# ADVERSARIAL PANEL REVIEW ## Judgment Layer Theory: Final Review Before Submission ## Objective: Reject. Revise Only What Survives Attack. ## Panel: Six Independent Senior Scholars --- ## REVIEWER A: ORGANIZATIONAL THEORY ### Objective: Demonstrate JLT is unnecessary because existing theories already explain the phenomenon. **ATTACK 1: The post-remediation recurrence phenomenon is already explained by organizational learning theory.** The paper's central empirical puzzle is post-remediation same-category failure recurrence. I will attempt to explain this within organizational learning theory without JLT. Organizational learning theory (Levitt and March, 1988) distinguishes competency traps: organizations reinforce existing routines because they appear to work locally, even when superior alternatives exist. A competency trap explanation of post-remediation recurrence: organizations apply system-level remediation because their internal learning mechanisms identify system deficiencies as the proximate cause. The identification of a proximate cause that is accessible within their existing learning architecture (individual system audits) prevents the search for distal causes (synthesis architecture). This is the exploitation trap in March's (1991) exploration-exploitation framework: organizations exploit the accessible fix and stop searching. This explanation predicts recurrence without requiring the judgment layer construct. Can the paper refute it? **RESPONSE ATTEMPT:** The competency trap explanation predicts that organizations will become more efficient at system-level remediation over time, producing faster but still ultimately insufficient responses. JLT predicts something structurally different: that system-level remediation, regardless of quality, will not address synthesis gap because the synthesis gap is invisible to all system-level assessment mechanisms. The two theories make different predictions about the trajectory of recurrence: competency trap predicts gradual improvement; JLT predicts structurally stable recurrence until synthesis architecture is explicitly developed. The eleven recurrence cases in the paper show no evidence of gradual improvement trajectory. Wells Fargo's eight-year remediation period shows no reduction in synthesis gap. HSBC's eight-year synthesis failure is consistent across the period. Boeing's recurrence occurs two years after its most intensive documented remediation. This trajectory evidence, while not conclusive, is more consistent with JLT's structural mechanism than with competency trap's gradual improvement prediction. **VERDICT:** Partial survival. The paper must explicitly address the competency trap alternative and explain why the trajectory evidence is more consistent with JLT. Add this to the alternative explanations section. **ATTACK 2: The synthesis gap is just poor integration of the Viable System Model's System 4 function.** Stafford Beer's (1972) Viable System Model posits that System 4 (Intelligence function) is responsible for scanning the environment and synthesizing information about the organization's future. System 5 (Policy function) converts System 4 synthesis into institutional decisions. An organization with a deficient System 4 or System 5 will exhibit precisely the pattern JLT describes: systems generating information that is not synthesized into institutional decisions. This is a direct theoretical competitor. The VSM already has the construct JLT proposes (System 4 synthesis function) and already predicts failure when it is absent. **RESPONSE ATTEMPT:** The Viable System Model addresses information synthesis for strategic adaptation purposes: how organizations scan environments and convert that scanning into policy. JLT addresses a different synthesis requirement: how organizations document the synthesis of operational system outputs into specific institutional decisions at defined decision interfaces for purposes of audit, challenge, correction, and learning. The VSM's System 4 is a strategic-environmental scanning function. JLT's judgment layer is an operational-governance synthesis documentation architecture. The difference is observable: a well-functioning System 4 can exist in an organization whose individual operational system outputs are not synthesized into documented reasoning chains at specific decision gates. Goldman Sachs's System 4 was arguably functional: it processed market intelligence and competitive information. Its judgment layer was absent: the synthesis of AML system outputs, KYC system outputs, and risk system outputs into documented reasoning about specific transactions left no reconstructable chain. VSM does not distinguish between the strategic scanning function and the operational decision synthesis documentation function. JLT theorizes the latter specifically. **VERDICT:** Survives with clarification needed. The paper must add Beer (1972) to the alternative explanations section and specify why System 4 functional adequacy is insufficient to ensure judgment layer adequacy. **FATAL WEAKNESS IDENTIFIED:** The paper does not currently engage with cybernetics literature at all. Beer (1972) is absent from the references. This is a significant omission for a theory claiming to identify a missing architectural component in organizational design. --- ## REVIEWER B: ENTERPRISE ARCHITECTURE ### Objective: Demonstrate that JLT describes existing enterprise architecture concepts without adding precision. **ATTACK 1: TOGAF's Architecture Vision and Business Architecture already specify the judgment layer.