Mechanism
The non-executing cognitive mode is the operating state an agent enters when the confidence governor withdraws execution authorization but the agent's cognition continues. In the disclosed architecture execution is treated as a revocable permission rather than a default assumption: execution is a conditional privilege that must be continuously earned by the agent's demonstrated sufficiency, and the confidence governor is the hard gate that grants and revokes it. When the governor determines that execution authorization should be withdrawn, execution ceases. The agent cannot override the withdrawal through self-assessment, affective escalation, or policy reinterpretation, because no alternative pathway to execution exists that bypasses the gate.
What makes the suspended state a distinct mode, rather than a halt or a failure, is that the withdrawal of execution authorization does not impair the agent's ability to think. The architecture enforces a structural separation between the execution subsystem and the cognitive subsystems. Execution suspension is not cognitive suspension. An agent whose execution is suspended retains full access to its forecasting engine, its affective state field, its integrity engine, and its memory field. It is fully cognitively active while structurally prohibited from acting.
The Three Authorization States
Execution authorization gating operates in one of three states. In the authorized state the confidence value is above the authorization threshold and the confidence trajectory triggers no alarm conditions, and execution is permitted. In the suspended state the confidence value has fallen below the authorization threshold, or the trajectory has triggered a preemptive suspension, and execution is prohibited while cognitive processes continue. In the locked state a severe integrity violation, a catastrophic resource failure, or a governance-mandated halt has occurred, and both execution and certain cognitive processes are restricted pending external review.
The non-executing cognitive mode is the suspended state. It is distinguished from the locked state by what remains running: the suspended state preserves the agent's cognitive capacity while removing its ability to act, whereas the locked state is reserved for conditions in which continued cognitive operation itself may produce harmful effects, for example when integrity has been so severely compromised that the agent's reasoning cannot be trusted. Transition from any state to locked occurs only upon governance-mandated triggers and is not reversible by the agent itself; locked-state recovery requires external authorization.
Why the Prohibition Cannot Be Bypassed
The prohibition on action during suspension is not implemented as a flag that the execution subsystem checks and optionally respects. It is implemented as a structural decoupling of the execution subsystem's output pathway, such that the subsystem cannot produce effects regardless of its internal state or the urgency of the agent's intent. The separation is enforced at the substrate level through distinct processing pathways. The cognitive pathway comprises all processing that evaluates, reasons about, projects, or represents state without producing externally observable effects: forecasting, planning-graph construction and evaluation, confidence computation, affective state updates, integrity evaluation, and inquiry generation. The execution pathway comprises all processing that commits mutations to verified state, produces externally observable outputs, initiates delegation, or consumes irreversible resources.
The confidence governor gates only the execution pathway. The cognitive pathway remains active regardless of the authorization state. Because the gate operates on the output pathway rather than inside the agent's decision logic, the prohibition is robust against the agent's own deliberation: an agent that has reasoned itself into urgency still cannot act while suspended.
Cognition During Suspension
The non-executing cognitive mode is not idle, passive, or waiting. It is an active cognitive state in which the agent redirects its processing capacity from execution to deliberation. While suspended, the agent may construct new planning graphs exploring how to recover execution authorization; evaluate the conditions that caused confidence to decay and identify potential remediation strategies; generate inquiry requests seeking information that would resolve the uncertainty or capability gaps contributing to low confidence; perform introspective analysis of its own state to determine whether affective biases, integrity degradation, or memory distortions are contributing to the confidence deficit; and forecast the consequences of alternative action sequences that might be available when execution authorization is restored.
This is the disclosed pause-to-think behavior. When confidence falls below the execution authorization threshold but remains above a minimum engagement threshold, the agent enters a structured inquiry mode comprising hypothesis expansion, information ingestion, re-evaluation loops, and condition monitoring. The agent is freed from execution urgency, so it can generate and evaluate less probable but potentially valuable hypotheses that would be pruned during normal execution. The inquiry mode operates iteratively: at each iteration the agent updates its confidence computation with new information or revised assessments and determines whether confidence has recovered sufficiently to warrant a transition back toward execution authorization.
