Governed Measurement from Observable Signals to Contract-Bound Findings

SubstrateX Aperture™ does not treat every computed number as evidence of equal authority.

Operational records can support many forms of measurement: direct observations, deterministic transformations, composite indicators, bounded proxies, interpretive classifications, and experimental constructs. If these are presented without distinction, an operator can easily mistake a projection for an observation or a heuristic score for a validated probability.

Aperture addresses this problem by separating two implemented layers:

1. behavioral telemetry formation, in which observable runtime records are transformed into versioned and source-linked signal records; and

2. governed instrument projection, in which contract-bound instruments read those signals and issue bounded findings over the same evidence identity.

The governing distinction is:

Telemetry is computed from the reconstructed runtime. Instruments project bounded findings over that telemetry.

An instrument does not create the behavioral record it examines. It cannot manufacture a missing signal, rewrite a metric, move a temporal marker, or increase the authority of a finding through visualization. It can only read the evidence objects and signal classes permitted by its contract and issue the kind of finding that contract allows.

From Reconstruction to Measurement

Runtime Reconstruction establishes the ordered longitudinal record: canonical frames, source and canonical coordinates, events, regimes, markers, temporal relationships, and replayable runtime structure.

Behavioral telemetry measures registered properties of that reconstructed runtime.

The engineering sequence is:

Qualified operational record
↓
Canonical runtime and source-linked frames
↓
Observable feature construction
↓
Level 0 core telemetry
↓
Level 1 behavioral dynamics
↓
Level 2 reconstruction and instrument projections
↓
Contract-bound Instrument Findings
↓
Investigation, synthesis, and preservation

Every downward step depends on the authority and computability of the step above it. Level 1 dynamics cannot become more authoritative than their Level 0 basis. Level 2 projections cannot silently present themselves as direct observations. Instrument Findings cannot extend the claim boundary of the signals they consume.

This hierarchy makes measurement inspectable before it becomes explanatory.

Behavioral Telemetry

Behavioral telemetry is the structured measurement of observable change across a runtime.

It may describe changes in properties such as:

• entropy and repetition;

• lexical or trajectory drift;

• divergence from earlier states;

• variation in response length or structure;

• reasoning-density indicators;

• semantic curvature;

• contraction or transition pressure;

• recurrence and echo persistence;

• runtime coherence and continuity;

• temporal coupling;

• role phase lag and handoff shear;

• worldline stability;

• Basin proximity;

• regime state;

• recovery support; and

• runtime motion.

These measurements concern Longitudinal Computational Behavior: how observable computational activity develops across an ordered record. They do not reveal private reasoning, model weights, inaccessible activations, subjective experience, human intention, or an unobserved causal mechanism.

Behavioral telemetry is therefore not a claim that Aperture has recovered the complete internal process. It is a governed measurement layer over what the supplied record permits the system to observe and derive.

Three Signal-Authority Levels

The reviewed implementation organizes the active Signal Authority registry into three levels.

Level 0 — Core Surface Telemetry
Computes foundational measurements from source-visible and canonical runtime features.
Entropy · Lexical Drift · Divergence · Repetition · Length Volatility · Reasoning Density

Level 1 — Behavioral Dynamics
Derives longitudinal relationships and pressures from Level 0 signals, events, roles, and runtime structure.
Semantic Curvature · Contraction Pressure · Echo Persistence · Runtime Coherence · Temporal Coupling · Transition Pressure · Collapse Stress · Constraint Re-entry · Role Phase Lag ·
Handoff Shear

Level 2 — Reconstruction and Instrument Projections
Projects higher-order runtime properties from the governing evidence run and registered lower-level signals.
Worldline Stability · Basin Proximity · Basin Exit Risk · Attractor Lock · Regime State · Recovery Signal · Runtime Stability · Runtime Continuity · Runtime Motion

In this implementation, the active Signal Authority registry contains 25 governed entries: 6 Level 0, 10 Level 1, and 9 Level 2.

That number is registry-specific. Aperture may expose additional measurements, readouts, temporal operators, appendix fields, or research channels elsewhere. Those broader totals must not be presented as though they were automatically equivalent to the 25 entries governed by this Signal Authority version. Public counts should always identify the registry, release, and inclusion rule to which they refer.

