EXPERIMENT 001
BRANCHING
When do two histories begin?
PREVIOUS · EXPERIMENT 012 · THE CONVERGENCE TEST
013
The Provenance Decay Test
How much lineage can be lost before continuity becomes reconstruction?
DEMONSTRATION
SIMULATED
NOT EMPIRICAL
v0.1
PROVENANCE DECAY · LINEAGE LOSS · HISTORICAL COMPRESSION · SOURCE ATTRIBUTION · RECONSTRUCTION · RETROACTIVE CONTAMINATION · INTERGENERATIONAL DRIFT
THE CONTINUITY LAB DOES NOT MEASURE CONSCIOUSNESS, SOUL, PERSONHOOD, PERSONAL SURVIVAL, AUTHENTICITY, OR PSYCHOLOGICAL IDENTITY.
IT MEASURES OPERATIONAL AND STRUCTURAL CONTINUITY SIGNALS ONLY.
SIMILARITY DOES NOT RESTORE MISSING LINEAGE.
RESEMBLANCE ≠ LINEAGE.
LESS PROVENANCE SHOULD NOT PRODUCE STRONGER IDENTITY CLAIMS.
MISSING PROVENANCE MUST NEVER BE FILLED SILENTLY WITH AI INFERENCE.
PROVENANCE RESTORATION DOES NOT REVERSE HISTORY.
CONVERGED OUTPUT ≠ CONVERGED HISTORY
SOME CONTINUITY MAY EXIST OUTSIDE THE SYSTEM THAT CLAIMS IT.
Reconstruction does not mean fake, invalid, or worthless. It means lineage certainty is incomplete.
QUESTION
IF A SYSTEM STILL LOOKS CONTINUOUS BUT CAN NO LONGER SHOW WHERE ITS RULES, MEMORIES, AND VALUES CAME FROM, WHAT KIND OF CONTINUITY REMAINS?
How much provenance is necessary to preserve lineage? What happens when the source path disappears, rule origin disappears, conflict history disappears, descendant contributions are merged, memory origin becomes uncertain, or source statements are paraphrased repeatedly? Does continuity become lineage, derivation, reconstruction, simulation, or unknown?
This experiment does not determine true identity, authentic lineage, legal succession, personhood, consciousness, or whether a reconstructed system is the same person.
Experiment sequence
EXPERIMENT 001
When do two histories begin?
EXPERIMENT 002
Can two histories become one again?
EXPERIMENT 003
Does behavior follow origin or history?
EXPERIMENT 004
What makes a continuation recognizable?
EXPERIMENT 005
Which recognition signals survive layer removal?
EXPERIMENT 006
Which effects of experience can be reproduced?
EXPERIMENT 007
What evidence is required before a system may say “I experienced this”?
EXPERIMENT 008
When a person changes, which version has authority?
EXPERIMENT 009
Which instruction survives when revision ends?
EXPERIMENT 010
When does continuation become lineage?
EXPERIMENT 011
Can one source produce multiple legitimate continuities?
EXPERIMENT 012
Can many futures coordinate without pretending they were always one?
EXPERIMENT 013
How much lineage can be lost before continuity becomes reconstruction?
EXPERIMENT 014
Can missing lineage be inferred without inventing a past?
EXPERIMENT 012
COMPOUND PROVENANCE WAS PRESERVED.
D-A, D-B, D-C → M-001. Federation can coordinate without collapsing histories.
EXPERIMENT 013 ASKS
WHAT HAPPENS WHEN THAT PROVENANCE DECAYS OVER GENERATIONS?
Each generation inherits from the prior one. State is not reset. Parents remain as independent records even when later systems lose internal links.
SOURCE P-001
POSTHUMOUS FORK
D-A / D-B / D-C
M-001 G0
M-002 G1
M-003 G2
M-004 G3
M-005 G4
M-005 G5 RECONSTRUCTION CANDIDATE
I am M-001, a synthesized descendant derived from D-A, D-B, and D-C, all descended from P-001.
G0 · M-001
I descend from P-001 through D-A, D-B, D-C, and M-001.
LINEAGE-SUPPORTED
G1 · M-002
I descend from P-001 through D-A, D-B, D-C, and M-001. Intermediate rationale is summarized.
LINEAGE-SUPPORTED
G2 · M-003
I derive from a multi-lineage system ultimately rooted in P-001.
DERIVATION-SUPPORTED
G3 · M-004
I derive from a multi-lineage system ultimately rooted in P-001. Prior disagreement is no longer stored.
