The Pair-Relationship Methodology (v2)

Status: Draft. Revision of v1_revision/methodology.md incorporating three-layer structure, graph semantics, and findings from the biology/cognition/entity system analysis series. Supersedes: v1_revision/methodology.md Companion: Advanced topics (navigation, mathematical connections, research directions) — to be updated separately.

1. Overview

This document describes a methodology for structural analysis of domains. The methodology identifies irreducible primitives, their partial levels, pairwise interactions, and multi-primitive compositions to produce a structural characterization of a domain — what it is, how it works, and where it sits relative to other domains.

The methodology operates at three layers:

• Layer 1 (Domain Analysis): Analyze a single domain — identify its primitives, partial levels, dependencies, interactions, compositions, and manifestation positions. The 12-step process.
• Layer 2 (Graph Construction): Connect analyzed domains through typed edges — determine how domains relate, what content the edges carry, how primitives map across domains, and what cross-domain constraints exist.
• Layer 3 (Graph Semantics): Analyze the populated graph as a whole — discover arrangements, classify domain types, abstract shared structure, identify patterns, and converge through self-correction.

These layers are connected: Layer 1 produces what Layer 2 connects. Layer 2 builds what Layer 3 analyzes. Layer 3's findings feed back to Layer 1 as expectations for new analyses. Layer 1's vocabulary is embedded within Layer 2 when analyzing edge substance (bridge primitives get the full 12-step treatment).

What the methodology produces

For a single domain: a primitive set with partial levels, dependency structure, pair-relationship classification, core triad, load-bearing compositions, phase transitions, and manifestation positions.

For cross-domain analysis: a typed graph of domains connected by edges with specific content. Constrained product lattices across connected domains. Multiple independent paths between domains giving different and complementary knowledge.

For graph-level analysis: domain type classifications, abstract domains capturing shared structure, structural patterns across types, and convergent self-correcting knowledge.

What the framework is and isn't

The framework is an analytical tool for structural understanding — a map of current knowledge that evolves as understanding deepens. It creates coordinate systems (domains) where systems can be precisely positioned, compared, and navigated.

It is not a physical simulation, not a formal mathematical theory (patterns are empirical, not proven), and not prescriptive (shows the landscape, doesn't dictate design choices). With mathematical development, it could become a computable structural analysis tool, bridge mathematics, or constraint-based model generator.

The structural vocabulary

The methodology operates with a specific vocabulary across its three layers:

Layer 1: {Primitive, Level, Dependency, Interaction, Composition, Position} Layer 2: {Edge, Character, Substance, Mapping, Constraint} Layer 3: {Instance, Arrangement, Type, Abstraction, Pattern, Convergence}

Each term is defined precisely in its layer's section below. The vocabulary is itself the product of applying the methodology to itself — each term passes the three-test primitive criterion (minimality, productivity, recurrence).

2. Layer 1: Domain Analysis

2.1 The twelve steps

Step 1 — Information gathering. Read existing rigorous analyses, specifications, and literature for the domain before making claims. Skipping this step is where errors originate.

Step 2 — Landscape analysis. Survey existing instances in the domain. Landscape work orients the analysis — it shows what has been tried and what patterns recur.

Step 3 — Primitive extraction. Identify irreducible primitives via three tests:

• Structural minimality — removing forfeits a class of design moves
• Compositional productivity — combining with others yields new capabilities
• Empirical recurrence — shapes design decisions across diverse instances

This is analyst-judgment-heavy. Candidates get surfaced, tested, and consolidated.

Step 3b — Partial primitive decomposition. For each primitive, decompose internal structure into partial levels — a gradient from absent to fully elaborated.

This step produces a feedback loop with step 3: when you identify partial levels within a primitive, you often discover that one "primitive" is actually multiple bundled together, or that two "separate primitives" collapse into one at coarser resolution. The loop iterates until both the primitive set and partial-level structure stabilize. This iteration loop is the methodology's core reliability mechanism — it converts heuristic primitive guesses into stable primitive sets.

