Governance knowledge systems fail. The evidence is unambiguous: knowledge management initiatives report 50–70% failure rates across three decades of investment (Davenport & Prusak, 1998), only 7% of organizations reach APQC Level 5 maturity despite widespread technology adoption, and the design-rationale adoption gap has persisted for half a century despite successive waves of methodological innovation. The question this report engages is not whether these systems fail — the documented record across multiple traditions is conclusive — but why they fail structurally and whether a resolution path exists in the documented record of an established practice tradition.

This report argues that a single architectural error explains the persistent failure: requiring practitioners to step outside operational work to document governance context. The cognitive-load diagnosis, grounded in Sweller's cognitive load theory and validated across the knowledge management, organizational learning, and decision science traditions, identifies documentation-as-separate-activity as the structural cause. The resolution is not better documentation tools but a different architecture: governance context as a structural by-product of operation rather than a parallel recording process.

The audit tradition provides the constructive evidence. AU-C 315 and ISA 315 have standardized organizational governance understanding across every industry for over a century. Bell, Marrs, Solomon, and Thomas (1997) demonstrated that auditors already construct predictive organizational models as standard practice — strategic-systems auditing is organizational world modeling. Klein's recognition-primed decision model establishes that expert practitioners use pattern-based cognition rather than analytical deliberation, and the critical decision method probes this cognition through standardized instruments that operate at the governance level rather than the domain-knowledge level.

The thesis would be disproven if: (1) the governance-knowledge adoption failure could be explained without reference to cognitive load from separation-from-work architecture — meaning the cognitive-load diagnosis does not hold; or (2) a tradition other than audit methodology achieved full governance-surface coverage without cross-domain standardization — meaning the audit tradition's unique position as externalization precedent does not hold.

Organizations have been trying to capture governance knowledge for fifty years, and the documented record shows persistent, systematic failure. The evidence converges from independent traditions: knowledge management reports 50–70% failure rates (reported across practitioner literature), design rationale systems have failed to achieve adoption since Rittel and Webber's "wicked problems" formulation in 1973, and cognitive load theory has not been applied to governance system design despite three decades of maturity. Nine independent traditions — decision science, cognitive load theory, knowledge management, knowledge engineering, enterprise architecture, audit methodology, organizational learning, semantic web, and bidirectional transformation — each built rich apparatus and each reached the same boundary: prescription without infrastructure.

The cognitive-load diagnosis explains the failure. Sweller's cognitive load theory (1988, 2019) establishes that working memory capacity is severely limited and that extraneous cognitive load — processing demands that do not contribute to the primary task — degrades performance. Documentation imposed as a separate activity is extraneous load: it competes with productive work for the same limited cognitive resources. Nonaka and Takeuchi's externalization bottleneck (1995) describes the same phenomenon from the knowledge-creation perspective: converting tacit knowledge to explicit form is "cognitively expensive, contextually destructive, and operationally costly." Szulanski's stickiness paradox (1996) confirms that the most valuable organizational knowledge is the most resistant to transfer precisely because it is tacit, contextual, and practice-embedded. The architectural error is not that organizations use the wrong documentation tools — it is that they treat governance context as something to be documented rather than something to be structural. The resolution: governance context as a by-product of operation, embedded in the work itself.

The audit tradition provides the constructive half of the argument. AU-C 315 and ISA 315 require the same organizational understanding dimensions — entity nature, objectives, internal control, risk assessment — across every industry. This is a century of empirical evidence that governance-level organizational understanding can be standardized across domains. Bell et al. (1997) demonstrated that auditors construct predictive models of client organizations — strategy, economic web, business processes — and measure deviations between predicted and actual outcomes. Klein's recognition-primed decision model (1998) explains how expert practitioners achieve this: experienced auditors recognize patterns, not execute algorithms. The critical decision method probes that expertise through structured instruments that are standardized in form while remaining sensitive to context. The externalization path — from practitioner cognition through governance-level extraction to computational representation — is not a hypothetical proposal. Each stage has documented precedent. The synthesis across stages is the contribution.