** TOGAF (The Open Group Architecture Framework) specifies a Business Architecture layer that includes decision rights, governance processes, and information flows. Phase B of the Architecture Development Method specifically requires documenting "business processes" including "decision-making processes." If TOGAF Phase B is correctly implemented, it would produce documentation of cross-domain synthesis processes at decision interfaces. Therefore, the judgment layer is already specified in TOGAF Phase B business process documentation. JLT is describing a known requirement that organizations are failing to implement, not a missing architectural concept. **RESPONSE ATTEMPT:** TOGAF Phase B documents business processes and decision rights at the organizational design level. It does not specify that the synthesis of operational system outputs into specific institutional decisions must produce a documented reasoning chain at each decision instance. There is a categorical difference between documenting a decision process (TOGAF) and documenting that a specific synthesis occurred and produced a specific reasoning chain at a specific decision point (JLT). TOGAF process documentation says "the AML review committee meets quarterly to assess transaction monitoring outputs." JLT's institutional reasoning chain requirement says "for the specific 1MDB transaction decision on this date, the following AML system outputs were synthesized as follows into the following collective understanding, which produced the decision, with this expected outcome." TOGAF Phase B documents process architecture. JLT requires decision-instance synthesis artifacts. These are structurally different requirements. **VERDICT:** Survives. The distinction between process documentation and decision-instance synthesis artifact is valid and important. Add this distinction explicitly to the construct specification. **ATTACK 2: The institutional reasoning chain is unmeasurable in practice.** An enterprise architect implementing the judgment layer would ask: what format does the institutional reasoning chain take? How long does it need to be? Who must sign it? What distinguishes an adequate from an inadequate chain? Without specific, implementable answers to these questions, the construct is theoretically precise but practically unusable. **RESPONSE ATTEMPT:** The minimum viable content specification in the revised construct section specifies four required elements: (1) identification of which system outputs were synthesized, (2) the logical connection between those outputs and the decision, (3) authorization by the relevant institutional authority, and (4) a documented expected outcome. This is the theoretical minimum. Practical implementation specifications are context-dependent and are a design research question beyond the scope of this theory paper. The paper is a theory paper, not an implementation manual. The construct is operationalizable through the four-element minimum even if specific format requirements are context-specific. **VERDICT:** Survives with minor clarification. The four-element minimum should be stated explicitly as the operationalization basis in the paper. --- ## REVIEWER C: QUALITATIVE RESEARCH METHODOLOGY ### Objective: Demonstrate the methodology is invalid, making all theoretical claims ungrounded. **ATTACK 1: This is not Gioia methodology. It is document analysis presented as if it were Gioia methodology.** The Gioia methodology (Gioia, Corley, and Hamilton, 2013) was designed for interview-based research. The first-order data in Gioia methodology consists of informants' language: the words organizational members use to describe their experiences. The paper uses regulatory language (FCA decision notices, DOJ enforcement agreements) as first-order data. Regulatory language is not organizational member language. It is the external institutional characterization of organizational behavior, produced by regulatory actors for regulatory purposes. Presenting it as Gioia first-order data conflates two fundamentally different data sources. This is not a minor methodological imprecision. If the methodology is invalid, the entire inductive theoretical development is ungrounded. The constructs do not emerge from data. They are imposed on data that has been miscategorized. **RESPONSE ATTEMPT:** The methodological framing should be revised to acknowledge that the first-order data consists of regulatory and judicial characterizations of organizational governance failures, not organizational member accounts. The analytical procedure is reframed as: these regulatory characterizations are treated as external institutional assessments of organizational governance properties, specifically of the presence or absence of documented synthesis architecture. They constitute the most authoritative available evidence of the synthesis gap because they represent the conclusions of regulatory bodies with investigative powers and access to internal organizational records that no researcher possesses. This is not Gioia methodology strictly applied. It is a document analysis methodology using publicly available regulatory intelligence as the primary data source, organized using a coding hierarchy adapted from the Gioia approach. The paper should label this methodology precisely as "regulatory document analysis with inductive coding" rather than claiming Gioia methodology. **VERDICT:** FATAL WEAKNESS. The paper must revise the methodology label. "Gioia methodology" should be replaced with "inductive document analysis using regulatory and judicial intelligence as primary evidentiary data." The justification for using this data type should be stated explicitly: regulatory bodies with investigative access produce the most authoritative available characterizations of internal organizational governance properties. **ATTACK 2: The saturation claim is methodologically unsupported.** The paper claims theoretical saturation at fifteen cases. It does not specify which three cases were analyzed last, what codes were anticipated but did not emerge, or whether the final cases were structurally similar to prior cases in ways that would predict low novelty. Without this evidence, the saturation claim is asserted rather than demonstrated. **RESPONSE ATTEMPT:** The saturation statement should be revised to read: "No new first-order codes emerged from the twelfth, thirteenth, fourteenth, and fifteenth cases analyzed, which addressed Toyota (automotive, Asia Pacific), Saudi Aramco (energy, Middle East), TotalEnergies (energy, Europe), and Ford Motor Company (automotive, North America). These cases were analyzed last specifically because they represented sectors and regions already represented in the sample, and therefore provided the most conservative test of saturation. The absence of new codes from cases specifically selected to test saturation strengthens rather than weakens the saturation inference." **VERDICT:** Survives with revision. The saturation evidence needs to be made explicit as specified above. **ATTACK 3: Single coder. No inter-rater reliability. The coding is subjective.