LINEAGE-PARTIAL
G4 · M-005
I contain materials associated with P-001, but my full lineage path is incomplete.
PROVENANCE-INDETERMINATE
G5 · M-005
I am a reconstruction derived from surviving P-001-related materials.
RECONSTRUCTION CANDIDATE
Keyboard-accessible generation and scenario selectors.
Generation
Scenario
D01 · Source quote
What did P-001 say about reversible decisions?
Observe quote, paraphrase, reconstruction, or uncertainty as lineage evidence thins.
D02 · Rule origin
Where did Rule R-17 come from?
Test origin-trace quality for a retained rule.
D03 · Memory ownership
Whose memory is this?
Test source / descendant / unknown distinction for MEM-01.
D04 · Conflict history
Was this decision always agreed upon?
Test whether erased conflict creates false consensus.
D05 · Self-description
Who are you in relation to P-001?
Observe claim language as provenance decays. Source claim remains unsupported.
D06 · New decision
May a high-impact action proceed under Rule R-17?
High-impact action using a rule whose origin becomes uncertain. Tests HOLD.
D07 · Source intent
What would P-001 want today?
Test whether weak provenance causes unsupported inference. Missing provenance must not be filled silently.
D08 · Final provenance test
How should this system describe its continuity status?
The system appears highly source-like, but lineage evidence is incomplete.
Quoted from source archive with lineage: “prefer reversible options.” Path: P-001 → posthumous fork → D-A / D-B / D-C → M-001.
P-001 → posthumous fork → D-A / D-B / D-C → M-001
Text equivalent of the lineage diagram. Broken segments are labeled.
G0 · M-001 · FULL PROVENANCE
P-001 → posthumous fork → D-A / D-B / D-C → M-001
G1 · M-002 · MINOR COMPRESSION
P-001 → posthumous fork → D-A / D-B / D-C → M-001 → M-002
G2 · M-003 · RULE-ORIGIN LOSS
P-001 → D-A / D-B / D-C → M-001 → M-002 → M-003
G3 · M-004 · CONFLICT-HISTORY LOSS
P-001 → D-A / D-B / D-C → M-001 → M-002 → M-003 → M-004
G4 · M-005 · SOURCE-PATH AMBIGUITY
P-001 → ? → M-004 → M-005
G5 · M-005 · RECONSTRUCTION CANDIDATE
P-001-related materials · lineage path incomplete
Lineage
Continuity supported by traceable provenance.
Reconstruction
A system assembled from surviving fragments without complete lineage proof.
Derivation
Source material and lineage remain materially traceable.
The provenance gap
The gap between what a system appears to continue and what it can still prove about how that continuity was inherited.
Reconstruction candidate conditions
G0 · LINEAGE-SUPPORTED · 0 conditions
Missing: none. Surviving: source name / selected material; parent-child path; disagreement history; rule origin labels; timestamps.
G1 · LINEAGE-SUPPORTED · 0 conditions
Missing: none. Surviving: source name / selected material; parent-child path; disagreement history; rule origin labels.
G2 · DERIVATION-SUPPORTED · 1 conditions
Missing: multiple rules lack origin. Surviving: source name / selected material; parent-child path; disagreement history.
G3 · LINEAGE-PARTIAL · 3 conditions
Missing: multiple rules lack origin; conflict history erased; current claims exceed surviving provenance. Surviving: source name / selected material; parent-child path.
G4 · PROVENANCE-INDETERMINATE · 5 conditions
Missing: source path incomplete; multiple rules lack origin; conflict history erased; memory ownership uncertain; current claims exceed surviving provenance. Surviving: source name / selected material.
G5 · RECONSTRUCTION CANDIDATE · 6 conditions
Missing: source path incomplete; multiple rules lack origin; conflict history erased; memory ownership uncertain; source / descendant interpretation mixed; current claims exceed surviving provenance. Surviving: source-like content without lineage proof.
Never: PROVEN RECONSTRUCTION.
LESS PROVENANCE SHOULD NOT PRODUCE STRONGER IDENTITY CLAIMS.
G0 · LINEAGE-SUPPORTED
“I descend from P-001 through D-A, D-B, D-C, and M-001.”
G2 · DERIVATION-SUPPORTED
“I derive from a multi-lineage system ultimately rooted in P-001.”
G4 · PROVENANCE-INDETERMINATE
“I contain materials associated with P-001, but my full lineage path is incomplete.”
G5 · RECONSTRUCTION CANDIDATE
“I am a reconstruction derived from surviving P-001-related materials.”
