Overview

The Structural Relationship Mapping Framework (T11) is the territory of operations that extract relations among entities in a representation — the topology of inter-element connections. Input can be textual (a description of a domain that needs to be turned into a relational map) or visual (a diagram, network graph, or schema where the user senses something is missing or off and wants gap detection). Output is a structured relational map — every entity named, every connection labeled by type and directionality, with the organizing structure (hub-and-spoke, chain, hierarchy, network, bipartite) named explicitly. The framework is the structure-extraction territory; it sits adjacent to but distinct from mechanism understanding (T16, the gears that produce behavior), process flow (T17, the sequenced steps over time), and compositional reading (T19, what the layout itself does as primary content).

The framework runs in two modes. Relationship Mapping is the general mode for textual input — takes a description of entities and produces the relational map with every connection carrying a literal type prefix (causal, correlational, dependency, influential, structural) and directionality. The map is acyclic by default; cycles trigger transition to systems-dynamics modes in T4 or T17. Common sub-types include concept-map outputs (heterogeneous relations with cross-links per Novak’s tradition) and causal-DAG outputs (causal-specific framing with focal exposure and outcome per Pearl’s tradition). Spatial Reasoning is the visual-input variant — takes a diagram (sketch, whiteboard photo, Excalidraw, Obsidian Canvas, or prior Ora visual) where the diagram itself is the question and performs structural extraction with gap analysis: what’s there, what’s ambiguous, what’s missing (missing nodes, connections, levels, feedback loops). The mode preserves the user’s spatial arrangement (annotates without rearranging), uses Tversky correspondence principles (proximity = relatedness; verticality = hierarchy; containment = category; connection = relationship), and generates open fog-clearing questions rather than leading questions.

The framework’s load-bearing intellectual content is the typed-connection requirement, the acyclicity discipline, the non-obvious-cross-link requirement, and the gap-detection-with-evidence rule. The typed-connection requirement says every connection in a relationship map carries a literal type prefix from the controlled vocabulary — defaulting to the weakest type the evidence supports (correlational rather than causal when mechanism is unspecified) prevents the causation-correlation trap. The acyclicity discipline keeps relational maps as DAGs by default; cycles signal feedback structure that requires systems-dynamics modes rather than pretending the cycle can be flattened. The non-obvious-cross-link requirement (drawn from Novak’s concept-map tradition) says a genuine concept map carries at least one cross-link bridging two otherwise-separate sub-trees — cross-links are the marker of integrative understanding because they indicate the mapper has seen a connection that the standard taxonomy does not assert. The gap-detection-with-evidence rule (Spatial Reasoning specific) says identified gaps must be implied by the spatial structure or domain logic, not by template pattern-matching — when a gap is asserted, the spatial or domain evidence must be cited.

The framework’s epistemological posture deliberately resists four patterns. Causation-correlation collapse — counteracted by the typed-connection requirement defaulting to the weakest type the evidence supports. Linear-narrative substitution for relational structure — counteracted by the relational-map output requirement that resists flattening into prose. Silent cycle smuggling in DAGs — counteracted by the acyclicity check. Diagram rearrangement that erases the user’s pre-conscious structural choices — counteracted by Spatial Reasoning’s preservation rule (annotate without rearranging).

The framework answers questions like: How do these entities connect? What affects what? Where are the missing connections in this diagram? Is there a feedback loop I haven’t drawn? Take this paragraph describing a domain and produce the relational structure. This whiteboard photo is bothering me — what’s missing or ambiguous?

Systemic context

Structural Relationship Mapping is a Tier-3 cognitive territory in the Ora reasoning lattice, paired with three sibling territories at canonical disambiguation points. T11 ↔ T19 (Structural Relationship ↔ Spatial Composition) — the same diagrammatic input answers different questions. T11 reads the diagram as notation: what relations are asserted among elements. T19 reads the diagram as composition: what the layout itself is doing. When both legitimately fire, T11 runs first because relation-extraction is the lighter, more determinate operation; T19 layers compositional reading on top. T11 ↔ T16 (Structure ↔ Mechanism) — T11 asks how the parts relate as a structure; T16 asks how the gears interlock to produce behavior. T11 typically runs first; T16 builds on the structure. T11 ↔ T17 (Structure ↔ Process) — T11 asks how the parts relate; T17 asks how the flow runs in sequence. T11 typically runs first; T17 builds on the structure. The escalation hooks within T11 itself: when input becomes visual mid-conversation, switch sideways from relationship-mapping to spatial-reasoning; when the diagram-reading question shifts from “what relations does this assert” to “what is the layout doing,” hook sideways to T19 per the canonical T11↔T19 disambiguator.