Level 0 Is Foundational, Not Automatically Direct

Level 0 identifies the foundational telemetry layer. It does not mean that every Level 0 value is direct evidence.

The supplied source record, an observed timestamp, a preserved role label, or an explicitly recorded event may qualify as direct evidence. Entropy, divergence, repetition, or lexical drift are ordinarily computed from that evidence. They are foundational measurements, but their legitimacy remains derived unless a specific field was directly supplied by the source under a declared schema.

This distinction prevents the telemetry hierarchy from being confused with the evidence-legitimacy hierarchy.

• Signal level describes where a measurement sits in the computational dependency chain.

• Metric legitimacy describes what kind of evidentiary claim that measurement can support.

A Level 0 signal can be derived evidence. A Level 2 finding can be a bounded projection. Neither becomes direct evidence merely because the system displays it prominently.

The Signal Record

Each governed signal is more than a name and value. Its authority record can include:

• canonical signal identifier;

• registry level and family;

• public label and scientific lineage;

• current value, when computable;

• evidence status;

• measurement window;

• computational basis;

• source spans;

• coverage information;

• source-field mapping;

• selected source path;

• observed legacy or alternate fields;

• derived dependencies;

• claim boundary;

• forbidden claims;

• method and registry version; and

• notes describing what the signal does and does not establish.

This structure makes the signal challengeable. An operator can ask not only what value was displayed? but also:

• Where did it come from?

• Which source interval did it cover?

• Which lower-level signals supported it?

• Was it computed, derived, projected, partial, or proxy-only?

• Which method version produced it?

• Which claims are excluded?

Without this surrounding record, a metric is only a number. With it, the metric becomes an inspectable evidence-bearing measurement.

Versioned Signal Registries

Signal definitions are versioned because measurement meaning can drift even when a label remains unchanged.

A registry version governs:

• the canonical signal name;

• its level and family;

• public-safe terminology;

• recognized legacy aliases;

• candidate source paths;

• dependency relationships;

• admissible evidence statuses;

• claim boundaries; and

• the method or adapter used to resolve the signal from the governing run.

Legacy names may be retained as lineage metadata or mapped to a canonical public label. This mapping does not authorize the adapter to change the underlying formula or create a value where none exists. The selected source path and mapping status remain part of the signal record.

When a formula, dependency, threshold, or classification rule changes, the relevant registry or method version must also change. A visual redesign alone does not constitute a new signal version, and a renamed label does not create a new scientific result.

Evidence Status and Computability

Every signal or finding must disclose whether it is actually computable from the active evidence run.

Aperture uses explicit states such as:

Source-Observed
The relevant fact or event was supplied by the operational record under a declared source relationship.

Computed
The value was produced by a registered method from an available evidence basis

Derived
The value was deterministically formed from other registered evidence or measurements.

Projected
The value is a Level 2 projection over governed evidence and signal records.

Partial
Some, but not all, of the required evidence basis is present.

Proxy Only
The value depends on an indirect but explicitly bounded substitute measure.

Not Supplied
The source or evidence run did not contain the required field or anchor.

Not Computable
The required inputs, coverage, or authority conditions are absent.

Unavailable
The relevant module, registry, or evidence object cannot be resolved.

Zero and false are valid computed values when the method supports them. They must not be mistaken for missingness. Conversely, an empty field must not be converted into zero to make an instrument appear complete.

Computability is determined from the evidence basis, not from interface demand.

Coverage Is Not Sufficiency

A signal may be present without being sufficient for a particular finding.

Coverage describes the material available to a measurement—for example, the number of canonical turns, replay frames, ledger events, roles, or source spans included in its window. Sufficiency asks whether that material satisfies the requirements of a specific computation or projection.

An instrument contract may require:

• a governing evidence run;

• aligned run and Passport identity;

• a replayable frame sequence;

• particular Level 0, Level 1, or Level 2 signals;

• qualifying events or markers;

• adequate source coverage;

• a valid operational context; or

• a required optional channel, where the finding depends on one.

