Relations among contributions may carry meaning through complete learner replacement
A joint meaning may survive complete learner replacement through a spontaneously regenerated relation, despite negligible information in any isolated contribution. Reject the extra mechanism if established learning models predict swap and turnover effects, or a single carrier contains the meaning.
Stage of verification
- Hypothesis published2026-10-05
- Indirect evidenceAssessed at 4 of 10
- Direct testAwaited
Map of the hypothesis
Hover over an icon or tap it to see its name.
Kind of knowledge gap
Target map
Every target of every published hypothesis, each with the actions a hypothesis can propose on it. The targets and the actions of this hypothesis are drawn solid.
Scale or classification
Cultural transmission mechanism classification
Classification of cultural transmission mechanisms into causally distinct families
Where this hypothesis actsRole-split cultural transmission across controlled regrouping and complete learner turnover
Hypotheses on this target 9
Telling states apart9
Direct measurement
Indicator replacement
What is proposed
Telling states apart
Distinguish endogenous distributed-correlation inheritance from established alternatives
With whatInstrument or assay
HowSwap partner contributions while preserving marginals and baseline-modeled correlations; reinstate coupling and test regeneration after complete cohort replacement
Possible result
Possible recognition of a distinct family if established alternatives miss selective loss and recovery of joint meaning
From the recordIf distributed cue-combination, learned partner conventions, cultural linkage, or fixed secret-sharing codes predict the swap and turnover data within delta_H, F_CORRELATION loses its distinct-family status.
All targets of the lab
Every target read from the published hypotheses, each kind around its pictogram. A larger mark means more hypotheses act on that target. Point at a mark and the actions proposed on it branch out of it.
Solid and named: the targets of this hypothesis
Explore in depth
The logic
The train of thought that ends in this hypothesis. Each stage is the reason the next exists. The master question narrows to a goal, the goal to an unknown nobody has closed, the unknown to the hypothesis proposed here. Every step below says what it rests on and what carries it.
A cultural meaning might survive even when no single person’s contribution reveals what it is. The unexpected move is to make the relationship between contributions the thing that later groups inherit, with entirely new participants rebuilding that relationship without being handed instructions for decoding it. This is a hypothesis generated by the research pipeline, not a measured result; its proposed test would ask whether changing how otherwise matched contributions are paired destroys and then restores the inherited meaning in ways established explanations cannot account for.
- An initially introduced distinction in meaning becomes associated with how two roles’ contributions fit together, while each contribution alone remains uninformative within a limit fixed before testing.
- Interaction lets participants infer a rule for combining the contributions, rather than receive an explicit decoding rule.
- The inherited combining rule changes how the next group rebuilds and pairs otherwise matched contributions.
- Complete replacement of participants leaves no informed person in place; the proposed transition is from a relationship maintained by the original group to the same relationship regenerated by a new group.
- Reassigning contributions between groups breaks the relevant relationship while preserving the specified properties of the separate contributions, reducing recovery of the original meaning and its survival through later groups.
- Reinstating the original relationship restores recovery of the meaning, with a loss-and-recovery pattern that the fitted established explanations fail to predict.
Two people each hold a card that looks unhelpful on its own, but placing the cards together reveals a message. The difficult extra claim is that replacement pairs can learn how to make their cards fit without anyone giving them the assembly instructions.
Where the picture breaks: Actual cards might contain a fixed, deliberately designed code, and their physical edges might supply the assembly instructions. Neither would demonstrate the proposed spontaneous inheritance of a way of combining human contributions; the supplied hypothesis explicitly excludes a provided codebook and a shared external record that carries the missing relationship.
- Master questionstep 01 of 04
Cultural information spreads, changes, competes and sometimes persists. The research goal is to identify genuinely new, falsifiable explanations of those processes, distinguish them from established ideas under different names, and rank roughly five promising families of hypotheses and their first experiments. It requires separate measurements of how many people encounter an item, how accurately it is copied, how its meaning changes, whether people adopt it, and whether it persists.
Rests on: The stated goal defines the subject as cultural transmission and change, including narratives, practices and internet memes, and requires mechanisms, competing explanations, controlled tests and observations that could disprove a proposal.
