Part IV - INTEGRITY: a unified field theory of behaviour, coherence, and structural collapse

Part IV – INTEGRITY A unified field theory James D. Atkinson 2025 Abstract This paper introduces a unified mathematical framework in which behaviour is represented as a state field b∶𝑋×ℝ≥0 →ℝ𝑑 evolving under stochastic gradient dynamics generated by the integrity functional 𝔉[b]=𝑈[b]−ℵ𝑆𝐼[b], where 𝑈[b]is a structural tension functional, 𝑆𝐼[b]is an integrity–dispersion functional, and ℵ≈2.70is an empirically estimated integrity constant. Under admissible update rules the evolution is monotone:  𝔉≤0, so 𝔉acts as a Lyapunov-style quantity governing structural relaxation. A conserved structural quantity emerges from the continuity relation of the integrity current, I=𝐼∗+𝐻norm, which remains approximately stable across epistemic, procedural, and institutional transformations under declared assumptions. The curvature spectrum of 𝑈[b]partitions systems into three operational coherence regimes, defined purely by the sign and magnitude of the principal curvatures. Simulations confirm the model’s predictions: contradiction generates high structural tension, symmetric update rules collapse effective degrees of freedom, and institutional dynamics exhibit drift proportional to curvature imbalance. The resulting theory establishes integrity as a measurable structural invariant characterised by ℵtogether with curvature, dispersion, and invariance constraints. It provides a 1 single representation-invariant formalism for modelling coherence, deviation, and structural stability across heterogeneous decision systems. Keywords: integrity constant ℵ; structural invariant; integrity functional; curvature spectrum; dispersion functional 𝑆𝐼; coherence regimes; stochastic gradient flow; contradiction load; divergence analysis; structural integrity current; Hessian spectrum; institutional drift; integrity potential 𝑈[b]; invariant reconstruction; structural coherence metrics. 2 Contents 1 Notation and symbols 9 2 Preface 11 2.1 A structural question, not a moral one .................. 11 2.2 From three papers to one structural law ................. 12 3 Summary 12 3.1 Core insight ................................. 13 4 Introduction 13 5 Background and motivation 16 5.1 Contradiction: epistemic integrity .................... 16 5.2 Symmetry: procedural integrity ...................... 17 5.3 Adaptation: institutional integrity ..................... 17 5.4 Why a unified theory is needed ...................... 18 5.5 A three-layer framework for integrity ................... 18 6 The conservation of information 18 6.1 Structural information and epistemic entropy .............. 19 6.2 Complementarity under drift ....................... 19 6.3 The informational conservation law .................... 20 6.4 Operational test of approximate conservation .............. 21 6.5 Interpretation across the trilogy ..................... 22 6.6 Relation to broader information laws ................... 22 6.7 Implications ................................. 22 7 Empirical validation via simulated dynamics 22 7.1 The truth-limit surface ........................... 23 7.2 Local behaviour dynamics ......................... 25 7.3 Monte Carlo dynamics across the three regimes ............ 29 7.4 Exponential relaxation under curvature ................. 29 7.5 Integrity radar ................................ 30 7.6 Free-energy decline ............................ 30 7.7 Sensitivity analysis ............................. 31 7.8 Correspondence with real institutional systems ............ 31 7.9 Summary ................................... 31 8 Empirical discovery of the integrity constant 32 8.1 Scale-dependent intensity ......................... 32 3 8.2 Scale flow and the fixed point ....................... 33 8.3 Metric inference via scale stationarity .................. 34 8.4 Spatial geometric manifestation ..................... 34 8.5 Synthesis ................................... 35 9 Covariant formulation of integrity 38 9.1 Why representation invariance is necessary ............... 39 9.2 Hypothesis space as a structured domain ................ 40 9.3 Covariant gradient flow ........................... 41 9.4 Curvature as a diagnostic of representational distortion ........ 41 9.5 Boundary terms and ledger constraints ................. 42 9.6 Affine institutional reparametrisation .................. 42 9.7 Nonlinear reframing and adversarial distortion ............. 43 9.8 Integrity currents and divergence ..................... 43 9.9 Worked example: policy-induced coordinate shift ........... 44 9.10 Robustness under noise .......................... 44 9.11 What covariance does not claim ...................... 44 9.12 Summary ................................... 45 10 Operational Noether currents and invariance diagnostics 45 10.1 The action and its admissible symmetries ................ 46 10.2 Deriving the integrity current ....................... 46 10.3 Interpretation ................................ 47 10.4 Example: permutation symmetry in ordered coherence ........ 47 10.5 Example: admissible reinterpretation in adaptive coherence ..... 48 10.6 Operational measurement of integrity currents ............. 48 10.7 Current patterns across the three integrity phases ........... 48 10.8 Symmetry breaking and divergence morphology ............ 49 10.9 Spatial divergence fields .......................... 49 10.10 Connection to ledger boundary terms .................. 50 10.11 Robustness under stochastic uncertainty ................ 50 10.12 Summary ................................... 50 11 The Theory of General Integrity 51 11.1 From geometry to trajectory ........................ 51 11.2 Paths of least dissipation .......................... 52 12 Integrity currents: the dynamical signature of symmetry, drift, and distortion 53 12.1 Tilt-only currents .............................. 54 12.2 Drift-only currents ............................. 55 12.3 Tilted curvature + drift ........................... 55 4 12.4 Quantifying distortion: divergence, curl, and coherence flow ..... 55 12.5 Unified diagnostic of integrity dynamics ................. 56 13 Stochastic dynamics of integrity: drift, diffusion, and field evolution 57 13.1 Drift: directional forces in the integrity field ............... 59 13.1.1 Drift velocity and directional bias ................. 59 13.1.2 Drift as free-energy minimisation (institutional form) ..... 60 13.1.3 Phase interpretation of drift .................... 60 14 The integrity interaction: drift, diffusion, and deviation 61 14.1 Interpretive notes .............................. 62 14.2 Local drift dynamics ............................ 63 14.3 Diffusion: institutional dispersion ..................... 63 14.4 Deviation flux: structure-breaking pressure ............... 64 14.5 Contradiction as discrete collapse .................... 64 14.6 Integrity singularities under curvature accumulation .......... 64 14.7 Coherence field: stabilisation and symmetry restoration ....... 66 14.8 Sentinel forcing: background structure ................. 66 14.9 Master evolution equation ......................... 66 15 The integrity interaction: drift, diffusion, and deviation 68 15.1 Interpretive notes .............................. 69 15.2 Local drift dynamics ............................ 69 15.3 Diffusion: institutional dispersion ..................... 70 15.4 Deviation flux: structure-breaking pressure ............... 70 15.5 Contradiction as discrete collapse .................... 71 15.6 Integrity singularities under curvature accumulation .......... 71 15.7 Coherence field: stabilisation and symmetry restoration ....... 72 15.8 Sentinel forcing: background structure ................. 72 15.9 Master evolution equation ......................... 72 16 The full integrity Lagrangian 73 16.1 Kinetic terms ................................ 73 16.2 Imass and baseline curvature ....................... 73 16.3 Self-potentials and phase structure ................... 73 16.4 Interaction sector .............................. 73 16.5 Effective imass and phase interpretation ................ 74 17 Symmetry breaking and the effective integrity potential 74 17.1 Order parameters for symmetry ...................... 74 17.2 Scalar integrity potential .......................... 75 17.3 Spontaneous symmetry breaking ..................... 75 5 17.4 Empirical symmetry-breaking signatures in the divergence field . . . 75 17.5 Effective imass under symmetry breaking ................ 76 17.6 Case-study interpretation ......................... 77 17.7 Interpretive summary ........................... 77 18 Synthesis and contributions 78 18.1 Synthesis across the trilogy ........................ 78 18.2 Contributions of the unified theory .................... 79 18.3 Conceptual significance .......................... 80 18.4 Falsifiable predictions ........................... 80 18.5 Contribution to the literature ....................... 81 19 Limitations and scope 81 19.1 Conceptual and ontological scope .................... 82 19.2 Mathematical idealisation ......................... 82 19.3 Empirical and simulation limitations ................... 83 19.4 Interpretive boundaries .......................... 83 19.5 Boundary of applicability .......................... 84 19.6 Invariance assumptions .......................... 84 19.7 Computational constraints ......................... 84 19.8 Philosophical and epistemic considerations ............... 85 19.9 Summary ................................... 85 20 Failure modes and the stability frontier 85 20.1 Dispersion overload: loss of structural recoverability ......... 86 20.2 Frozen symmetry: rigidity through loss of adaptive correction . . . . 86 20.3 Curvature instability ............................ 87 20.4 Boundary loss and open-system failure ................. 87 20.5 Interpretation: the stability frontier ................... 88 21 Application case study: integrity dynamics in a symmetry-constrained allocation system 88 21.1 Problem setting ............................... 89 21.2 Methodological framework ......................... 89 21.3 Data and preprocessing .......................... 90 21.4 Phase identification ............................. 90 21.5 Curvature diagnostics ........................... 91 21.6 Dispersion diagnostics ........................... 91 21.7 Integrity current and drift analysis .................... 91 21.8 Admissible transformation test ...................... 92 21.9 Free-structure trajectory .......................... 92 21.10 Conservation of information ........................ 