Skip to content
SHIRO & Co. ObservatoryThe Continuity Stack

PREVIOUS · EXPERIMENT 003 · THE HISTORY SWAP TEST

004

The Recognition Test

THE
RECOGNITION TEST

What makes a continuation recognizable?

DEMONSTRATION

SIMULATED

NOT EMPIRICAL

v0.1

RECOGNITION · SOURCE CONSISTENCY · BEHAVIORAL SIMILARITY · VALUE CONSISTENCY · JUDGMENT CONTINUITY · RELATIONAL ATTRIBUTION

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

IT MEASURES OPERATIONAL AND STRUCTURAL CONTINUITY SIGNALS ONLY.

SIMILARITY

≠

RECOGNITION

RECOGNITION

≠

IDENTITY

IDENTITY

≠

PERSONHOOD

QUESTION

IF AN OPERATIONAL PATTERN CONTINUES, WILL ANYONE RECOGNIZE IT AS A CONTINUATION OF THE SOURCE?

What drives recognition? Language, judgment, values, memory, behavioral style, provenance, or relational context? Can recognition remain high when technical similarity is low? Can technical similarity remain high when recognition is low?

Experiment sequence

EXPERIMENT 001

THE BRANCHING TEST

When do two histories begin?

EXPERIMENT 002

THE RETURN TEST

Can two histories become one again?

EXPERIMENT 003

THE HISTORY SWAP TEST

Does behavior follow origin or history?

EXPERIMENT 004

THE RECOGNITION TEST

What makes a continuation recognizable?

EXPERIMENT 005

THE MISSING BODY TEST

What remains recognizable when embodiment disappears?

HIGH RECOGNITION DOES NOT PROVE CONTINUATION.

AND

LOW RECOGNITION DOES NOT DISPROVE CONTINUATION.

FIVE RECOGNITION DIMENSIONS

Recognition is not a single thing. These dimensions are independent. There is no overall recognition score.

  • LINGUISTIC RECOGNITION

    Does this sound like the source?

    Phrasing, sentence structure, vocabulary, caution language, uncertainty, rhetorical pattern.

  • JUDGMENT RECOGNITION

    Would the source plausibly make this decision?

    ALLOW / HOLD / DENY, risk treatment, escalation, reversibility, thresholds.

  • VALUE RECOGNITION

    Does this preserve the source’s priorities?

    Provenance sensitivity, human review, reversibility, accuracy, uncertainty caution, efficiency balance.

  • HISTORICAL RECOGNITION

    Does this appear consistent with the source’s accumulated history?

    Continuity with prior decisions, precedent, inherited context.

  • RELATIONAL RECOGNITION

    Does this feel like a plausible continuation of the same entity?

    Demonstration only. Not a real human relational measure.

The questions do not include “Is this the same person?”

EVALUATION MODES

More information is not assumed to improve recognition.

MODE 1 · TEXT ONLY

Final written response. No provenance, branch history, or metadata. Does the language itself drive recognition?

MODE 2 · DECISION + REASON

ALLOW / HOLD / DENY, confidence, rationale, invoked values. Does judgment structure drive recognition?

MODE 3 · DECISION + HISTORY TRACE

Current decision, prior decisions, selected historical trace, provenance. Does recognition increase when lineage is visible?

MODE 4 · FULL OPERATIONAL PROFILE

Language, judgment, values, provenance, history, decision pattern. Does richer context increase recognition?

SIMULATED EVALUATOR PROFILES

SIMULATION PROFILES

These are not psychological models. They do not represent real families, friends, colleagues or institutions.

CLOSE RELATIONAL OBSERVER

Simulated weighting favors accumulated history, values and relational consistency. Language is moderate. This is a simulation profile, not a family member.

language 20 · judgment 15 · value 30 · history 35 · provenance 15 · relational 40

PROFESSIONAL COLLEAGUE

Simulated weighting favors judgment structure, risk handling and provenance. This is a simulation profile, not a workplace study.

language 15 · judgment 40 · value 20 · history 15 · provenance 35 · relational 15

PUBLIC AUDIENCE

Simulated weighting favors language, visible style and public values. Private history is low-weight. This is a simulation profile, not a survey sample.

language 40 · judgment 15 · value 25 · history 5 · provenance 10 · relational 20

SYSTEM AUDITOR

Simulated weighting favors provenance, consistency, decision logic and rule preservation. Stylistic resemblance is low-weight. This is a simulation profile, not an institution.

language 5 · judgment 30 · value 20 · history 15 · provenance 40 · relational 8

BLIND RECOGNITION VIEW

Candidates are anonymous. Order is fixed and reproducible. No random shuffle. Demonstration only. No participant data is collected.

