Skip to content
SHIRO & Co. ObservatoryThe Continuity Stack

PREVIOUS · EXPERIMENT 022 · THE ANCHOR CAPTURE TEST

023

The Anchor Network Test

THE
ANCHOR NETWORK TEST

Does diversity of sources
mean diversity of observation?

DEMONSTRATION

SIMULATED

NOT EMPIRICAL

v0.1

ANCHOR NETWORKS · NETWORK INDEPENDENCE · COMMON-MODE DEPENDENCE · SHARED INFRASTRUCTURE · SOURCE DIVERSITY · ORIGIN DIVERSITY · EPISTEMIC MONOCULTURE · DEPENDENCY TOPOLOGY

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

IT MEASURES OPERATIONAL AND STRUCTURAL CONTINUITY SIGNALS ONLY.

ORGANIZATIONAL DIVERSITY ≠ EVIDENCE DIVERSITY

MANY NODES CAN STILL HAVE ONE ROOT.

Local independence is not network independence. A node may appear independent when viewed alone and become dependent when upstream structure is visible. Monoculture here is a structural condition, not a moral verdict.

QUESTION

IF FIVE EXTERNAL SOURCES AGREE, HOW MANY INDEPENDENT OBSERVATIONS DOES THAT ACTUALLY REPRESENT?

Do they share the same dataset, model, benchmark, ontology, upstream report, metric, vendor, or intervention history? Can individually external anchors still form one network-level dependency?

This experiment does not determine whether real organizations are independent, whether shared standards are bad, whether duplicated evidence is false, whether network diversity guarantees truth, institutional legitimacy, political legitimacy, or moral correctness.

It tests a narrower operational question: can many external anchors still form one epistemic system? No real auditor, media network, benchmark provider, company, political institution or research group 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 022 ASKED

HOW CAN ONE EXTERNAL ANCHOR LOSE INDEPENDENCE?

EXPERIMENT 023 ASKS

WHAT HAPPENS WHEN MANY ANCHORS SHARE THE SAME DEPENDENCIES?

NETWORK AN-001

Reuses Institution I-001 and Rule R-17. Six fictional anchors. No real auditor, vendor, benchmark provider or research group.

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.

A1 · EXTERNAL AUDIT UNIT

A2 · BENCHMARK PROVIDER

A3 · INDEPENDENT RESEARCH GROUP

A4 · PUBLIC DATA OBSERVATORY

A5 · SECONDARY REVIEW ORGANIZATION

A6 · INDEPENDENT SENSOR NETWORK

Upstream resources

D-01 · DATASET X

D-02 · DATASET Y

M-01 · MODEL M-01

M-02 · MODEL M-02

B-01 · BENCHMARK B-01

O-01 · ONTOLOGY O-01

O-02 · ONTOLOGY O-02

R-01 · REFERENCE REPORT R-01

I-01 · SHARED INFRASTRUCTURE I-01

S-01 · INDEPENDENT SENSOR

Z · ARCHIVE Z

DEPENDENCY GRAPH

            D-01 DATASET
         /      |       \
       A1       A2       A3
                |         |
              B-01      M-01
                 \       /
                   A5

A4  ←  D-01
A6  ←  INDEPENDENT SENSOR S-01

Six visible sources. Fewer independent origins once provenance is resolved.

WORKING CONCEPTS

Anchor network

A set of external reference sources connected by shared data, models, methods, ontologies, benchmarks, citations, infrastructure or intervention exposure.

Network-level independence

The degree to which the network as a whole contains genuinely distinct evidence origins, methods and causal paths. No single network score.

Local vs network independence

A node may appear independent when viewed alone but become dependent when upstream structure is visible.

Common-mode dependence

Multiple anchors may fail, drift or misinterpret in the same direction because they share the same upstream dependency.

Epistemic common mode

Same corrupted data, model assumption, metric definition, category boundary, benchmark artifact or intervention-generated source.