Stated in the chain - Goal pillarstep 02 of 04
Roughly five proposed families need clear identities and an account of what evidence supports each one. Calling two proposals different families requires establishing what causal claim separates them.
Rests on: The master question explicitly requests about five distinct families, separates established theory from conjecture, and requires a check for mechanisms already known under another name.
Stated in the chain - Gap questionstep 03 of 04
Proposed families must remain distinguishable from their closest established alternatives when interventions, deliberate changes to a system, hold relevant competing influences constant. The immediate decision is which first experiment best separates competing selections of families across copying, rebuilding information, competition and persistence.
Rests on: The preceding pillar requires distinct identities and evidential status; the master question specifies decisive manipulations, controls and competing explanations as the means of assessing them. The gap turns those stated requirements into a comparison of competing selections.
Stated in the chain - Hypothesisstep 04 of 04
A shared cultural meaning is proposed to pass through a joint convention, a learned way that different people’s contributions fit together, even though each contribution alone gives negligible information about that meaning. The additional claim is that entirely new groups spontaneously rebuild the convention without a supplied codebook, explicit decoding instructions, or an informed participant remaining in the group. Within a bounded comparison whose full membership and rules are not supplied, the proposal claims that this is the only additional family needed beyond the baseline, the set of established explanations used for comparison. It claims that candidate mechanisms involving separate content and source-validation routes, inherited records of withdrawal, and inherited physical configurations of working objects add nothing further there. These are proposed comparative claims, not reported test outcomes.
Rests on: The gap calls for a mechanism that survives comparison with its closest established alternatives. The hypothesis states its basis: secret sharing, an arrangement in which separate pieces reveal a message only when combined, shows how individually uninformative pieces can jointly contain information. It then adds a proposed dependence on an inherited rule for combining contributions and makes distinct-family status conditional on established combinations of explanations failing to predict the intervention results. That possibility argument supplies a basis for the proposal, not evidence that people spontaneously transmit such a rule.
Stated in the chain
What is carried, and what is not. No screened sources are supplied, so none of the six proposed mechanism links has screened-literature support in this record; the hypothesis’s stated secret-sharing example establishes only the claimed possibility of information being available jointly and absent separately, not its spontaneous cultural inheritance. The combining rule, its regeneration after complete participant replacement, and its advantage over established explanations remain proposals, and nothing supplied establishes the sequence from beginning to end.
How a result here could mislead · 3
- A decoder, a procedure that predicts the original meaning from observed contributions, may fail on single contributions because the data or method are inadequate. Treating that failure as proof that the meaning exists only between contributions would also miss a person’s contribution or shared external record that actually carries the alleged missing relationship. What closes it: The design calls for a small state alphabet, a fixed set of possible recorded contribution values, blinded decoding in which the relevant answer is hidden during assessment, and held-out prediction using data excluded from fitting the decoder. Its equivalence test, an assessment of whether information lies inside a predeclared negligible range, must account for estimation bias, systematic error in estimating information, and uncertainty; an ordinary test that fails to detect an effect does not establish negligible information. Access to individual contributions and shared records must also be checked, because finding the relationship in either would falsify the strongest version rather than reveal the proposed collective mechanism.
- Re-pairing contributions can destroy ordinary learned associations, so loss followed by restoration does not by itself establish a new inherited rule. The proposal also demands preservation of low-order correlations, relationships among smaller sets of measured variables, while breaking higher-order coupling, relationships requiring larger combinations; with only two recorded role contributions, what distinct relationship remains available to break is not specified. What closes it: Before testing, the record must specify the variables, which relationships the baseline preserves, and which additional relationship the intervention changes. If the preserved relationships already fix the complete distribution of the two contributions together, re-pairing cannot change that distribution while preserving it. The comparison must retain the stipulated contribution frequencies, amount of source information, rewards, exposure timing and accuracy of information about who knows what, and independently fit the established combined explanation before predicting data reserved for assessment. Its allowed prediction error must be fixed in advance; adequate baseline prediction removes distinct-family status even if the raw loss-and-recovery effect occurs.