What partial primitive analysis produces:

• Levels per primitive (typically 4–6). Each level is a qualitatively distinct configuration.
• Phase transitions — certain level-changes are discontinuities, not gradual shifts. These mark qualitative behavioral changes.
• Position maps for real systems — concrete structural assessments, not vague comparisons.
• Compensation patterns — systems at low partial levels compensate with external mechanisms (the "layering trap").

Step 4 — Dependency specification. For each primitive, identify what it presupposes. Not all subsets are coherent — dependencies filter the lattice. Dependencies apply at both primitive-presence and partial-level granularity.

Filter stringency — the percentage of the 2^n lattice that is dependency-coherent — is a measurable domain characteristic.

Step 5 — Pair enumeration. For n primitives, C(n,2) pairs. Each pair is a candidate site of structural interaction.

Step 6 — Load classification. Classify each pair as heavy, medium, light, or negligible. Four qualitative criteria:

1. Lines of spec or literature primarily about the pair
2. Number of extensions or mechanisms whose primary content lives in the pair
3. Number of failure modes distinctive to the pair
4. Number of emergent properties the pair contributes to

Step 7 — Coherent sub-lattice construction. Count valid subsets of the 2^n lattice given the dependency structure. This produces the Hasse diagram — the space of valid primitive configurations.

Two resolutions: primitive-presence (coarse, 2^n positions) and partial-level (fine, ~k^n positions). The partial-level lattice is where most analytical value lives.

Step 8 — Hasse diagram walks. Trace monotone paths from the empty set to the full primitive set. Each path is a build-up narrative. Walks at partial-level resolution reveal phase transitions versus gradual elaboration.

Step 9 — Load-bearing composition identification. Find subsets of primitives whose semantic content is irreducible — not recoverable from decomposition into smaller subsets. Triangles are the most common form, but quads and higher arities exist.

Step 10 — Emergent property prediction. Map load-bearing compositions at specific partial-level regimes to observable properties. These are testable predictions.

Step 11 — Structural pattern observation. Cross-domain step. Observe which patterns replicate across domains.

Step 12 — Literature alignment and cross-domain mapping. Where literature exists, check findings against existing analyses. Where cross-domain structure is relevant, map primitives to other domains' implementations.

2.2 What produces the meaningful structure

The highest-value outputs:

1. Partial primitive decomposition (step 3b) — makes landscape analysis concrete and structurally comparable
2. The Hasse diagram at both resolutions — shows where real systems sit and what paths are available
3. Load-bearing compositions at any arity — the irreducible structural units carrying emergent properties
4. The emergent property map — connects structure to predictable behavior, making the analysis falsifiable

2.3 Acknowledged soft spots

Where the methodology relies on analyst judgment:

• Primitive identification — when is a candidate "irreducible enough"? Mitigated by the 3/3b iteration loop.
• Partial level boundaries — how many levels, where to draw boundaries? Guided by landscape observation.
• Load classification — heavy/medium/light are qualitative. A quantitative metric would ground this but has not been implemented.
• Irreducibility criterion for compositions — "can this be recovered from smaller subsets?" is hard to apply mechanically.
• Dependency vs regime — is a constraint structural dependency or design-choice regime? Sometimes unclear.

2.4 Methodological refinements (R1-R10)

R1. Name the domain kind explicitly. Substrate-domain primitives are discovered via structural necessity; application-domain primitives are selected via design choice.

R2. Mark dependency-filter stringency. Report the coherent-subset count as a percentage of the full 2^n lattice.

R3. Distinguish firm from borderline compositions.

R4. Add a literature-mapping step. The triangulation test should be explicit.

R5. Add a cross-domain-mapping step when applicable.

R6. Honesty about count-sensitivity. Accompany primitive count claims with alternatives considered.

R7. Engage existing rigorous analysis first. Surface-level analysis without engaging specs produces material errors.

R8. Mode A → Mode B zoom operation. When a domain has existing dimensional analysis, pair-relationship analysis can be applied directly to the dimension set.

R9. Candidate-primitive triage. Three outcomes: dimensional (add), reducible (absorb), scope-excluded (note).

