The Scientific Observatory Behind Fieldglass®

The Runtime Instrument Atlas is the public reference for the nine scientific instruments implemented within the Fieldglass Runtime Evidence Observatory.

It is not a catalogue of dashboards or analytical widgets. It documents a coordinated measurement system in which functionally distinct instruments examine different properties of one source-bound runtime.

Every instrument reads from the same Current Evidence Run.

Every instrument is connected to the same versioned Signal Authority.

Every instrument remains governed by the same runtime identity, provenance, temporal coordinates, and claim boundary.

Together, they provide complementary scientific projections of one reconstructed runtime without producing competing accounts of what occurred.

One Runtime. One Evidence Authority. Multiple Bounded Instruments.

Purpose of the Atlas

The Atlas establishes:

• what each instrument observes;

• which evidence and signals it may read;

• which property of runtime behavior it measures;

• what kind of finding it may produce;

• what authority governs that finding;

• when the measurement is partial or unavailable;

• and what the instrument is explicitly prohibited from claiming.

This makes the instrumentation layer inspectable.

An operator should not have to trust an instrument merely because it displays a score, classification, or compelling visualization. The instrument must disclose where its measurements came from, how they relate to the evidence run, and where their authority ends.

The Atlas provides that reference.

One Runtime, Many Projections

Fieldglass reconstructs the runtime once.

A completed computation establishes the Current Evidence Run, or CER: the authoritative evidence object connecting the qualified source, canonical runtime spine, event ledger, temporal coordinates, registered signals, markers, disclosures, and claim boundaries.

The instruments do not independently reconstruct this evidence.

They receive bounded read access to the shared runtime and project defined measurements onto it.

Current Evidence Run
↓
Signal Authority and Shared Measurement Context
↓
Instrument Contract
↓
Instrument Finding
↓
Investigation and Preservation

No instrument owns a separate runtime.

No instrument may create source facts.

No instrument may rewrite canonical telemetry or authoritative markers.

No instrument may modify the evidence identity, Passport, or preservation authority.

Each instrument may write only its own bounded Instrument Finding.

This allows the observatory to increase measurement resolution without multiplying the reality of the run.

The Signal Architecture

Fieldglass instruments consume registered signals across three measurement levels.

Level 0 — Source-Derived Features

Foundational computational features derived from the qualified record, including measures such as entropy, lexical drift, divergence, repetition, length volatility, and reasoning density.

Level 1 — Derived Runtime Dynamics

Measurements constructed from relationships among source-derived features, including semantic curvature, contraction pressure, echo persistence, runtime coherence, temporal coupling, transition pressure, collapse stress, constraint re-entry, role-phase lag, and handoff shear.

Level 2 — Instrument Projections

Higher-order, evidence-bound projections such as worldline stability, basin proximity, Basin Exit risk, attractor lock, regime state, recovery signal, runtime stability, runtime continuity, and runtime motion.

Level 2 readings are projections—not new source observations. Their legitimacy depends on the availability and quality of the signals beneath them.

An unavailable signal must remain unavailable. A missing value cannot silently become zero. A proxy cannot be represented as a calibrated probability unless separate validation supports that interpretation.

The Scientific Contract

Every Fieldglass instrument is governed by the same contract structure.

Sensor
The source objects, runtime frames, signals, markers, and contextual records the instrument is permitted to read.

Measures
The observable property or relationship the instrument is designed to examine.

Produces
The bounded Instrument Finding written into the run’s instrument-finding layer.

Evidence Status
Whether the finding is computed, derived, projected, partial, proxy-only, not supplied, not computable, or unavailable.

Cannot Claim
The interpretations prohibited by the evidence boundary.

Instrument findings retain their:

• instrument and finding identities;

• run and Passport references;

• signal links;

• source basis and source spans;

• operational context;

• projection confidence;

• claim boundary;

• and forbidden claims.

This contract separates measurement from presentation. An instrument may change how a finding is visualized without changing the evidence or authority behind it.

The Instruments

∿ Seismo

Runtime Trajectory and Boundary Motion

Seismo examines how the runtime trajectory moves across stability conditions.

Observes: The canonical runtime spine, regime timeline, runtime frames, transition pressure, collapse stress, coherence, and temporal coupling.

Measures: Worldline stability, disturbance formation, runtime motion, regime transition, boundary evidence, Basin Exit support, and recovery posture.

