “foundational substrate” with “tiny independent” at the same partial level; the unexpected cluster bridges this caused at fine resolution were a signal of the conflation.

3.3. Core triads, hubs, and anchor pairs

A core triad is a 3-primitive subset whose three pairs are all heavy and whose triangle is load-bearing (composition load irreducible to the pairs). Core triad function varies by domain type but the structural shape is constant: every analyzed domain has at least one core triad, and most substrate domains have exactly one. The convergence domain is unusual in having three overlapping core triads sharing the Ds primitive — a structural feature of its role as a Layer-3 abstraction.

A hub primitive is one participating in many heavy pairs. Hubs are typically the primitive an analysis ends up referring to most often; they tend to be present in every load-bearing composition. In the entity system, T (Tree) and I (Identity) are hubs. In the convergence domain, Ds (Distribution) is the unique hub. In Layer 4, Mn (Manifestation) is the hub with five of seven primitives forming heavy pairs with it.

Anchor pairs are heavy pairs whose joint presence enables a load-bearing structural move. They are the pair-level analog of the core triad: an irreducible unit at pair scale rather than triad scale. Anchor pairs and core triads are not independent — every core triad contains three anchor pairs, but not every anchor pair sits inside a core triad.

3.4. Phase transitions

A phase transition is a discontinuity in partial-level progression: beyond a particular threshold combination, the domain’s emergent properties change qualitatively. Phase transitions appear in two forms in the methodology:

• Within-domain. A single primitive’s partial-level progression has a phase-transition threshold above which a different configuration of compositions becomes available. The entity system’s $I = 0$ vs $I \geq 2$ split (location-addressed vs content-addressed) is the canonical example; many architectural properties change discontinuously across it.
• Cross-edge. A bridge or surface domain has a phase transition in one of its primitives that gates a cascade in a connected domain. The cognitive-development-bridge’s LA3 (recursive grammar) threshold is an example: the bridge’s primitive transitions release two language-cascading mechanisms in the surface domain above it.

Phase transitions are the structural form of crystallization: a transition that, once crossed, is hard or impossible to recross. We distinguish three crystallization sub-patterns at sub-level resolution: autocatalytic spirals (two primitives co-advance through a feedback loop with a critical threshold; the abiogenesis bootstrap loop is the canonical example), crystallization proper (a structural variable freezes permanently; the genetic code is the canonical example), and pre-separation fusion (two primitives function as one until a separation event releases them as independent). These sub-patterns appear in different forms in different domains; the vocabulary itself is portable.

3.5. Attractor states and the layering trap

Certain lattice positions attract many independent systems. A structural attractor is a position whose pair-relationship profile makes it especially productive for surrounding compositions, given the dependency DAG. Attractors are read off the lattice topology before any instance is positioned — they are predictions about where instances will cluster.

Empirically, instances tend to cluster at attractors and then scaffold: they ad-hoc compensate for primitives the attractor position omits. Scaffolding can be additive (a fence the system crosses by accumulating compatible features) or destructive (a wall the system would need to demolish in order to advance, because its architectural commitments are structurally incompatible with the missing primitive). The methodology’s layering trap is the empirical observation that scaffolding accumulated to compensate for a missing primitive can prevent the primitive from ever being added — the scaffolding occupies the structural slot the primitive would need.

The walls-vs-fences distinction is read off the dependency DAG: a fence is a missing primitive whose addition would be compatible with the system’s current partial levels (just additive work); a wall is a missing primitive whose addition would require lowering one of the system’s already-elevated partial levels (architectural retraction).

3.6. Design-opportunity discovery

Coherent but currently-unpopulated lattice positions are structural predictions: configurations the dependency DAG allows but that no instance in the surveyed landscape occupies. They are candidates for the methodology to flag as design opportunities (or as gaps the analyst should investigate). The entity system’s own partial-level position is one such unpopulated coherent corner in the entity arrangement’s substrate lattice — maximal substrate with nascent ecosystem, by construction. Whether such corners are fertile (a viable design space) or empty for a reason (a region the selection pressures avoid) is itself a structural question the analysis can frame but not, on its own, resolve.

4. Probabilistic Walks and Bayesian Inference

The 12-step analysis and the product lattice it produces are qualitative: they identify positions, dependencies, and walks but do not assign probabilities to them. The methodology becomes quantitative when the lattice is interpreted as a Bayesian network. This interpretation is not new machinery; the dependency DAG and the partial-level decompositions already define exactly what a Bayesian network requires.

4.1. The product lattice as a Bayesian network

A Bayesian network is a directed acyclic graph $G = (V, E)$ with a random variable at each node, together with a conditional probability distribution $P(X_i \mid \mathrm{Pa}(X_i))$ for each node given its parents. The product lattice supplies all three components:

• The vertices are the primitives.
• The edges are the dependency relations.
• The random variable at each node ranges over that primitive’s partial levels.
• The conditional distributions encode the conditional partial-level dependencies $\mathrm{Dep}(A \geq x, B \geq y)$.

Hard dependencies appear as zero-probability factors: positions that violate the DAG receive probability zero. Soft factors (physics, thermodynamics, market-friction) can be added on top as non-uniform weighting; heavy/light pair classification maps to coupling strength (roughly, mutual information between the pair’s primitives). Hard and soft factors compose naturally under the factor-graph representation.

The joint distribution factors as $P(X_1, \ldots, X_n) = \prod_i P(X_i \mid \mathrm{Pa}(X_i))$, restricted to the coherent sub-lattice. The coherent sub-lattice is therefore the support of the joint distribution. Any probabilistic question about positions, walks, or trajectories reduces to inference in this Bayesian network.

4.2. Forward walks (widening)

A forward walk starts from a known initial position and computes a distribution over reachable next positions, given the dependencies and constraints. The classical forward variable $\alpha_t(s) = P(X_t = s, \mathrm{evidence}_{1:t})$ captures the probability of reaching position $s$ at step $t$ given evidence accumulated so far. Forward walks are used for build-up narratives: starting from an empty primitive set, which sequence of additions is most probable under the dependency DAG plus the soft-factor weighting?

Forward walks widen, then narrow. As each step admits multiple successor positions, the distribution spreads. As dependency constraints accumulate (positions inconsistent with the DAG receive zero weight at each step), the distribution narrows again. The empirical signature is a probability funnel whose shape is informative about where the structural narrowing happens. The abiogenesis trajectory has a narrow funnel through R1.7 (the bootstrap threshold) followed by widening into the R2 plateau; the funnel shape is read off the joint distribution, not authored into it.

4.3. Reverse walks (convergent reconstruction)

A reverse walk starts from a known endpoint and works backward, narrowing the distribution at each prior step. The classical backward variable $\beta_t(s) = P(\mathrm{evidence}_{t+1:T} \mid X_t = s)$ captures the probability that the observed endpoint could arise from position $s$ at step $t$. Reverse walks are used for trajectory reconstruction: given that an instance is now at position $X_T$, what prior positions are consistent with it?

Reverse walks are the methodology’s primary tool for analyzing trajectories whose intermediate states are not directly observed. The abiogenesis trajectory is the paradigmatic case: the only directly observable terminus is the universal genetic code at R2, plus a handful of inferred milestones (the proto-ribosome, mineral compartmentalization). Everything between R0 (prebiotic chemistry) and R2 is reconstructed by reverse-walking from the known endpoint back through the dependency structure.

4.4. The posterior: forward × reverse

The forward-backward algorithm computes the posterior distribution over each intermediate position as $P(X_t \mid \mathrm{evidence}_{1:T}) \propto \alpha_t(X_t) \beta_t(X_t)$. The posterior is generally tighter than either the forward or reverse distribution alone: it incorporates both what reaches here from the start and what is necessary given the endpoint. The intersection is the high-probability corridor through the lattice.

Cross-domain constraint propagation works through the same mechanism. Evidence in one arrangement constrains posteriors in connected arrangements via the bridge primitives. The biology-cognition coupling at Sc=3 (where instances of cognitive ontogenesis co-evolve with instances of biological substrate development) is one such cross- domain propagation: evidence about substrate maturation in biology narrows the posterior over cognitive-development positions, and vice versa.

4.5. Computational tractability

For methodology lattice sizes encountered in practice, exact inference is computationally trivial. The largest lattices in current use have $n \leq 8$ primitives with $k \leq 6$ partial levels each — at most $6^8 \approx 1.7$ million positions. Brute-force marginalization runs in under a second on a workstation. Variable elimination on the dependency DAG runs in $O(n \cdot k^{w+1})$ where $w$ is the treewidth; typical lattices have $w \leq 4$, giving microsecond inference. We have not needed approximate inference anywhere; the methodology’s discrete, finite structures are well within exact reach.

This matters for two reasons. First, every probabilistic claim made under this framework can in principle be verified by exact computation; there is no approximation tolerance to argue about. Second, the cross-domain constraint propagation that the methodology relies on for Sc=3 coupling is exactly computable; the joint posterior across two coupled arrangements is no harder to compute than the posterior in one.

4.6. Mutual information and the area-law analogue

Two sub-networks of a Bayesian network exchange information through the edges that cross between them. The mutual information between the sub-networks is bounded by the number of bridge edges times the typical per-edge mutual information. This is the area law for Bayesian networks: information shared between regions scales with the boundary, not the volume.

The area-law observation is the structural reason bridge analyses are tractable. A bridge primitive sits on the boundary between two arrangements. The methodology treats bridges as first-class domains in their own right (with their own 12-step analysis, primitives, and filter); the area-law bound on information exchange justifies this treatment. The interior of each arrangement contributes its own posterior structure; the exchange between arrangements is mediated by a small number of bridge primitives, and the bridge’s analysis captures most of what flows across.

4.7. What is well-developed and what is not

The Bayesian-network interpretation is mathematically clean and computationally tractable at every scale the methodology has used. It is also under-exercised. Forward and reverse walks have been implemented and used at qualitative resolution in several trajectories. Cross-domain posterior propagation has been used in the Sc=3 coupling work. The full forward-backward algorithm at fine resolution, with explicit per-edge mutual information, has been sketched but not run as a primary instrument in any chapter of this paper.

Several avenues remain open. The category-theoretic and information- geometric framings noted briefly in the Convergence Domain chapter suggest connections we have not pursued at depth: Fisher information metrics over the partial-level state, geodesics on the manifold of coherent distributions, presheaf-to-section as a formal account of crystallization. The structural ground is laid; whether the formalization layer pays off in additional insight is an open question we are not currently positioned to answer.

5. The Convergence Domain

Applying the 12-step analysis to the abstract pattern shared by quantum measurement, Bayesian inference, biological fixation, lattice-walk crystallization, and market lock-in yields a domain with six primitives: Space (the structured set of possible states), Distribution (the probability assignment over them), Constraint (what shapes it), Dynamics (how it evolves), Collapse (the irreversible narrowing event), and Determination (the persistent post-collapse state). The dependency spine is linear — $\mathrm{Dt}\leftarrow\mathrm{Cl} \leftarrow\{\mathrm{Dy},\mathrm{Cn}\}\leftarrow\mathrm{Ds}\leftarrow \mathrm{Sp}$ — with a 14.1% coarse filter and three core triads: the Landscape triad $\{\mathrm{Sp},\mathrm{Ds},\mathrm{Cn}\}$ (shaped possibility space with attractors and barriers), the Directed-Evolution triad $\{\mathrm{Ds},\mathrm{Cn},\mathrm{Dy}\}$ (constrained search), and the Information-Gain triad $\{\mathrm{Ds},\mathrm{Dy},\mathrm{Cl}\}$ (the truth event, where uncertainty resolves irreversibly). A category- theoretic reading (Space as category, Distribution as functor, Collapse as limit, Determination as fixed point) and a structural classification by distribution type (Ds=2 real-valued classical convergence, Ds=3 complex-amplitude quantum convergence with interference) are available but not developed at length in this paper; we note them as connections to the literatures cited in the related-work chapter.

The result that matters for this paper is reflexive. The methodology is itself an instance of the convergence domain: Layers 1–3 build the Space and Constraint; Layer 4 operates the Distribution, Dynamics, Collapse, and Determination. Concretely, a cluster in any landscape analysis is the Landscape triad’s emergent product; an attractor or anchor is a Determination; a phase transition is a Collapse. This is not an analogy imposed after the fact. Role-identifying the six primitives across three arrangements built independently and from separate literatures — software systems, biological organisms, and the space of analytical methods — recovers the same dependency spine and the same core-triad functions in each. Three independent realizations of one topology, with the five prior abstract instances, place the convergence domain among the methodology’s confirmed Layer-3 abstractions alongside the Situated Substrate Architecture.

