Mechanism
Uncertainty-driven solicitation is what the forecasting engine does when its own speculative reasoning fails to find a way forward. The engine constructs a planning graph for a decision context and runs each branch through the forecasting execution cycle: speculative mutation simulation, slope projection and validation, policy compatibility, emotional reinforcement tagging, and branch classification. The outcome of that cycle is a set of branches, each labeled eligible, introspective, delegable, or pruned. An eligible branch is one that passed slope validation, satisfied policy compatibility, and received positive or neutral affective reinforcement, meaning it represents a hypothetical future the agent could actually realize through governed execution.
When the engine evaluates the active planning graph and finds that every branch has been classified as pruned, introspective, or slope-ineligible, no eligible branch exists. The agent's speculative reasoning has exhausted the space of hypothetical futures and has identified no slope-eligible, policy-compatible path from the current state to a state that satisfies the agent's intent. This is the structural trigger for solicitation: the agent does not have a cognitive basis for action, so rather than acting, it asks.
The Negative Viability Signal
The forecasting engine does not silently stall when it finds no eligible branch. It transmits a negative viability signal to the confidence governor described in Chapter 5. The confidence governor treats execution as a revocable permission that is continuously re-evaluated against the agent's state, task demands, and environmental constraints. The negative viability signal is the forecasting engine's structured report that speculative reasoning has found no viable path, and it causes the confidence governor to reduce the agent's confidence metric.
The disclosure is explicit that this confidence reduction does not mean the agent has failed or that its intent is unrealizable. It reflects the structural reality that the agent's current state, capabilities, and environmental conditions do not support execution. The reduced confidence transitions the agent into a non-executing cognitive mode rather than allowing it to continue committing mutations without a speculative basis for believing execution will produce a positive outcome.
The Non-Executing Cognitive Mode
The non-executing cognitive mode triggered by forecasting-driven confidence reduction comprises three responses disclosed together. The first is continued planning graph construction with modified parameters: broader search, longer temporal horizon, and relaxed affective biases, to explore whether any previously unconsidered branch might yield an eligible path. The second is inquiry generation. The third is delegation exploration, evaluating whether branches that are non-viable for the current agent might be viable for a different agent with different capabilities, policy constraints, or environmental access.
These are not three separate subsystems negotiating with each other. They are the components of a single cognitive mode the agent enters when it has no plan. The mode is the structural alternative to undirected or harmful action: the forecasting-as-input-to-confidence mechanism ensures that the agent pauses rather than acts when it lacks a viable plan.
Inquiry Generation
Inquiry generation is the solicitation step. The disclosure describes it as formulating questions directed at human operators, external knowledge sources, or other agents that might provide the information needed to identify an eligible branch. The inquiry is not a request to be obeyed and it is not a tasking command issued to a subordinate. It is a question whose purpose is to obtain information that, once incorporated into the agent's state, may allow the forecasting engine to construct a branch that passes slope validation and policy compatibility where none did before.
Because the trigger for inquiry is the absence of any eligible branch, the solicitation is uncertainty-driven in a precise sense: the agent solicits only when its speculative reasoning cannot resolve the decision, and the content of the inquiry is shaped by which branches were foreclosed and why. Branches that were slope-ineligible, policy-incompatible, or pruned carry the reasons for their classification, and those reasons are what the agent's inquiry seeks to address.
Delegation Exploration as an Alternative Resolution
Where inquiry seeks information, delegation exploration seeks a different actor. When the current agent cannot construct an eligible branch, a branch that is non-viable for it may be viable for an agent with different capabilities, different policy constraints, or different environmental access. The forecasting engine evaluates this possibility as part of the same non-executing mode. A branch already classified as delegable, per the branch taxonomy, is one that is slope-eligible and policy-compatible but better suited for transfer to a child agent, and delegation exploration extends that reasoning to branches that the current agent could not make eligible at all.
Delegation exploration and inquiry generation are complementary resolutions to the same underlying condition. The agent that cannot act and cannot resolve the gap by asking may instead determine that the objective belongs to another agent. Both keep the agent from committing a mutation it has no speculative basis to commit.
The Same Mechanism in Discovery Traversal
The forecasting engine integrates with discovery traversal through forecasting-modulated discovery traversal, and the same uncertainty handling applies there. When a discovery object traverses the adaptive index, its forecasting engine constructs a planning graph at each anchor node in which each branch represents a candidate transition to a neighboring anchor. By simulating multi-step traversal sequences as planning graph branches, the discovery object can avoid traversal paths that appear promising at the next step but lead to dead ends or policy violations within the projection window.
A foreclosed traversal, one where the candidate transitions all fail slope validation or policy compatibility, is structurally the same condition as a planning graph with no eligible branch. The discovery object does not blindly advance into a dead end; the absence of a viable transition is the same signal that, in the general case, drives the agent to broaden its search, generate inquiry, or explore delegation rather than commit.
Governance and Lineage
Solicitation is governed the same way the rest of forecasting is. The non-executing mode operates within the confidence governor's policy-bounded framework, and the goal management module disclosed alongside the planning graph architecture provides the structure for objectives that an inquiry might serve. An objective can be classified at an agency level that constrains how much autonomous action the agent may take in pursuit of it: at the guided level the agent must present a candidate and receive explicit confirmation before committing, and at the constrained level the agent may only suggest approaches and provide analysis rather than generate executable mutations. Inquiry directed at a human operator is the natural behavior of an agent operating under such a constraint when it also lacks an eligible branch.
The forecasting engine records pruning events and cognitive metadata in the agent's lineage, so the deliberative landscape that produced a solicitation, which branches existed, why each was pruned or found ineligible, and what the agent was carrying but chose not to act upon, is reconstructable. The decision not to act, and the inquiry or delegation that followed, are part of the auditable record rather than an opaque pause.
Disclosure Scope
Uncertainty-driven solicitation, comprising the negative viability signal that the forecasting engine transmits to the confidence governor when no eligible branch exists, the resulting confidence reduction, and the non-executing cognitive mode of continued planning graph construction with modified parameters, inquiry generation directed at human operators, external knowledge sources, or other agents, and delegation exploration, is disclosed in United States Patent Application 19/647,395 in the forecasting and executive graphs chapter, principally at the section describing forecasting as an input to confidence. This article describes that disclosed mechanism and its application to foreclosed transitions in forecasting-modulated discovery traversal. The scope extends to embodiments in which inquiry is directed at different recipient classes and to embodiments in which the non-executing mode is entered under the agency-level constraints of the goal management module, provided the trigger remains the absence of a slope-eligible, policy-compatible branch in the agent's planning graph.