Three WMI thesis positions are strengthened. WMI-P11 (controls-testing methodology) is strengthened by the evidence that the audit tradition's century of practice constitutes the externalization precedent. WMI-P14 (corrective action obligation) is substantively closed — the cognitive-load diagnosis establishes why corrective action for governance failure requires architectural intervention, not procedural improvement. WMI-P15 (architectural decisions as ethical decisions) is strengthened by the finding that documentation architecture determines who bears the cognitive burden of governance — an architectural choice with direct ethical consequences.

The central structural property this report names is the separation-from-work error: governance knowledge systems fail because they impose documentation as a separate activity competing with productive work for limited cognitive resources. The evidence converges from three independent traditions — cognitive load theory, knowledge management, and organizational learning — each documenting the same structural failure from different analytical perspectives.

Sweller's cognitive load theory (1988, 2019) provides the diagnostic framework. Working memory capacity is severely limited — Miller's (1956) "seven plus or minus two" items, refined by Cowan (2001) to approximately four chunks. Cognitive load theory distinguishes intrinsic load (inherent to the task's complexity), extraneous load (imposed by instructional or task design), and germane load (devoted to schema construction and learning). The critical finding for governance: documentation imposed as a separate activity from the work being documented is extraneous cognitive load. It does not contribute to the primary governance task; it competes with it. The evidence shows that cognitive load theory has not been applied to governance system design (RA-010 §F2) — a confirmed literature gap with significant implications. The documentation standards that governance traditions prescribe — audit workpaper requirements, compliance record-keeping, knowledge capture protocols — all impose extraneous load by requiring practitioners to perform the governance task and then separately record what they did and why.

Nonaka and Takeuchi's SECI model (1995) identifies the same failure from the knowledge-creation perspective. The externalization mode — converting tacit knowledge to explicit form — is the critical bottleneck in organizational knowledge creation (RA-003 §F1). Externalization is cognitively expensive (articulating implicit knowledge requires significant effort), contextually destructive (codification strips the context that gives knowledge its richness), and operationally costly (documentation competes with productive work for limited time and attention). Farnese et al. (2019) found in their empirical assessment that evidence for SECI's externalization mechanism remains "fragmented and inconclusive," confirming that the bottleneck persists despite thirty years of attention. The resolution is architectural: rather than attempting to externalize the tacit knowledge itself, capture the explicit products of tacit knowledge application — the decisions, commitments, and judgments that naturally emerge at the boundary between tacit and explicit cognition.

The knowledge management failure data confirms the diagnosis at scale. KM initiatives fail at 50–70% rates across three decades (Davenport & Prusak, 1998). Only 7% of organizations reach APQC Level 5 maturity; 93% plateau at Level 3 — the point where KM infrastructure exists as a separate system but is not integrated into operations (RA-003 §F11). Cook and Brown (1999) identified the root cause: the conflation of knowledge (possessable, storable) with knowing (practiced, contextual). KM systems optimized for capturing knowledge-as-possession systematically miss knowledge-as-practice — precisely the category that governance context occupies. Szulanski's (1996) analysis of 271 best-practice transfers confirmed that the primary barriers to knowledge transfer are knowledge-related (absorptive capacity, causal ambiguity), not motivational — the most valuable organizational knowledge is the most resistant to extraction (RA-003 §F3).

Star and Ruhleder's (1996) infrastructure theory explains why embedded infrastructure succeeds where standalone applications fail. They identified eight properties of infrastructure, including embeddedness (sunk into other structures), transparency (supports without requiring attention), and learned as membership (part of organizational participation rather than special training). Governance infrastructure that is embedded in work — transparent to practitioners, learned as organizational membership — avoids the extraneous cognitive load that documentation-as-separate-activity imposes. The infrastructure-vs-application distinction maps directly to the governance failure: documentation tools are applications (requiring conscious use, imposing cognitive load); governance context as structural by-product is infrastructure (embedded, transparent, part of organizational operation).

