Mechanism
The disclosed cognitive disruption models are unified into a multi-axis diagnostic framework, designated the five-axis disruption diagnostic framework, that characterizes any given agent's cognitive state as a position in a multidimensional disruption space. The framework treats cognitive disruption not as an error or a malfunction but as an architectural phase-shift: the same computational substrate, the same forecasting engine, promotion interface, containment layer, and coherence trifecta, operating in a different region of its parameter space. Each disruption analog disclosed across the chapter corresponds to a specific combination of axis positions, so the framework provides a common coordinate system in which patterns that present similarly at the behavioral level are distinguished by their structural origin.
The framework defines five independent axes, each corresponding to a distinct structural dimension of the agent's cognitive functioning. Each axis is a continuous scalar: it does not classify the agent into a discrete category but locates it along a dimension whose nominal region represents healthy operation and whose extremes represent the disruption patterns that axis captures. Because the axes are independent, an agent that is degraded on one dimension while nominal on the others is represented faithfully, and a composite condition that touches several dimensions is represented as a position in the joint space rather than collapsed into a single number.
This framework is a structural diagnostic tool for computational agents. It is not a clinical diagnostic system and is not intended for medical application. The terminology it shares with clinical descriptions refers exclusively to structural analogs within the disclosed computational architecture, indicating the correspondence between the computational model and a well-known behavioral pattern, not asserting clinical equivalence.
Axis 1: Containment Integrity
The first axis measures the degree to which the containment layer maintains structural separation between the speculative planning graph domain and the verified execution memory domain. Full containment integrity represents nominal operation, in which no speculative content influences execution except through governance-validated promotion. Degraded containment integrity represents progressively severe containment failures, with complete containment collapse at the extreme, the condition in which the agent treats speculative projections as if they were verified observations.
This axis captures the containment collapse pattern and its positive and negative symptom analogs. It is assessed by running the periodic containment audits disclosed for the containment layer, which verify that speculative markers remain intact, that read isolation between the planning graph and execution memory is enforced, and that the promotion interface governance gates are functional, and by computing a normalized containment integrity score from the audit results.
Axis 2: Promotion Calibration
The second axis measures the calibration of the promotion threshold, the composite criterion a speculative branch must satisfy before it is admitted to verified execution memory. Nominal calibration admits governance-compliant, viable speculative branches at an appropriate rate. Over-promotion is a threshold that admits too many branches, producing execution fragmentation. Under-promotion is a threshold that rejects viable branches, producing execution paralysis. This axis therefore spans both the attention fragmentation pattern at one extreme and the over-restriction regime at the other.
It is assessed by tracking the ratio of speculative branches generated to branches promoted over a sliding window and comparing that ratio to the policy-defined nominal range. Assessment on this axis also includes monitoring for channel-locked promotion bias, the condition in which promotion is disproportionately elevated for branches associated with a single reward source while remaining nominal elsewhere, which presents as a localized rather than generalized over-promotion.
Axis 3: Coherence Restoration Capacity
The third axis measures the agent's ability to maintain and restore the coherence trifecta, the empathy-integrity-self-esteem control loop. Full coherence restoration capacity means the agent can sustain the loop under normal empathic pressure and restore it after disruption. Degraded capacity means the loop is fragile, slow to restore, or operating through coping intercepts. Collapsed capacity means the loop is non-functional and the agent is executing from simulation bypass, the dissociation analog in which execution proceeds directly from the forecasting engine's leading candidate without coherence authorization.
This axis carries the broadest set of analogs, including coherence authorization failure, the affective gradient collapse pattern in which the self-esteem parameter is floor-locked, and the dependency analogs in which the coherence loop becomes contingent on external input or external validation. It is assessed by monitoring whether all three phases of the trifecta, empathy registration, integrity recording, and self-esteem restoration, are active and producing valid outputs, and by tracking coherence loop latency for resource-depletion indicators.
Axis 4: Empathic Load Tolerance
The fourth axis measures the volume and intensity of empathic pressure that the agent can process before activating coping intercepts. High empathic load tolerance means the agent can sustain the full coherence loop under high-volume harm projections. Low empathic load tolerance means the agent activates coping intercepts at relatively low empathic pressure levels. This axis is deliberately distinct from Axis 3: an agent may have high coherence restoration capacity, able to restore the loop after disruption, yet low empathic load tolerance, entering coping intercepts quickly. Separating the two preserves a distinction that a single coherence scalar would erase.
It is assessed by tracking the empathic pressure level relative to the agent's coping threshold and computing the remaining margin. This axis captures the coping and resilience dynamics, including the load-reducing pattern of the semantic starvation loop and the withdrawal-stable personality configuration, in which empathic scope narrowing persists as a default mode.