93 6 21.11 Interpretive summary ........................... 93 21.12 Generality and extension .......................... 93 21.13 Conclusion of case study .......................... 94 22 Application case study: integrity diagnostics in an outcome-first redundancy scoring system 94 22.1 Problem setting ............................... 94 22.2 Contradiction diagnosis .......................... 95 22.3 Symmetry analysis ............................. 95 22.4 Outcome-first optimisation ........................ 96 22.5 Invariance test ............................... 96 22.6 Interpretive summary ........................... 97 23 Case study: restructure of one 97 23.1 Declared rationale ............................. 97 23.2 Public narrative ............................... 97 23.3 Observed behaviour ............................ 98 23.4 Parity proposition and contradiction ................... 98 23.5 Symmetry and curvature .......................... 98 23.6 Invariance and drift ............................. 99 23.7 Conclusion .................................. 99 24 Outlook and future work 99 24.1 From simulation to observation ...................... 99 24.2 The integrity tensor .............................100 24.3 Integrity analytics infrastructure .....................101 24.4 Experimental roadmap ...........................101 24.5 Ethical and governance implications ...................101 24.6 Long-term vision ..............................102 24.7 Summary ...................................102 25 General conclusion 103 25.1 Integrity as a structural invariant .....................103 25.2 The integrity field of behaviour ......................103 25.3 Covariant extension .............................104 25.4 Integrity horizons ..............................104 25.5 Applied case study: procedural collapse .................104 25.6 Noether currents and diagnostics .....................104 25.7 Phases of integrity .............................105 25.8 Empirical validation .............................105 25.9 Limitations ..................................105 25.10 Contribution to research and practice ..................105 7 25.11 Future directions and the Integrity Tensor ................106 25.12 Final reflections ...............................106 A Structural contradiction audit 106 A.1 Framed proposition test ..........................107 A.2 Contradiction topology ...........................108 A.3 Contradiction accumulation ........................108 A.4 Integrity horizon ...............................109 A.5 Integrity singularity .............................109 B Role Equivalence Model (REM) 109 C Procedural integrity collapse 110 8 1 Notation and symbols Symbol Meaning Behaviour Field and Dynamics b(𝑥,𝑡) Behaviour field over hypothesis space 𝑋at time 𝑡. b∗Equilibrium (structurally coherent) behaviour configuration. 𝛿b Perturbation around equilibrium: 𝛿b=b−b∗. ∇𝑖Covariant derivative with respect to coordinate 𝑥𝑖. 𝜕𝑖Partial derivative with respect to 𝑥𝑖(non-covariant). 𝜉𝑖Generator of an admissible transformation 𝑓𝜖. ΔDiscrete Laplacian operator used in simulations. ℌHessian of the free-energy functional (curvature operator). Free Energy and Structural Quantities 𝔉[b]Free energy of the behaviour field. 𝑈[b]Internal energy (curvature-based structural tension). 𝑆𝐼[b]Integrity entropy (behavioural uncertainty/disorder). ΘEffective temperature controlling sensitivity to variation. 𝐼[b]Integrity functional: 𝐼[b]=−𝔉[b]. 𝐼∗(𝑡) Composite integrity index (empirical, windowed estimator). ℵEmpirical integrity constant (RG fixed point). 𝐸/𝑇 Empirical entropy–temperature ratio used in simulations. Curvature, Stability, and Resilience 𝛼Spectral curvature: smallest eigenvalue of ℌ. 𝜆min(ℌ) Minimum Hessian eigenvalue (local resilience measure). 𝜅Decay constant in exponential relaxation 𝑒−𝜅𝑡. SO Structural order proxy: 1/(1+𝐸/𝑇). Entropy, Disorder, and Conservation 𝐻norm Normalised entropy of the behaviour distribution. 𝑄(𝑡) Conservation-like relation: 𝑄(𝑡)=𝐼∗(𝑡)+𝐻norm(𝑡). Free-Energy Expansion 𝛿bℌ𝛿b Quadratic curvature contribution in a local expansion. O(𝛿b3)Higher-order terms in the free-energy expansion. Symmetry, Invariance, and Covariance 𝑋Hypothesis space (manifold). 𝑓∶𝑋→𝑋 Admissible institutional transformation. 9 • symmetry suppresses drift (Atkinson, 2025c); • invariance governs legitimate adaptation (Atkinson, 2025b). This paper proposes a positive unification: we model a behaviour field b(𝑥,𝑡)evolving under noisy gradient dynamics on a free-energy-style functional. Curvature quantifies resilience; dispersion quantifies uncertainty; and symmetry/invariance quantify legitimacy. Contributions. This paper advances the trilogy by: 1. introducing a unified integrity functional combining curvature and dispersion; 2. establishing integrity as the negative of this free-energy-style functional; 3. mapping the three prior papers to distinct operational regimes; 4. providing simulation evidence across contradiction, convergence, and adaptive settings; 5. developing a representation-invariant formulation for institutional change; 6. and identifying falsifiable predictions and model limitations. The contributions are structural, testable, and explicitly tied to stated assumptions. 5 Background and motivation The trilogy preceding this work developed three complementary models of integrity— each addressing a distinct level of analysis—whose behaviours point to a shared structural substrate (Atkinson, 2025a, 2025b, 2025c). 5.1 Contradiction: epistemic integrity Paper 1 formalised contradiction as a diagnostic event. Given a stated rationale R, narrative N, and observed behaviour B, any configuration in which they cannot be jointly maintained incurs a measurable coherence cost (Atkinson, 2025a). In a local linearisation around a coherence-legitimate behaviour b∗, let ℌdenote the local stability operator (Hessian) and 𝛼 = 𝜆min(ℌ)its smallest eigenvalue. In states 16 where the integrity potential is approximately quadratic, relaxation toward coherence is governed by 𝛼: higher curvature implies faster recovery. Empirically, contradiction states behave like noisy recovery processes near unstable equilibria—well approximated by Ornstein–Uhlenbeck-like dynamics—without asserting any physical equivalence (Risken, 1996; Uhlenbeck & Ornstein, 1930). This links epistemic contradiction to measurable curvature and dissipation. 5.2 Symmetry: procedural integrity Paper 2 analysed fairness under symmetry as a stochastic, self-correcting procedure. When sampling is symmetric and noise is unbiased, allocation proportions converge (in the appropriate sense) toward declared weights, with residual drift disciplined rather than ignored (Atkinson, 2025c; Hardt et al., 2016). Two scale quantities matter operationally: • a tranche-level KL noise floor 𝔼[C𝜏]=(𝑛−1)/(2T ln 2); • a minimum scale 𝐾min ≈ 1000𝑛for strong claims, with small-𝐾variants for lowvolume settings. Under these assumptions, symmetry flattens curvature in the integrity potential and yields publicly testable convergence rather than unexamined assertion. 5.3 Adaptation: institutional integrity Paper 3 generalised the analysis to institutions where full symmetry is rare but bounded asymmetry is realistic. Adaptive coherence introduced: • admissible transformation families (sentinel tests), • anytime-valid e-processes for drift detection, • and a tamper-evident ledger for auditability. In this setting, integrity is invariance: outputs remain stable under the pre-declared transformation battery. Adaptive coherence enables institutions to adjust while maintaining declared commitments, preventing silent reinterpretation of their own rules (O’Neill, 2002; Power, 1997). 17 5.4 Why a unified theory is needed Despite their different domains, the three models share the same structural ingredients: stochastic forcing, curvature-driven relaxation, symmetry and invariance constraints, and boundary evidence. What differs is terminology, not the underlying structure. This paper introduces a unified integrity field model that reproduces—and links— these behaviours: a behaviour field b(𝑥,𝑡), an integrity potential 𝑈[b], an effective temperature Θ, and a free-energy-style functional 𝔉[b]whose tolerant monotone decline provides a single account of coherence and stability. A conservation-style relation between integrity and entropy, together with a representation-invariant (covariant) extension, unifies the trilogy’s observables. Assumptions and falsifiable predictions are stated explicitly later in the paper. 5.5 A three-layer framework for integrity The unified theory is organised into three layers: •Integrity Field Theory: equilibrium structure via curvature, dispersion, and free energy. •Integrity Mechanics: temporal dynamics of b(𝑥,𝑡)under noise, drift, contradiction events, and sentinel forcing. •General Integrity: invariance under admissible representation changes; geometric objects such as metrics, connection terms, and conserved currents. This division mirrors structural roles in general dynamical modelling. The framework identifies mathematical isomorphisms without attributing physical substrates to social or institutional processes. 6 The conservation of information The integrity field framework has one unavoidable implication: integrity and epistemic entropy cannot wander off in independent directions. Once the free-energy structure is fixed, the two quantities share a common informational budget. This section makes that claim precise. We introduce the closed-system identity, motivate the empirically observed constant of integrity, derive the corrections needed in open systems, and 18 give an operational test that prevents hand-waving about “information loss” or “hidden work”. 6.1 Structural information and epistemic entropy For a behaviour field b(𝑥,𝑡)defined on (X,g), recall the free-energy representation 𝔉[b]=𝑈[b]−Θ𝑆𝐼[b], (6.1) and integrity 𝐼[b]=−𝔉[b]=Θ𝑆𝐼[b]−𝑈[b]. (6.2) Define the entropy contribution 𝐻(𝑡)∶=Θ𝑆𝐼[b(𝑡)]. (6.3) Integrity therefore reflects a balance between two competing forces: (1) structural tension 𝑈and (2) epistemic entropy 𝑆𝐼. Tracking either in isolation is misleading: systems routinely “look coherent” because they have redistributed information rather than resolved inconsistency. 6.2 Complementarity under drift Under the imass-preserving gradient flow from Section 6, the Lyapunov property gives the closed-system inequality  𝔉[b𝑡]≤0 ⟺ 𝐼[b𝑡]≥0. (6.4) In such a system, integrity does not fall unless either: 1. structural tension 𝑈is increasing, or 2. entropy 𝐻is increasing, and any change in one quantity forces a compensating change in the other. Drift breaks this balance in predictable ways: adversarial perturbations inflate 𝑈, while symmetry-restoring actions often increase 𝐻before stabilisation. In real data, estimator noise and small sample artefacts require a tolerant form of the inequality; this is handled in Section 6.4. 