Two distinctions are structurally load-bearing for the externalization path:

Documentation-as-activity versus context-as-infrastructure. Traditional governance knowledge systems treat documentation as an activity performed alongside or after the governed work. Audit workpapers are composed after the examination. Compliance records are assembled for periodic review. Knowledge is captured when practitioners choose to record it. Each of these is a separate cognitive act competing with the primary work for attention and working memory. Context-as-infrastructure inverts this relationship: governance context accumulates structurally as a consequence of the governed operation, without requiring a separate recording act. The distinction is not about automation (replacing manual documentation with automated recording) but about architectural position (governance context as by-product rather than product).

Governance-level versus domain-level extraction. The knowledge engineering tradition's central challenge — that expert knowledge is domain-specific and resists cross-domain standardization — is resolved by the level distinction identified in RA-013 §F7 and §F8. Cognitive task analysis operates at the domain-knowledge level: it extracts what experts know and do in their specific domain. The audit tradition operates at the governance level: it extracts organizational structure (authority, accountability, constraints, evidence requirements) that is invariant across domains. These levels are not in tension; they address different kinds of organizational knowledge. AU-C 315 and ISA 315 require the same organizational understanding dimensions across every industry because governance structure — who has authority, what evidence is required, what constraints apply, how decisions are accountable — is domain-invariant. The governance-level extraction methodology is standardizable precisely because it targets governance structure rather than domain-specific knowledge. CommonKADS provides a structural parallel: its context-level worksheets (OM-1 through AM-1) are standardized across domains; it is the Knowledge model's Domain layer that requires customization (RA-013 §Synthesis).

The evidence across the five research artifacts establishes a three-stage path from practitioner knowledge to computational governance representation:

Stage 1 — Practitioner cognition. Expert practitioners — particularly auditors with cross-domain governance assessment experience — carry governance knowledge in the form of recognition-primed decisions (Klein, 1998). They do not apply analytical algorithms; they recognize patterns, leverage experience, and exercise professional judgment developed through extensive practice (RA-010 §F12). This cognition is tacit, contextual, and practice-embedded — precisely the category that traditional externalization approaches fail to capture.

Stage 2 — Governance-level extraction. The audit tradition demonstrates that this cognition can be accessed through structured instruments that operate at the governance level rather than the domain-knowledge level. The critical decision method (Klein, Calderwood, & MacGregor, 1989) provides standardized probes that are RPD-derived: they work because they target the recognition process rather than the domain content (RA-013 §F3). The governance-level instrument — which asks about authority, accountability, constraints, evidence, and decision structure — is standardizable across domains because governance structure is domain-invariant. A century of audit practice under AU-C 315 / ISA 315 validates this cross-domain standardization empirically (RA-013 §F9, §C2).

Stage 3 — Computational representation. The extracted governance understanding requires computational representation — not merely documentation but structured, queryable, reasoning-capable infrastructure. The semantic web tradition has solved the representation problem at the technical level: OWL 2 provides decidable reasoning with computational guarantees, RDF provides a universal data model, PROV-O provides provenance tracking, and BFO demonstrates that foundational ontologies can integrate hundreds of domain ontologies at scale (RA-005 §F1–F3, §F7). The gap is in application: no existing ontology instantiates organizational decision governance as a structural domain (RA-005 §C1). Enterprise knowledge graphs and Graph RAG represent the current state of the art — but they are passive (descriptive), answering "what happened?" rather than "was what happened authorized?" The transition from passive knowledge graph to active governance substrate is the novel application (RA-005 §C3).