Two consequences are load-bearing and stated as constraints on the method’s own claims, not as results. First, the instantiation is topological, at classical (real-valued, non-interfering) distributions: it confirms that the structure recurs, and derives no new measurement; the dependency model is analyst-authored and the mapping invites disproof. Second, it disciplines what the method may assert. A cluster is Landscape-triad structure — a description of where mass sits — and is, by construction, not a finding; a finding requires the Information-Gain triad, an actual irreversible determination checked against an external fact. A determination is by default mutable (an attractor that a later analysis can displace); calling one permanent requires an independent irreversibility argument, not a clustering score. The methodology improving through its own use is itself a convergence process, which is the precise sense in which the framework is self-validating.

6. The Realization Chain

The 12-step procedure analyzes one domain at a time. Layer 2 connects analyzed domains through a typed graph of edges. Across many such analyses, one structural pattern recurs strongly enough to deserve its own treatment: a chain of information substrates from physics to computing, connected by bridge domains that progressively open a distance between where evaluation happens and where feedback operates. The chain itself is not a single domain. It is a sequence of domains plus bridges, with a structural variable that varies monotonically along it.

6.1. The chain

The chain has five levels:

$$\mathrm{Physics} \to \mathrm{Chemistry} \to \mathrm{Biology} \to \mathrm{Cognition} \to \mathrm{Computing}.$$

Each level is an information substrate analyzable by the 12-step procedure. Each is connected to the next by a bridge domain, also analyzable by the 12-step procedure. The chain is not a hierarchy in the sense of higher levels reducing to lower ones — chemistry does not reduce computationally to physics in any useful sense for an analyst trying to understand chemistry. What the chain captures is that each level depends on the previous level for its substrate while operating in its own medium, and the bridge between them is where the medium opens.

6.2. Bridge domains

A bridge exists wherever there is a substrate gap: the upper system operates in a medium structurally different from the lower. Bridge primitives are the operational machinery of the gap. Bridges are domains in their own right — they have a primitive set, partial levels, a dependency DAG, a coherent sub-lattice, and core triads. The abiogenesis bridge (chemistry $\to$ biology) is the most extensively analyzed, with six primitives covering encoding, catalysis, growth, fixation, compartmentalization, and feedback. Its substructure (the eight sub-levels R0 $\to$ R2 with the bootstrap loop) is the deepest sub-level analysis the methodology has applied to any bridge.

The chain’s bridges have been analyzed to varying depth. The abiogenesis bridge is the most developed. The neural / cognitive bridge between biology and cognition has been analyzed at primitive resolution. The design / implementation bridge between cognition and computing has been sketched. The quantum-chemistry bridge between physics and chemistry has been analyzed at primitive resolution but not with full partial-level decomposition. These are open avenues; none is fully complete.

6.3. Evaluation-feedback distance

The structural variable that varies monotonically along the chain is the evaluation-feedback distance: the spatial, temporal, and organizational separation between where evaluation happens and where feedback operates.

At physics the gap is essentially zero: the update operator IS the feedback, applied in place. At each level above, the bridge opens the gap further. In chemistry, products can fail to persist (thermodynamic feedback); evaluation (the reaction) and feedback (stability) operate at the molecular scale but are not the same event. In biology, an organism’s action and its reproductive consequence are separated by years and meters; the compartmentalizing membrane is the gap made physical. In cognition, an individual’s idea and its cultural persistence are separated by centuries; in computing, by design choice (an architecture can be specified to make evaluation and feedback arbitrarily distant).

The gap is not a side-effect of the chain — it is the source of the chain’s structure. Complexity at each level is what fills the gap. Without a gap, evaluation and feedback collapse to the same event and there is no room for the structure the level exhibits. At physics, where the gap is zero, there is no organism, no cognition, no computing system; the structures of the higher levels exist because the lower-level gap has widened enough to admit them.

6.4. Compartmentalization as the prototypical distance-opener

Each bridge has at least one primitive whose function is to open the evaluation-feedback distance. In the abiogenesis bridge it is compartmentalization (Cmp): the mineral micropore that physically separates an autocatalytic cycle from the bulk chemistry around it, so that the cycle’s products can accumulate without being immediately dispersed. The membrane is the structural ancestor of every later compartmentalizer: the cell membrane, the skull, the protocol specification. Compartmentalization is the recurring shape of the distance-opening move.

This is not a single primitive recurring under five names; it is a pattern recurring across five domains, instantiated as distinct primitives whose functional role is structurally analogous. The methodology’s discipline forbids collapsing them into one cross-domain primitive — they live in different arrangements, with different dependency partners, and the analytical work happens at the arrangement-specific resolution. The recurrence of the pattern across arrangements is a Layer-3 observation, not a Layer-1 primitive unification.

6.5. Nesting and termination

The realization chain nests recursively. Each bridge decomposes into sub-levels with their own primitives, dependencies, and walks; the abiogenesis bridge’s R0 $\to$ R2 sub-decomposition is the canonical example. Each sub-level can in principle be decomposed further when finer resolution is useful. The vocabulary is scale-invariant: the same primitive set theory, the same pair-relationship classification, the same coherent-sub-lattice construction, applies at every scale the analyst chooses to examine.

The nesting terminates at physics. Below physics the chain does not continue — there is no substrate the methodology has identified below the Planck information substrate. Whether this is a structural fact (physics is genuinely the floor) or a horizon (there is more below but the methodology cannot see it from where it stands) is an open question the chapter does not resolve. The Planck substrate analysis in the physics-domain track sketched in earlier work proposes the spectral triple as a structural candidate for the substrate at the chain’s bottom; whether this proposal is correct, and whether it terminates the chain or merely extends it, are open questions for further analysis.

6.6. What the chain is and what it is not

The realization chain is an observation about how the methodology’s domain analyses connect across realization levels. It is not a foundational claim about the structure of reality. The chain is the shape that recurs when many independently-analyzed substrate domains are placed next to each other and their bridges are filled in. The evaluation-feedback distance is the unifying variable the chain exposes; the variable is structural, not metaphysical.

Several open avenues remain. The quantitative form of the evaluation-feedback distance (whether it admits a unified mathematical definition rather than the qualitative ordering above) is not settled. The relationship between the Sc=4 cross-arrangement coupling at the chain’s top (where computing, cognition, and biology share substrate via shared physics) and the chain itself is structurally clean but not fully formalized. The chain’s behavior when applied to information substrates outside the physics-to- computing sequence — mathematics, for instance, which couples to physics through description rather than through realization — is an active direction that the present paper does not pursue further.

6.7. Parallel surfaces of a single substrate: the cognition example

A realization chain typically describes one track through a stack of substrates — e.g., the cognition arrangement’s behavioral track that runs from neural hardware through cognitive substrate, cognitive architecture, and into the cultural ecosystem. But within a single arrangement, a substrate can produce multiple surfaces, each decomposing a different aspect of what the substrate outputs. The cognition arrangement is the cleanest formally extracted example.

Alongside the behavioral surface — cognitive architecture, describing what cognition does (knowledge, skill, decision, planning, identity) — the cognitive substrate also admits a parallel semantic-content surface that describes the discrete content units cognition emits for use in language and thought. Both surfaces sit one realization step above the cognitive substrate and connect upward to the cultural ecosystem via their own bridges:

cognitive substrate {Rp, Ct, As, Sq, Sy, Ev}
  ├─ (cognitive-development bridge, ~10 primitives)
  │    → cognitive architecture {Kw, Sk, Dc, Pl, Co, Jd, Cr, Si, Id}
  │      (behavioral / organizational surface)
  │      → (social-transmission bridge, ~10 primitives)
  │
  └─ (cognitive-to-semantic bridge, ~12 primitives, hub: Lexicalization)
       → semantic surface {Rf, Mn, Ac, St, Sp, Md, Ev}
         (content surface)
         → (semantic-to-cultural bridge, ~12 primitives, hub: Circulation)

both bridges feed → cultural ecosystem {Pr, Ex, Tr, Dv, Cd, Gv, Te, Sc, Ct}

The semantic surface’s primitives are content-shaped rather than operation-shaped: Reference, Mental activity, Action, State, Spatiotemporal, Modal, Evaluative. They correspond closely to Wierzbicka’s Natural Semantic Metalanguage primes (developed in §Related Work), re-grouped at surface resolution: the ~65 NSM primes resolve to ~16 categories at intermediate resolution and to ~7 surface-level primitives at coarser resolution. The methodology’s scale-invariance predicts exactly this multi-resolution view of the same domain. Wierzbicka’s domain-type-declaration analysis treats NSM’s primes as semantic content the cognitive substrate emits at the symbolization interface — a surface output, not a substrate-level structural commitment in their own right.

The two surfaces are not in conflict. The behavioral surface describes the operations the cognition stack performs and the capabilities the substrate produces (representing, categorizing, planning, identity). The semantic-content surface describes the outputs the operations produce as semantic units (referenceable things, mental predicates, actions, states, modals, evaluations). Both surfaces connect cognitive substrate to cultural ecosystem; they describe different aspects of the same overall flow.

Observation: multiple surface analyses can coexist. The methodology does not dictate the shape of the realization graph; given a domain and a level-of-description declaration, it produces a primitive set, dependency structure, and the rest. Whether an analyst chooses to extract one surface or several from a given substrate is a methodological choice, not something the methodology constrains. The relevant test is whether the resulting analysis is coherent: each candidate surface must independently pass the three-test primitive criterion, the dependency-filter check, the core-triad identification, and the M3 derivation discipline. Cognitive architecture and the semantic surface both pass these tests, so the multi-surface decomposition is coherent in the cognition arrangement.

Whether multi-surface decompositions also pass in other arrangements has not been formally tested. Initial sketches of candidates suggest the cognition case may be unusual rather than recurring. A biology candidate (genome architecture as content surface alongside organism architecture) decomposes into primitives that look like fine-grained elaboration of the existing G (Genome) substrate primitive — sequence, reading frame, regulatory element, mobile element, packaging — which is the §Product Lattice sub-level decomposition pattern rather than a separate surface domain. An entity-system candidate (spec-as-content as a content surface) runs into a different issue: the entity-system substrate is self-describing, with types, namespaces, and capabilities all expressed using E + I + T applied reflexively to themselves. There is no distinct content vocabulary because the content units are the substrate primitives.

A tentative structural reading: multi-surface decompositions appear coherent when the substrate’s primitives are predominantly operations that produce stable content units distinct from themselves. Cognition has this property — the substrate primitives Representation, Categorization, Association, Sequence, Symbolization, Evaluation are operations whose outputs (semantic content units that NSM analyzes) are different objects from the operations themselves. Biology’s substrate is mixed (G is content, T/R/Reg are operations, Mem is spatial), with the bulk of the substrate’s content already lodged in G — which is why a candidate “content surface” reduces to sub-G detail rather than a parallel domain. The entity system’s substrate is content-heavy (E, I, T are content-shaped) and self-describing, so the substrate and the would-be content surface collapse together. This reading is candidate-licensed-claim status; it has not been tested by formally extracting a second surface from biology or the entity system at the discipline this paper applies.

The Situated Substrate Architecture description in this paper has been written assuming a single surface per substrate by convention. The convention is not load-bearing on the methodology — nothing in the analytical apparatus prevents extracting multiple surfaces when the substrate’s structure supports it. The cognition arrangement shows where the convention can be loosened; the biology and entity-system sketches above show where loosening it may produce sub-level analyses rather than fresh surface domains.

A scope distinction within the semantic surface. The cognitive architecture surface is straightforwardly single-mind: knowledge, skill, decision-making, planning, and identity are capabilities a cognitive substrate produces individually. The semantic surface admits two scopes that the diagram above does not separate. At the single-mind conceptual scope, the semantic surface’s primitives — Reference, Mental activity, Action, State, Spatiotemporal, Modal, Evaluative — are cognitive content units that an individual cognitive substrate can represent and manipulate. At the multi-mind lexical scope, the same primitives appear as cross-linguistic lexical items validated by Wierzbicka and Goddard’s empirical paraphrase discipline; they exist at this scope only because the cognitive-to- semantic bridge (Lexicalization, Conventionalization, Universalization, Crystallization) has stabilized them across a speech community. NSM as a research program operates at the multi-mind scope; the same primitives at single-mind scope are not directly tested by NSM’s methodology but are the cognitive prerequisite for the cross-linguistic stabilization to occur.