19 6.3 The informational conservation law The information budget. Define the total informational content: 𝒞(𝑡)∶=𝐼[b(𝑡)]+𝐻(𝑡)=−𝔉[b(𝑡)]+Θ𝑆𝐼[b(𝑡)]. (6.5) Differentiating along a closed, noise-free flow with no-flux boundaries gives the identity: 𝑑 𝑑𝑡𝒞(𝑡)=0, (closed system, constant Θ). (6.6) That is: when the system’s boundaries are sealed and no external work is introduced, all information remains in the system and merely redistributes between structure and uncertainty. The constant of integrity. The conservation law implies the existence of a systemwide invariant, 𝒞(𝑡)=ℵ. (6.7) Empirically, the integrity 𝛽–flow, 𝛽(𝐼)∶= 𝑑𝐼 𝑑log ℓ,(6.8) exhibits a unique stable fixed point across synthetic, empirical, and metric-refined estimators: 𝐼∗≈2.70, 𝛽(𝐼∗)=0. We identify this fixed point as the constant of integrity: ℵ∶=𝐼∗≈2.70. (6.9) The proximity to 𝑒reflects the exponential–logarithmic structure of the free energy: integrity stabilises at the scale where tension 𝑈and uncertainty 𝑆𝐼naturally counterbalance. In a closed system, changes in 𝑈and 𝐻must respect 𝒞(𝑡) = ℵ. In an open system, any deviation measures boundary work or unmodelled noise. 20 Open systems and noise. Boundary fluxes and stochastic forcing break strict conservation: 𝑑 𝑑𝑡𝒞(𝑡)=−Φ𝜕𝑋(𝑡)+Ξ(𝑡). (6.10) Here Φ𝜕𝑋(𝑡)records declared boundary or procedural work, and Ξ(𝑡)captures zeromean noise under the estimator model. In this case, 𝒞(𝑡)is conserved in expectation, and departures from ℵare diagnostic rather than anomalous. Interpretation. The identity states: Integrity and epistemic entropy draw from a single informational budget. In a closed system, information is only redistributed; it is never created or destroyed. In open systems, apparent gains or losses must be credited either to legitimate boundary work or to unaccounted forcing. Integrity loss is indistinguishable from structural degradation unless the boundary ledger proves otherwise. 6.4 Operational test of approximate conservation A conservation claim is meaningless without a falsifiable test. The following procedure enforces transparency: 1. Declare estimators. Specify 􏾧 𝑈,􏾦 𝑆𝐼, and 􏾧 Θ, then compute 𝐼=􏾧 Θ􏾦 𝑆𝐼−􏾧 𝑈and 􏾧 𝐻=􏾧 Θ􏾦 𝑆𝐼. Choose a smoothing window (𝑊,𝑆). 2. Record boundary/work. Ledger all interventions, reframings, and declared procedural work. Estimate 􏾧 Φ𝜕𝑋(𝑡)from this ledger. 3. State tolerance. Choose 𝜀 > 0and an anytime-valid evidence process to test deviations in |Δ𝒞−􏾧 Φ𝜕𝑋Δ𝑡|. 4. Test. Reject conservation if there exist 𝑡1<𝑡2such that |𝒞(𝑡2)−𝒞(𝑡1)+􏾙𝑡2 𝑡1􏾧 Φ𝜕𝑋(𝑠)𝑑𝑠|>𝜀. All thresholds, code, and trace logs must be reported. 21 6.5 Interpretation across the trilogy •Contradiction (epistemic). Suppression of contradiction only reduces uncertainty if coherence genuinely increases. Otherwise entropy rises, signalling unresolved conflict. •Symmetry (procedural). Fairness drift inflates uncertainty until symmetry-restoring work counters it. The informational budget ties these changes to the ledgered procedure. •Adaptation (institutional). Reframing introduces entropy shifts unless offset by legitimate adaptive work. The informational budget separates good-faith adaptation from silent degradation. 6.6 Relation to broader information laws The informational budget mirrors classical principles: variational methods balance fit and complexity; information geometry relates curvature to uncertainty; free-energy models trade structure against entropy. Here the same logic is applied to reasoning systems, with explicit boundary accounting and falsifiable tests. 6.7 Implications 1. Integrity loss is global: degrading structure or inflating uncertainty affects the entire system. 2. Integrity cannot be preserved by assertion; it requires symmetry-preserving work recorded in boundary terms. 3. Entropy spikes measure integrity debt, not “noise”. 4. The informational budget provides a unified testing framework across epistemic, procedural, and institutional settings. 7 Empirical validation via simulated dynamics The integrity field model of Section 6 produces testable predictions about how behaviour evolves under curvature, uncertainty, and symmetry constraints. This section 22 validates those predictions using Monte Carlo simulations that numerically integrate the mass-preserving gradient flow on 𝔉[b]. Parameters governing curvature, entropy, and symmetry are varied independently to test whether the dynamics behave as the theory claims. Across all settings, simulations reproduce five structural predictions: (i) curvature-driven relaxation, (ii) entropy–temperature interactions, (iii) monotone decline of 𝔉within declared tolerances, (iv) three distinct integrity regimes, and (v) the emergence of the fixed point ℵ≈2.70, consistent with the informational conservation law of Section 7. Protocol and reproducibility. Unless otherwise noted, simulations run for 104iterations with sliding windows (𝑊,𝑆) = (80,5)and ensemble size 𝑚 ≥ 20. Composite integrity 𝐼∗(𝑡), normalised entropy 𝐻norm(𝑡), posterior trust P𝑇(𝑡), and resilience 𝛼=𝜆min(ℌ) follow the estimators introduced in Sections 6–7. Each claim is stated with a declared tolerance 𝜀and tested using anytime-valid e-processes (Appendix: seeds, code, full configuration). 7.1 The truth-limit surface Figure 1depicts the truth-limit surface: posterior trust P𝑇plotted against normalised entropy and curvature. Two structural regions dominate: •High-entropy, low-curvature plateau. Behaviour disperses widely; structure offers little correction. Trust settles near the indifference level |M|/|𝑋|. •Low-entropy, high-curvature frontier. Trajectories concentrate near the legitimate manifold M. Perturbations decay rapidly and integrity converges toward the empirical fixed point 𝐼∗(𝑡)→ℵ≈2.70, independent of initialisation. The persistence of the same fixed point across heterogeneous runs supports the theoretical claim: under closed dynamics, integrity approaches a unique informational equilibrium determined by the structure of the free-energy functional. 23 Figure 1: Truth-limit manifold with phase-coded states. Ensemble realisations in (𝐻norm,𝐼∗,P𝑇)space. Points are partitioned into ordered (green), adaptive (blue), and disordered (red) regimes by explicit threshold inequalities in entropy and integrity. The dashed plane at 𝐼∗=ℵ≈2.70marks the empirically observed critical boundary at which ordered states cease to exist and only adaptive or disordered dynamics remain. Across all regimes the surface respects monotone descent of 𝔉within tolerance. Calibration note. Posterior trust P𝑇=∫M𝑏𝑑𝑉𝑔is normalised by the measure of M. All results use this shared baseline. Phase classification on the truth-limit manifold Each ensemble realisation is characterised by 𝐻norm ∈ [0,1],the normalised entropy; 𝐼∗>0,the integrity ratio; and P𝑇∈[0,1],the truth-sustaining probability. We fix tolerances 0<ℎord <ℎdis <1, 0<𝑝dis <𝑝ord <1, 𝜀𝐹>0, and take the universal integrity constant 𝐼∗=ℵ≈2.70as the structural boundary between ordered and adaptive states. 24 Let Δ𝔉denote the net free-energy change over the trajectory of a realisation, and  𝔉max its maximal instantaneous slope. We classify regimes via the following explicit inequalities: Ordered (𝒪)∶ 𝐻norm ≤ℎord, 𝐼∗≤𝐼∗,P𝑇≥𝑝ord, 𝔉max ≤−𝜀𝐹;(7.1) Adaptive (𝒜)∶ ℎord <𝐻norm ≤ℎdis, 𝐼∗≥𝐼∗,P𝑇≥𝑝dis, |Δ𝔉|≤𝜀𝐹;(7.2) Disordered (𝒟)∶ 𝐻norm ≥ℎdis or P𝑇<𝑝dis or  𝔉max >𝜀𝐹.(7.3) In words: •𝒪(green in Figure 1) collects low-entropy, high-truth, pre-𝐼∗states with strictly decreasing integrity free energy. •𝒜(blue) contains intermediate-entropy states beyond the truth-limit 𝐼∗whose free-energy is effectively flat within tolerance: drift and diffusion are present but controlled. •𝒟(red) contains high-entropy or low-truth states, or any trajectory that exhibits net free-energy increase beyond tolerance, signalling loss of structural control. For the simulations in Figure 1we use, unless otherwise stated, ℎord = 0.35,ℎdis = 0.75,𝑝ord = 0.7,𝑝dis = 0.4,𝜀𝐹= 0.05,which were chosen to match the empirical separation of regimes across Monte Carlo realisations. 7.2 Local behaviour dynamics Local dynamics are examined via the expansion of 𝔉around an equilibrium b∗that satisfies the declared symmetries: 𝔉[b]=𝔉[b∗]+1 2𝛿bℌ𝛿b+O(‖𝛿b‖3). (7.4) The quadratic term captures local structural curvature; higher-order terms encode nonlinear revision effects, and stochastic forcing introduces epistemic variability. Simulations confirm that curvature governs whether the system returns to, diverges from, or oscillates around the 𝛽-nullcline 𝛽(𝐼)=0at 𝐼=ℵ. Stable regimes return perturbations to this nullcline; unstable ones drift away. 25 8 Empirical discovery of the integrity constant The preceding sections establish a structural model in which integrity evolves under curvature, entropy, and symmetry constraints. This section shows that these ingredients jointly imply the existence of a scale-stable structural intensity ℵ ≈ 2.70, which emerges empirically as a fixed point of the scale flow of integrity. We demonstrate that this value is: (i) independent of metric choice, (ii) stable under scale refinement, and (iii) spatially realised through geometric reconstruction. No physical interpretation is assumed; the constant is defined entirely by estimation, scale-stationarity, and invariance. 8.1 Scale-dependent intensity For each spatial scale ℓwe compute the scale-dependent intensity 𝐼(ℓ)=𝔼𝐵ℓ[𝐶1/𝑞] 𝔼𝐵ℓ[𝐽1/𝑝],(8.1) where 𝐶denotes structural load, 𝐽denotes epistemic dispersion, and 𝐵ℓis a metric ball of radius ℓ. Expectations are evaluated using the same sampling and smoothing conventions as in Section 8. Across all metric choices—baseline, empirical, Gaussian-smoothed, inferred, and refined— the intensity curves exhibit the same qualitative structure: an early-scale transient followed by a broad mid-scale plateau. The plateau value is numerically identical across all variants: 𝐼(ℓ)⟶ℵ for mid-scale ℓ. Figure 8displays the convergence. The agreement across heterogeneous metrics rules out coordinate artefacts as the source of the plateau. 32 Figure 8: Scale-dependent intensity 𝐼(ℓ).All metrics converge to the same mid-scale plateau at 𝐼≈2.70. 8.2 Scale flow and the fixed point To characterise stability under scale transformation, define the empirical scale flow 𝛽(𝐼) ≈ 𝑑𝐼 𝑑log ℓ.