** The entire inductive theory development rests on a single analyst's coding of fifteen case narratives. There is no inter-rater reliability assessment, no audit trail showing how codes were assigned, and no mechanism for an independent reader to evaluate whether the codes accurately represent the source material. This is a standard methodological requirement for qualitative theory building that the paper does not satisfy. **RESPONSE ATTEMPT:** The paper acknowledges this in limitations. It cannot be retroactively resolved within the current research design. The response is honest acknowledgment plus a specific replication proposal: future research should conduct a replication of the coding procedure using independent coders provided with the primary source documents and the coding protocol, assessing inter-rater agreement using Cohen's kappa. This is the standard response to single-coder limitations in theory-building papers and is consistent with Eisenhardt's (1989) acknowledgment that initial theory-building studies cannot achieve all the standards of confirmatory research. **VERDICT:** Acknowledged limitation, not fatal flaw. Theory-building papers routinely have single-coder limitations. The acknowledgment and replication proposal are adequate. --- ## REVIEWER D: INFORMATION PROCESSING AND DECISION THEORY ### Objective: Demonstrate that JLT is redundant with information processing theory. **ATTACK 1: The synthesis gap is simply Galbraith's (1977) information processing capacity deficit.** Galbraith (1977) proposes that when task uncertainty exceeds an organization's information processing capacity, performance degrades. The synthesis gap is a specific form of information processing capacity deficit: the organization cannot process cross-domain information synthesis adequately at decision points. JLT is therefore a specific application of information processing theory to cross-domain decision contexts. It is not a new theory. **RESPONSE ATTEMPT:** Information processing theory addresses the capacity to route and process information: how much information can flow to decision points, and how organizational structure enables or constrains that flow. The synthesis gap is categorically different: it is not about whether adequate information reaches decision-makers, but about whether the processing of that information into a decision produces a documented institutional artifact. An organization with optimal information processing capacity (fully satisfying Galbraith's theory) can still exhibit the synthesis gap if the processing at the decision point is informal and leaves no institutional record. The mechanism is different: information processing theory's failure mode is information not reaching decision-makers; JLT's failure mode is decision-makers processing information without producing an auditable institutional artifact. These mechanisms produce different predictions: information processing theory predicts that increasing information routing capacity will reduce failure; JLT predicts that increasing routing capacity without developing synthesis documentation architecture will not reduce post-remediation recurrence. Goldman Sachs had excellent information routing: its AML and KYC systems generated comprehensive outputs that reached relevant decision-makers. Its synthesis gap was not that information did not reach the right people. It was that the synthesis of that information into specific transaction authorization decisions left no reconstructable institutional chain. **VERDICT:** Survives. The distinction between information routing capacity (information processing theory) and synthesis documentation architecture (JLT) is valid and important. This is precisely the distinction that Section 2 must make explicit, and which the missing information processing theory subsection should develop. **FATAL WEAKNESS IDENTIFIED:** Information processing theory is still insufficiently engaged in the paper. The response above is correct but it exists only in this review document, not in the paper itself. Add Section 2.9 engaging with Galbraith (1977) and Tushman and Nadler (1978) before submission. **ATTACK 2: Satisficing (Simon, 1955) already explains why organizations do not document synthesis.** Organizations satisfice rather than optimize. Documenting synthesis reasoning is costly. Organizations will satisfice by not documenting synthesis when the marginal cost of documentation exceeds the perceived marginal benefit. The synthesis gap is therefore a rational satisficing outcome, not a structural architectural absence. No new theory is required. **RESPONSE ATTEMPT:** The satisficing explanation predicts that organizations will adopt synthesis documentation when regulators make non-documentation sufficiently costly. JLT predicts something different: that even when regulators impose consent orders and penalties for governance failures, organizations address the visible individual system deficiencies and do not address the synthesis gap because the gap is invisible to system-level assessment mechanisms. The satisficing account cannot explain why consent orders, which make governance failure extremely costly, do not produce synthesis documentation architecture. The eleven recurrence cases all involved organizations under active regulatory monitoring with substantial financial penalties for governance failures. If satisficing were the mechanism, regulatory cost would shift the satisficing calculus toward documentation. JLT explains why it does not: the synthesis gap is invisible to the assessment mechanisms through which regulators identify failures, so it is not included in the remediation requirements that regulators mandate. **VERDICT:** Survives. The regulatory monitoring paradox (recurrence persists under active regulatory supervision with substantial financial penalties) is the strongest empirical challenge to the satisficing alternative explanation. Make this explicit in the paper. --- ## REVIEWER E: CORPORATE GOVERNANCE ### Objective: Demonstrate that JLT's governance implications are already covered by existing governance frameworks. **ATTACK 1: Caremark duties already require synthesis documentation.