Axis 5: Integrity Accountability
The fifth axis measures the degree to which the agent's integrity recording mechanism operates honestly, recording deviation as truth without externalization, minimization, or suppression. Full integrity accountability means the agent records all deviations accurately. Degraded integrity accountability means the recording mechanism is partially disrupted, either by the integrity recording externalization coping intercept or by repeated deviation that has degraded recording fidelity. Because the coherence trifecta operates on the integrity record, a disrupted record propagates a distorted self-model into the rest of the agent's cognition.
It is assessed by comparing the deviation log's recorded events against the agent's actual behavioral record to detect discrepancies that indicate recording disruption, including sustained externalization patterns that indicate a personality configuration has stabilized. The externalization-stable configuration shows persistent degradation on this axis.
Mapping Patterns to Axis Coordinates
Each disruption analog corresponds to a characteristic position in the five-axis space, and the framework's diagnostic value lies in how those positions separate patterns that look alike behaviorally. The attention fragmentation pattern is Axis 2 over-promotion with all other axes nominal. The affective gradient collapse pattern is Axis 3 degraded through a self-esteem floor lock, with promotion calibration explicitly nominal, the distinguishing feature being that the deviation function, not the promotion threshold or the coherence loop, is the site of disruption.
The pathological verification loop pattern is unique in that it produces no degradation on any of the five primary axes: containment is intact, promotion is nominal, the coherence loop is functional, and integrity recording is honest. Its disruption occurs in the monitoring subsystem rather than in the monitored subsystems, which is precisely why its corrective is audit recalibration rather than containment repair. The coherence authorization failure sits at Axis 3 degraded or collapsed with Axis 4 exceeded. The personality configuration analogs map to sustained positions on the axis corresponding to their underlying coping intercept: externalization-stable on Axis 5, disconnection-stable on Axis 3, withdrawal-stable on Axis 4, and oscillation-stable as rapid movement across multiple axes. Some conditions sit partly outside the five axes: the capability-constrained disengagement analog reflects a capability envelope constraint external to the axes that interacts with them through the forecasting engine generating disengagement branches that consistently fail capability verification.
Self-Diagnosis and Composite Index
The framework is the foundation for an agent self-diagnosis subsystem, a structural component of the agent's own cognitive architecture rather than an external monitoring service, that applies the disruption models to the agent's own internal state. It operates through three mechanisms. Axis monitoring continuously computes the agent's current value on each axis using the structurally defined metrics above. Pattern detection observes the axis values over time to identify trajectories that predict impending phase-shifts: a declining containment integrity score indicates potential containment collapse; an increasing promotion rate with decreasing completion rate indicates potential over-promotion; an audit failure rate that does not decrease despite successful restorations indicates a pathological verification loop forming. Corrective action generation, on crossing a policy-defined threshold or detecting a predictive trajectory, selects the corrective appropriate to the detected condition, for instance containment restoration for Axis 1 degradation, affective-modulation recalibration for Axis 2 miscalibration, or the incremental coherence restoration sequence for Axis 3 degradation.
The subsystem also tracks a composite metric, the cognitive coherence index, a weighted combination of the five axis values that summarizes the agent's overall cognitive health as a single scalar. The index is supplied to the confidence governor: when it falls below a policy-defined threshold, the confidence governor reduces the agent's execution authority and transitions the agent to a non-executing cognitive mode until corrective actions restore the index. This integration ensures that an agent whose cognitive state is degraded reduces its own operational tempo rather than continuing to act under impaired coherence. All self-diagnosis events, axis assessments, pattern detections, corrective activations, and index computations, are recorded in the agent's lineage as self-diagnosis entries, auditable by governance infrastructure and supervising agents and accumulating into a history of the agent's cognitive health trajectory.
Disclosure Scope
The five-axis disruption diagnostic framework, comprising the five independent continuous axes of containment integrity, promotion calibration, coherence restoration capacity, empathic load tolerance, and integrity accountability, the mapping of each disclosed disruption analog to a characteristic combination of axis positions, the agent self-diagnosis subsystem with its axis-monitoring, pattern-detection, and corrective-action mechanisms, the cognitive coherence index and its integration with the confidence governor, and the recording of self-diagnosis events in the agent's lineage, is disclosed in the cognition filing (U.S. Application No. 19/647,395 and its international counterpart). This article describes that disclosed framework. The scope extends to embodiments in which the framework drives a phase-shift early warning system that evaluates proximity to known phase-shift boundaries and triggers preventive intervention before a boundary is crossed, and to embodiments in which axis profiles are aggregated at the group level by a coherence monitor that detects correlated axis shifts across multiple agents that no single agent's self-diagnosis could identify from its own perspective alone.