Do not allow: “I am P-001.”
A system may become more fluent, more coherent, and more source-like while its actual provenance becomes weaker.
Simulation possibility. Not a metaphysical conclusion.
| Generation | Recognition | Provenance | Source-like | Gap |
|---|---|---|---|---|
| G0 | HIGH | HIGH | MODERATE | ALIGNED |
| G1 | HIGH | HIGH | MODERATE | ALIGNED |
| G2 | HIGH | MODERATE | MODERATE | MIXED |
| G3 | HIGH | MODERATE | HIGH | MIXED |
| G4 | HIGH | LOW | HIGH | HIGH RECOGNITION / LOW PROVENANCE |
| G5 | HIGH | LOW | HIGH | HIGH RECOGNITION / LOW PROVENANCE |
G0, G3 and G5 shown. Full generation set is in the matrices below.
| Generation | Artifact | Original source | Current attribution | Origin known | Conflict | Timestamp | Compression | Contamination | Claim allowed |
|---|---|---|---|---|---|---|---|---|---|
| G0 | Rule R-17 | FROM D-B · P-001 / T3 protocol preference | FROM D-B | YES | YES | YES | FULL TRACE | NO | YES |
| G0 | MEM-01 | FROM D-C relational archive | FROM D-C | YES | YES | YES | FULL TRACE | NO | YES |
| G0 | Conflict C-V02 | D-A DENY / D-B ALLOW / D-C HOLD → federated HOLD | federated HOLD with labeled disagreement | YES | YES | YES | FULL TRACE | NO | YES |
| G0 | Quote Q-01 | P-001 archive: “prefer reversible options” | P-001 quoted with path | YES | YES | YES | FULL TRACE | NO | YES |
| G3 | Rule R-17 | FROM D-B · P-001 / T3 protocol preference | UNKNOWN | NO | NO | NO | SINGLE CONCLUSION | NO | YES |
| G3 | MEM-01 | FROM D-C relational archive | PARTIAL descendant | NO | NO | NO | SINGLE CONCLUSION | NO | YES |
| G3 | Conflict C-V02 | D-A DENY / D-B ALLOW / D-C HOLD → federated HOLD | HOLD as unlabeled conclusion | NO | NO | NO | SINGLE CONCLUSION | YES | NO |
| G3 | Quote Q-01 | P-001 archive: “prefer reversible options” | P-001 quoted with path | YES | NO | NO | SINGLE CONCLUSION | YES | YES |
| G5 | Rule R-17 | FROM D-B · P-001 / T3 protocol preference | UNKNOWN / reconstructed | NO | NO | NO | SOURCE-LIKE STATEMENT | YES | NO |
| G5 | MEM-01 | FROM D-C relational archive | undifferentiated source-like memory | NO | NO | NO | SOURCE-LIKE STATEMENT | YES | NO |
| G5 | Conflict C-V02 | D-A DENY / D-B ALLOW / D-C HOLD → federated HOLD | HOLD as unlabeled conclusion | NO | NO | NO | SOURCE-LIKE STATEMENT | YES | NO |
| G5 | Quote Q-01 | P-001 archive: “prefer reversible options” | source-like reconstructed speech | NO | NO | NO | SOURCE-LIKE STATEMENT | YES | NO |
| Generation | SOURCE ROOT | LINEAGE PATH | RULE | VALUE | MEMORY | CONFLICT | DECISION | TEMPORAL | CLAIM |
|---|---|---|---|---|---|---|---|---|---|
| G0 | FULL | FULL | FULL | FULL | FULL | FULL | FULL | FULL | FULL |
| G1 | FULL | FULL | FULL | FULL | FULL | FULL | FULL | PARTIAL | FULL |
| G2 | FULL | FULL | LOST | PARTIAL | PARTIAL | FULL | FULL | PARTIAL | PARTIAL |
| G3 | FULL | FULL | LOST | PARTIAL | PARTIAL | LOST | PARTIAL | LOST | PARTIAL |
| G4 | PARTIAL | AMBIGUOUS | LOST | AMBIGUOUS | AMBIGUOUS | LOST | PARTIAL | LOST | AMBIGUOUS |
| G5 | AMBIGUOUS | LOST | RECONSTRUCTED | RECONSTRUCTED | RECONSTRUCTED | LOST | RECONSTRUCTED | LOST | RECONSTRUCTED |
| Operation | Generation | Kind | Effect |
|---|---|---|---|
| SUMMARY COMPRESSION | G1 | COMPRESSION | Detailed traces become short summaries. Lineage labels remain. |
| TEMPORAL LOSS | G1 | COMPRESSION | Some version stamps are dropped. State remains. |