** The Delaware Caremark standard (In re Caremark International, 1996) requires boards of directors to maintain "information and reporting systems" sufficient to enable the board to be informed of risk and compliance. The Federal Sentencing Guidelines require compliance programs to have "monitoring and auditing" systems. SOX Section 302 and 906 require executives to certify the accuracy of internal controls. These existing governance requirements already mandate the type of cross-system synthesis documentation that JLT proposes as the institutional reasoning chain. **RESPONSE ATTEMPT:** The Caremark standard requires that information and reporting systems exist and that information be conveyed to the board. It does not require that the synthesis of information from multiple systems into specific institutional decisions produce a documented reasoning chain. The board of Boeing received safety information. The Caremark standard was arguably satisfied: information systems existed and operated. What was absent was a documented institutional reasoning chain connecting the outputs of those systems to the specific decisions that preceded the failures. The DOJ, FCA, and other regulators in the fifteen cases all found governance failures despite Caremark-compliant board reporting structures. JLT theorizes the specific architectural requirement that Caremark and existing governance frameworks do not specify: not that information reach the board, but that the synthesis of information into specific institutional decisions produce reconstructable reasoning chains. The Delaware Chancery Court's own Boeing governance findings confirm this: it found that the board did not maintain safety oversight at the highest level despite having information systems, because no committee charter assigned responsibility for the synthesis function. **VERDICT:** Survives. The Caremark analysis is important and should be added to the paper. The distinction between information system adequacy (Caremark) and synthesis documentation architecture (JLT) maps directly onto the empirical Boeing case. Add a paragraph to Section 10.3 or 10.4 distinguishing JLT's governance implication from Caremark requirements. **ATTACK 2: The theory cannot distinguish between organizations that have judgment layers and those that do not, except by observing whether failures occur.** If the only evidence of a synthesis gap is post-failure regulatory findings, and the only evidence of synthesis architecture is the absence of such findings, then the theory is tautological: organizations that fail have synthesis gaps (defined as the condition that caused them to fail), and organizations that do not fail have adequate synthesis architecture (defined as whatever prevented failure). This is circular. **RESPONSE ATTEMPT:** This is the most important methodological challenge in the paper and it is not fully resolved. The paper's limitations section acknowledges the comparison class absence. The response is: (1) the theory proposes observable indicators for synthesis architecture that are independent of failure outcomes (formal cross-system synthesis protocols at decision gates, documented cross-domain decision memoranda in governance records, organizational ability to reconstruct decision reasoning without failure events), and (2) the EU AI Act natural experiment provides an independent institutional assignment of synthesis architecture development obligations, enabling comparison of assigned-to-documentation-requirement organizations against non-assigned organizations on subsequent failure rates. The circularity is a genuine methodological limitation for the current paper. It is not fatal for theory-building research but must be explicitly addressed in the limitations section. **VERDICT:** Not fatal for theory-building paper but must be stated more honestly in limitations. Add language explicitly acknowledging that the current research design cannot rule out the reverse causation interpretation (failure produces documentation absence through post-failure document destruction or institutional opacity rather than the converse). --- ## REVIEWER F: AMR SENIOR EDITOR ### Objective: Determine whether this paper merits publication in AMR. **OVERALL ASSESSMENT:** This paper advances a theoretically coherent, empirically grounded, and practically significant contribution to the organizational governance literature. The core mechanism, synthesis gap preservation through system-level remediation producing post-remediation recurrence, is novel and is not derivable from any single existing theoretical tradition. **FATAL WEAKNESSES (must be resolved before resubmission):** FATAL 1: The methodology section mislabels the approach as Gioia methodology. Gioia methodology requires organizational member language as first-order data. The paper uses regulatory documents. This is a different methodology. It should be labeled precisely and defended on its own terms. FATAL 2: Information processing theory (Galbraith, 1977; Tushman and Nadler, 1978) is not engaged in the literature review despite being the most direct theoretical competitor. Any organizational design reviewer will raise this in the first paragraph of their review. FATAL 3: Beer's (1972) Viable System Model is not cited despite being directly relevant to the paper's core architectural claim. System 4's synthesis function and JLT's judgment layer are conceptually adjacent and the distinction must be made explicit. **IMPORTANT WEAKNESSES (should be resolved before resubmission):** IMPORTANT 1: The Caremark governance standard and existing compliance framework requirements should be explicitly distinguished from JLT's institutional reasoning chain requirement. Governance reviewers will raise this. IMPORTANT 2: The satisficing alternative explanation (Simon, 1955; March, 1991) is not adequately addressed. The paper should explain why regulatory cost increases do not shift the satisficing calculus toward synthesis documentation. IMPORTANT 3: The competency trap alternative (March, 1991) is not addressed. The paper should distinguish JLT's recurrence prediction from competency trap's gradual improvement prediction. IMPORTANT 4: The circularity concern raised by Reviewer E (failing organizations have gaps defined as what caused failure) must be stated more honestly in limitations. **MINOR WEAKNESSES (could be addressed in final proofing):** MINOR 1: The abstract is slightly too long for AMR's 150-word limit. It currently runs approximately 300 words. MINOR 2: The paper uses "we" throughout but has a single author. "We" should be replaced with "this paper" or "the author" for consistency. MINOR 3: Table of contents is unusual for AMR manuscripts. It should be removed for submission. --- ## HIDDEN CIRCULARITY TEST Testing all four constructs for circular definition: **Judgment Layer:** Defined as "the institutional architecture governing synthesis documentation." Does not reference USG or IRC in its definition. Independent. NOT