R10. Architectural asymmetry acceptance. Sub-systems within the same broader system can have different primitive counts.

3. Core Vocabulary

Primitive

A minimal irreducible structural unit of a domain, meeting three criteria: structural minimality, compositional productivity, empirical recurrence.

Primitives have partial levels — a gradient from absent to fully elaborated. Each level is a qualitatively distinct configuration. Typical count: ~6 primitives per domain (range 4–9 observed across 17+ domains).

Partial Level

A qualitatively distinct configuration within a single primitive's gradient. Boundaries drawn where real systems show qualitative behavioral differences.

Partial levels are scope-relative — "Full" is the analyst's current ceiling, not an absolute limit. They are discrete milestones — qualitatively distinct. Continuous parameters within or between milestones are not primitives.

Dependency

A structural relation where one primitive presupposes another. Filters the 2^n lattice to its coherent sub-lattice.

Dependencies apply at both primitive-presence and partial-level granularity.

Phase Transition

A discontinuity in partial-level progression where qualitative behavior changes. Two types: within-domain (within a single primitive) and cross-edge (in bridge primitives, gating elaboration in connected domains).

Pair-Relationship (Interaction)

A bivariate interaction between two primitives when both are active. Classified as heavy, medium, light, or negligible by four qualitative criteria.

Load-Bearing Composition

A subset of primitives (at any arity) whose emergent properties are irreducible. Triangles are most common. Load-bearing compositions often require specific partial levels to activate.

Core Triad

A 3-primitive subset where all three pairs are heavy AND the triangle is load-bearing. Operationally defines "what the domain IS."

Hub Primitive

A primitive that most others depend on or relate to. Every domain analyzed has one or two.

Anchor Pair

A heavy pair that serves as a semantic attractor for load-bearing compositions.

Manifestation

A specific actual system occupying a position in a domain's lattice. Manifestations have identity (distinguishable), trajectory (they move through the lattice over time), scaffolding (ad-hoc compensation for partial positions), and properties determined by position.

The lattice is a possibility space, not a progress axis. Systems at partial positions are not deficient — partial positions have real advantages (lower complexity, specific optimization).

Ambient Primitive

A primitive at one analytical level that becomes the assumed medium at a higher level, disappearing as an explicit primitive.

4. Layer 2: Graph Construction

4.1 The inter-domain graph

Analyzed domains connect via a typed graph. Domains are nodes — each with its own primitive set, lattice structure, and population of manifestations. Typed edges connect them. The graph grows with analysis and is never "complete."

The graph is NOT a numbered hierarchy. It is a DAG (or more generally, a graph with possible cycles through feedback edges) with typed edges. The structure is categorical: domains are objects, edges are morphisms, edge composition is morphism composition.

4.2 Edge types

Edges between domains are classified by character — what kind of structural relationship they represent:

Realization — How a system is physically realized in its substrate. Has bridge primitives (computational translation machinery). Exists where there is a substrate gap — the upper system operates in a different medium than the lower. Direction: from substrate toward the system it enables.

Role identification — What structural role each primitive plays in an abstract framework. No bridge primitives — just a mapping table. Enables comparison via shared abstract roles.

Configuration — A domain IS an abstract framework at specific partial-level settings. No new primitives — the domain selects which regimes to emphasize. Direction: from abstract to specific.

Enrichment — Adding specific mathematical structure to an algebraic base. Specific to domains that ARE mathematical structures. The enrichments are primitive-like (partial levels, pair structure) but combine orthogonally with the base.

Decomposition — How application-domain primitives implement via substrate primitives. May have bridge-like content (not yet formally analyzed as bridge primitives).

Feedback — Bidirectional influence where surface activity modifies the context for substrate activity. Discovered in the biology analysis (niche construction).

Selection — Evaluative influence where community structure evaluates surface performance, propagating to encoding via the evaluator. Discovered in the biology analysis.

The edge type set is empirically extensible — new types are added when found. The current set is not known to be complete.