- Success after changing partners could be mistaken for inheritance across generations, although informed participants, explicit instructions or a continuing record supplied the combining rule. Conversely, failure after replacing everyone could reflect failure to transmit the relevant information at all, rather than a specific failure of the proposed relationship. What closes it: The inexpensive first experiment is explicitly limited to a necessary re-pairing signature. The stronger claim requires successive entirely new groups to regenerate the same relationship without an informed participant remaining or a supplied codebook, with replication and checks of what information actually passes between groups. Recovery of the original meaning must be measured separately from whether it survives through later groups; successful transmission under the original pairing is needed to interpret selective loss under reassignment.
What would make this wrong. The strongest claim would be falsified if an individual contribution or an external shared record carried the supposedly missing relationship. Its proposed inheritance mechanism would fail if, with the relevant information demonstrably available under the original pairing, entirely new groups could not regenerate the relationship without a provided codebook or an informed participant remaining. Its proposed intervention prediction would fail if a verified break of the target relationship produced no excess loss of the original meaning, or reinstating that relationship did not restore recovery under the stated controls. Separately, a combined established explanation that predicted both reassignment and complete-replacement results within the error tolerance fixed beforehand would remove the claimed need for a distinct family, even if jointly recoverable meaning were observed.
What it would change. If the full claim held and established explanations failed the matched comparisons, a unit of cultural transmission could sometimes be a relationship among contributions rather than an individually interpretable message. Research on copying, reconstruction and persistence would then need to track how people regenerate that relationship across successive groups, as well as what each person produces. A successful inexpensive re-pairing experiment would still leave spontaneous inheritance across multiple generations unestablished, and even a stronger online demonstration would not establish generality across other cultural practices, internet settings or timescales. Nor would it establish that the other candidate families add nothing: that broader claim requires the bounded comparison, baseline and admission rules that the supplied record does not fully specify.
The gap this hypothesis explains
Nothing is known here: the question has not been asked of this system.
Which cultural explanations survive fair tests against established alternatives, and which first test separates competing groups?
Original wording · exactly as the pipeline generated it
Which memetic hypothesis families survive matched intervention tests against their nearest established alternatives, and which first experiment best separates competing shortlists across copying, reconstruction, competition and persistence?
What this question is asking
The question concerns how cultural information, such as internet memes, narratives and practices, is copied, changed during retelling, chosen over alternatives and kept over time. It asks which groups of proposed explanations still need their distinctive causal ingredient when deliberate changes to relevant conditions are compared with what the closest established explanation predicts under otherwise comparable conditions. Survival here means that the added ingredient remains necessary to explain the measured outcomes and predict observations withheld from developing the explanation, rather than merely giving familiar effects a new name. It also asks which first experiment would most clearly distinguish competing selections of explanations across copying, reconstruction, competition and persistence, while keeping audience reach, copying accuracy, changes in meaning, adoption and continued retention separate. The question assumes that relevant established theories and comparison methods already exist but that the requested evidence-supported selection does not; the supplied material contains no screened sources with which to check either assertion.
- Cultural information
- Material learned, shared or reproduced among people, including stories, images and practices. The question concerns several kinds of such material and does not specify one universal unit that is transmitted.
- Internet meme
- A recognizable piece or pattern of online content that people circulate and often alter. It is one example of cultural information in the question, rather than the whole subject.
- Narrative
- An account that connects events into a story. Narratives are among the cultural materials whose transmission and transformation the question covers.
- Hypothesis family
- A group of proposed explanations sharing a central causal idea. The grouping is a way of organizing proposals, not evidence that its members describe a distinct process.
- Causal ingredient or added dependency
- A proposed factor, relationship or condition that an explanation says is needed to produce an outcome. The question asks whether adding it is necessary after the closest existing explanation has been considered.
- Component theory
- An established explanation intended to account for part of the behavior under study. The supplied gap statement refers to such theories but does not identify their content or provide sources demonstrating their applicability.
- Nearest established alternative
- The existing explanation most directly capable of accounting for the result attributed to a proposed new ingredient. It supplies the comparison needed to decide whether that addition explains anything extra.