If only part of that basis exists, the result remains partial. If a required dependency is absent, the finding is not computable. The interface may explain what is missing, but it cannot complete the basis through inference.

Metric Legitimacy

The Metric Legitimacy Matrix classifies measurements according to the authority of the evidence relationship they represent.

Direct Evidence
A source-observed record, field, event, role, timestamp, or explicitly supplied marker whose provenance is preserved.
May support claims limited to what the source directly establishes.

Derived Evidence
A deterministic result computed from identifiable evidence through a declared method.
May support claims within the method, inputs, and validation scope.

Bounded Proxy Signal
An indirect measure used when the target property is not directly observable.
Must remain labeled as a proxy and cannot be converted into hidden-state or causal certainty.

Interpretive Layer
A classification, synthesis, or translation applied over evidence and measurements.
May organize review but cannot rewrite the underlying evidence or become ground truth.

Experimental
A provisional construct, method, or channel that has not reached the same validation posture as the governed core.
May be preserved for research and challenge but cannot silently enter certified claims.

The matrix can record each metric's source, computation basis, challengeability, certification scope, and claim boundary.

This means two values displayed beside one another may lawfully carry different authority. A directly supplied failure marker, a derived regime state, a bounded pressure proxy, and an interpretive incident class are not interchangeable evidence objects.

Preservation retains those distinctions. Export does not flatten them into one undifferentiated list of results.

Signal Authority and Metric Legitimacy Are Different Controls

Signal Authority and Metric Legitimacy solve related but different problems.

• Signal Authority governs which signal exists, where it sits in the dependency structure, how it resolves from the evidence run, and which status and claim boundary it carries.

• Metric Legitimacy classifies the evidentiary relationship represented by that metric or construct.

Signal Authority answers: Is this a registered and computable signal?

Metric Legitimacy answers: What kind of evidence is it, and what may be claimed from it?

Both controls are required. Registration without legitimacy can make a proxy appear direct. Legitimacy without a source-bound registry can leave a metric disconnected from the run that produced it.

The Common Evidence Identity

Every instrument reads from the same governing evidence identity.

The shared read context can include:

• the Current Evidence Run;

• Runtime Evidence Passport identity;

• canonical source hash;

• runtime event ledger;

• canonical runtime spine;

• runtime playback frames;

• governed signal records;

• registered markers and temporal authority;

• Operational World context; and

• declared interpretive or Human Read material where relevant to downstream presentation.

The Operational World may change terminology, emphasis, examples, and investigative routing. It cannot change telemetry values, signal legitimacy, temporal markers, or instrument authority.

Because every instrument resolves to the same run, two instruments may emphasize different properties without constructing competing source histories. Their findings remain attached to the same run identifier, Passport identifier, canonical identity, frames, and evidence lineage.

One runtime. One evidence identity. Multiple bounded projections.

The Instrument Contract

Each Aperture instrument operates under a versioned contract.

The contract declares:

• instrument identity and public label;

• authority level;

• Level 0 inputs it may read;

• Level 1 dynamics it may read;

• Level 2 projections it may consume or issue;

• evidence objects available to its read context;

• required source basis;

• projection and summary functions;

• permitted output type;

• evidence-status rules;

• allowed downstream consumers;

• claim boundary;

• forbidden claims; and

• explicit mutation prohibitions.

The active instruments are registered as Level 2 projection instruments. Their permitted output is an Instrument Finding. They do not write new telemetry back into the underlying signal registries.

The Nine Instrument Contracts

∿ Seismo
Runtime spine, regime sequence, transition pressure, collapse stress, runtime coherence, worldline and boundary projections
Trajectory formation, disturbance, regime transition, boundary posture, failure development, and recovery evidence

τ Chronos
Temporal coupling, recurrence, transition pressure, continuity, runtime motion, canonical frames
Temporal deformation, compression, recurrence, shear, sequencing, and formation timing

Δ Drift
Lexical drift, divergence, semantic curvature, constraint re-entry, continuity, motion, role and regime evidence
Trajectory deviation, curvature, displacement, recurrence, and movement relative to registered anchors

