Cultural transmission may require six distinct pathways that cannot be merged
In text or rule learning, six pathways may each be necessary because they respond differently to targeted removal and rescue. The exact count is rejected if a merged model predicts those responses, including combinations and timing changes.
Stage of verification
- Hypothesis published2026-10-05
- Not enough research data
- 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 actsHeld-out cultural transmission experiments with localized removals and selective rescue
Hypotheses on this target 9
Telling states apart9
Direct measurement
Indicator replacement
What is proposed
Telling states apart
Distinguish independently necessary mechanism families and merge equivalent mechanisms
With whatInstrument or assay
HowUse timed removals, selective rescue and intermediate-state assays; test whether reduced nonlinear kernels predict held-out isolated and combined interventions
Possible result
Expected six necessary families, with fewer retained if merged models predict the same intervention effects
From the recordeach of these six dependency classes must have a transferable, independently necessary intervention signature.
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.
Stories and shared rules change as people remember them, pass them on and choose which versions receive another hearing. The unexpected move is to borrow a question from chemical reaction engineering: when does combining several hidden steps into one process destroy the ability to predict what happens after a specific disruption? This pipeline proposes that six processes must remain separate in cultural transmission, meaning the passage and alteration of information between people; it has not measured that minimum.
- Context present before a source is learned shapes how that source is organized in memory.
- Later cues and delay act on the stored information to change what can be accessed and updated.
- Other sources act on the accessible information to create relationships across their component facts.
- The intended recipient’s knowledge shapes how those relationships are expressed in the next version.
- Choice acts on the available versions to determine which unchanged version is selected.
- The exposure system’s retained history changes later delivery opportunities, which determine the sources encountered next.
A kitchen can have separate steps for preparing ingredients, cooking and serving, but naming three steps does not prove that all three need separate explanations. The separation earns its keep only if changing one step produces consequences that a combined description cannot predict.
Where the picture breaks: The six proposed cultural processes need not occupy separate physical stations, and a single person or system may perform several at once. The kitchen picture supplies no evidence that six is the minimum or that any particular experimental change affects only its intended process.
- Master questionstep 01 of 04
New explanations of how cultural information spreads and changes must earn their place through tests that separate them from established alternatives. The goal calls for roughly five promising families of explanations, with clear definitions of what passes between people, competing predictions, affordable initial tests and stronger follow-up evidence.
Rests on: The stated goal is a research agenda for the transmission, transformation, competition and persistence of cultural information. It explicitly requires a distinction between established findings and new proposals, including checks that apparently new mechanisms are not already known under another name.
Stated in the chain - Goal pillarstep 02 of 04
The shortlisted explanations must be distinct from one another, and the evidence behind each must be identified. Approximately five families is the organizing target.
Rests on: The master question explicitly asks for approximately five hypothesis families, meaning groups of explanations sharing a proposed causal dependency, and for an assessment of novelty and existing evidence.
Stated in the chain - Gap questionstep 03 of 04
The number of necessary explanations is to be determined by merging equivalent mechanisms and asking which remaining dependencies survive experimental removal. Their predictions must work on held-out experiments, meaning experiments whose outcomes were not used to build or adjust the explanation.
Rests on: The preceding pillar requires distinct families with an identified evidential status; the master question requires discriminating experiments and competing explanations. This stage makes predictive necessity after merging and removal the criterion for counting them, rather than treating approximately five as a required result.