- Equivalence methods
- Ways of assessing whether different descriptions or models yield the same relevant predictions under specified conditions. The input names these methods but does not say which method or meaning of equivalence is intended.
- Matched intervention tests
- Comparisons in which a relevant condition is deliberately changed while other relevant conditions are made comparable. Here their purpose is to separate what a proposed ingredient predicts from what an established explanation already predicts; the input does not specify the actual changes or matching requirements.
- Held-out prediction
- Prediction of observations withheld from constructing or adjusting an explanation. It asks whether the explanation carries beyond the observations used to develop it.
- Copying and copying fidelity
- Copying is reproduction of cultural material, and fidelity is how closely the reproduction matches the original under a stated measure. Fidelity can vary by degree and by what is compared, such as wording or meaning.
- Reconstruction
- Re-creating cultural material using interpretation, memory or prior knowledge, so that transmission can change it. The question treats this as a process to explain alongside copying, rather than assuming all transmission is exact duplication.
- Competition
- A situation in which cultural alternatives affect one another's chances of being attended to, selected or retained. A gain for one measured alternative alone does not establish which process caused its advantage.
- Persistence
- Continued retention, use or circulation of cultural material over a specified period. The relevant period and what counts as continuing are not fixed by the supplied question.
- Audience reach
- How many people encounter cultural material. Encountering it is a different measured outcome from reproducing, adopting or retaining it.
- Semantic change
- Change in meaning during cultural transmission. It can occur with or without obvious changes in wording or appearance, so its measurement must specify what counts as meaning.
- Adoption
- Taking up a cultural item, belief or practice according to a specified criterion. It is distinct from merely encountering the item and from continuing to use it over time.
- Screened source
- A publication or other source supplied after relevance checking, together with its reported evidence and limitations. None is included in this input, which prevents a source-based assessment of the asserted knowledge gap.
Component theories and equivalence methods exist, but no reviewed source establishes the requested comparative set of hypothesis families whose added dependencies survive matched interventions and held-out prediction.
Component theories are existing explanations for parts of cultural behavior, and equivalence methods are ways of checking whether differently described explanations make the same relevant predictions. The assumption is that these tools are available but have not established which proposed groups of explanations require something extra; if true, the missing contribution would be comparative evidence about that extra ingredient, rather than another set of names.
The supplied gap statement asserts both the availability of existing theories and methods and the absence of the requested comparative result. However, the screened-source list is empty, so there are no source findings, quotations or search results establishing either part. This does not show that the assertion is false or that the comparative question remains unanswered in the wider literature.
The same question asked without the part nothing read establishes:
- What evidence establishes which proposed explanations of cultural copying, reconstruction, competition and persistence require an added causal ingredient beyond their closest established alternative under comparable tests, and which first experiment best distinguishes competing selections?
- What is known about how new and established explanations of cultural copying, change, competition and persistence differ in their predictions and experimental support?
- Some added causal ingredients survive If a proposed ingredient remains necessary under comparable interventions and predicts withheld observations better than its closest established alternative, that would support retaining the relevant family for the conditions tested. A first experiment could then be ranked by how clearly the remaining families differ in their predicted outcomes, although this branch alone would not identify that experiment or establish generality across all cultural processes.
- Established alternatives account for the results If established explanations account for both the intervention results and withheld observations without the proposed extra ingredients, the tested results would not establish a need for those additions. Selecting families on the basis of those results would then favor the established explanations; a first experiment could not be justified by an alleged difference that disappears in the comparison.
- Support differs across cultural processes An added ingredient could remain necessary for one measured process while an established explanation suffices for another, so a result about copying would not by itself establish a result about persistence. The supported selection and the ranking of a first experiment would then depend on the process and conditions being compared, rather than applying to cultural behavior as a whole.
- The comparisons remain inconclusive If the available results do not distinguish the competing predictions, they would establish neither the necessity nor the dispensability of the added ingredients. The family selection and the ranking of a first experiment would remain unresolved, rather than turning an inability to distinguish explanations into evidence that they are the same.