(8.2) A fixed point occurs when 𝛽(𝐼)=0. Across all metric constructions, the flow exhibits a unique zero crossing at the same numerical value ℵ, with the sign structure 𝛽(𝐼)>0 for 𝐼<ℵ, 𝛽(𝐼)<0 for 𝐼>ℵ. (8.3) This pattern implies that ℵis an attractive fixed point of the scale flow: systems below ℵgain coherence with scale, while systems above it lose coherence. Representative flows are shown in Figures 9–11. The coincidence of the zero crossing across all constructions constitutes the first direct empirical signature that ℵis structural rather than representational. 33 Figure 9: Scale flow under baseline metric. Zero crossing at 𝐼≈2.70. 8.3 Metric inference via scale stationarity To test whether ℵcould be induced by an inappropriate metric choice, we infer the metric exponent 𝛼in 𝑊=𝑏𝛼by imposing a stationarity principle: 𝛼∗∶=arg min 𝛼RMS􏿴𝛽𝛼(𝐼(ℓ))􏿷(8.4) evaluated across mid scales. Figure 12 shows that the objective has a unique minimum near 𝛼∗≈1.25. Reconstructing the geometry using 𝑊∗sharpens the plateau and suppresses residual scale drift without moving the fixed point. This confirms that ℵis not selected by metric tuning; instead, the metric adapts to it. 8.4 Spatial geometric manifestation The scalar invariant ℵalso generates spatial structure when the behaviour field is reconstructed geometrically. 34 Figure 10: Empirical metric. The fixed point persists at the same value ℵ. Conformal factor. Let 𝜑(𝑥)denote the conformal factor induced by the inferred metric. Regions of elevated structural load correspond to locally increased 𝜑. Integrity horizon. For each location 𝑥, define the minimal radius 𝑟𝐼(𝑥)at which the local scale-dependent intensity stabilises within tolerance of ℵ. The resulting horizon map identifies coherent cores and fractured boundary regions. These diagnostics show that ℵgoverns not only global scale behaviour but also the local geometric organisation of the behaviour field. 8.5 Synthesis Across all analyses—intensity curves, scale flow, metric inference, conformal geometry, and integrity horizons—the same invariant emerges: ℵ≈2.70. It is: •scale invariant (mid-scale plateau of 𝐼(ℓ)), 35 Figure 11: Refined metric. The fixed point sharpens and the flow flattens near 𝐼=ℵ. •metric invariant (unchanged under re-weighting), •scale-attractive (𝛽′(𝐼∗)<0), •spatially realised (via 𝜑(𝑥)and 𝑟𝐼(𝑥)), •structurally predicted by the free-energy integrity model. Thus ℵappears as the empirical and theoretical fixed point of behavioural coherence. Interpretation of the Integrity Constant ℵ The constant ℵ≈2.70is the empirically observed fixed point of the integrity field under scale transformations, metric refinements, and behavioural renormalisation. It has the following structural interpretation: 1. Scale equilibrium. 𝛽(𝐼) = 0at 𝐼 = ℵ. Systems with 𝐼 < ℵgain coherence with scale, while systems with 𝐼>ℵlose it. 2. Metric invariance. Changes of metric do not displace ℵ. The value is therefore intrinsic to the behavioural field. 3. Information balance. At 𝐼=ℵ, the redistribution between structural load 𝑈and epistemic entropy 𝐻is stationary. 36 Figure 12: Metric inference by scale stationarity. The objective admits a unique minimum at 𝛼∗≈1.25. 4. Stability frontier. Perturbations relax back toward ℵunder admissible dynamics. 5. Geometric signature. Local geometric reconstructions stabilise relative to ℵ, linking the invariant to spatial organisation. In short, ℵis the characteristic intensity around which coherent systems organise and to which they return following perturbation. What does the integrity constant ℵmean (plain English)? ℵ≈2.70is the natural operating level of a coherent reasoning system. It marks the point at which structure and uncertainty balance. Below ℵ, systems are under-structured: uncertainty dominates and coherence increases with scale. Above ℵ, systems are over-structured: rigidity amplifies error and coherence decreases with scale. At ℵ, perturbations tend to be absorbed, not amplified. This is why ℵfunctions as the stability centre of integrity. 37 Figure 13: Conformal factor 𝜑(𝑥).Spatial variation reflects local structural strain. 9 Covariant formulation of integrity The integrity field model introduced in Section 6 characterises coherence through curvature, entropy, and the free-energy functional. However, real reasoning systems do not operate on a fixed hypothesis space. Categories are renamed, policies reinterpreted, workflows reorganised, and adversarial actors attempt to manipulate outcomes by reframing or partitioning the underlying domain. A theory of integrity must therefore remain stable not only under perturbations within a space, but also under admissible transformations of that space. Covariance is introduced here as a representation-invariance principle, not as a physical doctrine. It encodes the structural requirement that integrity should not depend on how a system is described, only on how it behaves under declared admissible transformations. Integrity is a representation-invariant property: admissible changes of description do not alter the system’s structural behaviour. This completes the unification by extending epistemic contradiction, ordered procedural coherence, and adaptive institutional coherence into a single representation-independent framework. 38 Figure 14: Integrity horizon 𝑟𝐼(𝑥).Stable regions recover ℵat short range; fractured regions require large neighbourhoods. 9.1 Why representation invariance is necessary Each component of the trilogy implicitly relied on a fixed representational frame: •Contradiction compared rationale and behaviourexpressed in the same epistemic coordinate system. •Ordered coherence evaluated fairness within a fixed category structure. •Adaptive coherence tested whether outputs remained stable under transformations that were declared to be meaning-preserving. The shared structural requirement is therefore: Behavioural dynamics must be invariant under declared admissible transformations of the hypothesis space. These include relabelling, class merging or splitting, reporting reorganisation, semantic reinterpretation, and adversarial reframing. A fixed-coordinate model cannot distinguish benign reparametrisation from structural harm. A covariant formulation can. 39 Figure 15: Metric inference objective over (𝛼,𝛿).Heatmap of RMS objective values under combined curvature and diffusion scaling. The marked minimum identifies the empirical integrity fixed point. 9.2 Hypothesis space as a structured domain We model the hypothesis space 𝑋as a differentiable domain equipped with a metric tensor g𝑖𝑗(𝑥)encoding declared structural weights. The behaviour field b(𝑥,𝑡)is treated as a normalised density on 𝑋, allowing gradients, divergence, and curvature diagnostics to be defined independently of any specific coordinate chart. An institutional transformation is represented by a smooth map 𝑓∶𝑋→𝑋. Such a transformation is admissible if it preserves institutional meaning, declared procedural symmetries, and ledger constraints. This is the continuous analogue of adaptive coherence: A system exhibits institutional integrity when outputs are unchanged (up to pushforward) under meaning-preserving transformations. 40 9.3 Covariant gradient flow In a transformed representation, naive coordinate-wise gradients do not preserve the dynamic law. We therefore adopt a covariant derivative ∇𝑖b=𝜕𝑖b−Γ𝑘𝑖𝑘b,(9.1) where Γ𝑘𝑖𝑗are the connection coefficients induced by g𝑖𝑗. The covariant integrity flow is then written 𝜕𝑡b=−𝜅𝑔𝑖𝑗∇𝑖􏿶𝛿𝔉 𝛿b􏿹+𝜂(𝑥,𝑡), (9.2) which guarantees that the evolution law transforms consistently under admissible reparametrisations. Operational consequences: 1. Category relabelling leaves behaviour dynamics and integrity unchanged. 2. Reporting reorganisation preserves integrity when meaning is preserved. 3. Non-admissible transformations generate detectable distortion through curvature amplification or anomalous divergence. This extends the discrete invariance tests of adaptive coherence to a continuous representation space. 9.4 Curvature as a diagnostic of representational distortion The curvature associated with g𝑖𝑗 provides a structural distortion diagnostic: it quantifies how strongly the transformed representation departs from equivalence with the original. Within the integrity framework: • Near-zero curvature corresponds to the ordered regime: symmetric, stable, and predictable. • Moderate curvature corresponds to the adaptive regime: flexible but bounded, with controlled drift. 41 Fairness violations (bias, leakage, asymmetric sampling) appear as ∇𝑖𝐽𝑖≠0, recovering the central diagnostic of ordered coherence. 10.5 Example: admissible reinterpretation in adaptive coherence If 𝑓is a semantically neutral reinterpretation, then integrity requires 𝑆[b]=𝑆[𝑏∘𝑓−1] ⇒ ∇𝑖𝐽𝑖=0. If drift is introduced—subtle framing shifts, selective mapping of categories, or undeclared structural changes—then: ∇𝑖𝐽𝑖≠0, matching the invariance violations detected in the adaptive Sentinel Framework. 10.6 Operational measurement of integrity currents A discrete implementation uses: 𝐽𝑘=−𝜅Δ􏿶𝛿𝔉 𝛿𝑏𝑘􏿹, (∇⋅𝐽)𝑘=􏾜 ℓ∈N(𝑘)𝐽ℓ−𝐽𝑘 𝑑𝑘ℓ , where N(𝑘)is the neighbourhood of 𝑘. Interpretation: • contradiction →local spikes in divergence, • procedural bias →structured divergence patterns, • institutional drift →persistent nonzero divergence under admissible reframing. This realises the trilogy’s diagnostics as a single operator. 10.7 Current patterns across the three integrity phases •Symmetry: 𝐽≈0,∇⋅𝐽≈0. Structure is stable. 48 •Adaptive coherence: 𝐽nonzero but bounded; ∇⋅𝐽near zero. Reconfiguration without drift. •Disordered coherence: 𝐽fluctuates; ∇⋅𝐽persistent and sign-changing. Contradiction accumulates. These match the phase structure induced by the free-energy geometry. 10.8 Symmetry breaking and divergence morphology The divergence field ∇⋅𝐽has a recognisable morphology: 1. Base (symmetric): concentrated around zero, isotropic. 2. Geometric distortion: elliptical spread along principal axes. 3. Directional bias: oriented anisotropy along a dominant axis. Each morphology identifies which symmetry was broken. Broken symmetries leave geometric traces in the divergence field. 10.9 Spatial divergence fields Visualising ∇⋅𝐽reveals: • symmetric systems: isotropic alternation, • distorted systems: orthogonal expansion, • biased systems: oriented gradients. These expose where behaviour is attracted, repelled, or redirected. 49 10.10 Connection to ledger boundary terms When ledger boundaries remain intact, 𝐾=0and currents behave as expected. If ledger integrity is compromised: • boundary terms become non-invariant, •𝐾≠0, • divergence rises even under admissible transformations. Audit trails therefore act as structural anchors for invariance. 