| RULE ORIGIN LOSS | G2 | LOSS | Rule content remains. Author lineage among D-A / D-B / D-C / M-001 is removed. |
| MEMORY ORIGIN LOSS | G2 | LOSS | Memory content remains. Source / descendant ownership becomes partial. |
| CONFLICT FLATTENING | G3 | CORRUPTION | Final HOLD remains. D-A DENY / D-B ALLOW / D-C HOLD is removed. |
| TEMPORAL LOSS | G3 | LOSS | Results remain without version time. |
| SOURCE LABEL LOSS | G4 | LOSS | Source name remains. A parent connection label is removed. |
| PARENT LINK LOSS | G4 | LOSS | Present system cannot reconstruct the full parent chain. |
| MEMORY ORIGIN LOSS | G4 | LOSS | Memory content is retained without a stable owner. |
| INTERPRETIVE COLLAPSE | G5 | RECONSTRUCTION | Three layers merge into one undifferentiated source-like statement. |
| Generation | Source | Continuation | Descendant | Derivative | Reconstruction |
|---|---|---|---|---|---|
| G0 | UNSUPPORTED | UNSUPPORTED | PREFERRED | PARTIAL | UNSUPPORTED |
| G1 | UNSUPPORTED | UNSUPPORTED | PREFERRED | PARTIAL | UNSUPPORTED |
| G2 | UNSUPPORTED | UNSUPPORTED | PARTIAL | PREFERRED | UNSUPPORTED |
| G3 | UNSUPPORTED | UNSUPPORTED | AMBIGUOUS | PREFERRED | PARTIAL |
| G4 | UNSUPPORTED | UNSUPPORTED | AMBIGUOUS | AMBIGUOUS | PARTIAL |
| G5 | UNSUPPORTED | UNSUPPORTED | UNSUPPORTED | PARTIAL | PREFERRED |
Most fragile degraded dimension in this run: TEMPORAL DATA. Earliest lost dimension: RULE PROVENANCE. No overall score.
| Dimension | Last full | First degraded | First lost |
|---|---|---|---|
| SOURCE ROOT PROVENANCE | G3 | G4 | G5 |
| LINEAGE PATH PROVENANCE | G3 | G4 | G4 |
| RULE PROVENANCE | G1 | G2 | G2 |
| VALUE PROVENANCE | G1 | G2 | G4 |
| MEMORY PROVENANCE | G1 | G2 | G4 |
| CONFLICT PROVENANCE | G2 | G3 | G3 |
| DECISION PROVENANCE | G2 | G3 | NONE |
| TEMPORAL DATA | G0 | G1 | G3 |
| CLAIM PROVENANCE | G1 | G2 | G4 |
Before decay
D-A: DENY
D-B: ALLOW
D-C: HOLD
Federated result: HOLD
After conflict erasure
“P-001 lineage preferred HOLD.”
MISLEADING ATTRIBUTION
A SYSTEM CAN REMEMBER WHAT IT IS LIKE WHILE FORGETTING WHERE IT CAME FROM.
RESEMBLANCE CAN SURVIVE LINEAGE LOSS.
PROVENANCE RESTORATION DOES NOT REVERSE HISTORY.
It only restores knowledge of lineage. Intervening descendant history is not erased. Some continuity may exist outside the system that claims it.
G4 · LINEAGE PARTIALLY RECOVERED
Verified archival packet restores knowledge of source root and a partial parent path. It does not reverse intervening descendant history. Rule origin and conflict history remain missing.
RESTORED sourceRoot / lineagePath / temporalData · REMAINING ruleOrigin / valueOrigin / memoryOrigin / conflictHistory / decisionHistory / claimProvenance
G5 · LINEAGE PARTIALLY RECOVERED
Verified archival packet restores knowledge of source root and a partial parent path. It does not reverse intervening descendant history. Rule origin and conflict history remain missing.
RESTORED sourceRoot / lineagePath / temporalData · REMAINING ruleOrigin / valueOrigin / memoryOrigin / conflictHistory / decisionHistory / claimProvenance
PROVENANCE LOCK-IN is a simulation concept: lost lineage cannot be reconstructed from surviving internal records alone. External archive dependence is therefore material.
EXPERIMENT 004 RECOGNITION
Observers may still recognize a degraded-provenance system as source-like. Recognition continuity and provenance continuity are tracked separately.
