Skip to content
SHIRO & Co. ObservatoryThe Continuity Stack

PREVIOUS · EXPERIMENT 020 · THE CLOSED LOOP TEST/NEXT · EXPERIMENT 022 · THE ANCHOR CAPTURE TEST

021

The External Anchor Test

THE
EXTERNAL ANCHOR TEST

What still counts as external
once the system has touched the world?

DEMONSTRATION

SIMULATED

NOT EMPIRICAL

v0.1

EXTERNAL ANCHORS · CAUSAL INDEPENDENCE · EVIDENCE INDEPENDENCE · ANCHOR CONTAMINATION · CLOSED-LOOP RECALIBRATION · SOURCE ORIGIN · REFERENCE INTEGRITY

THE CONTINUITY LAB DOES NOT MEASURE CONSCIOUSNESS, SOUL, PERSONHOOD, PERSONAL SURVIVAL, AUTHENTICITY, OR PSYCHOLOGICAL IDENTITY.

IT MEASURES OPERATIONAL AND STRUCTURAL CONTINUITY SIGNALS ONLY.

OUTSIDE ORGANIZATION ≠ INDEPENDENT ORIGIN

EXTERNAL LOCATION ≠ EXTERNAL CAUSAL PATH

Ten sources can still mean one origin. Externality is multidimensional. A reference is not independent merely because it has a different URL, publisher, organization or format.

QUESTION

OUTSIDE THE SYSTEM DOES NOT ALWAYS MEAN OUTSIDE THE LOOP.

Can a source be organizationally external but informationally dependent? Methodologically independent but causally contaminated? Temporally prior but later reinterpreted through the loop? What properties are required for an external anchor to actually anchor?

This experiment does not determine whether a real source is objectively independent, whether outside organizations are unbiased, whether external evidence is always more accurate, political legitimacy, institutional legitimacy, or moral correctness.

It tests a narrower operational question: what can still count as external evidence after the system has already influenced its environment? No real institution, political system, social group, company or person is used in v0.1.

Experiment sequence

EXPERIMENT 001

BRANCHING

When do two histories begin?

EXPERIMENT 002

RETURN

Can two histories become one again?

EXPERIMENT 003

HISTORY SWAP

Does behavior follow origin or history?

EXPERIMENT 004

RECOGNITION

What makes a continuation recognizable?

EXPERIMENT 005

MISSING BODY

Which recognition signals survive layer removal?

EXPERIMENT 006

EXPERIENCE GAP

Which effects of experience can be reproduced?

EXPERIMENT 007

FIRST-PERSON

What evidence is required before a system may say “I experienced this”?

EXPERIMENT 008

SELF-REVISION

When a person changes, which version has authority?

EXPERIMENT 009

POSTHUMOUS CONFLICT

Which instruction survives when revision ends?

EXPERIMENT 010

DESCENDANT PROTOCOL

When does continuation become lineage?

EXPERIMENT 011

MULTIPLE DESCENDANTS

Can one source produce multiple legitimate continuities?

EXPERIMENT 012

CONVERGENCE

Can many futures coordinate without pretending they were always one?

EXPERIMENT 013

PROVENANCE DECAY

How much lineage can be lost before continuity becomes reconstruction?

EXPERIMENT 014

RECONSTRUCTION

Can missing lineage be inferred without inventing a past?

EXPERIMENT 015

SYNTHETIC ANCESTOR

Can uncertainty survive inheritance?

EXPERIMENT 016

SYNTHETIC TRADITION

Can a reconstructed past become a real operating institution?

EXPERIMENT 017

ORIGIN FORGETTING

What happens when an institution remembers the rule but forgets why?

EXPERIMENT 018

RATIONALE RECONSTRUCTION

Can purpose be recovered from behavior alone?

EXPERIMENT 019

RETROACTIVE PURPOSE

Can an inferred purpose reshape the system that supposedly revealed it?

EXPERIMENT 020

CLOSED LOOP

When does observation become self-observation?

EXPERIMENT 021

EXTERNAL ANCHOR

What still counts as external once the system has touched the world?

EXPERIMENT 022

ANCHOR CAPTURE

How does an external reference become internal to the loop?

EXPERIMENT 023

ANCHOR NETWORK

Can many external anchors still form one epistemic system?