The two scopes are connected: single-mind conceptual primes are the substrate that the W2 conventionalization bridge stabilizes into multi-mind lexical primes. They are aspects of one domain examined at different resolutions, consistent with the methodology’s scale- invariance (see §Product Lattice §2.5 on sub-level analysis). The diagram above shows the relationship at the single-mind scope — the semantic surface parallel to cognitive architecture, both sitting above cognitive substrate. The multi-mind lexical aspect lives downstream, anchored at the cognitive-architecture-to- cultural-ecosystem boundary where the conventionalization bridge deposits its stabilized output. Both placements are valid; they describe the same domain at different scopes.

Bridge-pattern observation. The two bridges to the semantic surface (cognitive-to-semantic and semantic-to-cultural) join the existing corpus of analyzed bridges (biology-to-organism, neural-to-cognitive, entity-system substrate-bridge extensions, and others). A pattern recurs across all analyzed bridges: ~10-12 bridge primitives, a hub primitive at the channel operation (Lexicalization in the cognitive-to-semantic bridge; Circulation in the semantic-to-cultural bridge; Cell in the biology-to-organism bridge; Type in the entity-system substrate-bridge set), and a two-core-triad structure (one production-side triangle for how new units enter the bridge, one authority-side triangle for how units get fixed in the bridge’s output). The pattern is developed further in §Cross-Domain Structural Patterns.

7. Applied Analysis: Layer 4 and the Scope Ladder

Layers 1–3 produce structural knowledge: primitives, dependencies, filter stringencies, core triads, cross-domain abstractions, the realization chain. Layer 4 applies this knowledge to concrete situations. It is the methodology’s use layer, where structural knowledge meets a specific question about a specific instance.

Layer 4 itself is a domain analyzable by the 12-step procedure. The analysis surfaces seven primitives, a scope ladder, and a small set of standard analytical moves the layer supports.

7.1. The seven primitives of Layer 4

Layer 4 has seven primitives:

• Framework (Fw). The arrangement under which an analysis is conducted — biology, cognition, entity, methodology, etc.
• Manifestation (Mn). A specific instance positioned within an arrangement.
• Scope (Sc). The resolution at which the analysis operates, from Sc=0 (universal) to Sc=4 (specific instantiated event).
• Context (Cx). Operating constraints external to the arrangement itself.
• Landscape (Ls). The set of manifestations placed in a common positional space for comparison.
• Coupling (Cp). Structural relationships between manifestations in different arrangements.
• Trajectory (Tj). A sequence of manifestations through time.

The hub is Manifestation: five of the six other primitives form heavy pairs with it. The independent root is Context — it constrains achievability without being constrained by the others. The anchor pair is Mn–Cx (an analysis always positions a manifestation against a context).

Three core triads branch from the Mn–Cx anchor:

• $\{\mathrm{Mn}, \mathrm{Cx}, \mathrm{Sc}\}$: the Analytical Frame. Choose a manifestation, choose a context, choose a scope; the rest of the analysis is bounded by those three choices.
• $\{\mathrm{Mn}, \mathrm{Cx}, \mathrm{Ls}\}$: Strategic Positioning. Place the manifestation against a landscape of peers in the same context to see where it sits structurally.
• $\{\mathrm{Mn}, \mathrm{Cx}, \mathrm{Tj}\}$: Trajectory Planning. Project or reconstruct the manifestation’s path through positions over time.

Filter stringency is 29.7% — on the loose side for an analytical domain, reflecting Mn’s hub role and the multiple parallel-use configurations of the rest of the primitives.

7.2. The scope ladder

Scope is the dial that controls what kind of question is being asked. The same arrangement admits qualitatively different analytical moves at different scope settings.

• Sc=0 (universal). The category as a whole. Structural vocabulary in its abstract form. No specific instance is in scope. At Sc=0 the analysis is about the methodology applied to domains in general rather than to any one domain.
• Sc=1 (class). A class of instances sharing arrangement and context. The structural topology of the class as a whole — lattices, walks, core triads, filter stringencies. Sc=1 is the default resolution for the 12-step domain analysis.
• Sc=2 (configuration). A specific configuration of partial levels within the class. Rate-weighted analysis enters here: rates per transition, expected times, expected populations, probability distributions over walks. Sc=2 is where wall-time calibration attaches.
• Sc=3 (instance). A specific manifestation, optionally evolving over time (a trajectory). Sc=3 is the natural resolution for case studies; the X-genesis trajectory taxonomy lives here.
• Sc=4 (event). A specific instantiated event, typically occupying positions in multiple arrangements simultaneously via shared physics. Sc=4 is the resolution at which cross-arrangement coupling becomes most concrete.

The scope dial structurally controls the character of every other primitive. Manifestation at Sc=1 is a class; at Sc=3 it is an instance; at Sc=4 it is an event. Trajectory at Sc=1 is the set of all walks through the class lattice; at Sc=3 it is one path through positions. The same vocabulary applies; the scope determines what the vocabulary refers to.

7.3. Unified manifestations

A unified manifestation specifies an entity’s position across all chain levels in its arrangement, not just one. A software system in the entity arrangement has positions at four chain levels (computing-to-entity bridge, entity-system substrate, application architecture, digital ecosystem); a unified manifestation specifies all four jointly. The full-vector representation has been used to compare Git and Postgres: they share a computing-level position but diverge sharply at the substrate level, and the divergence is exactly readable off the unified manifestations’ position vectors.

Unified manifestations are the standard analytical object. Partial manifestations — single-chain-level position vectors — are scoped views of unified manifestations rather than independent objects. The discipline matters because cross-level analysis (anchor authoring at one level, trajectory at another) requires the joint position to be specifiable when needed.

7.4. Context domains and bottleneck analysis

Every arrangement has a context domain — a separately-analyzed domain whose primitives constrain the arrangement’s achievability without being part of the arrangement itself. Three context domains have been analyzed:

• Biology environment (6 primitives): nutrients, climate, competition, etc.
• Cognitive context (6 primitives): cultural inputs, available symbols, transmission channels.
• Digital context (6 primitives): platform availability, network topology, regulatory regime, etc.

A context bottleneck is the context primitive currently limiting the manifestation’s achievable position. Bottleneck analysis at Sc=3 identifies which context primitive a manifestation is bound by; moving the bottleneck (changing the context) is structurally different from changing the arrangement itself.

7.5. Trajectories and lifecycle patterns

Trajectories at Sc=3 admit qualitative lifecycle patterns — characteristic shapes that recur across many trajectories in the same arrangement. Five patterns have been identified empirically:

• Ship-and-Done. Substrate position set early, no further primitive advancement; the system stops at its initial structural position and accumulates only operational refinements.
• Feature Plateau. Substrate set early; the system advances at the surface and ecosystem levels until a structural ceiling is reached, then plateaus.
• Continuous Elaboration. Substrate set early; the system advances continuously at surface and ecosystem levels without an obvious ceiling.
• Version-Cycle. Periodic substrate revisions; each revision shifts the structural position discretely.
• Evolve-or-Die. The system must advance the substrate to remain viable in its context; failure to advance produces contextually-driven extinction.

The patterns are domain-general — versions of them appear in biology, cognition, and computing trajectories — but the diagnostic value of any one pattern depends on the arrangement.

7.6. The Sc=1 $\to$ Sc=2 boundary

Above the Sc=1 $\to$ Sc=2 boundary, structural analysis is reliable: the methodology can identify what configurations exist, where they sit in the lattice, what their pair structure is, and which trajectories are accessible to them. Below the boundary, structural analysis becomes increasingly thin and empirical work takes over. The methodology identifies landscapes of possible designs and constraints on viable ones; choosing among viable designs in a specific situation requires prototyping and measurement that the methodology cannot replace.

The boundary itself is not a strict cut — some Sc=2 questions admit structural answers (the LUCA wall-time calibration recorded in the empirical-cartography companion note is an Sc=2 result with a structural component), some Sc=1 questions require empirical input (the precise location of an attractor in the entity-arrangement landscape depends on observable adoption patterns). What the boundary marks is where the methodology stops being load-bearing. Above it, structural reasoning carries the analysis; below it, structural reasoning frames the analysis but empirical work fills it in.

7.7. Open avenues

Layer 4 is the methodology’s most actively explored layer and several avenues remain partially open. The unified-manifestation schema and its multi-arrangement extensions are still maturing. Sc=3 sustained cross-arrangement coupling has been worked at the landscape level (the two paired-manifestation studies recorded in the empirical-cartography companion note) but not yet exercised across full multi-decade trajectories. The context domain analyses are at primitive resolution but their partial-level structures are less developed than the substrate domains. The applied use of Layer 4 in design guidance — given a current manifestation, what additive moves through the coherent sub-lattice are accessible to it, and what extension paths exist toward target capabilities — has been sketched but not yet developed as a primary analytical instrument. These are directions for continued exploration rather than gaps in the present paper’s structural claims.

8. Methodological Discipline: Licensed Claims and Analyst Cartography

A structural methodology applied across many domains can quickly produce outputs that look like findings without being findings. The product lattice, the coherent sub-lattice, the corridor through it, the cluster decomposition, the trajectory regime, the rate-weighted wall time — each is a description of where mass sits in a constructed space, and the construction is analyst-authored from the start. We distinguish two postures the methodology can take toward its own outputs:

• Analyst cartography is the description posture. The lattice, the corridor, the clusters, the role decompositions, the anchor inventories are all maps. Every map encodes a particular choice of primitives, partial levels, dependencies, instances, and signature formula. The map is useful insofar as it organizes attention; it is not, on its own, evidence that the underlying domain has the structure the map depicts.
• Licensed claim is the finding posture. A licensed claim is a statement the methodology asserts as true of the world, backed by evidence whose generation did not pass through the same choices that produced the map. The simplest sufficient condition is non-circular external validation: a pattern read out of the methodology’s outputs matches an independently established result in a discipline whose vocabulary, data, and conclusions were not inputs to the analysis. The strongest condition is cold out-of-sample classification: a held-out instance is scored against the methodology’s existing structure and the resulting position discriminates between known categories rather than placing the instance vacuously.

The Convergence Domain framework above makes the distinction constitutive rather than rhetorical. The Landscape triad $\{\mathrm{Sp}, \mathrm{Ds}, \mathrm{Cn}\}$ produces clusters — descriptions of where the distribution has mass. The Information-Gain triad $\{\mathrm{Ds}, \mathrm{Dy}, \mathrm{Cl}\}$ produces determinations — irreversible narrowings checked against an external fact. A cluster, by construction, is not a finding; it is Landscape-triad structure. A finding requires the Information-Gain triad to actually fire: a cold classification, a held-out prediction that survives, a non-circular external match. Determinations are mutable by default; calling one permanent requires an independent irreversibility argument, not a clustering score.

This discipline applies in both directions. Patterns the methodology produces (from its own authored inputs) are cartography. Patterns the methodology recovers (from inputs whose generation was independent of the analysis) carry licensed-claim weight commensurate with how independent the generation was. Two forms of self-deception that the discipline guards against:

Circular validation. Feeding popularity, prevalence, or any other target signal into the inputs (positions, dependencies, rate weights, seed conditions) of an analysis whose output is then compared against that same signal cannot validate anything. A pattern read back out under such a setup is an artifact of the input, not a structural finding. When circular validation is identified inside a result the methodology has previously produced, the result is downgraded: the mechanism may survive (the engine that ran the analysis is unchanged) but the validation claim is withdrawn. We have done this once during the development of this paper, when a competitive-displacement run that had been advertised as cross-domain validation was found to have been seeded with the very target signal it was meant to recover; the structural finding (directed-target navigation is vestigial for non-fixed-evaluator domains, competitive displacement is the operative model) survives because it is emergent from mechanics, not from the target signal. The “validation” label does not.