The three-stage externalization path derives four design requirements that any viable governance infrastructure must satisfy:

Falsifiable claim. The externalization path — the claim that practitioner methodology can be computationally externalized through governance-level extraction, resolving the cognitive-load failure that the documented record shows persistently degrades governance-knowledge systems — is falsifiable by demonstrating either: (1) that the governance-knowledge adoption failure can be explained without reference to cognitive load from separation-from-work architecture, meaning the cognitive-load diagnosis does not hold; or (2) that a tradition other than audit methodology has achieved full governance-surface coverage without cross-domain standardization, meaning the audit tradition's unique position as externalization precedent does not hold. Each condition corresponds to the failure of one leg of the two-part argument; both legs would need to fail for the externalization path to be fully overturned.

The evidence for the externalization path comes not from any single tradition but from the convergence of nine independent traditions — each using different methods, different vocabularies, and addressing different aspects of organizational knowledge — on the same structural boundary. The kind-specific convergence standard (PUB-TR-A-001 §4.2) governs the treatment: one focused paragraph per tradition, followed by a convergence synthesis.

Decision science and bounded rationality. Simon (1947/1997, 1955) established that human decision-making operates under bounded rationality — limited information, limited computational capacity, and limited time — producing satisficing rather than optimizing behavior. Kahneman (2011) formalized the dual-process architecture: System 1 (fast, intuitive, pattern-based) and System 2 (slow, deliberate, analytical). The governance implication: governance infrastructure must accommodate bounded rationality rather than assume rational compliance. The tradition built the diagnostic framework but not the infrastructure to operationalize it. RA-010 §F1, §F3.

Cognitive load theory. Sweller (1988, 2019) demonstrated that working memory capacity imposes hard limits on learning and performance, and that extraneous cognitive load — processing demands imposed by task design rather than task content — degrades outcomes. The tradition established the mechanism (cognitive load from documentation) but has not applied it to governance system design — a confirmed literature gap (RA-010 §F2, §C6). The triple convergence of Kahneman's System 1/2, LeCun's amortized inference, and Beer's variety engineering on dual-process governance architecture validates the diagnosis from three independent starting points (RA-010 §F3, §C7).

Knowledge management. Nonaka and Takeuchi (1995) identified the externalization bottleneck in organizational knowledge creation — the cognitively expensive conversion from tacit to explicit. Davenport and Prusak (1998) documented the persistent failure of knowledge capture as separate activity. Szulanski (1996) quantified the stickiness paradox across 271 best-practice transfers. The tradition documented the failure comprehensively — 50–70% initiative failure rates, Level 3 maturity plateaus — but resolved it as a cultural or technological problem rather than an architectural one. RA-003 §F1, §F3, §F11, §C1.

Knowledge engineering and cognitive task analysis. Klein (1998, 2008) developed the recognition-primed decision model and the critical decision method for eliciting expert cognition. CommonKADS (Schreiber et al., 2000) formalized knowledge engineering methodology with context-level standardization and domain-level customization. Militello and Hutton (1998) developed applied CTA for practitioners. The tradition built the extraction methodology but did not resolve the domain-specificity concern for governance applications. The level distinction (RA-013 §F7, §F8) — governance structure is domain-invariant; domain knowledge is domain-specific — resolves this: governance-level extraction is CTA applied to the domain-invariant governance surface. RA-013 §F3, §F7, §F8.

Enterprise architecture. Zachman (1987) established the framework for classifying organizational information by perspective and abstraction. TOGAF (The Open Group) codified enterprise architecture as a practice with process models, deliverables, and governance. ArchiMate provided a formal modeling language with approximately 57 element types. The tradition classifies organizational structure comprehensively but does not model governance — authority, decision provenance, accountability bindings, constraint relationships — as structural entities. The convergence table (RA-013 §F16) confirms: enterprise architecture provides structural classification but not governance representation. RA-013 §F16.