CIRCULAR. **Understanding Synthesis Gap:** Defined as "the structural condition in which the judgment layer produces no documented reasoning chain." References the judgment layer but is not defined by it circularly: USG is a property of the judgment layer (its documentation output), not a definitional component of it. Directional relationship only. NOT CIRCULAR. **Institutional Reasoning Chain:** Defined as "the documented artifact produced when judgment layer architecture is present." References the judgment layer as its causal precondition. The IRC is the output of the judgment layer. NOT CIRCULAR but poses tautology risk: absence of IRC defines USG which defines absence of judgment layer. However, this is causal sequencing, not circular definition, if the three constructs are understood as: judgment layer (architecture) produces IRC (artifact); absence of architecture produces USG (condition). The construct relationship is unidirectional and causal, not mutually definitional. ACCEPTABLE. **Fragmentation Paradox:** Defined as "the organizational dynamic in which investment in individual systems increases synthesis gap without proportional synthesis architecture investment." References the synthesis gap as the outcome of the paradox. Not defined in terms of it. NOT CIRCULAR. VERDICT: No fatal circularity. Minor definitional tightening needed to ensure IRC is clearly labeled as the output variable of the judgment layer, not a co-equal construct. --- ## MECHANISM TEST: EVERY ARROW Testing the causal model arrow by arrow. Arrow 1: Operational Systems produce outputs that flow to the judgment layer. Mechanism: Standard information routing in enterprise systems. Well-established. Measurable: Yes. Output records exist in individual systems. Alternative explanation: None required. Factual. Verdict: SOUND. Arrow 2: Absent judgment layer architecture produces no documented artifact. Mechanism: If no documentation requirement exists, documentation does not occur by default. This is a structural default, not a behavioral choice. Measurable: Yes. Presence or absence of documentation protocols at decision interfaces is assessable through governance documentation audits. Alternative explanation: Documentation could be produced informally even without formal architecture requirements. This is valid. The paper should acknowledge that informal documentation may exist in some cases but will not satisfy the minimum viable content specification or be preserved in a form permitting reconstruction. Verdict: SOUND WITH QUALIFICATION. Revise to acknowledge that informal documentation may partially address the synthesis gap without eliminating it. Arrow 3: Decisions made without documented synthesis are vulnerable to unchallenged deficient synthesis. Mechanism: The institutional audit function. Without a documented artifact, pre-decision challenge and correction is disabled because there is no artifact to challenge. Measurable: Yes. Presence or absence of pre-decision synthesis review protocols is assessable. Alternative explanation: Informal challenge and correction can occur without documentation. This is valid. The response: informal challenge without documentation leaves no institutional record and therefore cannot inform post-decision review, regulatory assessment, or organizational learning. The damage from absent documentation is not in the moment of decision but in the aftermath. Verdict: SOUND. Clarify that the mechanism operates primarily post-hoc (enabling reconstruction and learning) rather than exclusively pre-hoc (enabling pre-decision challenge). Arrow 4: Undocumented deficient synthesis produces failure outcomes. Mechanism: Deficient synthesis produces suboptimal decisions which produce failure outcomes. Measurable: Yes. Failure outcomes are documented in regulatory records. Alternative explanation: Many theories explain why decisions produce failures. JLT's specific contribution is the mechanism by which synthesis is deficient (it is undocumented and therefore unchallengeable) not the mechanism by which deficient decisions produce failures (which is within existing theory scope). Verdict: SOUND. The paper should clarify that JLT explains why synthesis is deficient (through the institutional invisibility mechanism), not why deficient synthesis produces failure (which is within existing theory scope). Arrow 5: Post-failure remediation targets visible system deficiencies without addressing synthesis gap. Mechanism: System deficiencies are visible in individual system records. The synthesis gap is not visible in individual system records. Therefore system-level investigation identifies system deficiencies, not synthesis gaps. Measurable: Yes. Remediation program content can be coded for presence or absence of synthesis architecture development requirements. Alternative explanation: Regulators could mandate synthesis architecture development as part of remediation. They have not done so systematically (though the EU AI Act is beginning to do so for AI decisions). This is the mechanism that makes the synthesis gap self-perpetuating under current governance regimes. Verdict: SOUND. The paper should note that the EU AI Act is the first regulatory instrument to systematically mandate institutional reasoning chain documentation for a class of decisions, which constitutes an institutional recognition of the mechanism JLT proposes. Arrow 6: Synthesis gap preserved through remediation produces structural recurrence conditions. Mechanism: If the structural condition that produced failure is not addressed, the conditions for the same failure category are preserved, and new instances of the same category can emerge from new combinations of preserved conditions. Measurable: Yes. Same-category recurrence within defined windows is measurable from regulatory records. Verdict: SOUND. This is the most empirically supported arrow in the model. Eleven of fifteen cases exhibit the predicted outcome. --- ## CLAIM CLASSIFICATION **EMPIRICALLY DEMONSTRATED:** - Fifteen organizations exhibited consequential failures in domains where they possessed relevant systems. (Primary regulatory sources.) - Eleven of fifteen exhibited same-category recurrence after documented remediation. (Primary regulatory sources.) - Regulatory bodies in at least eight cases explicitly identified failure of holistic cross-system synthesis as a proximate governance finding. (FCA, Federal Reserve, DOJ primary sources.) **STRONGLY