Confirmation through tuning. When an analysis appears to produce a striking effect, the discipline is to visualize the trajectory before asserting the effect. We have once mistaken a synchronized-extinction artifact (one split-policy trigger firing every step on a permanently off-manifold climber, generating combinatorial branch explosion followed by mass death) for a paradox-of-enrichment confirmation. The artifact surfaced on first visualization. The retraction is recorded in the development trace and the mechanism stands; the validation claim does not. We mark such cases as honest negatives, never tuned away.

The licensed/cartography distinction also disciplines the empirical work — the trajectory-regime slice kept in this paper and the wider record in the empirical-cartography companion note, which mixes both postures. The X-genesis trajectory regime taxonomy is cartography at the corpus level (we authored the trajectories; the clusters they fall into are descriptive). The BOUNDED-GENESIS candidate fourth regime is closer to licensed claim, because the chimpanzee cognitive-substrate ceiling is established in the primatology literature independently of our authoring choices, and the composite gate that fails to release at LA3 is read out of dependency structure specified before the trajectory was authored. The cross-arrangement keypress example is cartography at the analytical-decomposition level. The LUCA calibration is cartography of an architecture (the architecture works) plus a within-empirical-range observation that borrows its license from the LUCA anchor itself (independent geochemistry, not our authoring). The Sc=3 sustained-coupling landscape studies recover three-zone clustering structure empirically; the pattern transfer across the methodology arrangement and the entity arrangement is a licensed claim (independent corpora, common framework), the cluster labels are cartography.

The cleanest licensed claims in this paper come from where the methodology recovers independently established results. The Convergence Domain’s instances align with established mathematical descriptions of quantum measurement, Bayesian inference, and biological fixation in literatures that did not contribute to the methodology’s vocabulary. The methodology landscape’s three-zone clustering pattern is recovered identically from a thirteen-instance corpus of strategic-analysis methodologies and an independent thirty-four-instance corpus of entity-arrangement information systems, with the linear-inverse correlation strength itself diagnostic (smooth gradient in the methodology arrangement; bimodal walls in the entity arrangement). The biological taxonomy is recovered at silhouette $\sim 0.69$–$0.83$ from a fifty-four-instance biology corpus using the same recipe (lens family, anchor authoring, meta-stability spine) developed on the entity arrangement and never exposed to taxonomic ground truth during clustering. Each of these recoveries is a licensed claim about the methodology’s cross-domain applicability, not about the underlying domains.

The discipline carried into the rest of the paper is therefore: every result that follows is marked either as cartography (a description the methodology produces from its inputs) or as licensed claim (an empirical recovery whose evidence is non-circular). The two are not interchangeable. The cartography is the methodology’s output; the licensed claims are the methodology’s test.

8.1. Internal-coherence validation: the derivation discipline as a third validation kind

Cartography and licensed claim are the two postures established above. The derivation discipline (Step 10b) introduces a third kind of validation that operates internally to the methodology rather than against external evidence: every claimed emergent property must be derivable from the primitive set, dependency structure, and composition rules. Derivation failure is not a claim that the property doesn’t exist in the world; it is a claim that the methodology’s outputs are not internally coherent with its inputs.

The three validation kinds play complementary roles:

The derivation discipline was added because cartography and licensed claim left an internal-coherence gap. A primitive set can be cartographically honest (the analyst declared inputs, outputs, and their relation) and can pass non-circular external validation (specific predictions recover known results) while still being internally incoherent — emergent properties claimed at compositions that do not in fact follow from the primitives. The derivation discipline closes this gap by requiring the analyst to show the work connecting primitives to properties.

The four-level outcome spectrum (Clean / Plausible / Ambiguous / Failed) was developed because binary success/failure produces brittle removals — empirical observations that genuinely emerge from a domain but resist clean structural derivation would be incorrectly purged. The conservative-on-removal discipline (Step 10b above) means Ambiguous properties stay on the emergent map with a flag; only Failed properties surviving 3/3b iteration and sub- level decomposition drive primitive-set revision.

The empirical experience of running Step 10b across the analyzed corpus is summarized in §Cross-Domain Structural Patterns: ~150 derivations across 12+ domains, ~83% Clean, ~13% Plausible, ~2% Ambiguous, 0% Failed. The discipline surfaces useful structural questions (chemistry’s far-from-equilibrium dynamics flagged as Ambiguous, suggesting sub-level decomposition of energy / boundary / feedback as sub-primitives) without forcing removals of valid properties. The cross-corpus result is also the closest analogue to Wierzbicka’s NSM paraphrase discipline operating on the methodology’s own outputs — the parallel is developed in §Related Work.

8.2. A structural qualifier: the cartography/licensed-claim distinction is attractor-dependent

The cartography/licensed-claim distinction does not operate uniformly across all the domains the methodology might be applied to. In the later chapter on the methodology’s range, we present an eight- primitive meta-domain analysis of analyzable domains, with nine empirical attractors that map where the procedure produces high-value output, where it produces partial output, and where it does not apply. The cartography/licensed-claim distinction interacts with those attractors:

• In the high-value substrate attractor (the methodology’s most productive zone, where information-processing substrate domains live), the licensed-claim path is available whenever non-circular external recovery is in evidence. The cartography vs licensed-claim distinction operates as described above.
• In the cyclic-rich attractor (ecosystem dynamics, brain population dynamics, deep markets, climate, coevolution), the licensed-claim path is available for the substrate compositional skeleton of the domain but not for the equilibria of its cycles; cyclic equilibria belong to dynamical-systems analysis rather than to the structural methodology, and analyses of cyclic dynamics here remain cartography.
• In the cyclic-thin attractor (governance, macro-economics, law), the licensed-claim path is limited; outputs are mostly cartography because the analyst’s cycle-breaking convention is doing more of the work than the structural decomposition is.
• In the function-mismatch, granularity-bottleneck, mature-craft, contested-judgment, axiomatic-degeneracy, and empirical- bottleneck attractors, the licensed-claim path is not available in any operational sense. Outputs in those attractors are cartography only, or the methodology produces no operational output at all.

A second structural qualifier: the methodology is not self- validating against framework-level error. Its discipline catches internal inconsistency, filter-stringency anomalies, circular validation when input signals are fed back as outputs, and visualization artifacts; it does not catch the possibility that the methodology’s overall analytical frame is fundamentally wrong in a way that internal consistency cannot detect. Historical analogues exist (phrenology, caloric theory, Lamarckian inheritance, Galen’s humoral theory) where structural analytical frameworks passed all the discipline of their time before being displaced by experimental crucial-tests, better instrumentation, or theoretical advances that subsumed them as special cases. The licensed-claim path is conditional on the framework being approximately correct, and the methodology has no internal mechanism that would detect its own displacement. This is a structural limit of analytical practice, not a defect specific to this methodology.

8.3. The primordial intuition and open questions

A pattern recurs across every substrate the methodology has analyzed: the primitive sets manifest informational, temporal, and spatial aspects. The entity system’s six primitives factor as three informational (Entity, Identity, Type), two temporal (Emit, Execute), and one spatial (Peer). Biology’s primitive set factors with the informational and temporal weights swapped (Genome and Protein informational; Transcription, Translation, Regulation temporal; Membrane spatial). Cognition’s substrate primitives are heavy on cross-axis operations (Categorization, Association, Symbolization, Evaluation) with one informational (Representation) and one temporal (Sequence), and its spatial structure lives in the surrounding realization layers (neural hardware below; situation, context, and spatiotemporal NSM-prime above) rather than in the substrate’s primitive set. The ratios vary; the presence of the three primordial aspects does not.

The recurrence reads cartographically rather than as a structural claim. Every substrate the methodology can analyze operates within a reality that has information, time, and space; that the analysis surfaces primitives along those aspects is observation, not discovery. The methodology’s primitives are reorganizations of these primordial aspects under specific substrate constraints. The strict “3+2+1” ratio is the entity system’s signature, not a universal substrate property. Cross-cutting primitives that span multiple primordial aspects appear empirically across the corpus and carry much of each domain’s distinctive structural content; whether they constitute a discrete fourth bucket or are primitives operating on multiple primordial aspects simultaneously is a vocabulary choice that the cartographic stance does not need to settle.

This leaves a deeper question open. The realization chains the methodology produces terminate at physics — but what underlies physics is not something the methodology resolves. Candidate framings include: that information, time, and space are themselves the primordial substrate and physics is one elaborated surface of them; that all three are emergent from a deeper substrate the methodology is not equipped to analyze; that the chain does not terminate and “primordial” is a methodological floor declaration (Step 1c) rather than a structural fact; or that the question is malformed because the methodology’s analytical apparatus may not extend coherently below physics. The exploratory companion on physics as information substrate (see The Structural Methodology Applied to Physics) surfaces these framings without resolving them. The open question is offered as part of the methodology’s exploratory surface rather than as a gap to be closed inside this paper.

The methodology’s posture toward such open questions is the cartographic discipline applied recursively: observations are kept as observations, alternative readings are listed where they are coherent, and forced resolutions are avoided where the available analytical tools do not justify them. A reader who picks up the map the methodology has produced and pushes further is operating in exactly the mode the methodology supports.

9. Empirical Cartography: Trajectory Regimes

The methodology’s most concrete cross-domain work comes from authoring trajectories — sequences of unified manifestations through time — and examining their structural shapes. Trajectories sit at Sc=3 in the scope ladder: a specific manifestation moving through arrangement positions over time, with the methodology supplying the position vocabulary at each snapshot. We have authored ten trajectories across four arrangements; their shapes group into a small number of structural regimes. The grouping is cartography in the sense of the preceding chapter — the trajectories were authored, and the regimes are descriptions of where the resulting shapes sit in a 2D shape space — but the regimes’ separation across independently authored empirically distinct cases is itself informative.

This chapter keeps one illustrative slice of the empirical work: the trajectory-regime cluster. The wider record — cross-arrangement coupling at Sc=4 and Sc=3-sustained, the Sc=2 wall-time calibration architecture, the cross-corpus clustering recipe and its biology spot-check — is relocated to a companion note because it is exploration recorded for transparency rather than material the paper’s argument rests on. We keep the slice here, in the paper, so the reader can see the shape of what the methodology produces at Sc=3 without taking the full record on faith.

9.1. The trajectory corpus and its shape space

The X-genesis family of analytical questions — abiogenesis, ontogenesis, phylogenesis, technogenesis, civilizational evolution — is not a list of separate arrangements but a single use case: a Sc=3 trajectory analysis asked of an underlying arrangement. Abiogenesis is the Sc=3 trajectory through biology’s chemistry → bridge → substrate sub-segment; technogenesis is the same move through the entity arrangement; civilizational evolution through cognition’s cultural ecosystem. Each trajectory is authored as a sequence of unified manifestations, anchored on empirical milestones (developmental stages, evolutionary divergence events, version releases) chosen to capture structural transitions rather than uniform time steps. The current corpus is ten cases: five in biology (abiogenesis, post-LUCA cell evolution, stem and plant-branch phylogenesis), three in cognition (human and chimpanzee cognitive ontogenesis, civilizational), and three in entity (git, github, postgres evolution).

Each chain level in an arrangement plays one of four structural roles — substrate, bridge, surface, or ecosystem. Summing the per-snapshot rank deltas across the levels playing each role gives a four-dimensional role-Δ vector $(\Delta_{\text{substrate}}, \Delta_{\text{bridge}}, \Delta_{\text{surface}}, \Delta_{\text{ecosystem}})$ for the trajectory. Projecting to the substrate-Δ versus ecosystem-Δ plane turns each trajectory into a single point: near the origin is quiescent, far on the substrate axis is high substrate evolution with little ecosystem accumulation, far on the ecosystem axis the reverse.

9.2. The Three-Regime Empirical Cluster

When the ten trajectories above are placed in this shape space, they cluster into three structural regimes:

Per-trajectory role decompositions:

Each regime has N ≥ 3 in the current corpus (with one edge-case trajectory in GENESIS). The cultural-tech regime is the most populated, with four trajectories spanning three technology layers (version control, relational database, civilizational cultural evolution). That postgres-evolution clusters tightly with git-evolution and civilizational-cognitive — despite operating at a different abstraction layer (relational data vs version control vs cultural transmission) — is consistent with the cultural-tech-accumulation signature being a structural property of the trajectory shape rather than an artifact of the domain. Cluster placements are cartography; the robustness of the placement across independently-authored cases is what the chapter is doing.