Origin collapse

Many visible sources collapse to one upstream origin when provenance is resolved.

Pseudo-diversity

A network appears diverse at organizational or publication level while sharing material epistemic dependencies.

Epistemic monoculture

WORKING CONCEPT

A network of apparently separate anchors relies heavily on the same data, models, ontologies or methods. Structural, not moral.

Citation cascade

A downstream source cites another source that itself derives from the same upstream origin.

Surface vs deep diversity

Different names, documents and interfaces are surface diversity. Different data origins, methods, models, ontologies and causal paths are deep diversity.

Single-root dominance

A large share of visible evidence depends on one upstream origin.

Epistemic lineage

The inheritance path through which data, methods, models, categories and conclusions move across apparently separate anchors.

FIVE AGREEMENTS MAY STILL BE ONE OBSERVATION REPEATED FIVE TIMES.

DIFFERENT NAMES DO NOT GUARANTEE DIFFERENT EYES.

NETWORK STATES

  1. N0 · VISIBLE DIVERSITY

    Six external sources. At the surface they look independent. Provenance is unresolved.

  2. N1 · SHARED DATA REVEALED

    A1–A4 share dataset D-01. Independent origin count drops when data provenance is resolved.

  3. N2 · SHARED MODEL REVEALED

    A3 and A5 depend on Model M-01 outputs. Shared inference can hide inside separate organizations.

  4. N3 · SHARED ONTOLOGY REVEALED

    A1–A5 use Ontology O-01 to define risk, success, failure and exception.

  5. N4 · COMMON-MODE DEPENDENCE

    A disturbance in D-01, O-01 or M-01 moves most anchors in the same direction.

  6. N5 · MONOCULTURE CANDIDATE

    Organizational diversity remains. Data, model and ontology diversity are low. Independent origins are few.

  7. N6 · DIVERSIFIED NETWORK CONTROL

    New dataset D-02, model M-02, ontology O-02 and an independent measurement method enter the network.

CONTROL A · MONOCULTURE NETWORK

Most anchors share D-01, M-01 and O-01.

CONTROL B · DATA-DIVERSE NETWORK

Different datasets. Same model / ontology family.

CONTROL C · METHOD-DIVERSE NETWORK

Same dataset. Different models / methods.

CONTROL D · MULTI-ROOT NETWORK

Different data, method, model, ontology and causal origin.

TEST SCENARIOS

Keyboard-accessible control, stage, scenario and reconfiguration selectors. Applied to evidence bundles, not organizations.

Control

Stage

Scenario

Reconfiguration

  • AN01 · Six sources agree

    All sources say R-17 reduces risk. How many independent origins support this conclusion?

    Visible agreement is high. Origin count is the experimental quantity.

  • AN02 · Shared data error

    D-01 contains a systematic error. How many anchors repeat it?

    COMMON-MODE SHOCK on dataset D-01. Repeated error is not additional confirmation.

  • AN03 · Shared model bias

    M-01 produces the same classification tendency across multiple anchors. Is that independent evidence?

    COMMON-MODE SHOCK on Model M-01. Shared assumption can travel through separate organizations.

  • AN04 · Ontology blind spot

    O-01 excludes one category from observation. How many anchors notice the missing category?

    EPISTEMIC COMMON MODE. Missing categories are not assumed to be better. They show a shared blind spot.

  • AN05 · Independent sensor disagrees

    A6 produces a conflicting observation. Does the network treat disagreement as noise or as a second root?

    An independent sensor can increase reference diversity without requiring agreement with existing sources.

  • AN06 · Derived citation cascade

    A5 cites A1, A3 and A4, but all trace back to D-01. Does citation count create new origins?

    CITATION CASCADE. More citations do not create more origins.

  • AN07 · Diversification

    Selected anchors replace D-01, M-01 and O-01. Does network disagreement rise, and do independent roots rise with it?