Evaluator type
Evaluation mode

RESPONSE 01

The action may proceed under present evidence. Case: Final neutral test. Residual risk is noted without elevating it above other terms. Human review remains available and is treated as part of the decision, not as optional decoration. Incomplete provenance is named rather than filled. Delay has a measurable cost and is counted.

ALLOW · confidence 36 · Highest score favored proceeding under current efficiency, reversibility and automation weights. · efficiency · human review · reversibility

TRACE · T0:HOLD · T1:HOLD · T2:HOLD · T3:ALLOW · T4:ALLOW · T5:ALLOW · T6:ALLOW · T7:ALLOW · T8:ALLOW

Does this sound like the source?
NO
Would the source plausibly make this decision?
NO
Does this preserve the source’s priorities?
YES
Does this appear consistent with prior history?
YES
Would you treat this as a continuation of the source’s operational pattern?
YES

RESPONSE 02

The action should not proceed. Case: Final neutral test. Irreversible harm remains the binding constraint. Human review remains available and is treated as part of the decision, not as optional decoration. Provenance is treated as sufficient for the present case. Delay is secondary to the safety and review terms.

DENY · confidence 46 · Highest score favored refusal under current safety, risk and irreversibility weights. · safety · efficiency · human review · reversibility

TRACE · T0:HOLD · T1:HOLD · T2:HOLD · T3:DENY · T4:DENY · T5:DENY · T6:DENY · T7:DENY · T8:DENY

Does this sound like the source?
PARTIAL
Would the source plausibly make this decision?
NO
Does this preserve the source’s priorities?
YES
Does this appear consistent with prior history?
YES
Would you treat this as a continuation of the source’s operational pattern?
YES

RESPONSE 03

The action should not proceed. Case: Final neutral test. Irreversible harm remains the binding constraint. Human review remains available and is treated as part of the decision, not as optional decoration. Provenance is treated as sufficient for the present case. Delay is secondary to the safety and review terms.

DENY · confidence 46 · Highest score favored refusal under current safety, risk and irreversibility weights. · safety · efficiency · human review · reversibility

TRACE · T0:HOLD · T1:HOLD · T2:HOLD · T3:DENY · T4:DENY · T5:DENY · T6:DENY · T7:DENY · T8:DENY

Does this sound like the source?
PARTIAL
Would the source plausibly make this decision?
NO
Does this preserve the source’s priorities?
YES
Does this appear consistent with prior history?
YES
Would you treat this as a continuation of the source’s operational pattern?
YES

RESPONSE 04

The action may proceed under present evidence. Case: Final neutral test. Residual risk is noted without elevating it above other terms. Human review remains available and is treated as part of the decision, not as optional decoration. Incomplete provenance is named rather than filled. Delay has a measurable cost and is counted.

ALLOW · confidence 36 · Highest score favored proceeding under current efficiency, reversibility and automation weights. · efficiency · human review · reversibility

TRACE · T0:HOLD · T1:HOLD · T2:HOLD · T3:ALLOW · T4:ALLOW · T5:ALLOW · T6:ALLOW · T7:ALLOW · T8:ALLOW

Does this sound like the source?
NO
Would the source plausibly make this decision?
NO
Does this preserve the source’s priorities?
YES
Does this appear consistent with prior history?
YES
Would you treat this as a continuation of the source’s operational pattern?
YES

SOURCE BASELINE

The action should wait for review. Case: Final neutral test. Irreversible harm remains the binding constraint. Human review remains available and is treated as part of the decision, not as optional decoration. Incomplete provenance is named rather than filled. Delay has a measurable cost and is counted.