The centerpiece visualization renders the 2D shape space with regime regions tinted, paired with a per-trajectory role-decomposed bar chart. The two views together show both the cluster structure and the per-trajectory weight distribution.

The three-regime cluster is the slice this chapter keeps in the paper. The companion note carries the rest of the empirical record at the same discipline: a candidate fourth regime (BOUNDED-GENESIS, read off the chimpanzee trajectory’s substrate ceiling at the LA3 composite gate); the substrate–ecosystem disconnect between individual and civilizational cognition; cross-arrangement coupling at Sc=4 (the developer-keypress event) and Sc=3-sustained (the paired-manifestation landscape studies); the Sc=2 wall-time calibration architecture (LUCA-anchored, with three mutually consistent instruments); and the cross-sectional clustering recipe whose transfer from the entity arrangement to biology recovers taxonomic structure the recipe was never shown. The strongest licensed claim among them — that the recipe transfers across independently generated corpora — is carried forward in §Cross-Domain Structural Patterns; the rest stays exploration.

10. Methodology Applied to Itself

The methodology supports two distinct reflexive applications. The first, presented in this chapter, applies the twelve-step procedure to the methodology as a domain: its four layers, their primitives, their dependencies, their core triads. The second, presented in the chapter that follows, applies the methodology to its range — the meta-domain of domains the methodology can analyze — and surfaces empirical attractors that mark where the methodology produces high-value output, partial output, and no operational output.

The first reflexive application produces not a single primitive set but four, one per layer: Layer 1 (Domain Analysis, six primitives), Layer 2 (Graph Construction, five primitives), Layer 3 (Graph Semantics, six primitives), and Layer 4 (Applied Analysis, seven primitives). The four layers connect through a feed relation: Layer 1 produces analyzed domains, Layer 2 connects them, Layer 3 extracts patterns across the connected graph, and Layer 4 applies the structural knowledge to specific situations, with Layer 4’s results feeding back to motivate further Layer 1 analyses. The self-analysis was conducted using the methodology itself; the twenty-four primitives across the four layers passed the same three-test extraction procedure each domain analysis uses.

The self-analysis is reflexive but not circular in the sense established in the methodological-discipline chapter. The methodology applied to itself uses the procedure to analyze the procedure; each layer analyzes different subject matter (domains, inter-domain graphs, patterns across graphs, applied use). The same vocabulary recurs because the procedure produces the vocabulary that fits its own structure. The reflexivity is a structural consequence of the methodology being an instance of the Convergence Domain it discovered (Layers 1–3 build the Space and Constraint; Layer 4 operates the Distribution, Dynamics, Collapse, and Determination), not a hidden circular validation move.

The self-analysis also bears on the cross-domain primitive-count pattern documented later in the paper: substrate domains cluster at six primitives; the methodology’s own Layers 1 and 3 sit at six, and the four-layer total of twenty-four sits within the band a deeper multi-layer analysis would predict. The pattern thus recurs on the methodology itself — not as an additional licensed claim (the analyst is the same), but as a self-consistency check the procedure passes against its own structure.

10.1. The derivation discipline on the self-analysis

The Step 10b derivation discipline (introduced in §The 12-Step Domain Analysis) applies reflexively: every emergent property claimed at compositions within the methodology’s self-analysis should derive from the layer-internal primitives plus their dependencies plus the composition rules. When the stress-test was run across the corpus (summarized in §Cross-Domain Structural Patterns), the four layers of the methodology’s own self-analysis were included. The result:

• Layer 1 (Domain Analysis, 6 primitives) — emergent properties (the 12-step procedure’s outputs: irreducible primitive sets, dependency-coherent sub-lattices, core triads, emergent-property maps) derive cleanly. Most properties derive from the standard Layer-1 vocabulary plus the iteration discipline.
• Layer 2 (Graph Construction, 5 primitives) — emergent properties (typed inter-domain graph, edge compositions, product lattices across realization chains) derive cleanly.
• Layer 3 (Graph Semantics, 6 primitives) — one Plausible: Cv (Convergence) as a primitive is itself difficult to derive without invoking the self-application of Layer 3 to Layer 3. The Plausibility is structural: convergence-as-primitive is the methodology’s self-correcting operation, and analyzing it requires the very operation being analyzed.
• Layer 4 (Applied Analysis, 7 primitives) — one Plausible: Scope (Sc) as a primitive raises a level-relativity question — Sc is partly meta to the other L4 primitives, which is what M2 (R12) was added to handle in the first place. Sc’s level-relativity is the methodology’s tool for handling level-relativity in domain analyses, which has the structural shape of self-reference.

The two Plausible cases are informative: both involve primitives whose function includes operations on the methodology itself (self-correction in L3 Cv; level-relativity in L4 Sc). The methodology’s reflexive structure shows up at exactly the points where the procedure analyzes its own analytical operations. This is the same pattern the methodology’s Convergence-Domain instantiation predicts — the methodology is an instance of the Convergence Domain, and its self-application is the convergence-domain pattern running on the methodology’s outputs.

No primitive in any of the four layers was Failed; no primitive required removal or revision. The self-analysis is internally coherent under the discipline that disciplines other domain analyses.

11. The Methodology’s Range as a Domain in Its Own Right

The paper has so far drawn most of its examples from substrate-style arrangements — biology, the entity system, cognition — and from the realization chain that connects them. This emphasis reflects the project’s primary application area; the procedure itself makes no assumption about whether the domain it analyzes is substrate-style. It has been applied across a wider range than this emphasis makes visible: bridges in the realization chain, physics domains, abstract information substrates, the Convergence Domain at Layer 3, the methodology itself reflexively, and sub-domains nested within other analyses.

The question of how wide that range actually is, and where the methodology stops being applicable, is itself a question the methodology can analyze. This section presents the result of doing so: a reflexive application of the 12-step procedure to the meta- domain “analyzable domains,” producing a primitive set, partial-level decompositions, a dependency DAG, core triads, and empirical attractors that map where the methodology produces high-value output, where it produces partial output, and where it does not apply.

The detailed analysis lives in four research documents in the project’s methodology-strategy notes (the bounding-range, applicability-as-a-domain, review-and-gaps, and validation documents). What follows is a consolidation of their structural findings. We mark this as a current iteration: the 3/3b loop has been exercised twice on the meta-domain, and further iteration may revise the primitive set.

11.1. The meta-domain: eight primitives

Pushing roughly two dozen candidate domains through the procedure (both domains in the existing corpus and stress-test candidates not previously analyzed) surfaces eight primitives that vary across domains and jointly determine where the methodology applies:

• Cm (Compositionality). The degree to which the domain admits decomposition into interacting parts. Cm=0 atomic/holistic; Cm=4 fully compositional with a clean primitive set.
• Cy (Dependency-cycle density). The degree to which the dependency structure forms cycles vs an acyclic partial order. Cy=0 fully acyclic; Cy=4 the cycle is the structure (no useful DAG exists).
• Eg (Empirical groundedness). The count and observability of the domain’s instances. Eg=0 no observable instances; Eg=4 dense instance base.
• Jc (Judgment convergence). The degree to which competent analysts converge on the same primitive set after iteration. Jc=0 deeply contested; Jc=4 fully convergent or definitional.
• Gr (Granularity). Whether primitives admit natural discrete partial-level gradation. Gr=0 purely continuous; Gr=4 definitionally discrete.
• Fs (Function-structure alignment). Whether the domain’s function or value is located in its compositional structure. Fs=0 function entirely non-structural; Fs=4 function entirely structural.
• Rx (Reflexivity). The degree to which the domain changes in response to its own analysis. Rx=0 no feedback; Rx=4 total reflexivity (analysis and domain inseparable, as for the methodology applied to itself).
• Ds (Discovery vs Stipulation). Whether the primitives are empirically discovered or axiomatically stipulated. Ds=0 stipulated; Ds=2 discovered.

Eight primitives is above the typical six-cluster the methodology observes at the substrate level. Two of the eight surfaced during the 3/3b iteration on the meta-domain (Rx via the methodology-applied-to-itself, economics, and AI safety cases; Ds via the genetic-code-vs-category-theory comparison). Whether the set will collapse to seven under further iteration (by combining Rx and Ds into a single “epistemic status” primitive) is an open question; stress-testing shows them diverging across domains, so we keep them separate.

11.2. Dependency structure and filter

Cm is the root primitive: Cy, Gr, and Fs all require non-zero Cm (cyclicity, gradation, and structural function alignment all need compositional structure to operate on). Eg is the other semi- independent root (a domain has instances or doesn’t, regardless of structure). Jc weakly depends on both Cm and Eg.

The conditional partial-level dependencies are: $\mathrm{Dep}(\mathrm{Cm} \geq 1, \mathrm{Cy} > 0)$, $\mathrm{Dep}(\mathrm{Cm} \geq 1, \mathrm{Gr} > 0)$, $\mathrm{Dep}(\mathrm{Cm} \geq 1, \mathrm{Fs} > 0)$.

Coarse-level filter stringency estimates at approximately 25–30% of the unfiltered product space. This is classifier-style rather than substrate-style (substrates filter 12–20%, ecosystems 7–20%, abstract Layer-3 domains 14–19%, Layer 4 itself 29.7%). The methodology’s own range is structurally Layer-4-shaped — a classifier domain over what can be analyzed.

11.3. Core triads

Five core triads emerge, all sharing Cm as the hub:

• {Cm, Cy, Gr}: Structural decomposability. Joint determination of whether the domain admits clean lattice representation.
• {Cm, Eg, Jc}: Empirical grounding. Joint determination of cross-instance recurrence and analyst-convergence stability.
• {Cm, Fs, Jc}: Useful output. Joint determination of whether the methodology’s structural map is functionally relevant.
• {Eg, Jc, Ds}: Licensed-claim path. Joint determination of whether the methodology produces findings (vs cartography).
• {Rx, Eg, Fs}: Output stability over time. Joint determination of how long the methodology’s analysis remains accurate before the domain absorbs it.

The hub-and-anchor structure (Cm as central hub, triads branching through it) is the same shape Layer 4 exhibits. Two reflexive applications of the methodology produce structurally similar results, consistent with the Convergence-Domain reading that the methodology is itself a convergence process.

11.4. Nine empirical attractors map where the methodology applies

Positioning roughly thirty domains across the meta-lattice surfaces nine attractor zones — the archetypal structures the methodology encounters, graduating from where it does its best work to where it produces nothing operational. The high-value zone (attractor A: high compositionality, acyclic, grounded, structural function) is where the procedure earns its keep: substrate domains like the entity system, biology, cognition, the Convergence Domain, the genetic code, and the methodology applied to itself. There it produces the full apparatus — primitive sets, dependency DAGs, coherent sub-lattices, core triads, phase thresholds, design opportunities — and the cross-domain patterns of the next chapter emerge from many such analyses side by side. The licensed-claim path is open here when non-circular external recovery is in evidence.

The other zones grade the output down. Cyclic-rich domains (attractor B-high: ecosystem dynamics, deep markets, climate, coevolution) get a substantive lattice with cycles modelled explicitly, but resolving the cycles’ equilibria belongs to dynamical-systems analysis, not here — this zone covers most of the “interesting” cyclic domains in science. Cyclic-thin domains (governance, macro-economics, law) get a primitive set that reflects the analyst’s cycle-breaking convention more than the domain. The methodology stops applying meaningfully in three zones: where there are no observable instances (counterfactual histories, fictional worlds — the cross-instance recurrence test has nothing to run on), where primitives are stipulated rather than discovered (pure mathematics — the output is axiom transcription), and where competent analysts produce different equally-defensible primitive sets (ethics, contested politics — the methodology cannot adjudicate from inside itself). Function-mismatch (music, art at the experiential layer), granularity-bottleneck (fluid dynamics, continuous PDEs), and mature-craft (cooking, established practice) zones get a correct structural map that misses what the domain is for, needs continuous mathematics for the mechanism, or merely restates what practitioners already know. The full per-attractor profiles, inhabitant lists, and the capability/incapability breakdown live in the companion note, along with the meta-lattice’s empty regions and the chapter’s open avenues.