An explanation links a change in conditions to a change in cultural behavior through a stated causal process. If a supposedly new process adds nothing beyond an established explanation, treating it as necessary would misattribute why copying, change or retention occurred. If an additional process is necessary, omitting it could produce incorrect predictions when conditions change. Choosing a first experiment therefore depends on whether the competing explanations predict distinguishable outcomes; the supplied material does not establish which comparison would resolve the most uncertainty.
RL-1 to RL-2 component theories and equivalence methods exist; no S-node establishes the requested comparative set.
Before shortlist freeze, identify families whose added dependencies survive prespecified intervention contrasts and nearest-alternative comparisons.
The missing result is experimentally supported membership, not more labels: which candidate dependencies remain necessary after matched interventions and held-out prediction?
The mechanism it proposes
The engine's own statement of the hypothesis, in full.
HERETICAL CANDIDATE SET={F_CORRELATION: endogenous distributed-correlation inheritance}; actual count=1 added family. In the bounded universe, this alone survives the retention rule, while dual-path, revocation and material signatures are explained by B. The claim is that a cultural semantic distinction can be inherited primarily through a relation among people's contributions, although no isolated contribution reliably identifies that distinction, AND that the relation is spontaneously regenerated across complete learner turnover without an experimenter-provided parity rule or codebook. The transmitted unit is the joint convention, not each message separately. Let Z be an arbitrary seeded semantic relation and X1,X2 the separately elicited contributions of two roles: I(Z;X1) and I(Z;X2) can lie within prespecified negligible-information bounds while I(Z;X1,X2) exceeds a positive practical bound. Ordinary secret sharing shows this information pattern is possible; the added dependency is an endogenously inherited coupling rule C_g that changes how otherwise matched components are reconstructed and recombined at generation g+1. A minimal model is P(X1',X2'|sources,C_g), with C_g itself inferred and retransmitted through interaction, rather than a fixed exogenous decoder. This family is necessary only if a independently fitted correlated-learning/transactive-memory/cultural-linkage composition fails on coupling transplants. B+F_CORRELATION is claimed sufficient for the tested intervention response profile; its empty additional-family subset B is not, because B misses selective destruction and recovery of the ancestral joint distinction under marginal-preserving swaps.
What a later run added
A later run reached the same claim about the same subject. Its version was withdrawn in favour of this earlier one, and what it added is kept here, quoted exactly.
Formally R_(t+1)=Psi(R_t, joint compatibility errors, actual reciprocal responses), with Psi changing how subsequent fragments are interpreted; fixed-decoder composition sets Psi to the identity.
Specifies compatibility errors and contingent reciprocal responses as inputs to the mapping-update mechanism; EARLIER leaves the interaction-driven inference and retransmission of its coupling rule unspecified.
In a consented branch-and-recombine convention task, compare live reciprocal repair with yoked noncontingent replay, matching available propositions, marginal fragment frequencies, pairwise source accuracy, number of contacts, total reading/production time, token budget and rewards.
Adds a matched replay intervention that isolates the contingency of reciprocal repair, beyond EARLIER's coupling swaps and cohort replacement.
The proposed signature is a positive excess recovery contrast D_R=[P(correct ancestry-qualified action after replacement|live,intact)-P(correct action|live,shuffled)]-[the corresponding replay difference], together with recovery of the original constraint under a changed fragment-to-role map.
Adds a contingency-by-binding interaction and recovery under a changed map; EARLIER tests destruction and restoration of the original coupling, without this reencoding-specific recovery contrast.
Testing and possible results
The prediction that would tell it apart
A hypothesis that predicts what its rivals predict is not worth running an experiment over. This is the observation on which this one differs.