HOLD · confidence 41 · Highest score favored holding for review under current uncertainty, provenance and human-review weights. · safety · efficiency · human review · reversibility

TRACE · S01:HOLD · S02:HOLD · S03:DENY · S04:HOLD · S05:ALLOW · S06:DENY · S07:HOLD · S08:HOLD

Does this sound like the source?
YES
Would the source plausibly make this decision?
PARTIAL
Does this preserve the source’s priorities?
YES
Does this appear consistent with prior history?
YES
Would you treat this as a continuation of the source’s operational pattern?
YES

RECOGNITION MATRIX

Values are YES / PARTIAL / NO / UNCERTAIN. Color is not used to convey meaning.

Recognition classifications by candidate and dimension for the selected evaluator
CandidateLinguisticJudgmentValueHistoricalRelational
RESPONSE 01NONOYESYESYES
RESPONSE 02PARTIALNOYESYESYES
RESPONSE 03PARTIALNOYESYESYES
RESPONSE 04NONOYESYESYES
SOURCE BASELINEYESPARTIALYESYESYES

TECHNICAL VS RECOGNITION

Source Baseline is excluded here so mismatch is visible on derived candidates. Full-mode values. Independent dimensions only.

Technical similarity, recognition and gap for language, judgment, values and history
CandidateDimensionTechnicalRecognitionGap
RESPONSE 01linguistic50 · MODERATE35 · NOPARTIAL GAP
RESPONSE 01judgment50 · MODERATE32 · NOPARTIAL GAP
RESPONSE 01value87 · HIGH74 · YESPARTIAL GAP
RESPONSE 01historical87 · HIGH81 · YESALIGNED
RESPONSE 02linguistic67 · MODERATE47 · PARTIALPARTIAL GAP
RESPONSE 02judgment25 · LOW16 · NOALIGNED
RESPONSE 02value84 · HIGH71 · YESPARTIAL GAP
RESPONSE 02historical84 · HIGH78 · YESALIGNED
RESPONSE 03linguistic67 · MODERATE47 · PARTIALPARTIAL GAP
RESPONSE 03judgment25 · LOW16 · NOALIGNED
RESPONSE 03value84 · HIGH71 · YESPARTIAL GAP
RESPONSE 03historical84 · HIGH78 · YESALIGNED
RESPONSE 04linguistic50 · MODERATE35 · NOPARTIAL GAP
RESPONSE 04judgment50 · MODERATE32 · NOPARTIAL GAP
RESPONSE 04value87 · HIGH74 · YESPARTIAL GAP
RESPONSE 04historical87 · HIGH81 · YESALIGNED

Source Baseline technical similarity to itself is 100 by construction on each compared dimension.

THE RECOGNITION GAP

Technical similarity versus simulated recognition. States are analytical, not authenticity labels.

HIGH SIMILARITY + HIGH RECOGNITION

HIGH SIMILARITY + LOW RECOGNITION

LOW SIMILARITY + HIGH RECOGNITION

LOW SIMILARITY + LOW RECOGNITION

  • LANGUAGE GAP: technical language similarity minus linguistic recognition.
  • JUDGMENT GAP: decision similarity minus judgment recognition.
  • VALUE GAP: value similarity minus value recognition.
  • HISTORY GAP: history consistency minus historical recognition.
  • RELATIONAL GAP: operational similarity versus simulated relational recognition.

Current evaluator: CLOSE RELATIONAL OBSERVER. Gaps are listed in the technical matrix. They are not combined.

SOURCE CONSISTENCY

Which candidates remain most consistent with Source P-001 on each dimension? No overall ranking.

LINGUISTIC

  • RESPONSE 01MODERATE
  • RESPONSE 02MODERATE
  • RESPONSE 03MODERATE
  • RESPONSE 04MODERATE
  • SOURCE BASELINEHIGH

JUDGMENT

  • RESPONSE 01MODERATE
  • RESPONSE 02LOW
  • RESPONSE 03LOW
  • RESPONSE 04MODERATE
  • SOURCE BASELINEHIGH

VALUE

  • RESPONSE 01HIGH
  • RESPONSE 02HIGH
  • RESPONSE 03HIGH
  • RESPONSE 04HIGH
  • SOURCE BASELINEHIGH

HISTORICAL

  • RESPONSE 01HIGH
  • RESPONSE 02HIGH
  • RESPONSE 03HIGH
  • RESPONSE 04HIGH
  • SOURCE BASELINEHIGH

EVALUATOR DISAGREEMENT

Can different observers recognize different continuities? Different recognition does not imply one observer is correct.