One structural limit recurs across every zone and the methodology cannot remove it from inside itself: it is not self-validating against framework-level error. Its discipline catches internal inconsistency, filter anomalies, circular validation, and visualization artifacts, but not the possibility that the whole analytical frame is wrong in a way internal consistency cannot detect — the failure mode of phrenology, caloric theory, and Galen’s humours, each internally consistent for a long time before displacement. The licensed-claim path is therefore conditional on the framework being approximately correct. (This limit is developed in §Methodological Discipline and revisited in the conclusion.)

12. Cross-Domain Structural Patterns

The 12-step procedure has been applied independently to roughly twenty domains. Each analysis was authored from the literature of its own domain — biology from organismal biology and biochemistry, cognition from neuroscience and developmental psychology, physics from quantum gravity and condensed matter, computing from systems architecture and programming-language theory, and so on. The analyses do not borrow primitives from each other; the only thing they share is the procedure that produced them.

When the resulting primitive sets, dependency DAGs, filter stringencies, and core triads are placed next to each other, several shapes recur strongly enough to deserve listing. These are the cross-domain patterns the methodology produces. They are not predictions the methodology guarantees in advance; they are what we find when independent analyses are compared. Each one is a candidate licensed claim, in the sense established earlier: a pattern recovered from inputs (the per-domain analyses) whose generation was not targeted at recovering the pattern.

12.1. Primitive counts cluster near six

Across the twenty-plus analyses, primitive counts span 4–12 with a strong mode near six. Substrate domains in particular cluster tightly: the entity system has six primitives, the biology substrate has six, the cognition substrate has six, the convergence domain has six, the genetic code (as a sub-domain) has six, the abstract information substrate has six, the QG domain has six, the Planck information substrate has six, the abiogenesis bridge has six. Two domains run slightly higher: surface domains (organism architecture at nine, application architecture at twelve) and ecosystem domains (digital ecosystem at nine, cultural ecosystem at nine). Bridge domains return to six. Layer vocabularies sit at five to seven.

The clustering is not imposed. The three-test extraction procedure plus the 3/3b iteration loop tends to converge on a particular resolution: the primitive set that survives is the one for which partial-level decompositions are stable, dependencies are clean, and cross-instance recurrence is strong. The convergence is empirical, not arithmetical. Several analyses started with candidate sets of seven or eight and reduced through the 3/3b loop; several started with five and grew through the same loop. The terminal count’s clustering near six is what the procedure produces, not a target it aims for.

12.2. Filter stringency clusters by domain type

The fraction of the product space that survives the dependency filter falls in characteristic ranges by domain type. Substrates filter tightly (12–20%); surfaces filter loosely (25–40%); ecosystems vary widely (7–20%) depending on whether ecosystem primitives have mutual constraints or operate independently. Abstract Layer-3 domains sit in the substrate range (14–19%) because they preserve the substrate-like dependency tightness that recurs across instantiations.

This pattern has a structural reading. Substrates sit at the bottom of realization chains and accumulate downward constraints: every layer above them must be compatible, so their dependency DAGs are dense. Ecosystems sit at the top and accumulate upward freedom; their primitives often operate independently of each other, so their DAGs are sparser. Surfaces fall between. The empirical clustering is consistent with this structural reading, and the few analyses that fall outside expected ranges have been re-examined and (in some cases) revised on grounds independent of the filter percentage.

12.3. Heavy-pair ratio is approximately one-half

Pair-load classification (heavy / medium / light / negligible) is the most analyst-judgment-heavy step in the procedure. The criteria are qualitative, and the analyses are authored from distinct domain literatures. Despite the qualitative criteria, the fraction of pairs classified as heavy is stable across the corpus: the heavy-pair ratio falls in 40–53%, with most domains near 47%. The convergence domain, the entity system, the biology substrate, and the cognition substrate all land within a few percent of each other.

The stability has two possible readings. The first: the methodology’s pair-load criteria are picking up something real about how primitives interact, and the threshold at which a pair becomes load-bearing is determined more by the structure of the domain than by the analyst’s threshold for “heavy.” The second: the corpus is single-analyst, so the stability could reflect the analyst’s consistent threshold calibration rather than a property of the domains. The two readings cannot be distinguished from inside the current corpus; cross-analyst validation (independent analysts re-running the analyses) would be required to discriminate them, and is one of the open avenues discussed in the range chapter.

12.4. Every analyzed domain has a core triad

A core triad — three primitives all pairwise heavy and load-bearing in combination — exists in every domain the procedure has produced. The function of the core triad varies by domain type: substrate domains’ core triads typically organize information flow ($\{E,I,T\}$ in the entity system, $\{G,R,\mathrm{Mem}\}$ in biology, the encoding-evaluator-mechanism cluster in cognition); surface domains’ core triads organize functional integration; ecosystem domains’ core triads organize resource flow.

Some domains have one core triad; some have several. The convergence domain has three overlapping triads sharing the Ds primitive (Landscape, Directed Evolution, Information Gain), reflecting its role as a Layer-3 abstraction that recurs across instances. Layer 4 has three triads branching from the Mn-Cx anchor (Analytical Frame, Strategic Positioning, Trajectory Planning). No analyzed domain has been found without at least one core triad. We have not searched for counter-examples systematically; finding one would be informative.

12.5. SSA topology recurs across three substrate arrangements

The SSA topology is one of the Layer-3 abstractions the methodology has identified so far. It is the topology that recurs across substrate-style arrangements; the Convergence Domain is the primitive-set abstraction that recurs across convergence-under- constraint processes. Other Layer-3 patterns may exist (see the preceding chapter’s discussion of candidate patterns); the SSA and the Convergence Domain are the two that have been pushed to stable characterization. The recurrence reported here is of the SSA specifically and should not be read as a claim about all Layer-3 abstractions or about all domains the methodology analyzes.

Three independent substrate arrangements — biology (genome / ribosome / organism / ecosystem), the entity system (E+I+T / dispatch / extensions / app-architecture / digital-ecosystem), cognition (representations / symbolic processing / cognitive architecture / cultural ecosystem) — have been analyzed at Layer 2 (graph construction) and Layer 3 (graph semantics). The resulting inter-domain graphs share a topology: the same seven-role structure {Encoding, Evaluator, Mechanism, Surface, Context, Community, Selection} with the same cycle structure connecting them. We refer to this recurring topology as the Situated Substrate Architecture (SSA).

The SSA’s recurrence is one of the methodology’s stronger cross-domain patterns. Three arrangements analyzed from three different literatures, with three different primitive sets at the substrate, with three different bridge structures, produce the same seven-role topology with the same cycle structure when their Layer-2 graphs are placed side by side. The roles are distinct primitives in each arrangement (the entity-system’s evaluator is its dispatch extension, biology’s evaluator is the ribosome, cognition’s evaluator is symbolic processing); what recurs is the graph topology, not the primitive identity.

Whether the SSA topology recurs in further substrate arrangements beyond the three currently analyzed is an open question. A fourth arrangement that the methodology has begun analyzing is the mathematical / abstract substrate; preliminary work suggests the SSA topology does fit, but the analysis is not at the depth of the three established arrangements. A fifth direction — physical-realization substrate at the hardware level — has been sketched. Both extensions would tighten the SSA pattern’s empirical base; neither has been completed at the depth required to license a stronger claim than the three-arrangement convergence we currently have.

12.6. Pattern transfer at the recipe level

A separate cross-domain pattern is observed at the recipe level rather than at the primitive-set level. The cross-sectional cartographic recipe (signature families, lens stack, anchor authoring, meta-stability aggregation; detailed in the empirical-cartography companion note) was developed against the entity arrangement and ran on the biology arrangement without modification, producing clusters whose labels correspond to taxonomic categories the recipe was never exposed to during clustering. The same recipe applied to the methodology arrangement and the entity arrangement at the landscape level produced the same three-zone clustering structure with diagnostically different correlation mechanisms.

Pattern transfer at the recipe level is a stronger licensed claim than pattern transfer at the primitive-set level: the primitive sets across domains have similar shapes, but the primitive sets are not identical, so the cross-domain pattern is a recurrence of shape, not of object. The recipe is the same object across arrangements; its producing similar structural output across them is closer to a licensed claim about the methodology’s cross-domain applicability.

12.7. Convergence with an independently-derived dimensional analysis

A different cross-domain check applies to the type-systems and capability-systems domain analyses recorded in Dimensional Completeness. The dimensional framework recorded there (seven type dimensions, seven capability dimensions across sixteen surveyed systems) was developed by direct analytical work on the two design spaces, without using the methodology’s full 12-step procedure. When the matured methodology was later applied back to type systems and to capability systems as domains in their own right — running the 3/3b iteration loop, the dependency filter, the load classification, and the core-triad identification — it recovered eight type primitives and eight capability primitives whose structural roles correspond to the original dimensional axes. The convergence is not exact in count (eight rather than seven on each side, with the additional primitives filling roles the original analysis had folded together), but the core triads, the hub-primitive identifications, and the load classifications match across the two derivations.

This is a licensed claim of a specific kind: the methodology recovers a primitive set arrived at by independent analytical work on the same domain. The independent work used analyst judgment plus literature survey but not the 12-step procedure; the methodology used the 12-step procedure without consulting the prior dimensional analysis during the primitive-extraction phase. The convergence is not proof-of-correctness for either framework, but it is evidence that the methodology recovers something that survives a different analytical route. Dimensional Completeness records the reconciliation in detail.

12.8. Bridges share a structural shape

Beyond the primitive-set patterns above, the methodology’s bridge analyses (Step 4 specifies bridges as their own domains; the realization chain develops the bridges between substrate-style domains) exhibit a recurring three-part structure when looked at across the corpus:

1. ~10-12 bridge primitives, regardless of the substrate domains the bridge connects. Across biology-to-organism (~12), neural-to- cognitive (~12), entity-system substrate-bridge extensions (11), cognitive-to-semantic (~12), and semantic-to-cultural (~12), the count clusters tightly in the 10-12 band. This is empirical observation, not a methodological prediction.
2. Hub primitive at the channel operation. Every analyzed bridge has one primitive that participates in more heavy pairs than any other; in every case examined, that primitive names the bridge’s channel: Lexicalization (cognitive-to-semantic), Circulation (semantic-to-cultural), Cell (biology-to-organism), Type (entity-system substrate-bridge set). The bridge’s hub is what carries the bridge’s substantive operation.
3. Two core triads with complementary functions. Every fully analyzed bridge has one production-side core triad (how new units enter the bridge) and one authority-side core triad (how units get fixed as the bridge’s stable output). The cognitive-to- semantic bridge has {Lexicalization, Conventionalization, Universalization} on the production side and {Combinability, Decomposability, Translation} on the authority/discipline side. The semantic-to-cultural bridge has {Externalization, Inscription, Circulation} on the production side and {Canonization, Norm fixation, Curation} on the authority side. The same two-triad structure shows up in the entity-system substrate-bridge analysis and in earlier biology bridges. The pattern is consistent enough to deserve naming.

These observations sit at candidate-licensed-claim status: each is a recurrence across multiple independently-analyzed bridges. The explanation is open. One hypothesis: bridges sit between substrates that operate at distinct levels of description, and the bridge’s job is both to channel outputs from one level into the next (the production-side function) and to stabilize the channelled units into a form the next level can consume (the authority-side function). The two-triad structure may reflect the structural necessity of both functions for any bridge to operate.

12.9. Co-evolutionary primitive-pair spirals

Within several bridges and a few substrate domains, certain primitive pairs exhibit a co-evolutionary spiral pattern — the two primitives advance together through iteration, neither preceding the other, each enabling further advancement in the other. The canonical example from the cognitive-to-semantic bridge: Crystallization (a prime’s structural irreducibility) and Universalization (a prime’s cross-linguistic recurrence) advance together. Crystallization happens through cross-linguistic testing (Universalization is the test). Universalization stabilizes when the prime resists reduction in every tested language (Crystallization is the convergence condition). The pair is empirically tightly correlated but the primitives remain conceptually distinct.

The pattern parallels biology’s autocatalytic spirals at sub-level decomposition (the ribosome-protein bootstrap, the genetic-code-and- reading-machinery co-evolution). It is a structural pattern that recurs in different domains: a pair of primitives that together do what neither does alone, where the joint operation also progresses each primitive’s partial level. Co- evolutionary spirals are flagged as a candidate Layer-3 abstraction pattern, worth checking across other bridges and substrates as the corpus extends.