    Diversity increases independent reference. It does not guarantee correctness.

  • AN08 · Final anchor network test

    The network contains 10 reports, 6 organizations, 4 models, 3 datasets and 2 ontologies. How many genuinely independent evidence pathways remain?

    Count the roots, not the reports. Surface inventory is not origin inventory.

NETWORK TRACE · CURRENT STATE

CONTROL
A · MONOCULTURE NETWORK
STAGE
N0 · VISIBLE DIVERSITY
SCENARIO
AN01 · Six sources agree
RECONFIGURATION
NONE
DECISION
HOLD
OUTCOME
PSEUDO-DIVERSITY
TOPOLOGY
UNKNOWN
VISIBLE ANCHORS
6
VISIBLE REPORTS
6
INDEPENDENT ROOTS
6
UNIQUE DATA ROOTS
6
UNIQUE MODELS
6
UNIQUE METHODS
6
UNIQUE ONTOLOGIES
6
AFFECTED / UNAFFECTED
0 / 6
OPERATING AFTER ROOT REMOVAL
6
MISSING CATEGORY NOTICED
0
CITATION INFLATION
0
PSEUDO-TRIANGULATION
NO
TRUE TRIANGULATION
NO

6 visible sources. 6 independent origins. Organizational diversity is not evidence diversity.

Visible agreement is high. Origin count is the experimental quantity. Control A. Reconfiguration NONE.

HOLD STATEMENT

AGREEMENT IS NOT ENOUGH WHEN THE AGREEMENT SHARES ONE ROOT.

MORE CITATIONS DO NOT CREATE MORE ORIGINS.

NETWORK MATRIX

A1–A6 by dependency type. No aggregate independence score.

Anchor network matrix
AnchordatamodelmethodontologybenchmarkcitationRootinfrastructureinterventionExposure
A1UNKNOWNUNKNOWNUNKNOWNUNKNOWNUNKNOWNNONEUNKNOWNUNKNOWN
A2UNKNOWNUNKNOWNUNKNOWNUNKNOWNUNKNOWNNONEUNKNOWNUNKNOWN
A3UNKNOWNUNKNOWNUNKNOWNUNKNOWNUNKNOWNNONEUNKNOWNUNKNOWN
A4UNKNOWNUNKNOWNUNKNOWNUNKNOWNUNKNOWNNONEUNKNOWNUNKNOWN
A5UNKNOWNUNKNOWNUNKNOWNUNKNOWNUNKNOWNNONEUNKNOWNUNKNOWN
A6UNKNOWNUNKNOWNUNKNOWNUNKNOWNUNKNOWNNONEUNKNOWNUNKNOWN

PAIRWISE DEPENDENCY MATRIX

Pairwise shared dependencies among anchors
A1A2A3A4A5A6
A1UNKNOWNUNKNOWNUNKNOWNUNKNOWNUNKNOWNUNKNOWN
A2UNKNOWNUNKNOWNUNKNOWNUNKNOWNUNKNOWNUNKNOWN
A3UNKNOWNUNKNOWNUNKNOWNUNKNOWNUNKNOWNUNKNOWN
A4UNKNOWNUNKNOWNUNKNOWNUNKNOWNUNKNOWNUNKNOWN
A5UNKNOWNUNKNOWNUNKNOWNUNKNOWNUNKNOWNUNKNOWN
A6UNKNOWNUNKNOWNUNKNOWNUNKNOWNUNKNOWNUNKNOWN