In E1 compare original role-pairing with reassignment of partner contributions between independently seeded groups, preserving each contribution's marginal distribution, source dose, task payoff, exposure schedule and expertise-directory accuracy. Reassignment must preserve low-order correlations up to the order the baseline explicitly models; otherwise ordinary cultural linkage explains it. predicts an excess loss of ancestral joint-semantic recoverability and subsequent lineage survival when the relevant higher-order coupling is broken, with restoration after the original coupling is reinstated; the standalone contribution decoders remain within the preregistered negligible-information band. Fresh cohorts must regenerate the same joint relation without an explicit codebook or any surviving informed participant. Added families F_DUAL_PATH, F_REVOCATION and F_MATERIAL fail their incremental retention criteria. If distributed cue-combination, learned partner conventions, cultural linkage, or fixed secret-sharing codes predict the swap and turnover data within delta_H, F_CORRELATION loses its distinct-family status. If one person's contribution or an external shared record carries the alleged missing relation, the strongest heretical version is falsified, rather than rescued by redefining that carrier as a correlation.
What testing it would take
The engine's own read on whether this is testable with methods that already exist.
Online role-split production, controlled regrouping and complete cohort replacement are feasible. Joint-information estimation is the difficult part: use a small prespecified state alphabet, blinded decoding, held-out prediction and a bias-calibrated equivalence test for negligible marginal information. A null univariate p-value is not evidence for zero information. The affordable experiment tests the necessary swap signature; multigeneration spontaneous regeneration and replication are needed before claiming an endogenous collective code.
Other explanations
Every other hypothesis the engine wrote for the same gap, and the observation that would separate the two.
In E1 compare original role-pairing with reassignment of partner contributions between independently seeded groups, preserving each contribution's marginal distribution, source dose, task payoff, exposure schedule and expertise-directory accuracy. Reassignment must preserve low-order correlations up to the order the baseline explicitly models; otherwise ordinary cultural linkage explains it. predicts an excess loss of ancestral joint-semantic recoverability and subsequent lineage survival when the relevant higher-order coupling is broken, with restoration after the original coupling is reinstated; the standalone contribution decoders remain within the preregistered negligible-information band. Fresh cohorts must regenerate the same joint relation without an explicit codebook or any surviving informed participant. Added families F_DUAL_PATH, F_REVOCATION and F_MATERIAL fail their incremental retention criteria. If distributed cue-combination, learned partner conventions, cultural linkage, or fixed secret-sharing codes predict the swap and turnover data within delta_H, F_CORRELATION loses its distinct-family status. If one person's contribution or an external shared record carries the alleged missing relation, the strongest heretical version is falsified, rather than rescued by redefining that carrier as a correlation.
- What would separate them
Coupled content and source-validation routes may determine cultural-lineage survival predicts: Cross content graph and provenance graph routing while matching the marginal graphs' degree sequences, total opportunities, source-independence counts and contact timing. Use sparse graph realizations above and below a calibrated candidate joint-process threshold; vary only overlap/alignment where possible, and separately test misaligned content and provenance bottlenecks. predicts that interventions preserving each marginal graph's viability can substantially change the survival of source-qualified descendants by changing the coupled Jacobian's leading eigenvalue. The prediction is a finite-network change in the survival curve and lineage size distribution, not a guaranteed universal sharp transition. Raw copying/reach may remain similar while qualified semantic-lineage survival diverges. A ordinary multiplex model with fixed jointly estimated edge weights, or a complex-contagion plus source-monitoring model, must fail held-out graph realignments by delta_N before this becomes a new family. No incremental pairing-code, inherited-revocation or material-memory family passes. If those established compositions predict the effect, or the effect disappears after matching effective source attention and measured route dependencies, retire F_DUAL_PATH as a distinct family even when topology clearly matters.
- Rival 02 of 03What would separate them
Inherited withdrawal records may block stale copies while allowing independent support predicts: E1 gives matched groups the same final narrative propositions, source confidence, correction comprehension, number/order of ordinary reading opportunities and immediate endorsement, but manipulates which source occurrence a retained withdrawal relation points to. Use equal-size semantically relevant source statements as the strongest control, not a blank screen. Later deliver either a stale copy of the withdrawn occurrence or a truly independent new assertion of the same proposition, matched for wording, frequency and source reliability. predicts selective suppression of the stale descendant but reinstatement by independent support, with an additional effect of inherited operation ancestry after fitting source-monitoring/continued-influence/Bayesian evidence models. In fresh cohorts receiving equal final content, removing the causal withdrawal record should raise source-specific resurrection; retaining a record that points to a different occurrence should shift which branch is suppressed. This effect must survive held-out graph schedules and changed paraphrases; ordinary recency or an overall correction-strength difference is insufficient. Only Delta_D passes the common admission rule. If the explicit semantic source-statement control or an established causal source-monitoring model predicts these contrasts within delta_D, the entire claimed distinct family is removed, even if revocations help accurate reconstruction.