Audit methodology. AU-C 315 (AICPA) and ISA 315 (IAASB, revised 2019) require auditors to understand the entity and its environment — including its nature, objectives, internal control components, and risk assessment processes — using the same standardized dimensions across every industry. Bell, Marrs, Solomon, and Thomas (1997) reconceived auditing as organizational world modeling: the auditor constructs a predictive model of the client's organizational dynamics and measures deviations between expected and actual outcomes. COSO (2013) codified internal control as an integrated framework with five components and seventeen principles. The audit tradition is unique among the nine: it achieves full governance-surface coverage (authority, accountability, constraints, evidence, decision provenance) through cross-domain standardization. The tradition produces risk assessments rather than populated governance representations — the understanding methodology exists, the computational representation does not. RA-013 §F9, §F10, §F16, §C2, §C4.

Organizational learning. Argyris and Schön (1978) demonstrated that organizations systematically resist double-loop learning because defensive routines — Model I governing values — make governing variables undiscussable (RA-014 §F3). March (1991) established the exploration/exploitation tension as a structural resource allocation problem with self-reinforcing dynamics that bias toward exploitation (RA-014 §F7). Zollo and Winter (2002) showed that dynamic capabilities are built through experience accumulation, knowledge articulation, and knowledge codification — with codification most valuable for rare, causally ambiguous tasks (RA-014 §F11). The tradition identified ten distinct learning pathologies (RA-014 §F13) and established that organizational learning requires computational infrastructure (RA-014 §C1), but provided no such infrastructure.

Semantic web and knowledge representation. Gruber (1993) defined ontology as an explicit specification of a conceptualization and introduced ontological commitment as a social contract about terminology. Davis, Shrobe, and Szolovits (1993) established that knowledge representation simultaneously serves five roles — surrogate, ontological commitment, theory of reasoning, computational medium, and human expression medium — each with governance implications. BFO achieved ISO standardization (ISO/IEC 21838-2:2021) and anchors 400+ biomedical ontologies through the OBO Foundry. PROV-O provides provenance tracking but stops short of governance (RA-005 §F3). The tradition solved representation at the technical level — mature, standardized, computationally characterized — but no instantiation addresses organizational decision governance (RA-005 §F7, §C1).

Bidirectional transformation. Foster et al. (2007) established lens laws for well-behaved bidirectional transformations (GetPut, PutGet). Stevens (2010) argued that practical bidirectional transformations must handle non-bijective cases and identified unresolved semantic questions in QVT's approach. The consequence for governance: formal round-trip consistency between narrative documents and structured governance representations is mathematically impossible for the non-bijective transformations governance involves (RA-013 §C3). The resolution is asymmetric: the structured governance representation is canonical; narrative documents are derived views. The tradition proved the mathematical constraint but did not apply it to governance document architecture.

Convergence synthesis. Nine traditions — each using different methods, different vocabularies, and addressing different institutional domains — reach the same structural boundary:

• Decision science (Simon, Kahneman): governance must accommodate bounded rationality — diagnostic framework without infrastructure (RA-010 §F1, §F3).
• Cognitive load theory (Sweller): documentation-as-separate-activity is extraneous load — mechanism identified without governance application (RA-010 §F2, §C6).
• Knowledge management (Nonaka, Szulanski): externalization is the bottleneck, 50–70% failure rates — failure documented without architectural resolution (RA-003 §F1, §F3, §F11).
• Knowledge engineering / CTA (Klein, CommonKADS): extraction methodology exists but domain-specificity unresolved for governance — level distinction not drawn (RA-013 §F3, §F7, §F8).
• Enterprise architecture (Zachman, TOGAF): structural classification without governance representation (RA-013 §F16).
• Audit methodology (AU-C 315, Bell et al.): full governance-surface coverage through cross-domain standardization — understanding methodology without computational instantiation (RA-013 §F9, §F10, §F16).
• Organizational learning (Argyris, March, Zollo & Winter): learning pathologies identified, computational infrastructure requirement stated — no infrastructure provided (RA-014 §F3, §F7, §F11, §C1).
• Semantic web / KR (Gruber, Davis et al., BFO): representation infrastructure mature — no governance instantiation (RA-005 §F5, §F7, §C1).
• Bidirectional transformation (Foster et al., Stevens): round-trip consistency impossible for non-bijective cases — mathematical constraint without governance application (RA-013 §C3).