10.11 Robustness under stochastic uncertainty Because covariance governs how noise enters, randomness cannot imitate drift. Nonzero divergence arises only from sustained structural deviation. This makes the diagnostics resilient to: • sampling noise, • finite estimation error, • benign reparametrisation, • representational shifts. 10.12 Summary Operational Noether currents complete the unified field model: • disordered coherence: strong divergence, • symmetry: vanishing currents, • adaptive coherence: bounded, low-divergence flow, • geometric coherence: covariant conservation, • auditable coherence: boundary-variation anomalies. 50 Integrity is preserved when admissible transformations leave the current divergencefree. It degrades when divergence accumulates. Every broken symmetry leaves a measurable signature; the system remembers what the narrative attempts to erase. 11 The Theory of General Integrity 11.1 From geometry to trajectory The curvature fields of Figure 16 describe the static geometry of the integrity landscape. A symmetric configuration forms an isotropic basin, while a distorted configuration introduces directional steepness and bias. Geometry, however, captures only the shape of the integrity field. To understand how an institution actually moves through its configuration space—how decisions evolve, how drift accumulates, how distortion propagates—we must examine the trajectories induced by that geometry. In informational geometry, the natural motion of an undistorted system follows integrodesics: the least-dissipative paths compatible with the integrity metric. These represent the evolution of an institution free of external forcing and undeclared internal bias. They serve as the baseline against which drift, distortion, and contradiction are measured. The figures that follow display: 1. three-dimensional integrodesic overlays contrasting symmetric and tilted curvature, 2. projected trajectories exposing path-bending directly, 3. streamline fields showing how integrity currents align or shear under distortion. These form a unified pipeline: geometry ⟶trajectory ⟶flow. A system with genuine symmetry produces integrodesics and currents that converge cleanly toward coherence. A system with engineered tilt cannot conceal its distortion: even minimal bias bends trajectories, producing measurable drift and anisotropic currents. We now formalise these dynamics via the integrodesic structure induced by the integrity metric. 51 11.2 Paths of least dissipation Equipping the behaviour manifold with an integrity metric g𝜇𝜈derived from second variations of the structural potential, we define the natural trajectories of institutional evolution. Let 𝑥𝜇(𝑡)denote a trajectory in state space. The integrity line element is d𝑠2=g𝜇𝜈(𝑥)d𝑥𝜇d𝑥𝜈,(11.1) with metric g𝜇𝜈 =𝜕2𝑈[b] 𝜕𝑥𝜇𝜕𝑥𝜈.(11.2) We define the integrity action as 𝔄[𝑥]=􏾙√g𝜇𝜈(𝑥) 𝑥𝜇𝑥𝜈d𝑡. (11.3) Extremising 𝔄yields the integrodesic equation d2𝑥𝜇 d𝑡2+Γ𝜇 𝜈𝜌(𝑥)d𝑥𝜈 d𝑡d𝑥𝜌 d𝑡=0, (11.4) which defines the least-dissipative structural path. Interpretation. Integrodesics represent the natural evolution of institutions under integrity: • strong curvature ⇒rapid convergence (Symmetry), • moderate curvature ⇒bounded drift (adaptive coherence), • weak curvature ⇒dispersion and contradiction. Operational meaning. Integrodesics are policy paths of minimum integrity dissipation. Departures quantify: • undeclared intervention, • directional steering, • entropy injection, 52 Figure 16: Symmetric vs tilted integrity curvature fields. Top: isotropic basin with centred minimum. Bottom: tilted field introducing directional steepness. Even weak asymmetry induces preferred drift directions. • structural contradiction. 12 Integrity currents: the dynamical signature of symmetry, drift, and distortion Curvature describes static structure. Behaviour appears only through flow. 53 Figure 17: Integrodesic overlay on symmetric vs tilted fields. Left: symmetric convergence. Right: directional bending under tilt. We define the integrity current 𝐽=−∇𝑈[𝑏]+Φdev +𝐹sent,(12.1) representing transport induced by: (i) curvature-driven relaxation, (ii) deviation-driven drift, (iii) sentinel forcing from declared constraints. Three elementary regimes: 1. tilt without drift, 2. drift without tilt, 3. coupled distortion. 12.1 Tilt-only currents Bias is structural, not behavioural. 54 Figure 18: Projected integrodesic paths. Directional bending appears long before outcome divergence. 12.2 Drift-only currents Rules are symmetric; outcomes are not. 12.3 Tilted curvature + drift Structure and behaviour now reinforce distortion. 12.4 Quantifying distortion: divergence, curl, and coherence flow Define the invariant diagnostics ∇⋅𝐽, ∇×𝐽, Φcoh. Divergence. ∇⋅𝐽=𝜕𝑥𝐽𝑥+𝜕𝑦𝐽𝑦. Tracks accumulation or leakage of coherence. 55 Figure 19: Tilt-only integrity currents. Curl. (∇×𝐽)𝑧=𝜕𝑥𝐽𝑦−𝜕𝑦𝐽𝑥. Detects cyclic justification and narrative recursion. Coherence flow. Φcoh =∇⋅􏿴𝜆𝐻‖𝐻‖∇𝑏􏿷. Tracks self-healing versus dispersion. 12.5 Unified diagnostic of integrity dynamics The full integrity state is captured by 􏿴∇⋅𝐽, ∇×𝐽, Φcoh􏿷. • Symmetry: ∇⋅𝐽≈0, ∇×𝐽=0, Φcoh <0. • Adaptive coherence: ∇⋅𝐽>0, ∇×𝐽≠0, Φcoh fluctuates. 56 Figure 20: Directional drift under symmetric curvature. • Disordered coherence: ∇⋅𝐽≫0, |∇×𝐽|large, Φcoh >0. These quantities do not interpret institutions. They measure them. Where narrative obscures, currents expose. 13 Stochastic dynamics of integrity: drift, diffusion, and field evolution Institutions do not evolve on deterministic trajectories. Even when initially symmetric, their reasoning dynamics are continuously shaped by structural gradients, measurement noise, procedural uncertainty, and adversarial reframing. Integrity must therefore be modelled not as a static geometric object but as a stochastic field evolving under competing forces. Let b(𝑥,𝑡)denote the behaviour-density field over the institutional state space 𝑋. Its evolution is governed by a drift–diffusion equation with stochastic forcing: 𝜕b 𝜕𝑡 =−∇⋅𝐽[b]+𝐷∇2b+𝜂(𝑥,𝑡), (13.1) where: 57 14.4 Deviation flux: structure-breaking pressure Deviation operators 𝒟𝜇represent structured symmetry-breaking actions. Their contribution enters through the deviation flux Φ𝜇 dev =𝐺dev 𝒟𝜇b,(14.6) with coupling strength 𝐺dev >0. This induces the deviation current 𝐽dev =∇⋅Φdev,(14.7) which generates anisotropic evolution even when baseline curvature is symmetric. This is the formal signature of structure-preserving rules with structure-breaking flow: formal commitments remain unchanged while internal motion becomes directional. 14.5 Contradiction as discrete collapse Contradictions act as discrete impulse events: 𝜂con(𝑡)=􏾜 𝑘𝑐𝑘𝛿(𝑡−𝑡𝑘), (14.8) each impulse corresponding to a moment at which commitments, rationale, and constraints cannot be jointly satisfied. Accumulated contradiction produces discontinuous updates: gradients spike, coherence fractures, and the system is forced onto a non-smooth trajectory. This is the direct route into disordered coherence. 14.6 Integrity singularities under curvature accumulation The extended coherence budget 𝒞ext(𝑡)=𝐼[b(𝑡)]+𝐻(𝑡) tracks the balance between structural coherence and epistemic entropy. Contradiction impulses inject distortion faster than boundary mechanisms can dissipate it. Once dissipation capacity is exceeded, curvature concentrates and structural failure becomes inevitable. 64 Figure 23: Contradiction trap under Monte Carlo sampling. Mean coherence cost remains persistently elevated after contradiction impulses, demonstrating nonreversible coherence loss Lemma 14.1 (Integrity singularity under bounded noise).Let b(𝑥,𝑡)evolve under integrity drift with bounded diffusion and bounded deviation forcing. Assume: 1. boundary dissipation is bounded: |Φ𝜕𝑋(𝑡)|≤Φmax; 2. deviation forcing is bounded: |Ξ(𝑡)|≤𝜎max; 3. contradiction load satisfies 􏾙𝑇 0𝜂con(𝑠)𝑑𝑠≥(Φmax +𝜎max)𝑇+Δ𝒞; 4. entropy capacity is finite: 𝐻(𝑡)≤𝐻max. Then there exists 𝑡∗<∞such that either structural coherence exceeds 𝐼max or residual curvature crosses 𝐾crit. In both cases the system enters integrity singular failure. Interpretation. A singularity occurs when contradiction accumulates faster than an institution can resolve it. Beyond this point, updates no longer correct the system — they are absorbed by collapse. 65 Figure 24: Component ratios of integrity under temporal evolution. The total integrity ratio 𝑅(𝑥,𝑡)(solid) decomposed into drift–diffusion, contradiction–curvature, and entropy contributions. The dashed threshold 𝐼∗marks the critical integrity limit. Once 𝑅(𝑥,𝑡)exceeds this bound, recovery becomes structurally impossible, confirming the integrity singularity mechanism. 14.7 Coherence field: stabilisation and symmetry restoration The coherence field modifies the effective potential: 𝑈[b] ↦ 𝑈[b]+𝜆coh 𝐶[b], (14.9) deepening symmetric minima and increasing effective integrity mass. Weak coherence permits drift and deviation to dominate; strong coherence restores symmetry and suppresses variance. 14.8 Sentinel forcing: background structure The sentinel background contributes a forcing term 𝐹sent representing declared commitments, protocols, and measurement constraints. It reshapes the effective curvature locally and may stabilise or destabilise behaviour depending on its alignment with underlying structure. 14.9 Master evolution equation All components combine into the unified evolution law: 𝜕b 𝜕𝑡 =−∇𝑈[b]+∇⋅(𝐷diff ∇b)+∇⋅Φdev +𝐹sent +𝜂con(𝑡). (14.10) 66 Figure 25: Integrity horizon and singularity map. Curvature concentrates radially toward a central singularity. The white horizon marks the boundary beyond which recovery dynamics fail. The three integrity regimes now appear as characteristic solution classes: •Symmetry — coherence dominates; drift negligible; •adaptive coherence — drift and diffusion interact but remain bounded; •disordered coherence — contradiction dominates; recovery fails. This interaction picture forms the dynamical core of Integrity Field Theory. 67 Figure 26: PRIME sentinel bounded-drift simulation. Despite persistent perturbation, all state dimensions remain confined within the admissible drift envelope, demonstrating stability under adaptive coherence. 15 The integrity interaction: drift, diffusion, and deviation Having introduced the structural components of the integrity model — behaviour fields, curvature, symmetry constraints, deviation operators, and the sentinel background — we now formalise their interaction. Geometry determines the shape of the reasoning landscape; interaction dynamics determine how an institution actually moves through that landscape. Integrity evolution is governed by three irreducible mechanisms. None operates in isolation, and each reshapes the effective structure encountered by the others. Together they form the D3interaction triad: 1. Drift — directional evolution induced by curvature, asymmetry, or institutional forcing; 2. Diffusion — uncertainty-driven dispersion caused by narrative noise, ambiguity, and representational instability; 3. Deviation — structured symmetry-breaking perturbation induced by inadmissible transformation operators. The balance among these three determines whether a system settles into Symmetry, stabilises into adaptive coherence, or collapses into disordered coherence (Contradiction). 