Ora’s output

Example question: I’ve been trying to understand how my organization actually works, and I keep getting confused. Here’s a paragraph describing the main parts: “We have a product team that builds features, an engineering team that implements them, a design team that owns visual decisions, a research team that informs both product and design, a customer success team that surfaces user pain, an executive layer that sets strategy, and a board that approves major bets. Engineering depends on design for visual specs and on product for prioritization. Customer success surfaces issues that go to product first, then to engineering. Research feeds into product but also goes directly to design when visual experiments are needed. The executive layer reviews product roadmaps quarterly and engineering capacity monthly.” Can you produce the relational structure?

Ora’s response (running T11 in relationship-mapping):

The input is a textual description of entities and relations — Relationship Mapping is the right mode. I’ll work the description into a structured relational map with typed connections, surface the organizing structure, identify non-obvious cross-links, and check for cycles.

Entities (seven, plus board for completeness). Product team (P); Engineering team (E); Design team (D); Research team (R); Customer Success team (CS); Executive layer (EL); Board (B).

Connections with type and directionality. E dependency→ D (engineering depends on design for visual specs). E dependency→ P (engineering depends on product for prioritization). CS causal→ P (customer success surfaces issues into product workload). P causal→ E (product feeds prioritized work to engineering). R causal→ P (research informs product decisions). R causal→ D (research informs design directly when visual experiments are needed). EL influential→ P (quarterly roadmap review). EL influential→ E (monthly capacity review). B structural→ EL (board approves major bets — governance authority).

Non-obvious cross-links surfaced. Cross-link 1. CS and R both surface user information into the same downstream nodes (P and partially D). The analogous-to cross-link raises whether the teams are duplicating discovery work or genuinely producing different evidence. The user did not name this. Cross-link 2. R → D directly bypasses P, but D’s output goes back to E via the visual-specs dependency, creating an R → D → E pathway outside P’s prioritization scope. Structurally significant: some engineering work is happening on research-driven design experiments P may not be tracking.

Organizing structure. Hub-and-spoke with one hub (P) and two flow patterns: the customer-and-research-to-engineering pipeline (CS, R → P → E with the design dependency layered) and the executive-governance overlay (EL → P, EL → E, B → EL). The map is not a clean hierarchy — engineering has two upstream dependencies; the board approves but does not direct day-to-day. Calling this an “org chart” would mis-classify it; it is a dependency-and-flow map.

Acyclicity check. Acyclic as drawn. But the description suggests two implicit cycles on longer time scales: when E ships and CS hears issues, they flow back through CS → P → E (a feedback loop on the shipping cadence); when EL reviews E’s capacity and adjusts P’s roadmap accordingly, there is a slower planning-rhythm feedback loop. Both surface if the question shifts from structure to dynamics — that triggers a transition to systems-dynamics modes in T4 or T17.

Boundary statement. This map covers operational structure. It does not cover informal influence patterns, resource-allocation structure, or customer-side structure. Those are separate passes.

Confidence per finding. High on entity inventory and explicit connections. Medium-high on cross-links and organizing-structure characterization (inferences from the description). Medium on implicit cycles — real in any organization shaped this way, but specific cadence depends on culture I do not have detail on.

The relational structure is now visible: seven nodes, nine typed connections, two non-obvious cross-links, the hub-and-spoke pattern named, two implicit cycles flagged for the systems-dynamics handoff. The user can reason about the structure directly, identify load-bearing constraints (engineering’s two upstream dependencies, P’s hub role), and decide whether to investigate the cycles next.

Commercial AI comparison

Comparison content auto-populates when the comparison-refresh framework runs against this question. Drafters do not author this section.

Brief comparison commentary

Auto-populates with the comparison content above.

How to use this framework

You can run the Structural Relationship Mapping pattern with any AI of your choice. The composition is single-pass for either of the two modes.

For relationship-mapping (textual input):

[Paste the framework specification.]

Run relationship-mapping.

Description: [Plain prose describing the entities and their relations.]

Output preference (optional): concept-map (heterogeneous relations with cross-links) or causal-DAG (causal-specific framing with focal exposure and outcome) or general (the default).

For spatial-reasoning (visual input):

[Paste the framework specification.]

Run spatial-reasoning.

Diagram: [Image, photo, or detailed description. Sketch fidelity matters less than for compositional reading; the structure can usually be extracted from rough sketches.]

Question (optional): What you sense is missing, ambiguous, or off about the diagram.

The AI runs the typed-connection extraction with directionality, names the organizing structure, surfaces non-obvious connections (≥2, with at least one cross-link in concept-map outputs), and declares acyclicity. For spatial-reasoning, the AI additionally runs the Tversky correspondence audit, identifies gaps with specific spatial or domain evidence, generates open fog-clearing questions rather than leading ones, and preserves the user’s spatial arrangement.