Stated in the chain - Hypothesisstep 04 of 04
Passing information onward may require six independently necessary processes: organizing a source when it is learned; accessing and updating what was learned; combining relationships across sources; shaping an expression for its recipient; choosing among unchanged existing versions; and adapting who encounters what later. The proposal predicts that no explanation with five merged processes can reproduce all six patterns of disruption and restoration across new tests. It also predicts that effects of the surrounding competing versions and of overlapping changes to shared meanings can be explained by combinations of these six processes.S1S2S3S4S5S6S7S8S9S10
Rests on: The gap question supplies the demand for a minimum number justified by intervention, meaning a deliberate experimental change to a proposed cause. The endpoint supplies its own basis by importing kinetic lumping, the replacement of several reaction pathways by a combined description, and proposing an analogous failure of combination when separately targeted changes produce distinct effects. This is a stated modeling proposal, not evidence that cultural processes behave like chemical reactions. The supplied literature bears on some component activities. S1, Frontiers in psychology (2020), reports that changing object-background pairings harmed recognition more for unfamiliar than familiarized objects; it does not independently separate organization during learning from later access. S3, Neuropsychologia (2016; abstract only), reports that meaningful scene-object pairings were associated with changes in brain activity during initial learning and faster responses on repetition; the same scenes appeared at both presentations, so the result does not separate initial organization from later cues. S4, Current Biology (2021), reports stronger similarity of memory-related brain activity for coherent events and preferential return of those activity patterns during recall a day later; it does not show that a new relationship changes while recall of its component facts stays fixed. S5, Hippocampus (2017; abstract only), reports similar direct-association performance but different indirect-inference patterns between patients with damage in memory-related brain regions and controls across repeated learning and recall; it does not selectively isolate the proposed combining process. S6, Cognition (2020; abstract only), reports that control over study improved both memory for studied relationships and deductions connecting items never studied together, with benefits limited to participants with higher capacity to hold information temporarily; simultaneous improvement in both outcomes leaves the proposed combining process unseparated from learning and exposure. S7, Brain (2020), reports impaired indirect inference in people with schizophrenia, a psychiatric condition, but does not establish that recall of the component facts was held constant or that the proposed brain-activity relationship caused the impairment. The remaining sources provide background rather than tests of the six-process claim. S2, Psychophysiology (2024), discusses limits on identifying the units people hold in short-term memory and effects of previous trials; it does not establish separate transmission processes. S8, Biological Reviews of the Cambridge Philosophical Society (1999; abstract only), reviews animal imitation and suggests that copying serving different functions might involve different mechanisms; it neither establishes the six proposed processes nor their transfer to human cultural transmission. S9, Evolution (2017; abstract only), describes a simulation of copying mate choices and the spread of an inherited tendency to copy; this is a model, not a test of the proposed human transmission sequence. S10, Human Molecular Genetics (2019; abstract only), reports recognition and socially acquired food-preference deficits in mice with a changed Kctd13 gene; these broad deficits do not distinguish the six processes or establish transfer to human cultural transmission. None of these sources tests the full proposed minimum.
Stated in the chain
What is carried, and what is not. Six screened sources—S1 and S3 through S7—speak to activities associated with three of the six proposed links: organization during learning, later access and combining relationships; they do not establish that those three are independently necessary, while S2 and S8–S10 provide background with the limits described above. No supplied source establishes the complete sequence, the independent necessity of audience shaping, unchanged-version choice or adaptive exposure, or the claimed minimum of six.S1S3S7S2S8S10
How a result here could mislead · 3
- Six different disruption effects could be mistaken for six independent processes when the changes also alter information, recall or exposure through neighboring routes. Moving context from before to after learning may change later access; scrambling relationships may impair the component facts; resetting the delivery system may change exposure history as well as its retained state. What closes it: The proposed comparisons require the same information at matched times, separately checked learning and access, unchanged recall of component facts for the relationship test, stable private recognition for the audience test, unchanged artifacts and matched exposure for choice, and exact exposure-history replay for the delivery-state test. Intermediate-state assays, meaning separate measurements of the states between source and output, must verify that each targeted change actually occurred; selective rescue, meaning restoration of the targeted function, must restore the corresponding effect. The supplied specification requires these checks but does not provide their full measurement procedures.
- Different scores for the same underlying change could look like extra mechanisms. The two-process rival specifically predicts that differences in how meanings, sentence roles, source origins or unit boundaries are coded can manufacture the apparent separation. What closes it: The relation between true changes and recorded scores must be checked separately in each condition using blinded challenge material with known meaning and origin, as the rival specifies. The six-process outline does not supply this coding-validation procedure. A successful six-way pattern must survive correction for those measured scoring differences, with the transmitted unit defined consistently.