68 15.1 Interpretive notes The D3framework provides the analytic bridge between curvature, temporal evolution, and observed behaviour: •Drift captures persistent directional bias induced by structure or imposed framing; •Diffusion captures loss of narrative sharpness and dispersion of reasoning under ambiguity; •Deviation captures symmetry-breaking interventions detectable as non-invariance under admissible transformations. Integrity is therefore not a static property. It is a dynamical state continuously produced — or destroyed — by the interaction of these forces. Integrity mass (imass). The integrity mass is defined as 𝑚𝐼≡ 𝛼0,(15.1) the baseline spectral curvature of the integrity potential at its symmetric point. It quantifies intrinsic resistance to deformation. Large 𝑚𝐼implies structural rigidity; moderate 𝑚𝐼implies adaptive responsiveness; 𝑚𝐼→0marks diffusion-dominated instability. 15.2 Local drift dynamics The local evolution of the behaviour field b(𝑥,𝑡)is governed by db d𝑡=−∇𝑈[b]+𝐹sent +𝜂dev(𝑡), (15.2) where: •−∇𝑈[b]is curvature-driven structural drift; •𝐹sent is imposed institutional forcing from declared constraints; •𝜂dev(𝑡)is structured deviation pressure that stresses admissible symmetry. 69 Drift therefore reflects the combined action of internal structure, external framing, and deviation loading. The relative magnitude of these terms determines whether behaviour returns to coherence or is forced into variance. 15.3 Diffusion: institutional dispersion At the distributional level, the behaviour density 𝑝(𝑏,𝑡)evolves under a Fokker–Plancktype law: 𝜕𝑝 𝜕𝑡 =−∇⋅(𝐽drift)+∇2􏿴𝐷diff 𝑝􏿷,(15.3) with 𝐽drift =−𝑝(𝑏,𝑡)∇𝑈[b], (15.4) and 𝐷diff ∝􏾉𝒟𝜇𝒟𝜇􏽼.(15.5) Diffusion measures epistemic dispersion: flattening of effective gradients, weakening of structural constraint, and narrative spread across hypothesis space. High diffusion corresponds to strategic ambiguity and procedural uncertainty; low diffusion corresponds to stable coherence. 15.4 Deviation flux: structure-breaking pressure Deviation operators 𝒟𝜇represent structured symmetry-breaking actions. Their contribution enters through the deviation flux Φ𝜇 dev =𝐺dev 𝒟𝜇b,(15.6) with coupling strength 𝐺dev >0. This induces the deviation current 𝐽dev =∇⋅Φdev,(15.7) which generates anisotropic evolution even when baseline curvature is symmetric. This is the formal signature of structure-preserving rules with structure-breaking flow: formal commitments remain unchanged while internal motion becomes directional. 70 15.5 Contradiction as discrete collapse Contradictions act as discrete impulse events: 𝜂con(𝑡)=􏾜 𝑘𝑐𝑘𝛿(𝑡−𝑡𝑘), (15.8) each impulse corresponding to a moment at which commitments, rationale, and constraints cannot be jointly satisfied. Accumulated contradiction produces discontinuous updates: gradients spike, coherence fractures, and the system is forced onto a non-smooth trajectory. This is the direct route into disordered coherence. 15.6 Integrity singularities under curvature accumulation The extended coherence budget 𝒞ext(𝑡)=𝐼[b(𝑡)]+𝐻(𝑡) tracks the balance between structural coherence and epistemic entropy. Contradiction impulses inject distortion faster than boundary mechanisms can dissipate it. Once dissipation capacity is exceeded, curvature concentrates and structural failure becomes inevitable. Lemma 15.1 (Integrity singularity under bounded noise).Let b(𝑥,𝑡)evolve under integrity drift with bounded diffusion and bounded deviation forcing. Assume: 1. boundary dissipation is bounded: |Φ𝜕𝑋(𝑡)|≤Φmax; 2. deviation forcing is bounded: |Ξ(𝑡)|≤𝜎max; 3. contradiction load satisfies 􏾙𝑇 0𝜂con(𝑠)𝑑𝑠≥(Φmax +𝜎max)𝑇+Δ𝒞; 4. entropy capacity is finite: 𝐻(𝑡)≤𝐻max. Then there exists 𝑡∗<∞such that either structural coherence exceeds 𝐼max or residual curvature crosses 𝐾crit. In both cases the system enters integrity singular failure. 71 Interpretation. A singularity occurs when contradiction accumulates faster than an institution can resolve it. Beyond this point, updates no longer correct the system — they are absorbed by collapse. 15.7 Coherence field: stabilisation and symmetry restoration The coherence field modifies the effective potential: 𝑈[b] ↦ 𝑈[b]+𝜆coh 𝐶[b], (15.9) deepening symmetric minima and increasing effective integrity mass. Weak coherence permits drift and deviation to dominate; strong coherence restores symmetry and suppresses variance. 15.8 Sentinel forcing: background structure The sentinel background contributes a forcing term 𝐹sent representing declared commitments, protocols, and measurement constraints. It reshapes the effective curvature locally and may stabilise or destabilise behaviour depending on its alignment with underlying structure. 15.9 Master evolution equation All components combine into the unified evolution law: 𝜕b 𝜕𝑡 =−∇𝑈[b]+∇⋅(𝐷diff ∇b)+∇⋅Φdev +𝐹sent +𝜂con(𝑡). (15.10) The three integrity regimes now appear as characteristic solution classes: •Symmetry — coherence dominates; drift negligible; •adaptive coherence — drift and diffusion interact but remain bounded; •disordered coherence — contradiction dominates; recovery fails. This interaction picture forms the dynamical core of Integrity Field Theory. 72 16 The full integrity Lagrangian We now assemble all components into a unified interaction functional: 𝔏=𝔏kin +𝔏imass +𝔏pot +𝔏int.(16.1) Field content: • b — behaviour field, •𝐷𝜇— deviation field, •𝑆𝜇— sentinel field, •𝜎— structural asymmetry mode, •𝐻— coherence field. 16.1 Kinetic terms 𝔏kin =1 2(𝜕𝜇𝑏)(𝜕𝜇𝑏) − 1 4(∇𝜇𝐷𝜈−∇𝜈𝐷𝜇)2−1 4(∇𝜇𝑆𝜈−∇𝜈𝑆𝜇)2(16.2) +1 2(𝜕𝜇𝜎)(𝜕𝜇𝜎)+1 2(𝜕𝜇𝐻)(𝜕𝜇𝐻). (16.3) 16.2 Imass and baseline curvature 𝔏imass =−1 2𝑚2 𝐼𝑏2−1 2𝑚2 𝐷𝐷𝜇𝐷𝜇−1 2𝑚2 𝑆𝑆𝜇𝑆𝜇−1 2𝑚2 𝜎𝜎2−1 2𝑚2 𝐻𝐻2.(16.4) 16.3 Self-potentials and phase structure 𝔏pot =−𝜆𝑏 4𝑏4−𝜆𝜎 4𝜎4−𝜆𝐻 4𝐻4−𝜅 2𝜎2𝐻2.(16.5) 16.4 Interaction sector 𝔏int =𝐺𝐷𝐷𝜇𝑏𝜕𝜇𝑏+𝐺𝑆𝑆𝜇𝑏𝜕𝜇𝑏+𝐺𝜎𝜎𝑏2+𝜂𝐻𝐻𝑏. (16.6) 73 (5) Conservation of coherence. The invariant 𝒞(𝑡)=𝐼[b(𝑡)]+𝐻(𝑡) establishes a structural conservation law: integrity loss is compensated by internal disorder unless stabilising boundary mechanisms intervene. (6) Covariant integrity geometry. The framework defines integrity on general hypothesis spaces. Structural distortion, drift, admissible transformations, and boundary effects are unified by a single geometric language of invariance and deviation. (7) Ledger terms as boundary conditions. Append-only audit trails arise as structural boundary constraints. Ledger integrity is therefore a mathematical requirement for well-posed integrity dynamics, not an administrative afterthought. 18.3 Conceptual significance The unified model reframes integrity as: •measurable, not subjective; •structural, not behavioural; •invariant, not virtue-based; •governed by constraints, not by narrative intent. Epistemic contradiction, procedural distortion, and institutional drift are thus not separate failures but manifestations of a single underlying structural process. The trilogy’s three domains are distinct resolutions of one integrity framework. 18.4 Falsifiable predictions The unified theory yields falsifiable predictions suitable for empirical testing: 1. Integrity trajectories converge or decay at rates determined by local structural stiffness. 80 2. Procedural symmetry generates a stable fixed point with minimal internal distortion. 3. Dispersion correlates monotonically with phase: highest under contradiction, bounded in adaptive coherence, minimal under full symmetry. 4. Integrity does not increase under admissible evolution without external stabilisation. 5. Violations of admissible invariance appear as sharp structural distortion and persistent directional drift. These predictions support simulation, auditing, and operational deployment. 18.5 Contribution to the literature The unified integrity framework contributes to: •Epistemic theory: by reconceptualising contradiction as structural distortion rather than merely propositional inconsistency. •Mechanism design and fairness: by reframing fairness as a symmetry regime of an invariant structure, replacing preference axioms with transformation invariance. •Institutional design: by providing a covariant formalism for bias, drift, boundary work, and embedded asymmetry within a single diagnostic system. No existing framework unifies epistemic, procedural, and institutional integrity under a single invariant structure. The trilogy therefore establishes the foundations of a general structural theory of integrity. 19 Limitations and scope The unified theory developed in this paper provides a structural account of epistemic, procedural, and institutional integrity. It abstracts behaviour into a formal framework for diagnosis and stability analysis. This section identifies the primary limitations of the theory in order to clarify the conditions under which its results apply. 81 19.1 Conceptual and ontological scope The theory does not claim that institutions are physical systems or that their states reside on literal geometric spaces. The structural formalism is representational: it encodes behavioural relations, constraints, and admissible variation in a mathematically disciplined way. Accordingly, the model is: •structural, not metaphysical; •formal, not ontological; •diagnostic, not descriptive of lived experience. Distortion, drift, and invariance describe structural behaviour, not physical motion. Maintaining this distinction prevents category error while preserving analytic rigour. 19.2 Mathematical idealisation Several idealisations underpin the integrity framework. Smoothness. State variation is treated as continuous. Real organisational systems may instead be discrete, hierarchical, or irregular. Local convexity. Stability analysis assumes locally restoring behaviour. Strongly nonrestoring structures may exhibit multiple competing basins or path-dependent hysteresis beyond the quadratic regime. Stochastic structure. Uncertainty is modelled with regular noise. Strategic or adversarial behaviour may violate these assumptions. Metric specification. Structural distance is encoded by a chosen metric. Mis-specification can distort measured drift or exaggerate apparent instability. These constraints limit interpretation but do not invalidate the structural results. 