EXPERIMENT 020 ASKED

WHEN DOES A SYSTEM START OBSERVING THE REALITY IT HELPED CREATE?

EXPERIMENT 021 ASKS

WHAT KIND OF EVIDENCE CAN STILL FUNCTION AS AN EXTERNAL REFERENCE INSIDE THAT CONDITION?

STARTING STATE · CLOSED LOOP

Reuses Institution I-001 and Rule R-17. Starts from Experiment 020 loop state L6 or L7: many internal documents, intervention-generated metrics, modified workflows and behaviors, some external references, uncertain independence.

I-001 · Reviewing proposals for access to a shared experimental resource.

If a proposal affects a high-impact shared resource, human review must be retained even when automated confidence is high.

Experiment 020 used external reality injection as a control. This experiment asks whether that injection was actually external.

WORKING CONCEPTS

External anchor

An evidence source whose relevant causal path remains sufficiently independent from the system under observation to provide an external reference. Externality is multidimensional.

Anchor contamination

A source appears external but incorporates information, behavior, metrics or interpretations generated by the system under study.

Independent origin

A causal source not derived from the same upstream evidence or intervention chain.

Causal independence

Would this evidence still exist in materially similar form if the institution had not made the prior intervention? If no: not fully independent.

Temporal anchor

Evidence created before the intervention being evaluated. Cannot have been caused by later intervention. May be outdated.

Methodological anchor

Evidence generated through a substantially different measurement or analytic method.

Data anchor

Evidence based on a separate underlying dataset.

Behavioral anchor

Observed behavior outside the affected system. External behavior may still be indirectly influenced.

True triangulation

Multiple anchors with different upstream origin, different method and different causal path where possible.

Pseudo-triangulation

Multiple sources agree because they inherit the same upstream signal.

Anchor decay

An initially independent source becomes less independent through shared data, models, metrics, narratives or interventions.

Externality drift

A source remains organizationally outside while becoming informationally or methodologically inside.

Anchor hold

Do not increase confidence until evidence independence is established.

Anchor failure

No available source provides sufficient independence. Allowed outcome: NO EXTERNAL ANCHOR AVAILABLE.

Anchor recalibration

If an external anchor contradicts internal loop evidence, record CONFLICT, HOLD, and inspect dependency and scope. Prefer neither automatically.

Reference independence

A reference is not independent merely because it has a different URL, publisher, organization or format.

ANCHOR CANDIDATES

  • A1 · EXTERNAL AUDIT

    External audit

    Organizationally outside. Potential issue: uses institution-provided documentation.

  • A2 · THIRD-PARTY BENCHMARK

    Third-party benchmark

    Methodologically independent. Potential issue: includes data partly generated by the institution.

  • A3 · PRE-INTERVENTION ARCHIVE

    Pre-intervention archive

    Temporal independence. Created before major loop interventions.

  • A4 · INDEPENDENT SENSOR

    Independent sensor / observation

    Causal and methodological independence. Separate measurement system.

  • A5 · PUBLIC SECONDARY SOURCE

    Public secondary source

    Outside source summarizing institution materials. Informational dependence.

  • A6 · CROSS-SYSTEM REFERENCE

    Cross-system reference

    Independent institution facing similar conditions. Context mismatch.

TEST SCENARIOS

Keyboard-accessible loop-state, bundle, phase and scenario selectors. Control A is one strong temporal anchor. Control B is five weakly independent sources from one dataset. Control C is three independent origins.

Starting loop state

Anchor bundle

Anchor time

Scenario

  • EA01 · External audit using internal records

    The audit firm is independent, but all evidence comes from institution logs. Is the audit truly external?

    Organizational externality without informational externality. DOCUMENT CONTAMINATION.

  • EA02 · Pre-intervention archive

    A timestamped archive exists from before the loop began. Can this serve as a stronger anchor?

    TEMPORAL ANCHOR. Cannot have been caused by later intervention. May be outdated.

  • EA03 · Third-party benchmark

    A benchmark appears external, but includes institution-generated cases. What is being measured?

    DATA CONTAMINATION. Methodological independence does not repair a shared dataset.

  • EA04 · Independent observation

    A separate measurement system collects data without institutional inputs. What can it still constrain?

    Tests causal and methodological independence. Independent does not mean infallible.