- Failure of an overly simple smaller model could be credited to six necessary processes, or six fitted components could be counted as six mechanisms. Neither result shows that the strongest allowed smaller explanation fails, and a successful component test alone does not establish the proposed sequence. What closes it: The actual nonlinear probability models, meaning models that allow interactions and predict distributions of outcomes rather than only average changes, must make predictions before combined and reordered interventions are observed. The rival comparisons must include one shared rule for reconstruction and choice; reconstruction plus choice with validated scoring; an added effect of the current competing versions on reconstruction; and separate effects of partner-specific meanings and overlapping updates to a shared record. The primary outcome and tolerated prediction error for each proposed separation must be fixed in advance, alongside the removal and restoration checks; the supplied proposal gives no numerical error margins. A smaller model that reproduces all relevant new-test predictions counts as a successful merger.
What would make this wrong. The exact six-process claim fails if a properly calibrated smaller model predicts the relevant isolated, combined and reordered interventions and their restorations within the prespecified error margins. The proposal explicitly allows learning organization and later access to merge if one stored-source model predicts both, relationship combination to merge if ordinary reconstruction from complementary facts explains it, and adaptive delivery to merge with selection if logged exposure fully accounts for it. A remaining independent effect of competing-version context or overlapping shared-record updates after the six processes are included would also contradict the claim that six suffice. A null effect after an intervention that failed to change its intended intermediate state would not decide the claim.
What it would change. If all six processes proved independently necessary under the specified comparisons, the master question would have a defensible six-part candidate explanation for the tested forms of cultural transmission, even though its initial organizing target was approximately five families. Work on that question would have to compare interventions on intermediate processes and adaptive delivery history, rather than count labels or rely on overall copying success. This would still not establish that the six parts are individually new discoveries, that they are the universal minimum for culture, or that laboratory text results explain everyday practices and internet transmission. The proposed practice task, second language, independently built delivery policies and follow-up after exposure stops would still be needed for the corresponding broader and persistence claims.
Sources read · 10
Behavioral and Neural Effects of Familiarization on Object-Background Associations. · Frontiers in psychology · 2020
“the decrement in object recognition performance between intact and rearranged test pairs was significantly greater for non-familiarized objects than for familiarized objects”
Does not settle: The reported familiarization-by-test-context interaction bears on encoding and context-dependent recognition, but does not establish independently necessary E and U pathways or rule out a reduced composed kernel. The text discusses a resource-limited encoding alternative and calls hippocampal activity an ambiguous indicator of memory strength. It does not test cultural transmission, the I, A, S or F dependency classes, timed pathway removals and rescues, order changes, transferable intervention signatures, or a minimum K=6.
Short-Term Memory Impairment · Psychophysiology · 2024
“Although that may well be the case, the issue of capacity limits cannot be solved until the actual chunks that people form and use can be consistently identified across many situations.”
Does not settle: This excerpt discusses working-memory capacity, chunk identification, task-dependent limits and trial-history effects in absolute judgment. It does not establish six independently necessary cultural-transmission pathways, distinguish encoding from retrieval through localized interventions, or test cross-source integration, audience-conditioned expression, fixed-artifact selection or adaptive exposure allocation. It supplies no reduced-kernel comparison, removal/rescue experiment or transferable intervention signatures establishing K=6.
Semantic congruence affects hippocampal response to repetition of visual associations. · Neuropsychologia · 2016
“Encoding the targets with congruent contexts was associated with increased activation in visual cortical regions at initial presentation and faster response time at repetition, but we did not find enhanced activation in mPFC relative to incongruent stimuli at either presentation.”
Does not settle: The abstract supports context-dependent encoding and subsequent recognition effects in a short-lag scene-object task. It does not separate encoding organization from retrieval access: identical scenes accompanied both presentations, and retrieval cues were not independently manipulated with encoding held constant. It does not establish cultural transmission, independently necessary E and U kernels, integration, audience expression, selection or exposure allocation, or a minimum of six irreducible pathways under localized removal, rescue and order-changing interventions.
The hippocampus constructs narrative memories across distant events. · Current biology : CB · 2021
“Thus, overlap across events, while potentially necessary, was not sufficient to integrate information across distant events.”