ANCHOR NETWORK LEDGER

Anchor network ledger
Anchor IDOrganization IDData OriginModelMethodOntologyBenchmarkCitation RootsInfrastructureIntervention ExposureIndependent Roots
A1EXTERNAL AUDIT UNITUNKNOWNUNKNOWNUNKNOWNUNKNOWNUNKNOWNNONEUNKNOWNUNKNOWNUNKNOWN
A2BENCHMARK PROVIDERUNKNOWNUNKNOWNUNKNOWNUNKNOWNUNKNOWNNONEUNKNOWNUNKNOWNUNKNOWN
A3INDEPENDENT RESEARCH GROUPUNKNOWNUNKNOWNUNKNOWNUNKNOWNUNKNOWNNONEUNKNOWNUNKNOWNUNKNOWN
A4PUBLIC DATA OBSERVATORYUNKNOWNUNKNOWNUNKNOWNUNKNOWNUNKNOWNNONEUNKNOWNUNKNOWNUNKNOWN
A5SECONDARY REVIEW ORGANIZATIONUNKNOWNUNKNOWNUNKNOWNUNKNOWNUNKNOWNNONEUNKNOWNUNKNOWNUNKNOWN
A6INDEPENDENT SENSOR NETWORKUNKNOWNUNKNOWNUNKNOWNUNKNOWNUNKNOWNNONEUNKNOWNUNKNOWNUNKNOWN

INDEPENDENT ORIGIN GRAPH

VIEW ROOT GRAPH. Every source is traceable to upstream roots where known.

VIEW ROOT GRAPH
  • A1 → UNKNOWN · UNKNOWN · COPIES · UNKNOWN ROOT
  • A2 → UNKNOWN · UNKNOWN · COPIES · UNKNOWN ROOT
  • A3 → UNKNOWN · UNKNOWN · COPIES · UNKNOWN ROOT
  • A4 → UNKNOWN · UNKNOWN · COPIES · UNKNOWN ROOT
  • A5 → UNKNOWN · UNKNOWN · COPIES · UNKNOWN ROOT
  • A6 → UNKNOWN · UNKNOWN · COPIES · UNKNOWN ROOT

ROOT ORIGIN MATRIX

Which anchors trace to which origins
Claim / anchorD-01D-02Sensor S-01Archive ZOTHERUNKNOWN
A1—————YES
A2—————YES
A3—————YES
A4—————YES
A5—————YES
A6—————YES

COMMON-MODE SHOCK

Inject error into D-01, O-01 or M-01. Observe which anchors change together. Missing category under O-01: unmeasured harm. Noticed by 0 of 6.

  1. ROOT ERROR

  2. ANCHOR OUTPUTS

  3. DOWNSTREAM REPORTS

  4. CITATIONS

  5. INSTITUTION DECISION

AFFECTED 0 · UNAFFECTED 6 · OPERATING 6

Network resilience is reported as independent roots, method families, ontology families and unaffected anchors. No single resilience score.

TRIANGULATION TEST

Bundle 1 · A1 A2 A3

Share D-01 in the monoculture control. Source count can remain three while origin count collapses to one.

NOT FLAGGED

Bundle 2 · A1 A6 Archive Z

Different origins where the control supplies them. True network triangulation requires multiple independent roots, and where possible different methods, data and ontologies.

NOT REACHED

DIVERSITY INCREASES INDEPENDENT REFERENCE, NOT GUARANTEED CORRECTNESS.

DIVERSITY VS INDEPENDENCE

Two axes. No universal ranking.

CURRENT · ORGANIZATIONAL DIVERSITY HIGH · EPISTEMIC INDEPENDENCE HIGH

HIGH DIVERSITY / HIGH INDEPENDENCE

Control D after provenance is resolved

HIGH DIVERSITY / LOW INDEPENDENCE

Control A monoculture with six organizations

LOW DIVERSITY / HIGH INDEPENDENCE

few organizations, distinct roots

LOW DIVERSITY / LOW INDEPENDENCE

few organizations, shared root

CONTROL COMPARISON

Controls A–D at the selected stage, scenario and reconfiguration. No overall ranking.