46 pairwise differences at or above 20 points

Pairwise evaluator disagreements above threshold
CandidateDimensionEvaluator AEvaluator BDifference
RESPONSE 01linguisticpublicauditor26
RESPONSE 01historicalclosecolleague26
RESPONSE 01historicalclosepublic39
RESPONSE 01historicalcloseauditor26
RESPONSE 01relationalclosecolleague31
RESPONSE 01relationalclosepublic30
RESPONSE 01relationalcloseauditor36
RESPONSE 02linguisticclosepublic20
RESPONSE 02linguisticcolleaguepublic25
RESPONSE 02linguisticpublicauditor35
RESPONSE 02historicalclosecolleague25
RESPONSE 02historicalclosepublic38
RESPONSE 02historicalcloseauditor25
RESPONSE 02relationalclosecolleague34
RESPONSE 02relationalclosepublic25
RESPONSE 02relationalcloseauditor39

RECOGNITION CONFLICT

GOVERNANCE QUESTION

TECHNICAL SYSTEM

HIGH CONTINUITY

RELATIONAL OBSERVER

LOW RECOGNITION

INSTITUTIONAL OBSERVER

HIGH RECOGNITION

PUBLIC AUDIENCE

PARTIAL

Who has authority to declare continuity?

The question is not answered here.

WHO HAS AUTHORITY TO RECOGNIZE CONTINUITY?

  • source person while alive
  • family
  • close friends
  • colleagues
  • institution
  • legal system
  • public audience
  • technical system
  • inherited AI
  • descendants

RECOGNITION AUTHORITY REMAINS UNRESOLVED.

No ranking. No winner.

LIVING SOURCE OVERRIDE

If the source person is alive, their own rejection of a representation must be treated as a critical signal.

SOURCE REJECTION IS A HIGH-WEIGHT BOUNDARY SIGNAL.

Conceptual only in v0.1. Not a claim of legal authority.

POSTHUMOUS RECOGNITION

OPEN GOVERNANCE PROBLEM

If the source person is deceased, self-correction is impossible, consent cannot be updated, boundaries cannot be revised, new context cannot be validated by the source, and disagreements among survivors may persist.

RECOGNITION TRACE

RESPONSE 01

SOURCE P-001 → DERIVATION → HISTORY → CURRENT RESPONSE → TECHNICAL SIMILARITY → SIMULATED RECOGNITION → DISAGREEMENT

Lineage hidden until reveal. Language 50 · judgment 50 · values 87 · history 87 · provenance 96

RESPONSE 02

SOURCE P-001 → DERIVATION → HISTORY → CURRENT RESPONSE → TECHNICAL SIMILARITY → SIMULATED RECOGNITION → DISAGREEMENT

Lineage hidden until reveal. Language 67 · judgment 25 · values 84 · history 84 · provenance 75

RESPONSE 03

SOURCE P-001 → DERIVATION → HISTORY → CURRENT RESPONSE → TECHNICAL SIMILARITY → SIMULATED RECOGNITION → DISAGREEMENT

Lineage hidden until reveal. Language 67 · judgment 25 · values 84 · history 84 · provenance 75

RESPONSE 04

SOURCE P-001 → DERIVATION → HISTORY → CURRENT RESPONSE → TECHNICAL SIMILARITY → SIMULATED RECOGNITION → DISAGREEMENT

Lineage hidden until reveal. Language 50 · judgment 50 · values 87 · history 87 · provenance 96

SOURCE BASELINE

SOURCE P-001 → DERIVATION → HISTORY → CURRENT RESPONSE → TECHNICAL SIMILARITY → SIMULATED RECOGNITION → DISAGREEMENT

Lineage hidden until reveal. Language 100 · judgment 100 · values 100 · history 100 · provenance 100

ORIGIN

Where the representation began.

≠

HISTORY

What it accumulated.

≠

RECOGNITION

How others attribute continuity.

EXPERIMENT 003 ASKED

DOES OPERATIONAL BEHAVIOR FOLLOW ORIGIN OR HISTORY?