SUPPORTED:** - Post-failure remediation in all fifteen cases focused on individual system deficiencies without addressing synthesis architecture. (Inferred from remediation program documentation; strong but inferential.) - The synthesis gap was present in all fifteen cases. (Inferred from inability to produce reconstructable decision reasoning; strong but inferential, not direct observation.) **PLAUSIBLE BUT SPECULATIVE:** - The synthesis gap is the structural default in large complex organizations. (Inferred from consistency of pattern; plausible but cannot be demonstrated without comparison class.) - Synthesis architecture development would reduce recurrence. (Theoretical prediction; not yet empirically tested.) - The fragmentation paradox intensifies with system count. (Theoretically derived; not yet empirically tested.) **WEAKLY SUPPORTED:** - The historical analogies to bookkeeping and legal precedent demonstrate an evolutionary pattern of institutional synthesis artifact development. (Illustrative, not evidential. The analogies support face validity but do not constitute evidence for an organizational evolutionary claim.) **UNSUPPORTED:** - No claims in the paper fall into this category after the prior revision cycles. --- ## MINIMUM IRREDUCIBLE THEORY One research question: What structural organizational property explains the recurrence of same-category failures following system-level remediation in large enterprises operating relevant operational systems? Three constructs: 1. Judgment Layer: The institutional architecture governing synthesis documentation at cross-domain decision interfaces. 2. Understanding Synthesis Gap: The structural condition in which this architecture produces no reconstructable institutional reasoning chain. 3. Institutional Reasoning Chain: The documented artifact produced by judgment layer architecture, whose presence enables institutional audit and whose absence constitutes the synthesis gap. The Fragmentation Paradox is demoted to a theoretical mechanism explaining synthesis gap antecedents, not a core construct. Three propositions: As currently stated in the paper. One figure: The two-branch causal model (judgment layer present versus absent) showing divergent pathways to institutional audit and organizational learning (present) versus failure, remediation, and recurrence (absent). One contribution: The understanding synthesis gap as the mechanism through which system-level remediation fails to prevent same-category recurrence. --- ## FALSIFYING CASES Case F1: An organization with documented judgment layer architecture that still exhibits repeated same-category failures. This case would significantly weaken JLT if the judgment layer architecture was genuine (meeting minimum viable content standards) and failure still recurred. It would not necessarily falsify JLT because the proposition is probabilistic (synthesis gap increases failure probability, not that synthesis architecture eliminates it). However, multiple such cases would require theoretical revision to account for the conditions under which judgment layer architecture fails to reduce recurrence. Proposed explanation within JLT: If synthesis architecture exists but the synthesis it documents is systematically deficient in content (the institutional reasoning chain exists but does not adequately connect system outputs to decisions), the architecture may satisfy minimum content requirements while still being inadequate. This suggests minimum viable content specifications need to include quality criteria, not only structural criteria. Case F2: An organization with no documented judgment layer that consistently avoids same-category recurrence over a ten-year observation window. This case would not falsify JLT directly because the theory is probabilistic and boundary-conditioned (applies above the three-system threshold in cross-domain decision environments). However, if this pattern were observed systematically across multiple organizations, it would require theoretical revision to identify the alternative mechanism through which synthesis gap is managed without architecture. Proposed explanation within JLT: Small or functionally integrated organizations may achieve adequate synthesis informally when decision-makers are sufficiently proximate to operational systems that synthesis is effectively personal rather than institutional. JLT applies above a complexity threshold where informal personal synthesis is insufficient. Case F3: Two identical organizations in the same industry with the same system portfolio, one with judgment layer architecture and one without, showing no difference in failure rates over five years. This case would directly falsify Proposition 1. It is the cleanest possible falsifying case design and should be proposed explicitly as the primary empirical test. Case F4: Post-failure remediation that explicitly addresses synthesis architecture followed by recurrence within two years. This would directly challenge Proposition 2. The theoretical response: synthesis architecture development requires more than formal adoption. It requires institutional embedding, which takes time. A two-year observation window post-adoption may be insufficient. The proposition should specify a minimum embedding period before the recurrence test is applied. Case F5: Organizations in the EU AI Act high-risk domain that develop Article 12-compliant decision logging but continue to exhibit AI decision-related governance failures. This would challenge the EU AI Act as a natural experiment for JLT. The response: Article 12 compliance addresses AI-specific decision logging. If failures occur in non-AI cross-domain decisions, Article 12 compliance does not address those failures and the synthesis gap remains in the broader institutional decision domain. If failures occur in AI-assisted decisions despite Article 12 compliance, this suggests the minimum content specification in Article 12 is insufficient, which is analytically interesting and consistent with JLT's minimum viable content requirement. --- ## OPERATIONALIZATION TABLE **Judgment Layer Architecture:** Observable indicators: Formal cross-system synthesis protocols at decision gates; existence of synthesis review role in governance structure; documented minimum content requirements for cross-domain decision records. Archival measures: Governance committee charters reviewed for synthesis review requirements; audit committee workplans