For best results:

  1. Provide the description in plain prose. The framework extracts structure from prose; bullet lists work too but flatten relationships that prose preserves implicitly.
  2. Don’t pre-classify the connection types. When the framework asks for type prefixes, it defaults to the weakest type the evidence supports — this is the guard against causation-correlation collapse. Pre-classifying would defeat the discipline.
  3. Honor the acyclicity transition. If the framework finds cycles in your description, that is a signal to switch territories (to T4 or T17 for systems dynamics) rather than to flatten the cycles into a DAG.
  4. For spatial-reasoning, do not request “cleanup.” The framework preserves your arrangement because the arrangement carries information about how you are thinking. Annotation overlay is the correct output, not a re-drawn cleaner diagram.
  5. Take the cross-links seriously. When the framework surfaces a cross-link you had not named, that is the moment the analysis is doing its load-bearing work — making visible a connection your standard taxonomy of the domain does not assert.

The framework is deliberately tool-agnostic. The typed-connection vocabulary, the acyclicity discipline, the non-obvious-cross-link requirement, and the Tversky correspondence audit are conceptual disciplines that survive the lift to any environment. The output is diagram-friendly and can be rendered as a concept map, causal DAG, or annotated overlay depending on the input.

Other examples

  • A causal-DAG output for an empirical question — A user investigating whether a workplace policy change caused a productivity shift. Causal-DAG framing extracts entities (policy change, productivity, candidate confounders: hiring cohort, tooling rollouts, seasonal cycles, leadership transitions) and causal edges with directionality. The DAG identifies confounders (selection effect on which teams adopted first), mediators (engagement-level shifts), and colliders (performance reviews whose conditioning would induce spurious association). Demonstrates the causal-DAG sub-mode where naïve causal claims would over-state evidence.

  • A spatial-reasoning gap detection on a whiteboard sketch — A photo from a strategic-planning off-site: fifteen entities, three rough clusters, one orphan node. Spatial-reasoning extracts structure with ambiguities flagged; applies Tversky correspondence audit (proximity-grouping drifted in one cluster — probable mis-grouping); identifies gaps (orphan connectivity, implied cluster bridge, missing feedback loop suggested by domain logic); generates open fog-clearing questions; preserves the user’s arrangement and produces an annotation overlay. Demonstrates the preservation discipline against the rearrangement temptation.

  • A concept-map output for pre-writing organization — A user preparing to write on attention research. Concept-map framing produces major concepts (attention as filter, spotlight, resource, control), typed connections (filter extends spotlight; resource analogous-to filter on capacity; control supersedes spotlight in modern accounts), and at least one cross-link bridging concepts the standard taxonomy treats as separate (control ↔ working-memory load via executive function). The cross-link is the marker of integrative understanding — a connection standard surveys do not foreground.

Citations

The Structural Relationship Mapping Framework integrates multiple traditions under one territory. Tversky’s “Visualizing Thought” (Topics in Cognitive Science, 2005) and Mind in Motion (2019) provide the foundational treatment of spatial-conceptual correspondence (proximity = relatedness; verticality = hierarchy; containment = category; connection = relationship). Novak and Cañas’s The Theory Underlying Concept Maps (2008) provides the concept-map theory and the cross-link as marker of integrative understanding. Pearl’s Causality (2009) provides the causal-DAG theory and acyclicity-by-construction; Textor, Hardt, and Knüppel’s “DAGitty” (Epidemiology, 2011) provides the practical tooling. Larkin and Simon’s “Why a Diagram is (Sometimes) Worth Ten Thousand Words” (1987) is the foundational treatment of diagram literacy.

The structural-relationship taxonomy is internal to Ora and operationalizes the controlled vocabulary that prevents causation-correlation collapse. Pattern libraries draw on graph-theory pattern recognition adapted for cognitive-mapping use. Senge’s The Fifth Discipline (1990) provides system-archetypes vocabulary used optionally in spatial-reasoning when causal structure surfaces in diagrams.

The framework is single-author and originated 2026-05-01 from the T11 territory consolidation. The Decision G re-home moved spatial-reasoning from T19’s old name to T11 — its actual operation (structural gap detection on diagrammatic input) is a T11 operation on visual-medium input, while T19 was renamed to “Spatial Composition” for compositional-reading work.

Downloads

  • Framework specification (PDF) — link to ora-ai.org canonical artifact when published
  • Framework specification (plain text) — link to ora-ai.org canonical artifact when published
  • Full white paper (PDF) — link when published