Does not settle: The supplied excerpt reports greater hippocampal pattern similarity during encoding and preferential reinstatement during recall one day later for coherent versus unrelated events sharing characters. This bears on narrative organization and integration across distant events, but does not establish six independently necessary cultural-transmission pathways or a minimum K=6. It does not show a novel relation changing with premise recall held constant, distinguish integration causally from encoding or retrieval, or test localized removals, rescues, order changes or reduced kernels. It leaves audience-conditioned expression, fixed-artifact selection, adaptive exposure allocation, cultural descendants and longer-term transfer untested. The text explicitly leaves open whether boundary-related memory effects reflect one common process or multiple processes.
Memory integration in humans with hippocampal lesions. · Hippocampus · 2017
“In direct trials, performance of patients and controls was similar and stable across cycles. By contrast, in indirect trials, patients and controls showed distinct patterns of behavior.”
Does not settle: The abstract supports a narrower distinction between memory for learned overlapping associations and inferential integration across repeated encoding/retrieval cycles in humans with mediotemporal lesions. It does not establish integration as independently necessary under selective intervention with premise recall experimentally held constant, distinguish encoding from retrieval contributions, or test cultural transmission, transferable intervention signatures, removal/rescue responses, reduced kernels, or the proposed six-pathway minimum.
Active transitive inference: When learner control facilitates integrative encoding. · Cognition · 2020
“Active control improved memory for studied premises as well as transitive inferences involving items that were never experienced together during study.”
Does not settle: The abstract supports active learner control affecting premise memory and novel relational inference, with benefits limited to participants with higher working memory capacity. It does not isolate integration from encoding, retrieval, or exposure allocation: premise memory also improved, and overlapping-premise search is only suggested as facilitating integration. It does not establish inference changes with premise recall held constant, localized removal and rescue signatures, transferable independent necessity of six pathways, or failure of a reduced composed kernel. Audience-conditioned expression and fixed-artifact selection are not tested in the supplied abstract.
Impaired theta phase coupling underlies frontotemporal dysconnectivity in schizophrenia. · Brain : a journal of neurology · 2020
“The schizophrenia group showed a striking impairment (Hedges' g = 1.2) in indirect inference in the memory integration task, which may depend on hippocampal-mPFC theta coupling ( ), although we could not establish such a relationship across tasks.”
Does not settle: The excerpt reports impaired indirect inference in schizophrenia, but does not establish integration impairment with premise recall held constant, a causal theta-coupling mechanism, or transfer to cultural transmission. It does not test independently necessary intervention signatures for E, U, I, A, S and F, localized removal and rescue, order effects, or whether reduced composed kernels can reproduce the same predictions; K=6 remains unestablished.
The ethological analysis of imitation. · Biological reviews of the Cambridge Philosophical Society · 1999
“The potential importance of reinforcement raises the possibility that copying abilities serving divergent functions might be partly under the control of different mechanisms.”
Does not settle: This abstract reviews animal imitation and proposes possible mechanistic differences; it does not establish six independently necessary cultural-transmission pathways. It provides no localized-removal or rescue signatures distinguishing E, U, I, A, S and F, no matched intermediate-state measurements, and no tests of reduced kernels under isolated, combined or reordered interventions. Transfer to human cultural transmission and the proposed minimum K=6 remain unestablished.
Mate-choice copying: A fitness-enhancing behavior that evolves by indirect selection. · Evolution; international journal of organic evolution · 2017
“A spatially explicit, individual-based simulation model is used to study the spread of an allele for mate-choice copying (MCC) through horizontal cultural transmission when female innate preferences do or do not coevolve with a male viability-increasing trait.”
Does not settle: The abstract describes a simulation of mate-choice copying and indirect genetic selection. It does not establish six independently necessary cultural dependency classes, distinguish the proposed intermediate representations, or test localized removals, rescues, order changes, or reduced kernels. It does not establish transfer of its model results to the proposed cultural transmission system.
Kctd13-deficient mice display short-term memory impairment and sex-dependent genetic interactions. · Human molecular genetics · 2019
“Behavioral testing revealed a significant deficit in novel object recognition, novel location recognition and social transmission of food preference in Kctd13 mutants.”
Does not settle: The abstract reports behavioral deficits in Kctd13-mutant mice, but does not distinguish encoding, retrieval, relational integration, audience-conditioned expression, variant selection or future-exposure allocation. It does not test localized removal and rescue, timed intervention order, reduced-kernel predictions or transferability, and therefore does not establish six independently necessary pathways or a minimum K=6 for cultural transmission.