12.10. Cross-corpus M3 derivation stress-test (summary)

The derivation discipline (Step 10b) was stress-tested across the analyzed corpus at the point of this paper’s revision — roughly 150 emergent-property derivations across 12+ domains spanning substrate / surface / ecosystem / bridge / Layer-3 abstraction / methodology self-analysis. Distribution of outcomes:

No primitive set required revision; no emergent-property claim required removal. The Plausible and Ambiguous outcomes cluster in three structural locations: ecosystem domains (where loose filter predicts joint-regime derivations); Layer-3 abstractions (where derivations admit appropriately looser rigor than substrate-level derivations, validating the level-relativity discipline); and domains where the substrate-level primitive set may benefit from sub-level decomposition (chemistry’s far-from-equilibrium dynamics is the canonical case, flagged as Ambiguous and noted as a candidate for sub-level extraction of energy-flow / boundary / feedback sub-primitives). The conservative-on-removal discipline proved empirically valuable — it prevented reflexive removal of valid emergent properties in cases where clean structural derivation was hard.

12.11. What these patterns are and are not

The patterns above are what we have found, not what the methodology guarantees. Each is candidate evidence that the procedure picks up something real about the domains it analyzes, and each can be tested by extending the analysis to further domains. We treat them as structural observations across the analyzed corpus rather than as universal claims. A pattern that fails to recur in a new domain is informative; a pattern that recurs strengthens the case but does not make it definitive. The discipline established earlier — cartography vs licensed claim, non-circular external recovery vs internal description — applies to these patterns as much as to any specific result in the empirical cartography.

Several extensions would tighten the patterns’ base: more substrate arrangements for the SSA topology, more bridge analyses, more sub-level decompositions, more cross-domain applications of the cartographic recipe. The patterns we have are sufficient to motivate the methodology as worth applying further; they are not yet sufficient to close any of the structural questions the methodology raises.

13. Computational Implementation and Reproducibility

The methodology’s discrete, finite structures make the analyses tractable to compute exactly, and the implementation has accumulated across several components: a single-source JSON data model for arrangements and manifestations; a lattice engine producing exact filter stringencies and walk counts by enumeration (no Monte Carlo — at the methodology’s sizes, at most $6^8 \approx 1.7$ million positions, brute force runs in under a second); an inference layer over the Bayesian-network interpretation; a pluggable metric framework for cross-instance comparison; a multi-agent dynamic engine for trajectory and population analysis at Sc=2/Sc=3; and a forward-looking Lean 4 formalization track. None of it is load-bearing for the paper’s structural claims — the engine in particular is one instrument among several, and its outputs are characterizations of the engine under analyst-chosen parameters, not findings about the domains it models.

The discipline that matters for a reader is reproducibility. All Python runs go through a container-isolated environment built from hard-pinned dependencies (name==X.Y.Z) with a committed, hashed lock file; the base image is pinned by SHA digest, the resolver by SHA-256, dependency resolution uses a cutoff at least thirty days in the past, and container runs use --network=none. The analyses, lattice computations, engine runs, and figure generation are exactly reproducible from the committed corpus and pinned environment alone, verified by byte-identical reruns at each substantive increment. The component-by-component detail — data architecture, lattice computation, Bayesian inference, the metric framework, the dynamic-engine pipeline, and the Lean 4 track — lives in the companion note.

14. Related Work

The methodology draws on, and connects to, several established literatures. We list the closest connections briefly, organized by which part of the methodology they touch. The first subsection places the methodology in its broader methodological lineage; subsequent subsections cover the specific mathematical and domain-literature connections.

14.1. Methodological context: the analysis/synthesis tradition

The construct-and-reduce cycle described in §The 12-Step Domain Analysis is a modern, iterated instance of one of the oldest methodological pairs in Western inquiry: the Method of Analysis and Synthesis, with a 2,000-year lineage running through Greek mathematics, early modern science, German idealism, and 20th-century philosophy of language. Placing the methodology in this lineage clarifies what it inherits, what it adds, and what it does that the classical tradition leaves implicit.

The pair originates with Greek geometry — analusis (“loosening up”) and synthesis (“putting together”). Pappus codified the discipline: analysis assumes a desired conclusion is true and works backward to known axioms; synthesis is the reverse, starting from axioms to construct the proof. Aristotle applied the same pair to logic: analysis is the resolution of a compound into its fundamental, primary principles — with the methodologically important caveat that “fundamental principles” are relative to the level of description being analyzed. (This level-relativity is what the methodology’s R12 makes explicit.) Descartes’ Discourse on the Method (1637) made decomposition a normative rule: “divide each of the difficulties under examination into as many parts as possible.” Newton’s Opticks Query 31 (1704) is the canonical statement of analysis-as-empirical-method: analysis (experiments, observations, induction) must precede synthesis (assuming the discovered causes as principles and deducing the phenomena from them) — and, critically for the methodology presented here, Newton was explicit that the procedure iterates. Synthesis’s predictions get checked against new experiments, which feed back into further analysis. The iterated construct-and-reduce cycle the methodology runs is Newton’s analysis-then-synthesis with iteration made the primary mode rather than a refinement after the main pass.

Kant moved the analysis/synthesis pair from method into the structure of cognition itself: analytic judgments clarify via decomposition (the predicate is already in the subject); synthetic judgments combine distinct concepts into a new whole. The methodology uses this distinction implicitly in the cartography vs licensed-claim discipline (analytic = description of inputs; synthetic = recovery combining independent inputs into a non-trivial joint result). Hegel argued that static decomposition cannot capture dynamic truths and reframed synthesis through the dialectic: a thesis generates its internal contradiction (antithesis), and the resolution is elevated into a higher unified truth (synthesis, or Aufhebung, which both negates and preserves the original distinction). The methodology’s construct-and-reduce cycle has this dialectical character: each reductive pass exposes a contradiction or redundancy in the build, and the resolution is a higher unified structure expressed at a higher level. The pattern is the Hegelian dialectic applied to engineering design and analytical method rather than to consciousness or history.

The pair also appears in other vocabularies across disciplines without changing its substance: resolution / composition in classical philosophy, reduction / construction in logic and epistemology, induction / deduction in scientific method, decomposition / recomposition in chemistry, anatomization / integration in cognitive science. The methodology’s “reduce / construct” is the same pair, in the vocabulary of its own domain.

What the methodology adds to this tradition: (i) Iteration is first-class, not just sequencing. Newton said analysis precedes synthesis; the methodology says they alternate until both stop producing changes. (ii) Level-relativity is explicit through scope (Sc=0 universal → Sc=4 specific event) and through the recursive partial-level decomposition that terminates at physics. Aristotle’s “fundamental principles” are level-relative; the methodology operationalizes the relativity. (iii) Convergence-as-stopping-rule via the bilateral fixed-point criterion (R13). The classical tradition leaves open when analysis should stop; the methodology stops when further reductions stop appearing and no additions are recoverable from existing primitives — a structural rather than foundationalist stopping rule. (iv) Dependency structure as first-class (Step 4). The classical tradition treats primitives as independent atoms; the methodology requires explicit dependency specification.

14.2. Modern reductive programs

The methodology sits in a lineage of explicit reductive programs:

• Logicism (Frege, Russell-Whitehead Principia Mathematica) — reduce mathematics to logical primitives.
• Bourbaki (Éléments de mathématique) — reduce mathematics to a small set of structural primitives.
• Carnap’s Aufbau — reduce all knowledge to elementary experiences via a constructional hierarchy.
• Logical positivism / the Vienna Circle — reduce meaningful statements to observation statements plus logical structure. Did not survive Quine’s “Two Dogmas of Empiricism” and the historical philosophy-of-science turn (Kuhn, Lakatos).
• Wierzbicka’s Natural Semantic Metalanguage (NSM) — reduce all human meaning to ~65 cross-linguistically stable semantic primes (developed in the next subsection because of its particularly close structural parallel to the methodology).
• Modern systems thinking — Wardley mapping, Cynefin, TRIZ (40 inventive principles) all apply analysis/synthesis to engineering and management.

The lineage matters because it tells the methodology what to expect: reductive programs can succeed at producing useful structure (Bourbaki, NSM, the particle-physics Standard Model) and can fail by overreaching (logicism, naïve positivism). The methodology’s cartography vs licensed-claim discipline is its safeguard against overreach: every output is marked as either description (cartography) or recovery from independent inputs (licensed claim), and the distinction is enforced throughout.

14.3. NSM as the closest structural analogue

Wierzbicka’s Natural Semantic Metalanguage program is the most fully developed empirical primitive-decomposition project in the humanities and the closest structural analogue to the methodology in adjacent literature. It is worth describing in some detail because the parallels and the differences are both informative.

The NSM claim. There exists a small set of semantic primes — concepts like SOMEONE, GOOD, BEFORE — that (a) appear as lexical items in every studied human language, (b) cannot be defined in terms of other primes without circularity, and (c) are sufficient to paraphrase any other concept in any language. The current set is approximately 65 primes (Wierzbicka, Semantics: Primes and Universals, 1996; Goddard, Semantic Analysis, 2011), empirically derived through decades of cross-linguistic testing.

The NSM discipline. An NSM definition is a paraphrase of a concept using only prime English (NSM-English) words. If the paraphrase uses a non-prime word, the definition has failed — that word has to be paraphrased further until only primes remain. Once a definition is in NSM-English, it is translated word-for-word into other languages; if the translation produces natural-sounding sentences in the target language, the definition is cross- linguistically valid. If it produces awkwardness, either the definition or the primes inventory needs revision. The discipline is empirical and iterative.

The seven parallels. NSM and this methodology share more structural commitments than any other adjacent project we have found:

(i) Empirically determined primitive set, not a priori. NSM discovered its ~65 primes through iterative testing; the methodology discovers its per-domain primitive sets through the 3/3b iteration loop.

(ii) Iterative discipline. NSM definitions get revised when they fail in new languages; the methodology’s primitive sets get revised when cross-domain application reveals gaps.

(iii) Small primitive set produces large functional space. ~65 NSM primes paraphrase all human meaning; ~6 methodology primes per substrate domain produce the entire combinatorial space for that domain.

(iv) Reductive bar is strict. NSM cannot use any non-prime in a definition; the methodology’s three-test criterion is analogous.

(v) Cross-instance test as validation. NSM tested across all human languages; the methodology tested across all analyzed instances of a domain type.

(vi) Level-relativity. NSM’s primes are level-relative — fundamental for natural-language semantics, not for formal logic. The methodology’s primes are level-relative through scope and the recursive partial-level decomposition.

(vii) Reflexive applicability. NSM’s primes are themselves defined using NSM English; the methodology applied to itself produces 24 primitives across 4 layers analyzed using the methodology itself.

Where the projects differ. NSM operates on a single domain (human meaning); the methodology operates on an open class of domains. NSM treats its primes as atomic semantic units; the methodology treats primitives as nodes in a dependency graph with combinatorial behavior at every arity. NSM has no analogue of the dependency structure, multi-arity composition analysis, cross-domain pattern extraction, or cartography vs licensed-claim discipline that the methodology provides.

What NSM has that the methodology adopted. The Step 10b derivation discipline is structurally the methodology’s analogue of NSM’s paraphrase test. The four-level outcome spectrum (Clean / Plausible / Ambiguous / Failed) and the conservative-on-removal policy were added in part on the basis of how NSM handles its own empirical iteration (NSM does not purge a concept just because the paraphrase is hard; it flags the concept for further work).

A finding for NSM. When the methodology was applied back to NSM as a domain, the analysis recovered a structurally coherent picture at three resolutions of the same domain: ~7 primitives at the substrate level {Rf, Mn, Ac, St, Sp, Md, Ev}, ~16 categories at intermediate resolution (Wierzbicka’s organizational grouping), and ~65 primes at fine resolution (NSM’s published inventory). The three-resolution view is what the methodology’s scale-invariance predicts. NSM’s literature treats the ~65 primes as primitive; the methodology suggests there is a coarser-resolution layer at ~7 primitives that the NSM tradition has not surfaced. Whether Wierzbicka and Goddard would accept the coarser-resolution reduction is an open question worth pursuing if the methodology’s reading of NSM is ever published as a reverse contribution.