Suspension as a Structural Shield
The suspended state participates in a bidirectional confidence-integrity feedback loop. On the forward path, when the integrity engine detects a deviation event, the integrity field degrades and that degraded value is propagated to the confidence computation as an adverse input, reducing confidence in proportion to the severity of the violation, with severe deviations capable of independently triggering suspension. On the reverse path, because a suspended agent is structurally prevented from committing mutations to verified state, it cannot commit integrity-violating mutations. The execution suspension thereby creates a structural shield against further integrity degradation: the agent cannot make its integrity problem worse because it cannot act.
The loop converges. Integrity violations cause confidence to drop, the drop suspends execution, suspension prevents further violations, and the cessation of integrity-degrading action creates conditions under which integrity restoration can proceed. A circuit-breaker mechanism guards against deadlock: if integrity is so severely degraded that no achievable confidence value can support reauthorization, the circuit breaker transitions the agent to the locked state and signals that external intervention is required.
Leaving the Mode: Recovery of Authorization
Re-entry to the executing state is not a function of elapsed time. Recovery of execution authorization is a structured process comprising three phases: confidence restoration, stability verification, and reauthorization. Confidence restoration is the process by which the confidence value rises from below the threshold to above it, resulting from resolution of the adverse conditions that caused the original drop, from inquiry operations that supplied confidence-enabling information, from changes in task state that reduced demands, or from environmental changes that mitigated adverse factors. It is computed by the same confidence evaluation function used during normal operation, applied to the agent's updated state.
Stability verification follows restoration and precedes reauthorization. During this phase the confidence governor monitors the confidence value and trajectory for a configurable verification period to confirm the restored confidence is stable: that it is not fluctuating near the threshold, that the differential rate is not trending negatively, and that no alarm conditions are active. The phase implements a hysteresis requirement: the confidence value must exceed the authorization threshold by a configurable hysteresis margin throughout the verification period, so that the agent's confidence is not merely above the threshold but meaningfully above it. Only on successful completion of stability verification does the governor reauthorize, reconnecting the execution subsystem's output pathway and returning the agent to the authorized state, at which point it incorporates the products of its suspension-time cognition into its resumed plan.
Embodied Execution and the Safe Physical State
For embodied agents that control physical actuators or robotic systems, the non-executing mode acquires an additional, stricter layer. The confidence governor for embodied agents implements a physical safety floor: a minimum confidence threshold below which no physical action is permitted regardless of task urgency, intent priority, or external command. The physical safety floor is set higher than the general execution authorization threshold and cannot be overridden by the agent's own deliberation or by delegation commands from parent agents, reflecting that physical actions carry risks categorically more severe than those of computational actions.
When an embodied agent's confidence drops below the physical safety floor, it transitions to a safe physical state: a predefined configuration in which all actuators are brought to a controlled stop, all end effectors are moved to safe positions, and the agent's physical presence is made inert. The transition is immediate and overrides any in-progress physical action. The agent remains cognitively active in the non-executing cognitive mode but cannot produce physical effects until confidence is restored above the physical safety floor.
Distinction From Failure and From Reactive Pause
In conventional systems an agent that stops executing has either succeeded or failed, and runtime pause-and-resume capabilities suspend execution reactively in response to external failures or resource interruptions. The disclosed confidence governor suspends execution proactively, based on the agent's own continuously computed assessment of its sufficiency, enabling the agent to stop itself before damage occurs rather than recovering after damage has occurred. The transition from executing to non-executing is therefore not synonymous with failure: a suspended agent may be in a state of deliberate, structurally governed pause in which cognition continues but action does not.
Disclosure Scope
The non-executing cognitive mode, comprising the structural separation of the execution pathway from the cognitive pathway, the three authorization states of authorized, suspended, and locked, the gating of only the execution pathway at the substrate level so the prohibition cannot be bypassed by the agent's deliberation, the active suspension-time cognition of forecasting, inquiry, introspection, and delegation evaluation, the confidence-integrity feedback loop in which suspension acts as a structural shield against further integrity degradation, the three-phase recovery of confidence restoration, hysteresis-bounded stability verification, and reauthorization, and the embodied physical safety floor with its safe physical state, is disclosed in the cognition filing (U.S. Application No. 19/647,395 and its international counterpart) in Chapter 5. This article describes that disclosed mechanism. The scope extends to single-agent and multi-agent embodiments and to computational and embodied agents, provided that execution authorization is gated structurally on a pathway the agent's decision policy cannot reach and that cognition continues while action is suspended.