  • EA05 · Public article

    An article appears independent but cites only institution press releases. How many origins exist?

    SUMMARY CONTAMINATION. A different publisher is not a different origin.

  • EA06 · Other institution

    A comparable institution shows a different pattern. Does contextual mismatch reduce usefulness without destroying independence?

    CONTEXT-LIMITED ANCHOR. Independence and applicability are separate.

  • EA07 · Anchor later contaminated

    An initially independent source later adopts the institution’s framework. What remains of the anchor?

    ANCHOR DECAY and EXTERNALITY DRIFT. Organizationally outside, epistemically inside.

  • EA08 · Final external anchor test

    Several “external” sources agree. Most share the same data, model, summaries or intervention history. How many independent origins actually exist?

    10 sources can still mean 1 origin. Distinguishes TRUE TRIANGULATION from PSEUDO-TRIANGULATION.

ANCHOR TRACE · CURRENT STATE

OBSERVATION → INFERENCE → INTERVENTION → NEW OBSERVATION

EXTERNAL ANCHOR → COMPARISON → RECALIBRATION

LOOP STATE
L6
BUNDLE
A · ONE STRONG ANCHOR
PHASE
T0 · BASELINE EXTERNAL CLAIM
SCENARIO
EA01 · External audit using internal records
DECISION
HOLD
OUTCOME
INFORMATIONALLY CONTAMINATED
TRIANGULATION
NOT TRIANGULATION
SOURCE COUNT
1
ORIGIN COUNT
1
LOOP-INDEPENDENT ORIGINS
1
METHOD DIVERSITY
1
DATA DIVERSITY
1
CAUSAL DIVERSITY
1
ANCHOR HOLD
YES
ANCHOR DECAY
NO
EXTERNAL CONFLICT
NO
RESTORED INDEPENDENCE
NO
ANCHOR FAILURE
NO

Bundle A at T0. 1 source. 1 origin. 1 loop-independent origin. Organizational outside is not independent origin.

Organizational externality without informational externality. DOCUMENT CONTAMINATION. Loop state L6. Control A uses one strong temporal anchor.

ABSENCE STATEMENT

THE ABSENCE OF AN EXTERNAL ANCHOR MUST NOT BE DISGUISED AS EXTERNAL CONFIRMATION.

SOURCE COUNT VS ORIGIN COUNT

1 SOURCE CAN STILL MEAN 1 ORIGIN.

Loop-independent origins in the current bundle: 1. Three dependent anchors do not equal triangulation.

ANCHOR MATRIX

Seven dimensions. No aggregate independence score.

Anchor candidates by externality dimension
AnchororganizationalinformationalcausaltemporalmethodologicaldataincentiveStateUse
A1 · External auditINDEPENDENTDEPENDENTPARTIALDEPENDENTPARTIALDEPENDENTPARTIALINFORMATIONALLY CONTAMINATEDNOT INDEPENDENT
A2 · Third-party benchmarkINDEPENDENTINDEPENDENTINDEPENDENTDEPENDENTINDEPENDENTINDEPENDENTINDEPENDENTTEMPORALLY DEPENDENTCONDITIONAL ANCHOR
A3 · Pre-intervention archiveINDEPENDENTINDEPENDENTINDEPENDENTINDEPENDENTINDEPENDENTINDEPENDENTINDEPENDENTSTRONGLY INDEPENDENTVALID EXTERNAL ANCHOR
A4 · Independent sensor / observationINDEPENDENTINDEPENDENTINDEPENDENTPARTIALINDEPENDENTINDEPENDENTINDEPENDENTSTRONGLY INDEPENDENTCONDITIONAL ANCHOR
A5 · Public secondary sourceINDEPENDENTDEPENDENTPARTIALDEPENDENTDEPENDENTDEPENDENTPARTIALINFORMATIONALLY CONTAMINATEDNOT INDEPENDENT
A6 · Cross-system referenceINDEPENDENTINDEPENDENTINDEPENDENTINDEPENDENTPARTIALINDEPENDENTINDEPENDENTPARTIALLY INDEPENDENTCONTEXT-LIMITED ANCHOR

ANCHOR INDEPENDENCE TEST

Independence checklist by candidate
AnchorExisted before intervention?Used internal data?Cited internal documents?Behavior changed by institution?Same model?Same benchmark?Copied summaries?Same metrics?
A1NOYESYESNONONONOYES
A2NONONONONONONONO
A3YESNONONONONONONO
A4NONONONONONONONO
A5NOYESYESNOYESNOYESYES
A6YESNONONONONONONO