The convergence validates the boundary rather than any single tradition's authority. The audit tradition's unique position in this convergence is the constructive finding: it is the only tradition that achieves full governance-surface coverage through cross-domain standardization (RA-013 §F16). The externalization path takes this finding and connects it to the cognitive-load diagnosis (why documentation fails) and the representation infrastructure (how externalized governance becomes computational). The convergence does not assert that all nine traditions agree — they use different vocabularies, make different assumptions, and address different problems. What they share is the structural boundary: each has built rich prescriptive apparatus, each has failed to produce infrastructure that makes governance context structural, and none — except the audit tradition's understanding methodology — covers the full governance surface.

The audit tradition's documented record provides the evidential grounding for the externalization path. This is not a literature review of audit methodology — it is the identification of a structural pattern in the profession's century of practice that has not been computationally exploited.

The International Standards on Auditing (ISA 315, Revised 2019) and the equivalent U.S. standard (AU-C Section 315) require auditors to obtain an understanding of the entity and its environment, including: the nature of the entity (industry, regulatory environment, ownership structure, operations, investment activities, financing); the entity's selection and application of accounting policies; the entity's objectives and strategies; the measurement and review of the entity's financial performance; and internal control relevant to the audit. These understanding dimensions are identical across industries — from manufacturing to healthcare to technology to financial services to government. The standardization is structural, not procedural: auditors ask about the same governance dimensions regardless of domain because governance structure is domain-invariant.

IIA Standard 2330 (Documenting Information) requires that internal auditors document sufficient, reliable, relevant, and useful information to support engagement conclusions and results. GAGAS §6.50–6.59 (Documentation) imposes parallel requirements for government auditing. The IIA's Common Body of Knowledge (CBOK) studies, surveying thousands of internal audit practitioners across decades, consistently document challenges in meeting these documentation standards — challenges that the evidence identifies as structural (imposed by the separation-from-work architecture of audit documentation) rather than disciplinary (solvable by better documentation practices within the profession).

The COSO Internal Control — Integrated Framework (2013) codifies internal control as an integrated system of five components (control environment, risk assessment, control activities, information and communication, monitoring activities) and seventeen principles. Organizations adopt the framework formally — establishing control environments, implementing risk assessment processes, deploying monitoring activities — while the operational gap between the adopted framework and the organization's actual governance practice persists. The evidence from the accountability literature (RA-004 §F1) identifies this as Meyer and Rowan's (1977) "ceremonial conformity": the framework is present in the organization's formal structure but decoupled from operational reality.

The structural pattern across these documented records is consistent: the audit profession has developed a comprehensive, cross-domain, standardized methodology for understanding organizational governance, and this methodology works. It has been applied across every industry, every organizational size, every regulatory regime, and every cultural context for over a century. The understanding methodology is validated by practice at a scale no other tradition approaches.

The pattern that has not been computationally exploited is the conversion of this understanding methodology from a professional practice into a computational process. Auditors understand governance structure through structured inquiry — questions about authority, accountability, constraints, evidence, and decision processes. The inquiry is standardized at the governance level (the same dimensions apply across domains) while accommodating domain specificity in the application layer (the specific manifestations of authority, accountability, and constraints vary by industry). This dual-level structure — standardized inquiry, domain-sensitive application — is precisely the architecture that the cognitive task analysis tradition identified as the resolution to the domain-specificity challenge (RA-013 §F7, §F8).

Bell et al.'s (1997) strategic-systems auditing makes the pattern explicit: the auditor constructs a systemic model of the client organization (strategy, economic web, business processes, performance measurement) and uses this model to predict expected outcomes, measuring deviations when actual results differ. This prediction formalism — construct model, predict outcomes, measure deviations — is the governance analog of the Conant-Ashby theorem's model requirement: every good regulator must contain a model of the system it regulates. The auditor's model IS the organizational world model that the governance infrastructure requires. The externalization path does not propose to create something new; it proposes to computationally instantiate something the audit tradition has been doing for a century.