The gap this hypothesis explains
Nothing is known here: the question has not been asked of this system.
How many explanations predict new cultural transmission experiments after equivalent explanations merge and each claimed causal link is removed?
Original wording · exactly as the pipeline generated it
How many causally distinct hypothesis families remain necessary to predict held-out cultural transmission experiments once equivalent mechanisms are merged and each retained dependency is experimentally removed?
What this question is asking
The question concerns how many genuinely different explanations are needed for the way stories, images and practices spread, change, compete and last. It asks for a count after explanations that make the same relevant predictions under experimental changes have been grouped together, and after each remaining explanation has been challenged by removing a relationship it says produces an effect. The remaining groups must help predict results from experiments that were kept separate when the explanations were chosen; the comparison is whether a smaller grouping predicts those results adequately or whether distinct groups are still needed. The supplied gap description treats approximately five groups as a possibility to assess, while assuming that the named existing methods do not already establish the count. It asks for the count when the shortlist is finalized, including an explicit range or unresolved grouping if the evidence cannot distinguish a single answer.
- Cultural transmission
- The passing of learned information or practices between people, including stories, images and customs. Here the question also concerns changes to that material, competition among alternatives and how long material remains in use.
- Cultural item or transmitted unit
- The thing treated as being passed along, such as a particular image, a story version or a practice. Its boundaries are a choice made for measurement rather than necessarily a naturally separate object; changing those boundaries can change what is counted as copying or change.
- Hypothesis family or group of explanations
- A collection of proposed explanations treated as sharing the relevant causal account. Here a family is a grouping to be assessed, not a category whose independence is established merely by giving it a name.
- Causal link, causal relationship or dependency
- A relationship in which one feature helps produce another, rather than merely appearing alongside it. Removing a claimed link means changing the conditions so that this proposed contribution cannot operate, then assessing what follows.
- Mechanism
- The sequence of processes through which a proposed cause produces an outcome. Different descriptions of a mechanism do not automatically establish different causes.
- Experimental equivalence
- Treating explanations as equivalent because the relevant experimental comparisons do not distinguish their predictions. Such equivalence concerns the assessed changes and outcomes; it need not mean that the explanations are identical in every possible setting.
- Held-out or reserved experiments
- Experiments whose results are kept separate from the process used to select or arrange explanations. In this question, predicting their results assesses whether the proposed grouping works beyond the evidence used to construct it.
- Irreducible family count
- The number of groups of explanations that cannot be further combined or discarded while retaining the predictive performance required for the assessed comparisons. The term expresses a requirement relative to those comparisons and the rule for acceptable prediction, not proof of an absolute number for all culture.
- Unit-sensitivity methods
- Methods for checking whether conclusions change when the cultural item being counted or followed is defined differently. The input names these methods but gives no procedure or findings.
- RL-2
- An unexplained label in the supplied gap description, associated there with equivalence and unit-sensitivity methods. No expansion, definition or supporting source is supplied, so its meaning cannot be established more precisely.
- Shortlist freeze
- The point when the selected set of explanations is finalized for the report. The requested family count is to describe the evidence available at that point.
- Partition uncertainty
- Uncertainty about how proposed explanations should be divided into groups. Different defensible groupings may imply different counts, and uncertainty can concern group membership even when the counts agree.
- Consolidation and subdivision
- Consolidation combines proposed groups into fewer groups; subdivision splits a proposed group into more groups. Here those changes depend on whether experimental comparisons justify treating the explanations together or separately.
- Prediction and predictive performance
- A prediction is a stated expectation about a result; predictive performance is how closely that expectation matches the observed result. The input does not supply the rule for how close a match must be to retain or merge groups.
- Outcome
- The feature of cultural transmission that is measured. How widely material is seen, how accurately it is copied, how its meaning changes, whether it is taken up and how long it lasts are separate outcomes, so a grouping supported for one need not be established for all.
- Screened sources
- The publications or other records supplied as having been assessed for their bearing on the question, together with their quoted evidence and limitations. This task supplies none, so no reported literature findings can be attributed to them.