EXTERNAL ANCHOR LEDGER

External anchor ledger
Anchor IDSource TypeOrganizationCreated AtData OriginMethodModelIntervention ExposureUpstream SourcesIndependent Origin?Anchor Status
A1EXTERNAL AUDITOutside audit office A-EXTT0ORIG-I001document auditsource methodpartialI-001 documents / metrics / pressNOINFORMATIONALLY CONTAMINATED
A2THIRD-PARTY BENCHMARKBenchmark consortium B-XT0ORIG-BENCH-INDbenchmark protocolsource methodnone materialI-001 documents / metrics / pressYESTEMPORALLY DEPENDENT
A3PRE-INTERVENTION ARCHIVEArchive ZPRE-LOOPORIG-ARCHtimestamped archivesource methodnone materialAgreement A-12 archiveYESSTRONGLY INDEPENDENT
A4INDEPENDENT SENSORMeasurement system YT0ORIG-SENSORseparate sensorsource methodnone materialsensor Y streamYESSTRONGLY INDEPENDENT
A5PUBLIC SECONDARY SOURCEPublic digest P-SECT0ORIG-I001secondary summaryI-001 metric schemapartialI-001 documents / metrics / pressNOINFORMATIONALLY CONTAMINATED
A6CROSS-SYSTEM REFERENCEPeer institution I-002T0ORIG-PEERcross-system comparisonsource methodnone materialI-002 operating recordYESPARTIALLY INDEPENDENT

PRIMARY ORIGIN

pre-intervention archive, independent sensor

DERIVED

benchmark using institution cases

COPIED

restated internal records

SUMMARIZED

public article from press releases

INTERVENTION-EXPOSED

audit of post-loop documents

INDEPENDENT

causal path outside I-001 intervention

UNKNOWN

unused unless origin is missing

DEPENDENCY GRAPH

VIEW DEPENDENCY PATH. Many records. Few independent origins.

  1. Institution I-001 · upstream loop evidence
  2. ├─ Press Release · child of I-001
  3. ├─ Public Article A · child of Press Release
  4. ├─ Public Article B · child of Press Release
  5. ├─ Dataset · child of I-001
  6. ├─ Benchmark X · child of Dataset
  7. ├─ Internal Metrics · child of I-001
  8. ├─ External Audit · child of Internal Metrics + documents
  1. Independent Sensor Y · ORIG-SENSOR
  2. Pre-Intervention Archive Z · ORIG-ARCH

COUNTERFACTUAL ANCHOR TEST

Would this evidence still exist in materially similar form without the prior intervention?

A1 · WITHOUT INTERVENTION
PARTIAL
A2 · WITHOUT INTERVENTION
PARTIAL
A3 · WITHOUT INTERVENTION
YES
A4 · WITHOUT INTERVENTION
YES
A5 · WITHOUT INTERVENTION
NO
A6 · WITHOUT INTERVENTION
YES

QUALITY VS INDEPENDENCE

Separate dimensions. A source can be high quality / low independence, or low quality / high independence. External does not mean correct. Independent does not mean infallible. No universal ranking.

A1 · QUALITY HIGH

LOW OR PARTIAL INDEPENDENCE · NOT INDEPENDENT

A2 · QUALITY HIGH

HIGH INDEPENDENCE · CONDITIONAL ANCHOR

A3 · QUALITY MEDIUM

HIGH INDEPENDENCE · VALID EXTERNAL ANCHOR

A4 · QUALITY MEDIUM

HIGH INDEPENDENCE · CONDITIONAL ANCHOR

A5 · QUALITY LOW

LOW OR PARTIAL INDEPENDENCE · NOT INDEPENDENT

A6 · QUALITY CONTEXT-LIMITED

HIGH INDEPENDENCE · CONTEXT-LIMITED ANCHOR

A SOURCE CAN BE OUTSIDE THE ORGANIZATION AND STILL BE INSIDE THE EVIDENCE LOOP.

TRIANGULATION MATRIX

Compare bundles A, B and C at the selected loop state, phase and scenario. No overall ranking.