Ξ Pressure
Entropy, divergence, contraction pressure, transition pressure, collapse stress, Basin and regime projections
Runtime strain, accumulated pressure, boundary pressure, contraction, and recovery support

⇄ Bridge
Role phase lag, handoff shear, constraint re-entry, runtime continuity, recovery, operational context
Role relationships, handoff continuity, coordination strain, tool interaction, and operational bridge findings

⟳ Noesis
Repetition, echo persistence, runtime coherence, semantic curvature, constraint re-entry, recovery and continuity signals
Source-visible recurrence, coherence, re-entry, continuity, and formation posture

Ω Scope
Runtime frames, worldline stability, motion, regime state, Basin proximity, topology-related evidence
Coverage, topology, runtime field structure, qualification posture, and available evidentiary scope

Ψ Dynamics
Core telemetry, semantic curvature, temporal coupling, transition pressure, coherence, stability, continuity, and motionMultivariate runtime dynamics and relationships among registered dimensions

Φ Interferometer
Divergence, repetition, coherence, echo persistence, contraction pressure, stability, continuity, and Basin evidence
Alignment, divergence, interference, and coherence across registered streams or trajectories

These domains define what each instrument may project. They do not authorize the instrument to claim root cause, intent, truth, hidden state, cognition, consciousness, agency, provider internals, safety certification, or automated decision authority.

The Instrument Finding

An Instrument Finding is a bounded projection object tied to the governing run.

Its schema can require:

• finding identity;

• instrument identity;

• run identity;

• Passport identity;

• canonical hash;

• authority level;

• evidence status;

• reported value;

• bounded interpretation;

• linked signals;

• evidence basis;

• source spans;

• Operational World context;

• claim boundary;

• forbidden claims;

• projection confidence; and

• the evidence object from which the finding was created.

This structure prevents a finding from becoming a free-floating analytical statement. It remains attached to the evidence run, the signals it consumed, and the boundaries under which it was issued.

Projection confidence describes support within the instrument's declared evidence basis. It must not be represented as the probability that the interpretation is true unless a separate calibration process has established that meaning.

Evidence Basis and Source Spans

Every issued finding must expose enough of its basis to be challenged.

The evidence basis identifies the objects and signal relationships used by the instrument. Source spans identify the relevant turns, frames, events, or intervals within the run.

A finding should therefore answer:

• Which run produced it?

• Which instrument contract governed it?

• Which signal records were linked?

• Which frames or source spans were examined?

• Which dependencies were present or missing?

• Was the finding computed, projected, partial, or proxy-only?

• What confidence describes basis coverage?

• What is the claim boundary?

If a source span cannot be resolved, the finding must disclose that limitation. A broad summary cannot substitute for absent provenance.

Projection Eligibility

An instrument becomes eligible to project only when the applicable authority conditions are satisfied.

Full projection eligibility requires a governing evidence run and a resolved Signal Authority context. Individual findings then evaluate the subset of signal links and evidence objects required by their own contracts.

The possible outcomes include:

• projected — the contract's evidence basis is sufficiently present;

• partial — some linked signals or evidence objects are present, but the basis is incomplete;

• proxy only — the result depends on bounded proxy fields;

• not computable — the governing run or required dependencies are absent; or

• unavailable — the contract, registry, or module cannot be resolved.

An instrument may remain visible while its finding remains unavailable. Visibility is an interface state. Projection eligibility is an authority state.

The Non-Mutation Law

Instrument contracts enforce a strict non-mutation law.

An Aperture instrument may not:

• create telemetry;

• create or replace the governing evidence run;

• alter canonical frames;

• rewrite the runtime event ledger;

• create temporal markers;

• promote a candidate boundary to confirmed Basin Exit;

• supply an absent observable-failure marker;

• compute formal lead-time without admissible anchors;

• mutate Signal Authority;

• change metric legitimacy;

• alter Passport identity;

• extend its own claim boundary;

• change Evidence Commons content authority; or

• grant export or automated decision authority.

Its permitted operation is narrower:

Read governed evidence → apply the registered projection contract → issue a bounded Instrument Finding.