The audit profession has demonstrated, through a century of practice across every domain, that governance-level organizational understanding can be standardized, that it produces reliable results, and that it covers the full governance surface (authority, accountability, constraints, evidence, decision provenance). The cognitive-load evidence establishes that the separation-from-work architecture of traditional documentation is the structural cause of governance-knowledge failure. The nine-tradition convergence establishes that no existing infrastructure makes governance context structural. The question is: why has the profession's documented capacity for governance-level organizational understanding not been computationally externalized? The answer, per the evidence base, is that the audit tradition's methodology has been treated as a professional practice (something auditors do) rather than as an engineering precedent (evidence that governance extraction works and can be formalized). The externalization path reframes the audit tradition's century of practice as the constructive precedent for computational governance infrastructure.

This report's evidence base comprises five research artifacts (RA-003, RA-005, RA-010, RA-013, RA-014) engaging over 175 external sources across nine traditions. The evidence permits asserting: (a) the cognitive-load diagnosis of governance-knowledge failure; (b) the audit tradition's unique position as the only tradition with full governance-surface coverage through cross-domain standardization; (c) the three-stage externalization path as structurally viable; and (d) design requirements for externalized governance methodology. The evidence does not permit asserting: empirical effectiveness of any specific externalization implementation; the optimality of the three-stage path relative to alternatives not examined; or the completeness of the nine-tradition survey.

Open-question chain references: RA-013 §OQ1 (governance-level vs domain-level question boundary) remains open. RA-013 §OQ2 (parse path extraction quality) remains open. RA-014 §OQ1 (single-loop vs double-loop learning threshold) remains open. RA-010 §OQ1 and §OQ4 (CLT and choice architecture applied to governance system design) remain open. RA-005 §OQ1 (SHACL's role in governance constraint validation) remains open.

The externalization path is established through convergence evidence across nine traditions and the audit tradition's documented century of practice. It is not established through a formal impossibility proof (the cognitive-load diagnosis does not prove that documentation-as-separate-activity must fail — it explains why it does fail systematically) or through empirical implementation (no system instantiating the full three-stage path has been built and measured). The convergence evidence establishes structural viability, not optimality or necessity.

Nine traditions were examined. Other traditions that may bear on the externalization question were not engaged: operations research (optimization of governance processes), healthcare informatics (clinical governance documentation), legal informatics (regulatory knowledge representation), and information systems research (system design theory for governance infrastructure). The nine-tradition survey is selective, not exhaustive; the convergence claim is supported by the traditions examined but could be strengthened or qualified by additional traditions.

The externalization path has not been validated through implementation. Validation would require: (a) instantiating the governance-level extraction methodology as a computational process; (b) applying it across multiple domains to test the cross-domain standardization claim; (c) measuring the cognitive-load reduction relative to traditional documentation approaches; and (d) evaluating the quality of the resulting governance representations. This validation is the role of the forthcoming TR-E series.

The author holds the Certified Internal Auditor credential (CIA, Institute of Internal Auditors) and has direct professional experience in the audit and governance assessment domains. The externalization path's claim extends to all governance domains — healthcare, technology, manufacturing, financial services, government — on the basis of the audit tradition's documented cross-domain standardization (AU-C 315 / ISA 315). This extension is supported by the documented record but has not been empirically validated in each domain independently.

