1. Short Definition

Signal-Proxy Scaling Risk occurs when a signal stops functioning as a passive measurement and becomes an active steering artifact that reshapes the system it was meant to describe.

At scale, signals become control surfaces.

2. Canonical Pattern

signal used for steering ⇒ signal shapes system behavior

Expanded:

measurement signal
+
selection / reward / punishment / ranking
⇒ actors adapt to signal
⇒ proxy behavior diverges from target

Plain form:

Scaled signals do not only report reality; they begin to produce behavior.

3. Mechanic Description

SCALE-055 defines the scaling risk of signals and proxies.

A signal may begin as a measurement:

• score
• metric
• grade
• ranking
• KPI
• benchmark
• diagnostic
• rating
• label
• compliance indicator
• risk class
• engagement number
• productivity measure
• safety score
• legitimacy signal

At small scale, the signal may provide useful information.

But when the signal becomes tied to reward, punishment, access, status, funding, enforcement, ranking, promotion, visibility, or legitimacy, actors begin adapting to the signal.

The signal becomes a control artifact.

The system may then optimize:

• the score rather than learning
• the metric rather than coherence
• the benchmark rather than capability
• the label rather than truth
• the compliance surface rather than safety
• the engagement number rather than meaning
• the risk class rather than actual risk
• the dashboard rather than reality

This is closely related to Goodhart behavior, but SCALE-055 is broader.

It emphasizes that signals become active steering structures under scale pressure.

Once a signal steers behavior, it must be treated as an operator, not merely an observation.

4. UTS Variable Mapping

5. Diagnostic Questions

1. What signal is being used?
2. Is it descriptive or steering behavior?
3. What rewards or penalties attach to the signal?
4. Are actors adapting to optimize the signal?
5. Does signal improvement correspond to coherence improvement?
6. What is the signal failing to measure?
7. Is hidden debt rising outside the measured surface?
8. Does the signal shape access, legitimacy, or ranking?
9. Can the signal be audited for back-action?
10. Should the signal be redesigned, bounded, or de-emphasized?

6. Failure Signatures

1. Proxy Capture

signal_target substituted for coherence_target

The system optimizes the signal instead of the real target.

2. Signal Back-Action

signal visibility + reward pressure ⇒ behavior changes

The act of measurement changes the system.

3. Measured-Surface Improvement

signal↑ while O↓

The signal improves while coherence declines.

4. Hidden Debt Outside Metric

metric success↑ + unmeasured burden↑

Cost moves outside the measured field.

5. Legibility Gaming

actors optimize what can be seen

The system becomes better at appearing successful.

7. Related Failure Modes

• proxy capture
• Goodhart collapse
• metric gaming
• signal back-action
• dashboard substitution
• compliance theater
• performance-coherence divergence
• hidden debt accumulation
• legitimacy distortion
• classification distortion
• pseudo-coherence

8. Related Diagnostics

9. Restoration Implications

If SCALE-055 is active, restoration requires signal governance.

Required actions:

1. Identify which signals are steering behavior.
2. Separate descriptive signals from control signals.
3. Audit signal-target alignment.
4. Track hidden debt outside the measured surface.
5. Reduce reward pressure on fragile proxies.
6. Add counter-metrics where necessary.
7. Rotate or contextualize metrics to reduce gaming.
8. Restore direct coherence checks.
9. Re-anchor signal use to O rather than Φ alone.
10. Validate over time through recurrence and ring-down.

Core restoration rule:

Treat scaled signals as control artifacts, not neutral mirrors.

10. Compact Registry Entry

id: SCALE-055
name: "Signal-Proxy Scaling Risk"
family: "SCALE-J — Attention, Signals, and Classification Mechanics"
type: "signal-proxy-control-mechanic"
status: "draft-ready"
short_definition: "Signals used for measurement, ranking, classification, governance, or optimization become steering artifacts that reshape system behavior at scale."
canonical_pattern: "signal used for steering ⇒ signal shapes system behavior"
failure_signature: "measurement signal + selection/reward/punishment/ranking ⇒ actors adapt to signal + proxy behavior diverges from target"
primary_variables:
  - O
  - H
  - ε
  - ι
  - Au
  - µᵢ
  - BΣ
  - K
  - R
  - Φ
primary_diagnostics:
  - signal_target_alignment
  - proxy_pressure
  - signal_back_action
  - Φ_signal
  - O_trend
  - H_unmeasured
  - Au_signal
  - gaming_rate
  - target_drift
  - affected_node_cost
related_failure_modes:
  - proxy_capture
  - goodhart_collapse
  - metric_gaming
  - signal_back_action
  - dashboard_substitution
  - compliance_theater
  - performance_coherence_divergence
  - hidden_debt_accumulation
  - legitimacy_distortion
restoration_implication: "Govern signals as steering artifacts, audit signal-target alignment, reduce fragile proxy pressure, track hidden debt outside metrics, and restore direct coherence checks."

11. One-Line Canon

At scale, a signal tied to selection stops being only a mirror and becomes a steering wheel.