Produces: A staged reconstruction of where weakening, candidate boundary formation, confirmed transition, observable failure, and recovery evidence appear along the runtime.

Cannot claim: Hidden model state, universal causation, root cause, prediction without prospective validation, or formal lead time without the required temporal anchors.

Seismo answers:

Where did the runtime begin to move differently, and when did that change become evidentially significant?

τ Chronos

Temporal Organization and Symbolic Time

Chronos examines how consequential change is organized through the runtime.

Observes: Event order, recurrence, temporal coupling, echo persistence, transition pressure, constraint re-entry, and the canonical runtime timeline.

Measures: Symbolic-time progression, temporal compression, dilation, shear, recurrence, continuity, anomaly windows, and recovery timing.

Produces: A temporal projection showing how densely or unevenly runtime structure developed and whether an admissible inspection window exists before an observable event.

Cannot claim: Hidden time, subjective time, physical time beyond the supplied record, prediction from retrospective information, or lead time when the failure anchor is absent.

Chronos answers:

How did the runtime develop through time—not merely how much clock time elapsed?

Δ Drift

Trajectory Deviation and Attractor Movement

Drift examines cumulative displacement across the runtime.

Observes: Lexical change, divergence, recurrence, semantic curvature, constraint re-entry, role topology, runtime coherence, and regime development.

Measures: Departure from established objectives, constraints, behavioral references, recurring formations, or stable trajectory conditions.

Produces: Drift trajectories, deviation postures, acceleration indicators, attractor-related movement, and evidence of re-alignment or continued displacement.

Cannot claim: Internal intent, permanent identity change, provider mechanism, or that every behavioral difference represents instability.

Drift answers:

What changed direction, how far did it move, and did the trajectory return?

Ξ Pressure

Runtime Strain and Boundary Pressure

Pressure examines how competing demands and destabilizing conditions accumulate.

Observes: Entropy, divergence, reasoning density, contraction pressure, transition pressure, collapse stress, semantic curvature, and boundary-related measurements.

Measures: Runtime strain, accumulated pressure, basin proximity, transition load, collapse-related pressure, and recovery reserve.

Produces: A bounded pressure projection identifying where the runtime appears increasingly constrained, deformed, or exposed to a regime transition.

Cannot claim: Literal physical energy, calibrated failure probability, inevitable collapse, root cause, or the internal forces operating within the model.

Pressure answers:

Where did runtime strain accumulate, and how did it relate to the observed boundary conditions?

⇄ Bridge

Role, Handoff, and Interaction Dynamics

Bridge examines how participants, tools, and operational roles shape the runtime.

Observes: Role topology, authority changes, handoff events, role-phase lag, handoff shear, constraint re-entry, tool-loop activity, and operational context.

Measures: Role continuity, coordination gaps, authority conflict, cross-role drift, handoff coherence, interaction pressure, and recovery across operational exchanges.

Produces: An operational projection connecting runtime measurements to the roles, handoffs, tools, and workflow conditions recorded in the source.

Cannot claim: Human intention, fault, blame, competence, organizational truth, or causal responsibility.

Bridge answers:

How did the interaction structure influence the recorded trajectory?

The selected Operational World may change Bridge’s language and investigative emphasis. It does not change the canonical telemetry.

⟳ Noesis

Recursive Formation and Re-Anchor Behavior

Noesis examines observable patterns of symbolic recurrence and formation.

Observes: Repetition, lexical movement, reasoning density, echo persistence, semantic curvature, constraint re-entry, coherence, and source-visible anchors.

Measures: Formation progression, recurrence, reinforcement, modulation, reflection, re-anchoring, continuity, and recovery-related return.

Its organizing sequence is:

Seed → Echo → Modulate → Reflect → Re-anchor → Loop

Produces: A formation projection describing how recurring structure appeared, persisted, changed, weakened, or returned across the runtime.

Cannot claim: Cognition, consciousness, subjective experience, stored memory, agency, or the presence of a literal mind.

Noesis answers:

What observable structure formed through recursion, and did that structure persist or re-anchor?

Ω Scope

Runtime Topology and Containment Geometry

Scope provides a topological projection of the runtime.

Observes: Worldline stability, runtime motion, regime state, basin proximity, transition pressure, coherence, collapse stress, and frame-level trajectory evidence.