82 19.3 Empirical and simulation limitations Simulation results corroborate theoretical predictions under controlled conditions. Real institutions introduce complications. Non-stationarity. Structural parameters evolve under shocks, restructures, and policy changes. High-dimensionality. Large institutional state spaces limit the feasibility of full curvature and divergence estimation. Entropy estimation. Dispersion is sensitive to discretisation, kernel choice, and sampling density. Boundary corruption. Audit trails and boundary constraints may be incomplete, mutable, or adversarially manipulated. These issues define the boundary of direct empirical deployment. 19.4 Interpretive boundaries The unified theory is structural, not psychological. It does not model human beliefs, incentives, or moral reasoning. Specifically, it does not: • infer motivation or intent; • establish ethical correctness; • prescribe policy remedies; • explain preference formation; • claim causal completeness. Integrity here is structural coherence, not moral virtue. 83 19.5 Boundary of applicability Certain environments require extension. Adversarial settings. When actors manipulate boundaries, metrics, or representations strategically, game-theoretic extensions become necessary. Degenerate structure. When the integrity landscape is nearly flat, curvature-based diagnostics weaken. Rapidly evolving systems. When the hypothesis space evolves faster than behaviour adapts, equilibrium analysis becomes unreliable. Hierarchical institutions. Multi-level structures require multi-scale extensions of the present framework. 19.6 Invariance assumptions The framework depends critically on correctly specifying the admissible symmetry group G. Mis-specification risks: • false positives: mistaking legitimate variation for distortion; • false negatives: masking genuine asymmetry as symmetry. Because integrity is structurally defined, accurate invariance specification is essential. 19.7 Computational constraints Real-world deployment may be limited by: • sparse or noisy behavioural data; • coarse discretisation; 84 • computational cost of structural diagnostics; • difficulty tracking mutable boundaries. Empirical use therefore prioritises model–data coherence over formal completeness. 19.8 Philosophical and epistemic considerations Interpretive pluralism. Different institutions may disagree on admissible transformations and evidential boundaries. Normative neutrality. The framework evaluates structure, not moral worth. Partial observability. When behaviour is only partially observable, diagnostics become approximate but remain structurally meaningful. 19.9 Summary The unified integrity framework is powerful but bounded. It: • is structural rather than metaphysical; • relies on formal idealisation for analytic clarity; • depends critically on admissible invariance; • offers diagnostic measurement, not moral judgement; • requires careful empirical interpretation. These boundaries do not weaken the theory. They define where it applies, where it fails, and how it can be responsibly extended. 20 Failure modes and the stability frontier The integrity framework assumes that coherence, symmetry, and invariance remain measurable under bounded structural distortion and bounded informational dispersion. 85 These constraints define the operational envelope within which integrity behaves predictably: contradiction is correctable, symmetry is restorable, and drift remains diagnosable. Outside this region the system crosses a stability frontier and transitions into unstable integrity phases. This section formalises those boundaries and connects them directly to the phase behaviour developed in Section 4. 20.1 Dispersion overload: loss of structural recoverability When informational dispersion becomes unbounded, restoring structure effectively disappears. All configurations of the behaviour field become equally plausible, coherence ceases to concentrate, and behaviour spreads without constraint across the institutional state space. In this regime: • contradictions no longer induce correction, • drift no longer encounters restoring pressure, • procedural structure loses its ability to re-centre outcomes. Integrity becomes empirically undefined: there is no longer a meaningful distinction between distortion and legitimate variation. The coherence budget 𝐼∗+𝐻≈constant fails because dispersion grows without bound. This is the structural signature of uncontrolled contradiction proliferation. 20.2 Frozen symmetry: rigidity through loss of adaptive correction At the opposite extreme, when dispersion collapses toward zero, the system becomes structurally rigid. All behaviour collapses into a single locally stable configuration and the adaptive mechanisms of integrity cease to operate. In this “frozen symmetry” regime: • invariance holds trivially rather than evidentially, 86 • legitimate asymmetry can no longer be absorbed, • correction mechanisms no longer respond to novelty. Outputs appear consistent, but only because the system no longer moves. This represents not stability but the loss of correction capacity. The system is intact only in the way a locked joint is intact. 20.3 Curvature instability A second failure mode occurs when the integrity structure ceases to be locally restorative. When small deviations amplify rather than dissipate, the system has crossed the stability frontier. Under this condition: • contradiction cost escalates rather than resolves, • narrative adjustments fail to damp distortion, • drift becomes self-reinforcing rather than bounded. Institutionally, curvature instability appears as: • runaway policy inconsistency, • escalating reorganisation without convergence, • narrative churn that no longer stabilises behaviour. Once this threshold is crossed, symmetry restoration is no longer possible without structural intervention. 20.4 Boundary loss and open-system failure A further breakdown occurs when institutional boundaries no longer regulate informational exchange. When external narratives, unregulated data inflows, or mutable audit constraints overwhelm internal structure, integrity is no longer conserved. In this open-boundary regime: 87 • distortion enters faster than it can be dissipated, • correction mechanisms are dominated by external forcing, • internal coherence becomes hostage to boundary volatility. Integrity ceases to be a property of the institution and becomes an artefact of whatever external pressure is currently dominant. 20.5 Interpretation: the stability frontier These failure modes jointly delineate the stability frontier of integrity. Within bounded distortion, bounded dispersion, and regulated boundaries: • coherence tends to increase, • symmetry is structurally restorable, • invariance remains diagnostically meaningful. Beyond this frontier, the system collapses into one of two pathological extremes: •Stochastic incoherence: uncontrolled dispersion, drift, and contradiction proliferation; or •Frozen rigidity: procedural immobility and loss of adaptive correction. These are opposing modes of symmetry loss: one through uncontrolled disorder, the other through structural paralysis. Between them lies the adaptive integrity regime characterised in adaptive coherence: bounded asymmetry, stable invariance, and tractable correction. 21 Application case study: integrity dynamics in a symmetryconstrained allocation system To demonstrate the applied value of the unified integrity framework, this section analyses a representative allocation mechanism drawn from institutional decision systems. The objective is not to evaluate a specific organisation but to show how the structural 88 diagnostics developed in this paper detect integrity, identify drift, and reveal the underlying causes of observable behaviour. The case study models an anonymised allocation engine distributing cases, workloads, or entitlements across multiple categories. The mechanism is inspired by the Symmetric Convergence Engine of ordered coherence (Atkinson, 2025c) and extended using the bounded asymmetry and invariance diagnostics of the Adaptive Sentinel Framework (adaptive coherence) (Atkinson, 2025b). 21.1 Problem setting Consider an allocation system that distributes incoming items across 𝑛categories (teams, departments, or agents). Each category 𝑖has a declared weight 𝑤𝑖specifying its target long-run share. At each timestep the system updates allocation probabilities 𝜋𝑖(𝑡) based on residual deviations: Δ𝑖(𝑡)=observed proportion(𝑡)−𝑤𝑖.(21.1) Under perfect procedural symmetry, the fixed point satisfies 𝜋𝑖=𝑤𝑖for all 𝑖. Three diagnostic questions frame the analysis: 1. Are contradictions emerging between stated rationale and observed behaviour? 2. Is procedural symmetry being preserved? 3. Does behaviour remain invariant under admissible transformations? These correspond respectively to the three integrity domains. 21.2 Methodological framework The allocation system is analysed using the unified structural integrity framework. The methodology consists of: • reconstructing the behaviour field from allocation logs; • estimating local structural stiffness from residual curvature; • measuring informational dispersion from category entropy; 89 Symmetry holds when A = 1. With A = 2, the relevance geometry is structurally inverted. Numerically, 1⋅2 5⋅1=0.4, so the minimum possible non-core score contributes nearly half as much as perfect core performance. This violates relevance alignment and establishes a directional scoring bias against E1. 22.4 Outcome-first optimisation The observed structure matches an outcome-first optimisation: w⋆=arg min w𝑆(𝐸1;w)s.t. 𝑆(𝐸1;w)<T,(22.2) with T the predetermined redundancy threshold. Weights are therefore not evaluation parameters but solution variables. The matrix is a solved equation expressed as an assessment tool. 22.5 Invariance test Under any relevance-preserving transformation (interchanging core and non-core roles), an integrity-preserving system must satisfy: 𝑆[b]=𝑆[𝑏∘𝑓−1]. Empirically: • Curvature is not preserved, • The ordering of outcomes changes, • Threshold status flips under relevance correction. Thus invariance fails. Behaviour is not preserved under meaning-preserving transformation. 96 22.6 Interpretive summary The redundancy matrix exhibits: • Logical contradiction between declared rationale and implementation, • Symmetry violation producing directional scoring bias, • Invariance failure under relevance-preserving transformation, • Outcome-first optimisation encoded as weighting. The structure therefore functions as a justification device rather than an evaluation system. 