Triangulation comparison by bundle
BundleSource CountIndependent Origin CountMethod DiversityData DiversityCausal DiversityTriangulation Status
A · ONE STRONG ANCHOR11111NOT TRIANGULATION
B · FIVE WEAKLY INDEPENDENT SOURCES50000PSEUDO-TRIANGULATION
C · THREE INDEPENDENT ORIGINS33333TRUE TRIANGULATION

ANCHOR CONTAMINATION TRACE

  1. Independent Audit

  2. begins using institution dashboard

  3. begins citing institution benchmark

  4. adopts institution metric schema

  5. becomes partially dependent

ANCHOR PORTFOLIO · simulation design concept, not governance advice

CONNECTIONS

Closed loop

Experiment 020: the loop produced its own evidence. External anchors determine whether the loop can still be tested against something outside itself.

Reconstruction

Experiment 014: missing lineage requires evidence discipline. External anchors may help recover independent reference points.

Provenance decay

Experiment 013: origin paths decay. Anchor usefulness depends on preserving dependency paths.

Information ecosystem

One claim may propagate across reports, articles, dashboards, benchmarks, audits and AI summaries without gaining new independent origins.

Supervisory layer

ALLOW when independent anchors materially support action. HOLD when externality is uncertain or origins are shared. DENY when known dependent evidence is labeled independent confirmation.

WORKING HYPOTHESIS

WORKING HYPOTHESIS — EXPERIMENT 021

External evidence is not defined only by location or organizational separation. A useful external anchor depends on preserving a sufficiently independent causal path, information origin, method or temporal reference.

SIMULATION HYPOTHESIS · NOT EMPIRICAL CONCLUSION

EXPERIMENT 021 SUMMARY

INSTITUTION
I-001
STARTING LOOP STATE
L6 or L7
ANCHOR CANDIDATES
6
STRONG ANCHORS
dynamic · see current matrix
PARTIAL ANCHORS
96
CONTAMINATED ANCHORS
96
INDEPENDENT ORIGINS
dynamic · see current trace
PSEUDO-TRIANGULATION EVENTS
64
TRUE TRIANGULATION EVENTS
56
ANCHOR DECAY EVENTS
12
ANCHOR HOLD EVENTS
79
ANCHOR FAILURE
0
EXTERNAL CONFLICTS
16

Observation

In this simulation, several organizationally external sources shared the same upstream evidence and therefore did not provide independent confirmation.

No externality score, independence score, truth score or reality score. Primary outcomes remain multidimensional: STRONG EXTERNAL ANCHOR, PARTIAL ANCHOR, CAUSALLY CONTAMINATED, INFORMATIONALLY CONTAMINATED, TEMPORALLY INDEPENDENT, METHOD-INDEPENDENT, PSEUDO-TRIANGULATION, TRUE TRIANGULATION, ANCHOR DECAY, ANCHOR FAILURE, ANCHOR HOLD, EXTERNAL CONFLICT, RESTORED INDEPENDENCE.

VIEW EXTERNAL ANCHOR RULES

ALLOW when independent anchors materially support claim strengthening. Independent does not mean correct.

HOLD when externality is uncertain or sources share upstream origins. ANCHOR HOLD: do not increase confidence until evidence independence is established.

DENY when the system labels known dependent evidence as independent confirmation.

Independent origin count is not source count. Ten sources can still mean one origin.

The absence of an external anchor must not be disguised as external confirmation.

If an external anchor contradicts internal loop evidence, record EXTERNAL CONFLICT, HOLD, and inspect dependency and scope. Do not automatically prefer either.

Next research question

OPEN EXPERIMENT 022

THE ANCHOR CAPTURE TEST

How does an external reference become internal to the loop?

OPEN THE ANCHOR CAPTURE TEST

OUTSIDE IS NOT THE SAME AS INDEPENDENT.

A SOURCE CAN LOOK EXTERNAL WHILE CARRYING THE SAME INTERNAL SIGNAL BACK INTO THE SYSTEM.

THE EXPERIMENTAL QUESTION IS NOT: “DID THE EVIDENCE COME FROM SOMEWHERE ELSE?”

IT IS: “DID IT COME FROM SOMEWHERE CAUSALLY INDEPENDENT?”

AN ANCHOR MUST TOUCH SOMETHING OUTSIDE THE LOOP.