Measures: Containment structure, basin relationships, trajectory deformation, boundary geometry, collapse corridors, and possible recovery arcs.

Produces: A navigable representation of how the runtime is organized as a bounded trajectory rather than a collection of disconnected measurements.

Cannot claim: A literal hidden manifold, physical geometry inside the model, objective truth, or direct observation of internal model space.

Scope answers:

What shape did the evidence-supported runtime trajectory take?

Its geometry is a computational representation of observable relationships—not a claim that the model contains a directly observed physical landscape.

Ψ Dynamics

Cross-Signal Runtime Dynamics

Dynamics examines how multiple runtime dimensions change together.

Observes: The broader registered signal architecture, including drift, divergence, curvature, temporal coupling, pressure, coherence, contraction, regime state, and runtime motion.

Measures: Cross-signal relationships, recursive dynamics, stability posture, continuity, worldline movement, and changing relationships among runtime dimensions.

Produces: A combined projection of the runtime’s dynamical state. Where optional embedding or trace evidence is supplied, Dynamics may provide additional comparative views without treating those traces as hidden-state authority.

Cannot claim: Unobserved internal dynamics, provider-controlled model state, causal mechanism, consciousness, agency, or universal diagnostic validity.

Dynamics answers:

How did the measured dimensions of the runtime move in relation to one another?

Dynamics remains a research-oriented instrument whose findings depend on the evidence channels available for a particular run.

Φ Interferometer

Coherence, Reinforcement, and Interference

Interferometer examines whether recurring behavioral structures reinforce or disrupt one another.

Observes: Divergence, repetition, reasoning density, runtime coherence, echo persistence, contraction pressure, basin proximity, continuity, and worldline stability.

Measures: Coherence, interference, attractor-related locking, competing formation patterns, and evidence consistent with the strengthening or weakening of a stable runtime configuration.

Produces: An experimental coherence projection showing how recurring structures align, compete, or lose organization across the trajectory.

Cannot claim: Literal field ignition, a directly observed internal substrate, consciousness, cognition, hidden state, or the existence of a physical interference field inside the model.

Interferometer answers:

Did recurring structures reinforce a coherent runtime configuration, or interfere with its persistence?

Interferometer remains an experimental, availability-gated instrument. Its projections must be interpreted as output-derived evidence rather than direct access to an internal field.

Complementary, Not Competing

No single instrument is expected to explain the complete runtime.

Seismo may locate a transition.

Chronos may establish its temporal organization.

Drift may show displacement.

Pressure may reveal accumulated strain.

Bridge may identify the relevant role or handoff structure.

Noesis may show whether a recurring formation re-anchored.

Scope may represent the surrounding topology.

Dynamics may examine cross-signal movement.

Interferometer may test whether patterns reinforced or disrupted one another.

These findings may support, qualify, or challenge one another, but they remain projections of the same evidence object.

Instrument agreement does not automatically establish truth. Instrument disagreement is not something to conceal. It may indicate different sensitivities, incomplete evidence, an unresolved boundary, or a construct requiring further calibration.

The Atlas makes those relationships visible.

The Measurement Boundary

Fieldglass instruments operate within an observable-runtime-behavior boundary.

They do not independently determine:

• root cause;

• intent;

• objective truth;

• hidden model state;

• provider internals;

• consciousness;

• cognition;

• agency;

• blame;

• automated decision authority;

• automated control authority;

• or safety certification.

They measure what the supplied record and declared methods permit them to measure.

Every finding remains subordinate to its source basis, evidence status, measurement contract, and claim boundary.

Why an Atlas?

An atlas does more than list instruments. It maps the measurement structure of the observatory.

It allows researchers, engineers, operators, and investigators to understand:

• which instrument addresses a particular question;

• how its measurements relate to the wider runtime;

• where two instruments overlap;

• which signals support a finding;

• where evidence is absent;

• and where interpretation must stop.

The existence of an Atlas does not, by itself, validate an instrument. It makes validation possible by exposing the instrument’s purpose, inputs, outputs, methods, and limits.

That is what distinguishes an evidence-governed instrument from an opaque analytical feature.

Fieldglass is designed as a scientific observatory because the runtime is not reduced to one score, one explanation, or one dashboard. It is reconstructed as a shared evidentiary object and examined through multiple bounded forms of measurement.

The runtime is the specimen. The instruments are the lenses. The evidence remains the authority.