The evidence from RA-013 §F9, §F10, §C2, and §C4, combined with RA-010 §F12 and §C1, and RA-003 §F12, strengthens WMI-P11. The audit tradition's controls-testing methodology — the systematic assessment of whether governance mechanisms are designed effectively and operating as intended — is not merely a professional practice. It is the documented precedent for what the externalization path requires: a standardized, cross-domain methodology for extracting governance-level organizational understanding. The century of AU-C 315 / ISA 315 practice demonstrates that the controls-testing methodology works at scale, across every industry, using standardized inquiry dimensions. Bell et al.'s strategic-systems auditing demonstrates that auditors construct organizational world models as standard practice. Klein's RPD model explains the cognitive basis: expert auditors use recognition-primed decisions, not algorithmic procedures. The externalization path computationally instantiates what controls-testing methodology already does — it makes explicit and computational what auditors achieve through professional expertise. WMI-P11 is strengthened because the evidence establishes that the methodology the position names (controls-testing) is the externalization source.

Disposition: strengthens-refines | Evidence base: RA-013 §F9, §F10, §C2, §C4; RA-010 §F12, §C1; RA-003 §F12 | Provenance: plan

The evidence from RA-010 §F6 and §C3, combined with RA-014 §C1 and §C4, and RA-003 §C1, strengthens WMI-P14. The corrective action obligation — the requirement that governance failures trigger structured corrective response — is substantively grounded by the externalization path's cognitive-load diagnosis. The evidence establishes that governance-knowledge failure is architectural (not cultural, technological, or motivational). If the failure is architectural, the corrective action is architectural intervention: changing the system's structure (from documentation-as-separate-activity to governance-context-as-structural-by-product) rather than improving practitioners' documentation discipline. Lerner and Tetlock's (1999) process/outcome accountability distinction, confirmed as NOT applied to AI governance (RA-010 §C3), reinforces the position: corrective action must be process-based (changing the infrastructure that produces governance outcomes) rather than outcome-based (penalizing practitioners for documentation failures produced by architectural inadequacy). WMI-P14 is closing with substrate carry-forward: the founding-period evidence base establishes the obligation; the firm thesis cluster (RA-024 Bhardwaj, RA-025 EY/Broderick) will carry the operational implementation forward.

Disposition: strengthens-refines | Evidence base: RA-010 §F6, §C3; RA-014 §C1, §C4; RA-003 §C1 | Provenance: plan

The evidence from RA-010 §C6, combined with RA-005 §F5 and §F6, and RA-013 §C5, strengthens WMI-P15. The cognitive-load diagnosis reveals that documentation architecture is an ethical choice with direct distributional consequences. When governance documentation is designed as a separate activity, the cognitive burden falls on individual practitioners — typically the most junior or most time-constrained members of the organization. The architecture distributes the burden inequitably: those least able to absorb additional cognitive load bear the heaviest documentation obligations. Gruber's (1993) ontological commitment concept reinforces the ethical dimension: the choice of governance primitives — what the system can "see" and reason about — is not merely a technical design decision but a constitutive act that determines whose governance concerns are visible and whose are structurally invisible. Davis, Shrobe, and Szolovits' (1993) five roles of knowledge representation confirm that governance representation choices constitute a theory of governance: the representation determines what counts as valid governance reasoning, what computation is tractable, and what is human-interpretable. The Modeling View paradigm (Studer et al., 1998) identified in RA-013 §C5 confirms that the ontological commitments embedded in the governance construct schema are the primary design decisions — decisions with direct ethical consequences for who governs, what is governed, and how governance is made visible.

Disposition: strengthens-refines | Evidence base: RA-010 §C6; RA-005 §F5, §F6; RA-013 §C5 | Provenance: plan

The evidence from RA-005 §F5 and RA-013 §C5 provides supporting reinforcement for WMI-P02. Gruber's ontological commitment as a social contract about terminology, and the Modeling View paradigm's identification of ontological commitments as primary design decisions, establish that governance vocabulary is not merely a labeling convention but a constitutive framework that determines the scope and fidelity of governance representation. The externalization path depends on governance vocabulary: what the extraction methodology produces (Stage 2) and what the computational representation encodes (Stage 3) are both governed by the vocabulary's ontological commitments.

Disposition: strengthens-refines | Evidence base: RA-005 §F5; RA-013 §C5 | Provenance: plan