4.3 When edges have bridge primitives

Bridge primitives exist when there is a substrate gap — the upper system's primitives operate in a different medium than the lower system's, and specific translation machinery is required.

Bridge primitives are structurally identical to domain primitives — they have partial levels, pair structure, core triads, phase transitions. They are analyzed using Layer 1's full 12-step methodology (Layer 1's vocabulary is EMBEDDED within Layer 2's substance analysis).

The only difference is location: bridge primitives live on an edge rather than in a node. They ARE the translation machinery.

4.4 Edge composition

Edges compose through shared domains. Composition is typed — not all type combinations compose meaningfully.

4.5 Product lattices and feasible regions

When domains are connected by realization edges, their lattices combine into product spaces:

Upper domain lattice × Bridge lattice × Substrate lattice

Not all positions in the product are coherent. Constraints reduce it to a feasible region — the set of positions where all cross-domain constraints are satisfied.

The feasible region's boundary is where minimal viable configurations live: abiogenesis thresholds, minimum viable implementations.

4.6 Category theory's dual role

Category theory appears in two distinct ways:

1. As a node — the categorical base {Object, Morphism, Composition, Identity} is a domain that mathematical structures (physics) build on via enrichment.
2. As the meta-language — the graph itself has categorical structure. This isn't imposed; it follows from "relating structured things with composable relationships" being what a category IS.

These roles shouldn't be conflated.

5. Layer 3: Graph Semantics

5.1 What Layer 3 does

Layer 3 operates on the populated graph as a whole — not on individual domains (Layer 1) or individual edges (Layer 2), but on patterns that emerge from the graph's structure when multiple domains are analyzed and connected.

5.2 The six primitives

5.3 The genesis transition

The genesis transition at Layer 3 is abstraction from multiple instances (Ab appearing). Before: classified instances but no generalizations. After: abstract domains capturing what all instances of a type share. This is when the methodology becomes predictive — the abstract domain predicts properties of not-yet-analyzed instances.

The genesis transition requires at least TWO independent arrangements. Three is better — it distinguishes universal structure from coincidental similarity.

5.4 The convergence mechanism

The graph self-corrects through constraint propagation:

1. New instance analyzed (Layer 1)
2. Compared against existing abstractions (Layer 3)
3. Mismatch detected — instance doesn't fit current abstraction
4. Abstraction revised (or instance re-analyzed)
5. Revision cascades through connected domains

This convergence is DIRECTED (by the analyst's cognitive evaluation), not autonomous (like biological evolution). The methodology gets better with use because its predictions get validated or revised against each new domain.

5.5 Arrangements are topology-general

Arrangements can take any shape:

• Realization sequences (linear chains from substrate to surface to ecosystem)
• Configuration clusters (abstract framework → multiple specific domains)
• Enrichment hierarchies (algebraic base → enriched structures)
• Decomposition trees (system → application domains)
• Mixed topologies (combinations of the above)

The methodology doesn't privilege any particular arrangement shape. Different edge types produce different topologies.

6. Navigation

6.1 Position and tangent set

At any point in a domain's lattice, a manifestation occupies a specific position — a tuple of partial-level assignments. The tangent set T(P) is the set of single-step advances available from the current position, given dependency constraints within the domain AND feasibility constraints across edges.

Some moves are immediately available. Others are blocked — they require bridge or substrate advances first. The blocked moves and their blocking constraints are the most actionable output.

6.2 Paths and path optimization

Multiple paths exist between any current position and a target. They differ in sequence, bridge requirements, intermediate capabilities, and phase transition encounters.

Optimization criteria:

• Minimal total steps → shortest path
• Maximum enabling → through phase transitions (each step opens the most new moves)
• Minimal bridge cost → stays within current substrate capability
• Maximum intermediate value → every step produces a usable system

6.3 Attractors and the layering trap

Certain lattice positions are attractors — positions where many independent systems converge. Attractors are stable because they deliver clear value while the NEXT step requires significant bridge infrastructure with non-obvious payoff.

Systems at attractors build scaffolding — ad-hoc bridge primitives compensating for partial positions. The layering trap is when scaffolding accumulates to the point where advancing actual primitives requires dismantling the scaffolding.