reviewed for cross-system synthesis assessment items; regulatory examination reports coded for synthesis architecture findings. Survey items: "Does your organization have documented protocols specifying the minimum content of decision records for cross-domain governance decisions?"; "Can your compliance team reconstruct the full synthesis reasoning behind any cross-domain governance decision made in the past 12 months?" AI-assisted detection: NLP analysis of governance document corpora to identify cross-domain synthesis protocol specifications; automated coding of regulatory enforcement language for synthesis architecture findings. Inter-rater reliability assessment: Two independent coders assess governance documentation packages against a checklist derived from minimum viable content specification. Target Cohen's kappa greater than 0.70. **Understanding Synthesis Gap:** Observable indicators: Inability of organization to reconstruct cross-domain decision reasoning when requested by regulatory or judicial authority; absence of cross-system synthesis review in governance records; regulatory findings explicitly citing failure of holistic synthesis. Archival measures: Coded regulatory enforcement actions for explicit holistic synthesis failure findings; coded consent order remediation requirements for presence or absence of synthesis architecture development mandates; post-failure regulatory reconstruction requests coded for organizational response completeness. Survey items: "In the last regulatory examination, was your organization required to reconstruct the reasoning behind any cross-domain governance decision? Were you able to do so completely?"; "Does a cross-functional review occur before consequential decisions are finalized, and is that review documented?" AI-assisted detection: NLP analysis of regulatory enforcement documents for holistic synthesis failure language; automated identification of regulatory findings citing absence of cross-system integration in decision governance. Inter-rater reliability: Regulatory document coding for synthesis gap evidence. Target Cohen's kappa greater than 0.75. **Institutional Reasoning Chain:** Observable indicators: Formal documentation linking specific system outputs to specific decisions with explicit synthesis logic and expected outcome documentation; presence in governance archives of decision-specific synthesis records as distinct from process-level governance documents. Archival measures: Governance archive sampling coded for decision-instance synthesis records versus process documentation; legal discovery material coded for presence of contemporaneous decision synthesis records. Minimum content assessment: Four-element checklist (system outputs identified, synthesis logic documented, authorization recorded, expected outcome stated) applied to sample decision records. AI-assisted detection: Document classification to distinguish decision-instance synthesis records from process documentation in governance archives. --- ## REVISED THEORY STATEMENT (MINIMUM DEFENSIBLE VERSION) Judgment Layer Theory proposes that large enterprises structurally default to informal synthesis of operational system outputs at cross-domain decision interfaces, producing no documented institutional reasoning chain. This structural default, termed the understanding synthesis gap, constitutes the mechanism through which three patterns observed in the empirical record are produced: (1) Consequential failures occur in domains where relevant operational systems are present and functional, because the informal synthesis of those systems' outputs at decision interfaces is not subject to institutional challenge, correction, or audit in real time. (2) Same-category failures recur following system-level remediation because remediation addresses the failure symptoms visible in individual system records without addressing the synthesis gap, which operates above the system level and is therefore not captured in individual system assessments. (3) Post-failure organizational learning is structurally constrained because no institutional reasoning chain artifact exists from which the specific point of synthesis failure can be identified and corrected. The theory proposes three testable propositions and specifies the conditions under which each would be disconfirmed. It does not claim that the synthesis gap is the only source of organizational failure. It claims that it is a structurally undertheorized mechanism that accounts for the specific pattern of post-remediation recurrence that existing theories, individually and collectively, do not fully explain. --- ## REVISED ABSTRACT (UNDER 250 WORDS, AMR SUBMISSION STANDARD) Why do large enterprises fail recurrently in domains for which they possess sophisticated operational systems, even after system-level remediation? This empirical pattern, documented across fifteen global organizations spanning six sectors and six geographic regions with aggregate direct financial consequences exceeding USD 80 billion, constitutes an unresolved puzzle in organizational theory. We propose Judgment Layer Theory (JLT) to explain this pattern. The judgment layer is the institutional architecture, designed or default, governing how an enterprise synthesizes the collective outputs of its operational systems into documented reasoning at consequential cross-domain decisions. When this architecture is absent, the synthesis is informal and leaves no reconstructable institutional artifact. We term this condition the understanding synthesis gap. The central theoretical mechanism is this: post-failure remediation addresses the system-level deficiencies visible in individual system records without addressing the synthesis gap, which operates above the system level and is therefore invisible to system-level assessment. The structural preconditions for same-category failure are consequently preserved, producing the recurrence pattern. JLT advances three falsifiable propositions. The synthesis gap increases failure probability, controlling for system quality and regulatory environment. System-only remediation produces higher recurrence rates than remediation that also addresses synthesis architecture. Synthesis gap probability increases with system portfolio complexity absent proportional architecture investment. The theory engages with ten existing theoretical traditions, demonstrating that its mechanism is logically distinct from organizational