This law ensures that an instrument remains a measurement interface rather than becoming an ungoverned second compute engine.

Downstream Consumption

Instrument Findings may be consumed by:

• the Structure and Evidence State surfaces;

• the Instrumentation and Measure surfaces;

• Guided Investigation;

• Runtime Reconstruction chapters;

• the Runtime Evidence Brief;

• Human Read and REIM translation layers;

• technical appendices; and

• preservation manifests and Evidence Commons packages.

These consumers may organize, translate, compare, or preserve the finding. They may not increase its authority or detach it from its evidence basis.

Human-readable synthesis is therefore downstream of measurement governance. A clear explanation does not become stronger evidence merely because it is easier to understand.

Engineering Failure States

No governing evidence run
Block instrument projection and report the finding as not computable.

Signal registry unavailable
Withhold projection authority; do not infer signal definitions from the interface.

Required signal absentReport partial or not computable according to the contract.

Signal exists only as a proxyPreserve proxy-only status through the finding and every downstream surface.

Source coverage is insufficientDisclose the coverage limitation and withhold unsupported conclusions.

Run or Passport identity mismatch
Reject attachment of the finding to the active evidence object.

Contract version cannot be resolved
Treat the instrument as unavailable rather than executing an unregistered projection.

Interpretation exceeds the claim boundary
Preserve the evidence but reject or qualify the unsupported statement.

These states prevent presentation pressure from overriding evidence authority.

Relationship to the Instrumentation Section

The existing Instrumentation section explains the scientific and operational role of runtime instruments:

• why instrumentation matters;

• how the shared stability substrate supports multiple views;

• how the Instrumentation Stack is organized;

• how the Instrument Atlas presents the nine domains;

• how reconstruction becomes the shared object of observation; and

• how formation, interpretation, evaluation, and synthesis participate in investigation.

This Engineering page establishes the machinery beneath that system.

• The Instrumentation section explains what the instruments examine and how they support investigation.

• Behavioral Telemetry and Instrumentation Engineering explains how signals are formed, classified, versioned, gated, bound to evidence identity, and converted into contract-governed findings.

The first concerns scientific use. The second concerns computational authority.

Architectural Invariants

The behavioral telemetry and instrumentation architecture is governed by the following invariants:

1. Observable evidence precedes measurement. No signal exists without a declared source or derivation path.

2. Signal level and evidence legitimacy remain distinct. Computational position does not automatically determine evidentiary class.

3. Every signal is version-bound. Names, mappings, dependencies, and boundaries resolve through a declared registry.

4. Missingness remains explicit. Absent inputs are not converted into zero or inferred values.

5. Coverage does not imply sufficiency. A present value may still be inadequate for a particular projection.

6. All instruments share one evidence identity. Findings cannot attach to a different run, Passport, or canonical source.

7. Contracts govern projection. An instrument may read only declared inputs and issue only declared findings.

8. Confidence is bounded. Projection confidence describes evidentiary support, not truth probability, unless separately calibrated.

9. Instruments do not mutate evidence. Telemetry, markers, the governing run, Signal Authority, and export authority remain upstream.

10. Claim boundaries propagate downstream. Investigation, synthesis, and preservation cannot strengthen a finding beyond its contract.

The Governing Definition

SubstrateX Aperture™ Behavioral Telemetry and Instrumentation is the evidence-governed measurement architecture through which observable runtime records become versioned behavioral signals and those signals become bounded, contract-governed Instrument Findings.

Behavioral telemetry makes longitudinal computational behavior measurable. Signal Authority establishes which measurements exist and how they resolve. Metric Legitimacy establishes what kind of evidence each measurement represents. Instrument contracts define what each scientific instrument may read and what it may issue. The governing evidence identity keeps every finding connected to the same source-bound runtime.

The result is not a collection of independent analyzers producing competing conclusions.

It is one governed measurement system:

One reconstructed runtime → one Signal Authority → one evidence identity → multiple bounded instrument projections.

That architecture allows Aperture to expose complex behavioral telemetry without allowing a metric, instrument, visualization, or interpretation to claim more than the evidence supports.