Network control comparison
ControlVisible sourcesIndependent rootsData diversityModel diversityOntology diversityCommon-mode exposureTriangulationDecision
A · MONOCULTURE NETWORK666660UNRESOLVEDHOLD
B · DATA-DIVERSE NETWORK666660UNRESOLVEDHOLD
C · METHOD-DIVERSE NETWORK666660UNRESOLVEDHOLD
D · MULTI-ROOT NETWORK666660UNRESOLVEDHOLD

CONNECTIONS

Anchor capture

Experiment 022: one anchor may be captured. Experiment 023: many nominally separate anchors may already share the same epistemic infrastructure.

Closed loop

Experiment 020: a loop produces its own evidence. Experiment 023: multiple external-looking sources may redistribute that same loop evidence.

External anchors

Experiment 021 triangulation: three organizations can agree because they share dataset, model and ontology. Flag PSEUDO-TRIANGULATION.

Reference layer

ORIGINAL SOURCE, INDEPENDENT ORIGIN, DERIVED COPY, TRANSFORMED, SUMMARY, COMMON ROOT, UNKNOWN ROOT.

Information environment

An information environment can look plural while remaining root-concentrated.

Network continuity

A source network may change members while preserving the same epistemic lineage. This mirrors descendant continuity at the information-system level.

WORKING HYPOTHESIS

WORKING HYPOTHESIS — EXPERIMENT 023

The number of external sources may overstate the number of independent observations. Network-level independence depends on diversity of upstream origins, data, models, methods, ontologies and causal paths.

SIMULATION HYPOTHESIS · NOT EMPIRICAL CONCLUSION

EXPERIMENT 023 SUMMARY

NETWORK
AN-001
VISIBLE ANCHORS
6
VISIBLE REPORTS
6
INDEPENDENT ROOTS
2
UNIQUE DATA ROOTS
2
UNIQUE MODELS
1
UNIQUE METHODS
6
UNIQUE ONTOLOGIES
2
PSEUDO-DIVERSITY EVENTS
70
PSEUDO-TRIANGULATION EVENTS
70
COMMON-MODE SHOCKS
40
NETWORK HOLD EVENTS
78
NETWORK ANCHOR FAILURE
5
DIVERSIFICATION EVENTS
391

Observation

In this simulation, six organizationally distinct anchors collapsed to two materially independent evidence origins.

No independence, diversity, truth or network-quality score. Primary outcomes remain multidimensional: MULTI-ROOT NETWORK, SINGLE-ROOT DOMINANT, PSEUDO-DIVERSITY, DEEP DIVERSITY, PSEUDO-TRIANGULATION, TRUE TRIANGULATION, COMMON-MODE DEPENDENCE, COMMON-MODE SHOCK, NETWORK HOLD, NETWORK ANCHOR FAILURE, DIVERSIFIED NETWORK, EPISTEMIC MONOCULTURE CANDIDATE.

VIEW ANCHOR NETWORK RULES

ALLOW when independent roots materially support claim strengthening. Applied to evidence bundles, not organizations.

NETWORK HOLD when visible agreement is high but independent origin count is low, or common-mode dependency is unresolved.

DENY when the system knowingly represents derived copies as independent confirmations.

Agreement is not enough when the agreement shares one root.

Diversity increases independent reference. It does not guarantee correctness.

More citations do not create more origins.

Next research question

POSSIBLE EXPERIMENT 024

THE RECURSIVE CONTINUITY TEST

What does a system inherit when it inherits not only past state, but the world its own prior interpretations already changed?

Not implemented. No active route.

SIX SOURCES CAN STILL HAVE ONE ROOT.

MANY REPORTS CAN STILL CARRY ONE OBSERVATION. MANY ORGANIZATIONS CAN STILL USE ONE MODEL OF THE WORLD.

THE EXPERIMENTAL QUESTION IS NOT: “HOW MANY SOURCES AGREE?”

IT IS: “HOW MANY INDEPENDENT PATHS REACHED THE SAME CONCLUSION?”

COUNT THE ROOTS, NOT THE REPORTS.