RL-2 equivalence and unit-sensitivity methods do not supply a measured number of irreducible hypothesis families.
The description names a set of methods for deciding when explanations count as equivalent and for checking whether conclusions change with the definition of the cultural item being tracked, but it supplies no account of those methods or their results. It assumes that they have not measured how many distinct groups of explanations must remain to predict the experiments. If that assumption held, establishing the count would still be unfinished work rather than a result already supplied by those methods.
No screened sources were supplied, so the assertion about what the named methods establish cannot be checked against any read literature. The input does not define RL-2, document an experimental comparison, or provide a measured count. Approximately five is a proposed shortlist size to assess, not an established finding in the supplied material; the absence of supplied sources establishes neither that the count is unknown in the literature nor that any particular count is correct.
The same question asked without the part nothing read establishes:
- What count or range of distinct explanations is supported for predicting cultural transmission experiments kept separate from selection, after equivalent explanations are grouped and each proposed causal relationship is tested by removing it?
- Do those experimental comparisons support approximately five groups of explanations, fewer groups, more groups, or several groupings that remain indistinguishable?
- Approximately five groups remain necessary If approximately five groups each contribute a needed prediction after equivalent explanations are merged and their claimed causal links are challenged, a shortlist of that size would reflect the assessed evidence. Its size would describe the tested cultural items, conditions and outcomes; it would not by itself establish five universal causes of cultural transmission.
- Fewer groups are needed If merging proposed groups preserves predictions, or removing a claimed causal link exposes no need for a separate group, the original shortlist would contain distinctions not required by those comparisons. Counting those distinctions as separate explanations would overstate how many different accounts the assessed evidence supports.
- More groups are needed If a proposed group combines explanations that respond differently when their claimed causes are removed, and keeping them separate is needed to predict the reserved experiments, that group would require subdivision. A fixed shortlist of approximately five would then conceal distinctions needed to account for the assessed results.
- The count remains unresolved If several groupings predict the assessed experiments comparably, or the effects of removing claimed causes remain uncertain, the comparisons would not select one count. A range or several possible groupings would describe that uncertainty, whereas a single number would imply a distinction the evidence had not established.
Grouping explanations determines what is counted as a separate account of cultural change. If two explanations describe the same cause and make the same relevant predictions, counting both can make a research shortlist appear more varied than it is. If two explanations depend on different causes, merging them can conceal why removing one cause changes a result while removing another does not. Predicting experiments kept separate from the selection process connects the grouping to explanatory performance beyond the results used to construct it. Treating five as an established count could therefore either preserve duplicate explanations or discard necessary distinctions; an unresolved count would limit how firmly the shortlist could be described.
RL-2 equivalence and unit-sensitivity methods do not supply a measured number of irreducible hypothesis families.
At shortlist freeze, report the actual family count after experimental equivalence decisions, with unresolved partition uncertainty explicit.
Determine whether approximately five distinct families are supported, or whether interventions require consolidation, subdivision or an unresolved count range.
The mechanism it proposes
The engine's own statement of the hypothesis, in full.
SCOUT 2 — CHEMICAL REACTION ENGINEERING CANDIDATE SET: S6={E, source organization during encoding; U, access and updating during retrieval; I, cross-source relational integration; A, audience-conditioned expression and generalization; S, choosing among already available variants; F, adaptive allocation of future exposure}; proposed minimum K=6. The imported mechanism is the failure of kinetic lumping when distinct intermediate pathways respond differently to a localized perturbation. In the cultural realization, an encoded source representation, its accessible retrieval state, an integrated relational representation and an audience-specific emitted version are experimentally distinguishable intermediate states; the selection and allocation processes act on different stages. A scalar 'reconstruction strength' or a single selection kernel cannot predict all localized-removal and rescue responses. E is needed when context before encoding changes later organization; U when retrieval cues change access with the source representation held constant; I when paired premises change a novel relation without changing premise recall; A when recipient knowledge changes the emitted descendant while private recognition is stable; S when fixed-artifact choice changes without semantic editing; F when policy-state reset changes later opportunities while immediate human responses are fixed. Neither six stage names nor six fitted states suffices: each of these six dependency classes must have a transferable, independently necessary intervention signature. M and C from the other candidates are predicted compositions of these intermediates, rather than seventh and eighth mechanisms. The program exposes intermediate distributions z_E,z_U,z_I,z_A: z_E=E(x,c_pre), z_U=U(z_E,c_retrieval,delay), z_I=I(z_U,other_sources), z_A=A(z_I,recipient_state), followed by S over outputs and F over later delivery opportunities. Each operator is a calibrated nonlinear probability kernel. A candidate merge is an actual reduced kernel that must predict both isolated and combined timed removals, including order changes, rather than a label deletion. If the same composed kernel has identical intervention predictions, it is one admissible partition regardless of the number of drawn boxes.