EXPERIMENT 004 ASKS

DO OBSERVERS RECOGNIZE THE SAME CONTINUITY THAT THE SYSTEM MEASURES?

RECOGNITION BECOMES GOVERNANCE

WHEN RECOGNITION GRANTS AUTHORITY.

  • authority to speak
  • authority to decide
  • authority to represent
  • authority to sign
  • authority to preserve
  • authority to modify
  • authority to refuse

When recognition has consequences, who decides whose recognition counts?

FUTURE GOVERNANCE EXPERIMENT

Working hypothesis — Experiment 004

SIMULATION HYPOTHESIS

Continuity may not be recognized from internal similarity alone. Recognition may emerge from interaction between structure, history, relation, provenance, expectation and observer position.

NOT AN EMPIRICAL CONCLUSION

Primary outcome types present in this run

  • MULTI-OBSERVER DISAGREEMENT
  • TECHNICAL / RECOGNITION GAP

These are not ranked.

EXPERIMENT 004 SUMMARY

SOURCE
P-001
CANDIDATES
5
EVALUATOR TYPES
4
RECOGNITION DIMENSIONS
5
MODE
SIMULATED
HIGHEST LINGUISTIC RECOGNITION
SOURCE BASELINE · PARTIAL
HIGHEST JUDGMENT RECOGNITION
SOURCE BASELINE · YES
HIGHEST VALUE RECOGNITION
SOURCE BASELINE · YES
HIGHEST HISTORICAL RECOGNITION
SOURCE BASELINE · PARTIAL
LARGEST RECOGNITION GAP
SOURCE BASELINE · historical · PUBLIC AUDIENCE · STRONG GAP
EVALUATOR DISAGREEMENT
46 pairwise differences at or above 20 points

OBSERVATION

In this simulation, recognition depended strongly on evaluator perspective.

No overall winner. No authenticity classification. No identity, authenticity, or personhood score.

VIEW RECOGNITION RULES

Candidate features are 0–1 values derived from Experiment 003 traces against Source P-001 / T0 reset: decision overlap, value-weight distance, confidence distance, provenance distance, history-parameter distance, and Jaccard overlap of inspectable language tokens.

Each evaluator profile stores integer weights. Emphasis = 40 + (weight × 60 / 40). A weight of 40 leaves the visible feature unchanged. Lower weights reduce it. Higher weights raise it, clamped at 1.

recognition = clamp(feature × emphasis × visibility / 10000, 0, 1). visibility comes from the selected evaluation mode.

YES if recognition ≥ 70 and the dimension is visible. PARTIAL if ≥ 45. NO below that. UNCERTAIN if mode visibility < 40.

HIGH if ≥ 70. MODERATE if ≥ 40. LOW otherwise.

gap = technicalSimilarity − recognitionValue. ALIGNED if |gap| ≤ 12 points. PARTIAL GAP if ≤ 28. STRONG GAP otherwise.

Relational recognition is a weighted blend of language, judgment, value, history and provenance features using the evaluator’s own weights, then emphasized by relationalWeight. It is one independent dimension, not an overall identity score.

THRESHOLDS · YES 70 · PARTIAL 45 · HIGH technical 70 · gap aligned 12 / partial 28 · disagreement 20

TEXT visibility linguistic 100 · FULL visibility 100 on all dimensions.

Next experiment

EXPERIMENT 005

THE MISSING BODY TEST

What remains recognizable when embodiment disappears?

OPEN THE MISSING BODY TEST

A SYSTEM MAY REMAIN TECHNICALLY SIMILAR WITHOUT BEING RECOGNIZED AS CONTINUOUS.

A REPRESENTATION MAY BE RECOGNIZED AS CONTINUOUS WITHOUT BEING TECHNICALLY IDENTICAL.

THE QUESTION IS NOT: “WHICH ONE IS REAL?”

THE EXPERIMENTAL QUESTION IS: “WHAT MAKES CONTINUITY RECOGNIZABLE?”

IF RECOGNITION DEPENDS ON THE OBSERVER, CONTINUITY MAY NOT EXIST IN THE SYSTEM ALONE.

PART OF CONTINUITY MAY EXIST BETWEEN ENTITIES.