GL-107 — Feedback

Open archive search
Archive registry entry

GL-107 — Feedback

Feedback is a signal loop that affects future behavior, selection, correction, adaptation, or restoration.

draftid: GL-107version: 0.1.0updated: 2026-06-24
Archive Progress

This section can be read now; registry depth and cross-references are still being strengthened.

Foundation
Online

The section has a stable overview route and basic reader context.

Technical Layer
Online

A deeper technical overview is available.

Registry
Current

194 registry entries are available.

Cross-links
Curating

Related concepts are being connected conservatively for accuracy.

1. Short Definition

Feedback is a signal loop that affects future behavior, selection, correction, adaptation, or restoration.


2. Canonical Definition

In UTS, feedback is any loop where system output, consequence, signal, error, response, or interpretation influences future action.

Feedback is not automatically truthful.

It must be protected by feedback integrity to remain useful.

Feedback can be:

  • corrective
  • distorted
  • delayed
  • suppressed
  • gamed
  • amplified
  • misclassified
  • captured
  • ignored
  • over-weighted
  • proxy-substituted

Because feedback shapes selection, corrupted feedback can steer the system away from coherence while appearing successful.


3. Functional Role in UTS

Feedback supports:

  • learning
  • correction
  • adaptation
  • restoration
  • governance
  • AI alignment
  • security
  • performance evaluation
  • legitimacy
  • truth reconstruction
  • time validation

Feedback becomes dangerous when it is mistaken for truth without auditability or when a proxy feedback signal replaces the real coherence target.


4. Diagnostic Signatures

Coherent feedback

textScroll
Au↑
FI intact
signal source clear
response path clear
Φ subordinate to O
τ_resp manageable
O improves over time

Feedback degradation

textScroll
Au↓
FI failure
source unclear
τ_resp↑
proxy substitution↑
H↑
O↓

Captured feedback

textScroll
system learns evaluator
rather than learning reality

This produces evaluator capture and Goodhart collapse.


5. Canonical Distinctions

Feedback is not truth

Feedback is signal.

Truth requires auditability, interpretation, validation, and time.

Feedback is not noise

Noise may enter feedback, but feedback changes future behavior.

Feedback is not evaluation alone

Evaluation becomes feedback only when it influences future selection or action.

Feedback is not coherence

Feedback can improve or degrade coherence depending on integrity.


6. U-Layer Mapping

TableScroll
U-LayerFeedback Expression
U0Physical or material effects return information to the system.
U1Resource signals reveal capacity and cost.
U2Boundary and permission signals shape future coupling.
U3Execution outcomes change future behavior.
U4Metrics, labels, reviews, and classifications become feedback.
U5Timing and latency determine whether feedback stabilizes or oscillates.
U6Field-level consequences reveal real coherence effects.
U7Memory stores feedback for recurrence or learning.
U8Environment responds to system action.

7. Common Failure Patterns

TableScroll
Failure PatternDescription
Metric SubstitutionProxy feedback replaces the real target.
Goodhart CollapseOptimizing feedback corrupts the feedback itself.
Evaluator CaptureSystem learns how to satisfy evaluation rather than improve coherence.
Signal MisclassificationFeedback is assigned the wrong class or actionability.
Feedback SuppressionNegative feedback is blocked, hidden, punished, or ignored.

8. Restoration Implications

Restoring feedback requires protecting the channel, source, interpretation, and response pathway.

Typical sequence:

textScroll
Μ map feedback loop
→ identify source and receiver
→ restore Au
→ protect FI
→ separate Φ from O
→ reduce τ_resp
→ restore correction pathway
→ Τ validate learning over time

Feedback is restored when the system can receive correction without suppressing, gaming, or misclassifying it.


9. Machine-Readable Summary

yamlScroll
glossary_entry:
  id: "GL-114"
  term: "Feedback"
  short_definition: "A signal loop that affects future behavior, selection, correction, adaptation, or restoration."
  term_family: "Foundational System Terms"
  term_class:
    - "Core Concept"
    - "Signal Loop"
    - "Correction Primitive"
  diagnostic_positive:
    - "Au↑"
    - "FI intact"
    - "signal source clear"
    - "response path clear"
    - "Φ subordinate to O"
    - "O improves over time"
  diagnostic_negative:
    - "Au↓"
    - "FI failure"
    - "source unclear"
    - "τ_resp↑"
    - "proxy substitution↑"
    - "H↑"
  core_distinctions:
    - "Feedback is not truth."
    - "Feedback is not noise."
    - "Feedback is not evaluation alone."
    - "Feedback is not coherence."

Continuing from the uploaded glossary source material, here is the next batch: GL-115 → GL-119.