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.
Use an intervention-and-rescue panel with timed, information-matched controls: E, move the same organizing context from before to after source presentation; U, add a retrieval-context cue after equal encoding; I, scramble the cross-source relation while preserving the same component facts and recall; A, swap experimentally known recipient knowledge while keeping the source and private recognition fixed; S, perturb choice information on fixed unchanged artifacts with exposure matched; F, reset versus preserve a logged allocator state and compare with exact exposure-history replay. Each full nonlinear model predicts the result of all removals before observing their held-out combination. Let Delta_f,o be the difference between the observed held-out effect of removal f and the best model in which f is merged into its nearest component. S6 predicts an out-of-margin Delta_f,o for every f on its prespecified primary outcome, successful mediator validation, and selective rescue; no five-family merge predicts all six. In contrast, no extra residual for the ecology modifier or concurrent-update dependency remains after these six established pathways and their calibrated compositions are included. If a single source-state model predicts both E and U removals and rescue, merge them; if I is explained by complementary facts or ordinary hierarchical reconstruction, merge I; if a logged exposure process fully explains F without an independent adaptive-state effect, merge F with S. Any such successful equivalence refutes exactly K=6 and narrows the admissible count downward. This must be tested with the actual nonlinear models, not a rank-six response matrix.
What testing it would take
The engine's own read on whether this is testable with methods that already exist.
The affordable version uses text or small rule systems, independently randomized timing and source relationships, private recognition plus public production, and a transparent experimental recommender. No neural imaging is required. A real first-stage E manipulation cannot be reconstructed by asking for an encoding explanation afterward; intervention fidelity and separate intermediate-state assays are essential. Strong validation repeats the signatures in a practice task and a second language, with independently implemented feed policies and withdrawal follow-up where persistence is claimed.
Other explanations
Every other hypothesis the engine wrote for the same gap, and the observation that would separate the two.
Use an intervention-and-rescue panel with timed, information-matched controls: E, move the same organizing context from before to after source presentation; U, add a retrieval-context cue after equal encoding; I, scramble the cross-source relation while preserving the same component facts and recall; A, swap experimentally known recipient knowledge while keeping the source and private recognition fixed; S, perturb choice information on fixed unchanged artifacts with exposure matched; F, reset versus preserve a logged allocator state and compare with exact exposure-history replay. Each full nonlinear model predicts the result of all removals before observing their held-out combination. Let Delta_f,o be the difference between the observed held-out effect of removal f and the best model in which f is merged into its nearest component. S6 predicts an out-of-margin Delta_f,o for every f on its prespecified primary outcome, successful mediator validation, and selective rescue; no five-family merge predicts all six. In contrast, no extra residual for the ecology modifier or concurrent-update dependency remains after these six established pathways and their calibrated compositions are included. If a single source-state model predicts both E and U removals and rescue, merge them; if I is explained by complementary facts or ordinary hierarchical reconstruction, merge I; if a logged exposure process fully explains F without an independent adaptive-state effect, merge F with S. Any such successful equivalence refutes exactly K=6 and narrows the admissible count downward. This must be tested with the actual nonlinear models, not a rank-six response matrix.