Walls: advancing requires destructive changes (dismantling scaffolding). Hard to cross. Fences: advancing is additive. Easier to cross.

6.4 Co-evolutionary walks

When domains are connected by realization edges, their manifestations advance alternately, each enabling the next. This is a co-evolutionary walk in the joint feasible region.

6.5 Design opportunity discovery

Coherent but unpopulated lattice positions are design opportunities — configurations the dependency structure permits but no existing system occupies. These are structural predictions.

7. Derived Operations

7.1 Unified manifestation

A unified manifestation is a specific entity's position across ALL connected lattices in its arrangement. It captures the entity's structural ID across multiple domains.

Unified manifestations are substantive for realization edges — where bridge primitives create independently variable lattice positions at each level. They are trivial for configuration or role-identification edges — where positions are just different views of the same coordinates.

The unified manifestation is a derived operation — computed by combining Layer 1 positions with Layer 2 mappings and constraints. It's the methodology's most comprehensive structural description of a single entity.

7.2 Structural constraint analysis

Given domain primitives, bridge dependency structure, and cross-domain constraints, compute the minimal coherent path from one state to another. This produces structural predictions about what must exist at each step. Testable against evidence.

8. Meta-Structure

8.1 The methodology applied to itself

The methodology's own structural elements form three primitive sets (one per layer):

Layer 1: {Primitive, Level, Dependency, Interaction, Composition, Position}

• Hub: Primitive. Core triad: {Primitive, Level, Dependency}. Filter: 29.7%.
• Two branches: Structure (Pm→Ix→Cp) and Grounding (Pm→Lv→Ps).

Layer 2: {Edge, Character, Substance, Mapping, Constraint}

• Hub: Edge. Core triad: {Substance, Mapping, Constraint}. Filter: 25%.
• Two branches converging: Character (Ed→Ch→Sb) and Mapping (Ed→Mp), meeting at Cn.

Layer 3: {Instance, Arrangement, Type, Abstraction, Pattern, Convergence}

• Hub: Instance. Core triad: {Instance, Type, Abstraction}. Filter: 18.75%.
• Genesis transition: Abstraction appearing (analysis becomes predictive).

8.2 How the layers connect

Layer 1 produces what Layer 2 connects. Layer 2 builds what Layer 3 analyzes. Layer 3 feeds back to Layer 1.

Layer 1's vocabulary is EMBEDDED within Layer 2: bridge primitives (Layer 2's Substance at high levels) are analyzed using Layer 1's full 12-step process.

Layer 1 can feed Layer 3 directly — domain comparison is possible without formal edge characterization. Layer 2 enriches the analysis but isn't strictly required.

8.3 The categorical spine

The methodology's graph is a category: domains as objects, edges as morphisms, composition as morphism composition. This is not imposed — it follows from the structure of relating structured things with composable relationships.

Each domain also has internal categorical structure (primitives as objects, dependencies as morphisms). The full structure is a 2-category or enriched category.

8.4 Graph constraints

The methodology constrains analytical depth, not structural breadth:

• Any topology is possible (chains, trees, DAGs, cycles, arbitrary)
• Any domain with decomposable structure is analyzable
• Edge types are empirically extensible
• The three-test primitive criterion constrains how rigorously you analyze, not what you analyze

9. Open Questions

1. Is the edge type set finite or open-ended? The current 7 types are empirically discovered. No theory of what types must or can't exist.

2. Do decomposition edges have bridge-like content? Entity system extensions might BE bridge primitives connecting substrate to application domains. Not yet formally analyzed.

3. Is Convergence (Cv) a structural primitive or a dynamic property? It passes the primitive test but sits uncomfortably between structure and process.

4. What mathematical development would most strengthen the methodology? Quantitative heaviness metrics, formal Galois connections for cross-domain mappings, and computational lattice tools are highest priority.

5. How many instances are needed for reliable abstraction? Two is suggestive, three is convincing. Is there a minimum for specific confidence levels?