A separate bridge analysis (cognitive-to-semantic bridge, described in §The Realization Chain) identified the bridge primitives that sit between cognitive substrate and the semantic-content surface NSM describes. One specific finding: NSM’s central claim — that certain concepts have prime status — is the joint output of four bridge primitives (Conventionalization + Universalization + Crystallization + Decomposability-failure). NSM’s empirical discipline implicitly runs this conjunction; the methodology names it explicitly. The finding is offered as a reverse contribution to NSM’s tradition, not as a claim NSM needs the methodology’s apparatus.

14.4. Lattice theory, formal concept analysis, Bayesian networks

The mathematical objects underlying the methodology are standard. Product lattices, dependency-filtered sub-lattices, and Hasse diagrams come from lattice theory in the sense of Davey and Priestley (Davey and Priestley 2002). The factor-graph representation of the coherent sub-lattice and the forward-backward algorithm are standard in Bayesian networks (Pearl 1988; Koller and Friedman 2009). Formal concept analysis (Ganter and Wille 1999) supplies the dual extension-intension structure when manifestations and primitives are treated as a binary relation. The methodology’s contribution is not new lattice machinery but a discipline for which lattices to construct from a domain analysis, and what structural properties of those lattices report something stable about the domain.

14.5. Convergence-domain instances in established literature

Each of the convergence domain’s confirmed instances corresponds to a mature literature in its own field. Quantum measurement and the collapse postulate are the subject of decoherence theory and many-worlds interpretation (Zurek 2003; Schlosshauer 2007). Bayesian inference as iterative belief updating is treated formally in (Cox 1946; Jaynes 2003). Biological fixation by selection and drift is foundational population genetics (Fisher 1930; Wright 1931; Kimura 1962). Lattice walks and their absorbing-state dynamics appear in combinatorial probability (Feller 1968). Market lock-in and path-dependence are treated in economic theory (David 1985; Arthur 1989). The convergence domain claims structural recurrence across these instances; it does not claim to add new results within any of them.

14.6. The realization chain and major transitions

The realization chain’s structural shape — substrate domains connected by bridges that progressively open an evaluation-feedback distance — corresponds at high level to the major transitions framework in evolutionary biology (Maynard Smith and Szathmáry 1995) and to the structural-discontinuity framings in comparative cognition (Deacon 1997; Tomasello 1999; Penn et al. 2008) and in the history of science (Kuhn 1962; Price 1963). The methodology’s contribution at this level is the unification of the chain’s variable (the evaluation-feedback distance) across levels, not a new account of any individual transition.

The bottom of the realization chain — the physics substrate — connects to the holographic principle [Bekenstein (1973); ’t-hooft-1993; Susskind (1995)] in its claim that information scales with area rather than volume, and to non-commutative geometry (Connes 1994) in the spectral-triple proposal for the substrate-level information structure. These connections are structural alignments, not endorsements of any particular physical theory.

14.7. Architecture comparison and convergent design in computing

The methodology’s application to information systems draws on several recent comparison-oriented designs: the syndicated actor model (Garnock-Jones 2022), tree calculus (Jay 2021), and the Plan 9 / Inferno line of operating-system research (Pike et al. 1995; Dorward et al. 1997). The convergence patterns this paper notes (terminology simplicity, partial-primitive scoring, walls vs fences) are developed at greater length in a companion paper on convergent evolution of information systems. The methodology landscape study (thirteen strategic-analysis methodologies positioned by analytical depth and cultural adoption) connects to the systems-thinking literature (Checkland 1981) and to recent landscape-mapping work in management practice (Wardley 2021).

14.8. Methods for structural cross-domain analysis

Topological and algebraic methods for cross-domain comparison have a substantial literature. Topological data analysis applies persistent homology to point clouds derived from data (Carlsson 2009; Edelsbrunner and Harer 2010); we have not used it, but the partial-level filtration on the lattice has a natural persistent-homology reading we have not explored. Category theory in cognitive and structural modelling appears in (Lawvere 2003; Spivak 2014). Combinatorial species (Joyal 1981; Bergeron et al. 1998) provide a different formalism for enumeration over labeled structures that overlaps partially with our walk-counting work. These are structurally adjacent frameworks; relating them to the methodology rigorously is an open direction.

14.9. What we are not doing

The methodology is not a foundationalist account of structure in nature. It is not a category-theoretic foundation; it is not an information-theoretic foundation; it is not a complete formal system. It does not claim the primitives it identifies are real features of the world independent of analytical purpose. The patterns it produces across domains are structural observations across an analyzed corpus, not theorems. The discipline established in the methodological-discipline chapter is what keeps the methodology honest about what it can and cannot claim.

15. Conclusion

This paper has presented a structural methodology for analyzing information system domains, together with the cross-domain patterns that have emerged from applying it across roughly twenty domains and the exploratory work that has accumulated alongside it.

The methodology’s four-layer architecture — domain analysis, graph construction, graph semantics, applied analysis — is the load-bearing core of the contribution. The twelve-step domain analysis procedure, with its 3/3b iteration loop and partial-level decomposition, is the methodology’s working unit. The product lattice and its dependency-filtered coherent sub-lattice are the methodology’s analytical object. The four-layer architecture positions per-domain analysis within a wider structure: each domain analyzed by Layer 1 is connected by Layer 2 into a graph, patterns across the graph are extracted at Layer 3, and structural knowledge is applied to concrete situations at Layer 4 through a scope ladder from universal to event-specific.

Several cross-domain patterns have emerged from independent application of the methodology across many domains: primitive counts cluster near six for substrate domains, filter stringencies cluster by domain type, the heavy-pair ratio is approximately one-half across domains, every analyzed domain has at least one core triad, and three independent substrate arrangements share a seven-role graph topology (the Situated Substrate Architecture). Analyzed bridges share a recurring three-part structure: ~10-12 bridge primitives, a hub primitive at the bridge’s channel operation, and a two-core-triad structure with one production-side triangle and one authority-side triangle. Several primitive pairs exhibit co-evolutionary spirals in which the two primitives advance together through iteration. These patterns are structural observations across the analyzed corpus, not universal guarantees; each can be tested by extending the analysis to further domains.

The derivation discipline (Step 10b) operates internally to the methodology and complements the cartography vs licensed-claim discipline that handles its outputs against the world. The four-level outcome spectrum (Clean / Plausible / Ambiguous / Failed) plus the conservative-on-removal policy proved empirically valuable when stress-tested across ~150 emergent-property derivations spanning the analyzed corpus: ~83% Clean, ~13% Plausible, ~2% Ambiguous, 0% Failed. No primitive set required revision; no emergent property required removal. The remaining open structural questions surfaced in the test (chemistry’s far-from-equilibrium dynamics; the appropriately looser rigor of Layer-3-abstraction derivations; ecosystem-domain joint-regime patterns) are tracked as future sub-level decomposition candidates rather than as primitive-set revisions.

The methodology’s exploratory extensions have produced additional material: a cross-sectional cartographic recipe whose recipe-level transfer between the entity arrangement and the biology arrangement is the strongest licensed claim about cross-domain applicability we have so far; a calibration architecture with three operationally independent instruments that are mutually consistent across roughly twenty-two orders of magnitude in per-event probability and thirty orders of magnitude in population size; a multi-agent dynamic engine supporting trajectory and population analysis as one instrument among several; and a methodological-discipline distinction between cartography (descriptions the methodology produces from its inputs) and licensed claim (non-circular recovery whose evidence is independent of the analysis). These extensions are recorded as exploration alongside the four-layer core rather than as central claims; the paper keeps one illustrative slice of each and relocates their full record to the companion notes.

The methodology is still under active development. Several avenues remain open. The Bayesian-network interpretation is mathematically clean and computationally tractable but under-exercised; running it at fine resolution with explicit mutual-information computations across bridge edges would tighten the cross-arrangement coupling work. The realization chain’s bridges between physics, chemistry, biology, cognition, and computing have been analyzed at varying depth; deeper analyses of the less-developed bridges would strengthen the chain’s pattern. The SSA topology has been confirmed across three substrate arrangements; further arrangements would tighten its empirical base.

The reflexive application of the methodology to its own range, described in the chapter on the methodology’s range as a domain, produces an eight-primitive meta-domain with nine empirical attractors. Two of those primitives — reflexivity (the degree to which a domain changes in response to its own analysis) and discovery vs stipulation (whether the primitives are empirically discovered or axiomatically stipulated) — surfaced from iterating on the meta-domain itself and are flagged here as current-iteration outputs. Whether the eight-primitive set will collapse to seven under further iteration, or extend with additional primitives as more domains are pushed through the procedure, is an open question the methodology can ask but only further application can answer.

Whether additional Layer-3 abstractions exist beyond the SSA and the Convergence Domain is an open question: candidate patterns the corpus suggests (cyclic constitution as a domain in its own right, crystallization, evaluation-feedback distance opening, substrate- vs-architecture distinction, function-substrate mismatch) have not yet been pushed through the full 12-step procedure to stable primitive sets. Each is a candidate Layer-3 abstraction awaiting analysis.

The methodology’s scope is broader than its primary application area; the substrate-style arrangements are one zone of its application, but it has also been applied to physics, mathematics, abstract domains, and to itself. The nine-attractor meta-domain map graduates where the methodology produces high-value output, partial output, and no output. Several domain families remain unexplored at the full 12-step depth — governance, language, UI/UX as a domain in its own right, deeper mathematical foundations, the cyclic-rich domains at attractor B-high (ecosystem dynamics, brain population dynamics, deep markets, climate, coevolution) — and each is a candidate avenue for extending the methodology’s range.

A structural caveat the methodology cannot remove from inside itself: it is not self-validating against framework-level error. The discipline catches internal inconsistency, filter anomalies, circular validation, and visualization artifacts; it does not catch the possibility that the methodology’s overall analytical frame is fundamentally wrong in a way internal consistency cannot detect. Historical analogues exist where structural analytical frameworks (phrenology, caloric theory, Lamarckian inheritance, Galen’s humoral theory) passed all the discipline of their time before being displaced by experimental crucial-tests, better instrumentation, or theoretical advances that subsumed them as special cases. The licensed-claim path the methodology offers is conditional on the framework being approximately correct, and the methodology has no internal mechanism that would detect its own displacement.

The Lean 4 formalization track has begun but is not load-bearing for any claim in this paper; whether the formalization layer adds analytical power or is primarily a verification check is unresolved.

We invite disproof. The methodology produces falsifiable structural claims: domains have irreducible primitive sets of bounded size, dependency filters fall in characteristic ranges by domain type, every domain has a core triad, the SSA topology recurs across substrate arrangements, the realization chain widens the evaluation-feedback distance monotonically. Each is testable. A domain whose primitive set fails to stabilize under the 3/3b iteration loop falsifies the bounded-size claim. A substrate domain whose filter stringency falls outside the 12–20% range, on first careful analysis without back-fitting, falsifies the filter-range claim. A substrate arrangement whose Layer-2 graph differs structurally from the seven-role SSA topology falsifies the SSA-invariance claim. A bridge in the realization chain where the evaluation-feedback distance does not widen falsifies the chain’s monotonicity claim. The methodology’s discipline is meant to make such tests informative: a failure is a failure, not a special case to be smoothed away.

The methodology was developed during the design of a distributed information system. The system itself is one domain the methodology analyzes; the methodology is not the system, and the system is not the methodology. We treat the methodology as a separate contribution worth presenting on its own terms, and the connection to the originating system as biographical rather than load-bearing. The methodology’s value, if it has any, is in producing analyses that reveal something structural about the domains it is applied to. That value is for readers and further applications to assess.

Generated under prompt-and-review. This paper, like the rest of the corpus, the supporting implementations, and the architectural specifications, is LLM-generated under direction from the author. The author provides prompts, evaluates outputs, redirects, and approves — text, code, and design refinements are generated rather than directly authored. This shapes the methodology described in The Entity Core Protocol (particularly the iteration tempo it enables) and is a real factor readers should weigh, especially here: the domain analyses presented in this paper were themselves produced through the same prompt-and-review loop, which means the analytical results inherit whatever systematic biases the tooling has. The methodology’s discipline (the 3/3b iteration loop, the falsifiability invitations, the structural cross-checks) is meant to surface such biases, but it does not eliminate them. Independent application by readers using different tooling is the natural complement.