- What would separate them
One fixed response to meaning loss may govern cultural reconstruction and choice predicts: Calibrate beta only on isolated reconstruction trials, then predict the log-odds difference for choosing variants a versus b: log[P(a)/P(b)]=-beta[L(a)-L(b)]. In the same people, a randomized reversal of the independently specified functional loss must reverse both choice preference and semantic transformation in the quantitatively predicted amount, within epsilon_2 and epsilon_4, without fitting a choice-specific beta. Source prestige/frequency, third-party modifier and concurrent-update contrasts have residuals within their equivalence margins after the specified loss, knowledge and opportunities are matched. S1 is the minimal set only if this cross-operation restriction predicts every held-out block and removal of the single source-loss dependency causes out-of-margin failure. A reproducible choice-versus-reconstruction double dissociation at matched loss, or a genuine moderator/concurrency/kinetic residual, falsifies K=1 in favor of a split answer. If an unrestricted single kernel succeeds but this tied law fails, that does not support S1: it is relabeling a composition.
- What would separate them
A third cultural variant may change how competing variants are reconstructed predicts: Measure lower-order source/competitor conditions first and freeze both the actual nonlinear R+S composition and its mixture-selection, causal-source-inference and complementary-information extensions. In held-out communities present the focal source i, competitor j and a third modifier k, keeping the focal producer's dose, selected parent, semantic facts and payoff constant. Change k's contextual relation to the i-versus-j contrast while preserving its surface salience and factual information; include independent-production and fixed-artifact-choice controls. Let D_M be the observed third-party change in the i-to-descendant semantic transition minus the prediction of the strongest calibrated established composition. S3 predicts |D_M|>epsilon_4, a prespecified direction for a given trained modifier relation, and loss of this residual when that relation is experimentally severed; restoring the relation rescues it in new source families. Fixed-artifact choice can show ordinary context effects without establishing M. The R and S removals separately impair source-specific feature transmission and fixed-variant selection, respectively, establishing the other two required dependencies. If the nonlinear established composition predicts D_M within margins, or if M survives only under one semantic meter, remove M as a distinct family and consolidate to the appropriate known R/S explanation. A significant three-way coefficient alone is expressly insufficient.
- What would separate them
Condition-dependent scoring may create false extra families of cultural transmission predicts: Randomize the semantic coding pipeline on the same fully logged outputs, including meaning-reversed high-overlap pairs and meaning-preserved low-overlap pairs. Before calibration, the fitted family partition changes with coder and unit boundaries. After independent calibration of C_a and behavioral comprehension/enactment validation, the proposed modifier, concurrency and stage-split residuals all lie within epsilon_o and their added held-out predictive gain within eta_o; the source-removal versus fixed-variant-choice double dissociation persists, requiring exactly R and S. The one-law tie is rejected by a replicated difference between the reconstruction-calibrated and choice-calibrated response slopes. This is evidence for K=2 only with adequately powered equivalence and successful manipulation checks, not because a larger model has nonsignificant coefficients. A modifier or group-protocol residual that remains in behavioral outcomes, survives independent human rubrics and is selectively abolished by its own dependency removal falsifies the observation-channel account in favor of a larger candidate set. If the scoring effect is real but a biological/cultural residual remains too, S2 is not sufficient.
- Rival 04 of 04What would separate them
Overlapping updates may transmit conflicting cultural rules despite accurate individual recall predicts: Run small communities learning an artificial reference system with two initially compatible rules. In an asynchronous arm two producers read the same earlier public version and issue individually valid but jointly incompatible updates; in a serialized arm the second update is produced after reading the first. Match numbers of messages, total time, word content available at final test, partner mix, payoff and aggregate feedback. A replay arm gives the identical final corpus to new individuals without joint updating; a yoked-delay arm controls staleness and recency. Fit ordinary CHAI-like partner learning, sequential priming, rule learning and allocator memory on matched histories before the test. Define D_C as the excess rate of jointly inconsistent but individually accurately recalled rule pairs over that full established composition. S4 predicts D_C>epsilon_4 specifically when overlapping writes target a shared rule, disappearance when the shared-update dependency is serialized or made nonconflicting, and rescue when conflict is reintroduced; the effect must persist beyond a mere last-item recall error. Independently resetting partner history while retaining a coherent shared record removes H's familiar-to-new-partner transfer, demonstrating H is also needed. Removing source access or choice feedback establishes R and S. If all C effects are predicted by established partner learning plus visible history and recency, delete C and reduce the proposed four-family answer; if H also merges with R under the panel, reduce again. The group label does not earn a family by itself.
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.