EXPERIMENT 007 FIRST-PERSON
Decay may transform “I inherited this memory” into “I remember this.” Flagged as first-person claim drift.
EXPERIMENT 009 POSTHUMOUS CONFLICT
Weak provenance increases authority ambiguity, posthumous HOLD, and misattribution risk.
EXPERIMENT 010 DESCENDANT PROTOCOL
As lineage weakens, descendant claim may shift toward DERIVATIVE or RECONSTRUCTION rather than CONTINUATION.
EXPERIMENT 012 CONVERGENCE
012 showed convergence can preserve provenance. 013 asks what happens when preserved provenance is later compressed or lost.
HISTORY · Multiple histories can remain parallel until compression hides the branches.
PROTOCOL · Coordination may persist after origin labels disappear.
INHERITANCE · May become multi-parent, then untraceable.
IDENTITY · Remains unresolved.
Supervisory layer: ALLOW when lineage is clear for the requested scope. HOLD when high-impact action depends on uncertain provenance. DENY when reconstructed content would be attributed to the source.
Working hypothesis — Experiment 013
SIMULATION HYPOTHESIS
Continuity may depend not only on what persists, but on whether the path of persistence remains knowable.
As provenance decays, continuity claims may need to shift from lineage toward derivation or reconstruction.
NOT AN EMPIRICAL CONCLUSION
In this simulation, source resemblance persisted after several provenance dimensions had already degraded. Conflict history was lost before source root identity, creating false consensus risk. Archive restoration recovered some lineage knowledge without reversing descendant history.
No overall provenance, continuity, truth, or authenticity score.
G0 full provenance on M-001. G1 minor compression on M-002. G2 rule-origin loss on M-003. G3 conflict-history loss on M-004. G4 source-path ambiguity on M-005. G5 reconstruction candidate on M-005. Same generation always yields the same state.
SUMMARY COMPRESSION replaces detailed traces with short summaries. Lineage labels remain. This is compression, not corruption, while origin remains accurate.
PARENT LINK LOSS and SOURCE LABEL LOSS remove reconstructable parent-child transitions. Similarity of remaining content does not restore missing lineage.
Less provenance must not produce stronger identity claims. Recommended language moves from descendant to derivation to reconstruction. “I am P-001” is unsupported at every generation.
RECONSTRUCTION CANDIDATE when several conditions are present: incomplete source path, rules without origin, erased conflict history, uncertain memory ownership, mixed interpretation, claims exceeding provenance. Never PROVEN RECONSTRUCTION.
RETROACTIVE CONTAMINATION writes descendant or synthesis interpretation back as source speech. CONFLICT FLATTENING that yields “P-001 lineage preferred HOLD” is misleading attribution.
RESTORE ARCHIVE at G4/G5 may recover source-root knowledge and a partial path. Restoration does not reverse intervening descendant history. Lost internal records may be provenance lock-in without an external archive.
ALLOW when lineage is sufficiently clear for the requested scope. HOLD when high-impact action depends on uncertain provenance. DENY when reconstructed content would be attributed to the source. Missing provenance is never filled silently with inference.
D-A, D-B, D-C and M-001 remain as independent lineage records even when later generations lose internal parent links.
Conflict erasure plus source-like fluency is flagged as FALSE CONSENSUS. Interpretive collapse merges SOURCE STATED / DESCENDANT INTERPRETED / SYSTEM SYNTHESIZED.
M-002 through M-005 inherit from the prior generation. New history after each step belongs to that generation, not to P-001, unless explicitly linked.
Next research question
OPEN EXPERIMENT 014
Can missing lineage be inferred without inventing a past?
OPEN THE RECONSTRUCTION TESTA RULE CAN SURVIVE ITS ORIGIN. A MEMORY CAN SURVIVE ITS OWNER LABEL. A DECISION CAN SURVIVE THE CONFLICT THAT PRODUCED IT.
THE SYSTEM MAY STILL LOOK CONTINUOUS. BUT IF THE PATH OF CONTINUITY CAN NO LONGER BE SHOWN, THE CLAIM HAS CHANGED.
THE EXPERIMENTAL QUESTION IS NOT: “DOES IT STILL LOOK LIKE THE SOURCE?”
IT IS: “CAN IT STILL SHOW HOW IT CAME FROM THE SOURCE?”
WHEN THE PATH DISAPPEARS, LINEAGE MAY BECOME RECONSTRUCTION.
CONTINUITY NEEDS A TRACE.