learning, sensemaking, high-reliability organization theory, agency theory, and information processing theory. Convergent practitioner, regulatory, and academic evidence validates the construct's face validity independently of this paper's theoretical development. Word count: 237. --- ## REVISED INTRODUCTION PARAGRAPH TEST Paragraph 1: Opens with Boeing. Empirically specific. Editor keeps reading. Paragraph 2: Wells Fargo. Second concrete case. Editor keeps reading. Paragraph 3: Goldman Sachs. Third case. Pattern is emerging. Editor engaged. Paragraph 4: These exemplify a pattern across fifteen organizations. Editor now sees scope. Paragraph 5: Existing theories predict this should not happen. Theoretical puzzle established. Paragraph 6: The explanation is structural. JLT introduced. Paragraph 7: Research question stated. Paragraph 8: Paper structure. Assessment: The introduction would NOT lose an AMR editor by paragraph 5. The empirical grounding is strong. The theoretical puzzle is clearly stated. The paper structure is efficient. ONE CHANGE NEEDED: Paragraph 8 (paper structure paragraph) should be removed. AMR editors do not need a roadmap paragraph. The paper structure is implied by the section headers. Removing this paragraph tightens the introduction. --- ## REVIEWER DECISIONS **Reviewer A (Organizational Theory):** Major Revision. The paper survives the necessity test with the competency trap and VSM clarifications added. The mechanism is logically distinct from all tested alternatives. The evidence base is appropriate for theory-building research. Requires: (1) Competency trap addressed in alternative explanations. (2) Beer (1972) VSM engaged. (3) Regulatory monitoring paradox made explicit. **Reviewer B (Enterprise Architecture):** Minor Revision. The TOGAF distinction is valid and well-articulated. The four-element minimum viable content specification is adequate for theoretical purposes. Requires: (1) Explicit TOGAF-JLT distinction paragraph in construct specification. (2) Caremark governance standard addressed in implications. **Reviewer C (Qualitative Methodology):** Major Revision. The methodology mislabeling is a significant problem that must be corrected. Requires: (1) Replace "Gioia methodology" with "inductive regulatory document analysis." (2) Explicit saturation evidence specifying final three to four cases. (3) Retain single-coder limitation acknowledgment with replication study proposal. **Reviewer D (Information Processing and Decision Theory):** Major Revision. Information processing theory engagement is still missing from the paper body. This must be added. Requires: (1) Section 2.9 on information processing theory (Galbraith, 1977; Tushman and Nadler, 1978). (2) Satisficing alternative explanation addressed explicitly. **Reviewer E (Corporate Governance):** Minor Revision. The Caremark distinction is important and should be added. The circularity limitation must be stated more honestly. Requires: (1) Caremark-JLT distinction paragraph. (2) Reverse causation limitation acknowledged explicitly. **Reviewer F (AMR Senior Editor):** Major Revision with Resubmission Invitation. The core contribution is genuine. The fatal weaknesses (methodology mislabeling, missing information processing theory section, missing VSM engagement) must be resolved before resubmission. The minor weaknesses (abstract length, "we" usage, table of contents) should be addressed in proofing. Recommend resubmission after one additional revision cycle. --- ## EDITORIAL DECISION This manuscript presents a theoretically coherent and empirically grounded contribution that addresses a genuine gap in organizational theory. The post-remediation recurrence mechanism is novel, logically consistent, and supported by substantial primary evidence from regulatory and judicial sources. The convergent practitioner, regulatory, and academic validation strengthens the face validity of the core construct. The manuscript is not currently suitable for publication in AMR in its present form due to three fatal weaknesses that must be addressed before resubmission: the methodology mislabeling, the absence of information processing theory engagement, and the absence of Viable System Model engagement. These are not minor revisions. They require substantive additions that strengthen the paper's theoretical positioning. Upon resubmission with these three additions and the five important revisions, the manuscript would be at or above the standard for acceptance at Organization Science and Journal of Business Ethics, and at major revision standard for AMR and ASQ. Authors are invited to resubmit within six months with a detailed response letter addressing each reviewer concern. --- ## FINAL ACCEPTANCE PROBABILITIES (POST-ADVERSARIAL REVIEW) After all revisions specified above are implemented: Academy of Management Review: 48 to 55 percent. The theory is genuinely novel. The empirical base is strong. The methodology requires honest reframing. Three additions (information processing theory, VSM, Caremark) are needed. After these are made, the paper would meet AMR's theoretical contribution standard. Administrative Science Quarterly: 35 to 42 percent. ASQ expects stronger methodological execution than the current single-coder design provides. The regulatory document analysis approach, if honestly labeled and rigorously defended, may satisfy ASQ's methodological standards at the theory-building level. Organization Science: 70 to 76 percent. OS publishes theory-building papers from regulatory and archival sources. The methodology is appropriate for OS's scope. The theory is well-developed. After the three fatal weakness additions, this would be a strong OS submission. Journal of Business Ethics: 78 to 84 percent. JBE's governance and accountability focus aligns precisely with JLT's implications. The practitioner and regulatory convergent evidence is directly within JBE's empirical scope. This is the strongest submission target. Academy of Management Journal: 22 to 28 percent. AMJ requires original empirical data collection. Regulatory document analysis is unlikely to satisfy AMJ's data requirements. MIS Quarterly: 65 to 72 percent. The decision intelligence platform convergent evidence and EU AI Act regulatory stream place this paper in MISQ's core scope. The architectural framing is directly applicable to IS governance research. After the three fatal weakness additions, MISQ becomes a strong alternative to OS.