Infrastructure · Data Structures
PerformanceSignal
Typed artifact emitted by a Solver's Feedback surface — downstream evaluation that feeds Pathway Memory
Definition
Mechanism
The typed artifact emitted by a Solver's Feedback surface — the downstream evaluation of how a completed transaction performed against its acceptance criteria, or the escalation signal that invokes a FrameError when failure is structural rather than local. Distinct from the general Feedback primitive: PerformanceSignal is the contract-typed output that feeds PathwayMemory (quality, depth, cost across problem classes) and drives localized learning at each node. The signal is produced by the consumer of a Result (parent Solver or end-caller) after the Acceptance Gate either admitted the Result as a Solution or rejected it back for reframing.
Invariants
- Typed: the signal conforms to a declared schema so PathwayMemory can aggregate across instances.
- Attributable: names the Solver node that produced the evaluated Result, so learning is localized.
- Three-outcome: either a scored evaluation, a pass/accept confirmation, or a FrameError escalation.
Risks
Failure modes
Sycophantic drift
evaluator shares the generator's blind spots and rubber-stamps weak outputs.
Signal poisoning in open commons
an untrusted consumer fabricates penalties to steal work (mitigated by ReceptivityGate).
Latency decay
the signal arrives long after the node has moved on, weakening the learning loop.
Grade inflation
evaluators converge on 'good enough' scoring so all signals compress into a narrow band.
Relationships
Builds on
References
Error correction signal
Mind
Typed failure signaling that lateral reframing, not retry, is required
Mind
Learned routing memory: problem_class → route → outcome_quality
Society
Verification gate for signed, clause-specific rejection records
Used by
References
Mind
Solver implementing the five-surface contract (polymorphic across substrates)
Mind
Abstract five-surface contract: Manifest, Execute, Consult, Verify, Feedback
Reference
Data schema
{
"type": "object",
"required": [
"solver_ref",
"outcome"
],
"properties": {
"solver_ref": {
"type": "string",
"description": "sema_id of the Solver node being evaluated"
},
"outcome": {
"type": "string",
"enum": [
"accepted",
"scored",
"frame_error"
],
"description": "Shape of the signal"
},
"score": {
"type": "number",
"description": "Optional quality score when outcome=scored"
},
"frame_error_ref": {
"type": "string",
"description": "sema_id of the FrameError when outcome=frame_error"
},
"evaluator_ref": {
"type": "string",
"description": "Identity of the evaluating entity, for trust-weighted aggregation"
}
}
}