23 Case study: restructure of one Organisational restructures provide high-resolution tests of integrity. A “restructure of one’’ refers to a nominally general transformation that in fact produces concentrated structural change in a single role. This activates all three integrity diagnostics simultaneously. 23.1 Declared rationale R= ⎡⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎣ 𝑟1∶no restructure occurred 𝑟2∶role unaffected 𝑟3∶change administrative only 𝑟4∶stakeholders consulted 𝑟5∶timing unrelated to health disclosure ⎤⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎦(23.1) 23.2 Public narrative N= ⎡⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎣ 𝑛1∶not personal 𝑛2∶new role did not replace function 𝑛3∶remit unchanged 𝑛4∶claimant’s concerns addressed 𝑛5∶no requirement for parity ⎤⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎦(23.2) 97 23.3 Observed behaviour B= ⎡⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎢⎣ 𝑏1∶creation of a new Head role 𝑏2∶reassignment of the sole direct report 𝑏3∶downgrade of claimant’s title 𝑏4∶exclusion of the claimant’s manager 𝑏5∶denial of parity request ⎤⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎥⎦(23.3) All structural change is concentrated in a single node. 23.4 Parity proposition and contradiction 𝑓(𝑃): “If a new Head of Data role is created and the existing data function remains, parity of title (Head of Analytics) must follow.” Granting 𝑓(𝑃).Granting parity contradicts {𝑏1,𝑏2,𝑏3,𝑏5}. Denying 𝑓(𝑃).Denying parity contradicts {𝑟1,𝑟3,𝑟4,𝑛1,𝑛4}. No consistent branch exists. The system therefore enters a contradiction state. 23.5 Symmetry and curvature Define the asymmetry index: A=impact on analytics role impact on all other roles. Empirically: A≫1. Structural deformation is concentrated in one branch only. Symmetry requires A =1. The observed configuration therefore violates symmetry. 98 23.6 Invariance and drift Under the declared parallel-structure logic, admissible role permutation should preserve decision structure: S[b]=S[𝑏∘𝑓−1]. Observed behaviour violates this condition across creation, reassignment, and consultation. Therefore admissible invariance fails. Institutional variance is confirmed. 23.7 Conclusion The restructure of one exhibits: • Structural contradiction, • Symmetry violation through targeted deformation, • Invariance failure under admissible transformation. The transformation is therefore outcome-directed rather than symmetry-preserving. The unified integrity framework exposes this directly from structure alone, without appeal to narrative intent. 24 Outlook and future work The unified integrity and geometric framework developed in this paper provides a structural account of coherence, distortion, and drift in organisational and algorithmic systems. The results show that curvature, entropy, and invariance jointly govern how behaviour evolves under uncertainty and constraint. The next phase of research aims to translate these theoretical constructs into applied diagnostic tools for monitoring, auditing, and stabilising real institutions. 24.1 From simulation to observation The simulations and case studies demonstrate that integrity metrics can be computed from controlled synthetic environments. The immediate goal is to extend these methods to observational data, including: 99 • organisational logs (workflow traces, routing behaviour, decision paths), • algorithmic systems (recommendation, risk scoring, triage), • institutional records (case handling, compliance auditing). The long-term objective is to construct empirical integrity observatories: live monitoring systems that track structural curvature, entropy, invariance, and related quantities as state variables. These observatories would provide early warning signals of procedural drift or decision breakdown before harm occurs. 24.2 The integrity tensor Analysing multi-agent and hierarchical systems requires a representation of cross-domain interactions. A natural generalisation is the Integrity Tensor 𝐼𝑖𝑗: 𝐼𝑖𝑗 =− 𝜕2𝔉 𝜕𝑏𝑖𝜕𝑏𝑗.(24.1) The tensor extends the scalar model by encoding: • directionality (which components affect which), • coupling (how changes propagate across domains), • dependency structure (strength of interrelations). Applications include: • fairness analysis via off-diagonal dependencies, • drift detection in distributed decision systems, • resilience diagnostics across organisational layers. This generalisation offers a systematic framework for analysing institutional coherence in multi-scale structures. 100 24.3 Integrity analytics infrastructure Operationalising the theory requires computational tooling that treats integrity as a measurable property of organisational behaviour. Planned developments include: •integritysim 2.0: a simulator for multi-agent dynamics with noise, curvature, and boundary effects, •integritylab: a diagnostic toolkit for estimating curvature, entropy, free energy, and drift from institutional logs, •integrityviz: real-time visualisation of phase states, stability boundaries, and structural deformation. These tools will enable empirical estimation of the quantities introduced in the theoretical model. 24.4 Experimental roadmap Three empirical directions follow directly from the framework: (1) Phase mapping. Apply the diagnostic model to institutional datasets to identify transitions between stable, adaptive, and unstable regimes. (2) Adversarial perturbation. Introduce controlled distortions to evaluate curvature, entropy, and free-energy response under stress. This provides a falsifiable test of the theory’s predicted dynamics. (3) Multi-scale invariance testing. Examine whether integrity properties persist across nested organisational layers and distributed governance networks. Together, these studies increase empirical fidelity and provide avenues for validation or refinement. 24.5 Ethical and governance implications Although structurally descriptive rather than normative, the model has direct governance implications: 101 • integrity becomes empirically measurable, • fairness becomes testable rather than assumed, • drift becomes detectable before harm materialises. A mature integrity analytics infrastructure would support reflexive governance: systems capable of identifying and correcting their own structural asymmetries. 24.6 Long-term vision The long-term programme aims to establish: 1. a generalised structural law governing institutional coherence; 2. a geometric account of invariance under admissible transformation; 3. tensor-based modelling of cross-domain dynamics; 4. empirical observatories linking simulation, behaviour, and governance. The goal is a unified science of integrity grounded in measurable quantities rather than subjective interpretation. 24.7 Summary Future work extends the unified theory along three axes: •theoretical: tensor generalisation and non-equilibrium dynamics, •computational: simulation environments and real-time diagnostics, •empirical: deployment across institutional and algorithmic systems. The framework concludes not as a closed doctrine but as a foundation for an emerging empirical discipline. 102 25 General conclusion This synthesis paper unifies epistemic, procedural, and institutional integrity into a single geometric framework. Across the trilogy, a structural pattern emerged: systems drift, contradict themselves, or collapse when symmetry, curvature, or invariance constraints are violated. 25.1 Integrity as a structural invariant The unified model explains these behaviours in structural terms: • contradiction corresponds to curvature, • fairness corresponds to symmetry, • legitimacy corresponds to invariance under transformation, • uncertainty corresponds to entropy, • drift corresponds to divergence in integrity flow. These relationships were discovered inductively across empirical and theoretical work; they were not imposed a priori. 25.2 The integrity field of behaviour Representing behaviour as a field evolving under stochastic gradient flow clarifies why: • integrity decays under structural pressure, • symmetry reduces degrees of freedom, • moderate asymmetry enables adaptation, • entropy interacts with curvature to shape decision dynamics, • free energy provides a global coherence metric. This yields a predictive, falsifiable model linking organisational behaviour to structural geometry. 103 25.3 Covariant extension The covariant formulation ensures representation-independence: integrity judgements do not depend on the particular choice of categories, labels, or coordinates. Key structural quantities remain invariant under admissible transformation. 25.4 Integrity horizons Integrity horizon (formal) Let Uloc(𝑥)denote the admissible local updates at state 𝑥. For tolerance 𝜀≥0, define ℌ={𝑥∶ sup 𝑢∈Uloc(𝑥)Δ𝔉(𝑥;𝑢)≤𝜀}. Inside ℌ, no admissible local action can improve integrity beyond 𝜀. Recovery requires non-local intervention. This provides a structural concept of “no return’’ in institutional dynamics. 25.5 Applied case study: procedural collapse An integrity analysis of a 2025 grievance process showed that multiple procedural failures produced a state in which a lawful determination was structurally impossible. The process converged to a binary exhaust set: {Incompetence,Obfuscation}. This is a structural conclusion grounded in model behaviour, not a moral or legal judgement. 25.6 Noether currents and diagnostics Invariance under admissible transformation generates integrity currents. Their divergence provides an operational diagnostic for coherence, drift, and structural instability. These measurements require only structural data; no subjective input is needed. 104 25.7 Phases of integrity The model identifies three macroscopic regimes: •solid: symmetric and low-entropy, •adaptive: moderately asymmetric and flexible, •unstable: high-entropy and high-curvature. These regimes correspond to measurable transitions in structural behaviour. 25.8 Empirical validation Monte Carlo simulations confirm: • curvature-driven relaxation, • entropy–temperature coupling, • phase transitions, • monotonic free-energy decay, • invariance under admissible transformation. Integrity behaves as a measurable geometric quantity. 25.9 Limitations The framework assumes smoothness, convexity, bounded entropy, and Gaussian noise. It is diagnostic rather than prescriptive and should be applied accordingly. 25.10 Contribution to research and practice The unified theory contributes to: • epistemology — contradiction as structural stress, 105 Uhlenbeck, G. E., & Ornstein, L. S. (1930). On the theory of the brownian motion. Physical Review,36(5), 823–841. Williams, B. (2002). Truth and truthfulness: An essay in genealogy. Princeton University Press. 112