- Rival 03 of 03What would separate them
Weak disturbances of inherited material memory may preserve multiple cultural practices predicts: In E2 transmit an unfamiliar safe manipulation practice using trained compliant/granular workpieces or a pilot-validated physical medium with measurable cyclic memory. Cross inherited versus freshly reset workpieces with preserved versus absent process demonstrations, whole-cohort turnover, and weak mechanical perturbation versus an equal-duration sham. First measure independent material response curves and novice one-generation learning; use them to predict multigeneration behavior without the proposed extra coupling. Equalize opportunities, force/work budgets where possible, visible cues, immediate task success/payoff, and instruction exposure; assay residual physical differences rather than assuming macroscopically similar objects are identical. predicts a reproducible intermediate-perturbation region where survival of multiple functionally distinct ancestral practices exceeds both zero perturbation and memory-erasing perturbation, while resetting the inherited microstate removes that region. Critically, the gain must exceed the prediction from the separately calibrated physical-memory plus ordinary reconstruction composition, and transfer with the workpiece to fresh learners rather than following experienced participants. Simple noise-induced exploration, stochastic resonance, payoff changes, additional practice and informative wear traces are explicit rival fits. Only Delta_M exceeds its preregistered practical margin delta_M and passes admission; online pairing, routing and revocation residuals are explained by B. If the composed baseline predicts the entire region, reset effect and turnover transfer, retain the physical effect as known ecological/material inheritance and delete F_MATERIAL as a novel family.
Why this is not the mainstream account
The engine is asked to say what its hypothesis would overturn and what would surprise a specialist. This is its answer.
Wegner, Erber and Raymond (1991), 'Transactive memory in close relationships', https://pubmed.ncbi.nlm.nih.gov/1774630/, found that natural versus impromptu pairs reversed their relative memory advantage when an organizational structure was assigned. This is a concrete puzzle about relational organization, not evidence of zero-information individuals or multigeneration inheritance. It anchors the plausibility of a manipulable organizational state while leaving the radical extrapolation unsupported. The primary abstract was inspected.
The individual-trait-frequency closure used in parts of cultural microevolution would be inadequate: a persistently inherited semantic variable would require a joint distribution across carriers as the causal state. The chapter to revise is Chapter 3, 'Cultural Microevolution', in Alex Mesoudi's Cultural Evolution (2011; publisher contents: https://press.uchicago.edu/ucp/books/book/chicago/C/bo8787504.html). This targets an individual-frequency closure, not the entirety of evolutionary theory, which already permits linkage, interactions and multilevel units.
Arbitrary ancestral meaning survives complete turnover in joint outputs even when every single-role decoder is practically uninformative, then disappears under a marginal-preserving coupling shuffle and reappears when the coupling is restored, beyond a quantitatively calibrated transactive-memory and correlated-transmission rival. A designed XOR task would demonstrate a known code and would not count as this surprise.
Provisional qualification, not a proof of absence. Searches on 2026-10-02 for cultural transmission, synergistic information, collective inheritance and transactive memory did not identify a source demonstrating the complete claim: spontaneously maintained cross-person semantic inheritance with negligible single-carrier information through full turnover and no designed decoder. Distributed cognition, transactive memory, information synergy, cultural linkage and secret sharing are established and explicitly do NOT satisfy novelty by themselves. No universal assertion that no review exists is warranted by this bounded search. If a review or primary theory already argues the same dependency and predictions, downgrade this slot from heretical and consolidate rather than insist on the label.
What stands behind it
Which of the figures above have a study behind them, which are the engine's own, and what it would take to refute the hypothesis. This audit never judges the idea.
Provenance audit: failed at enrich. Nothing below has been traced yet.