1. Recognition
Definition. Identifying something as already known or encountered.
Example. Recognising a reasoning error seen before.
Limit. RECOGNISING ≠ EXPLAINING ≠ TRANSFERRING.
The emergence of understanding
Unpacking what the word “understand” encompasses, tracking what becomes functionally available, and testing what holds when conditions change.
Why this field?
The Laboratory does not begin with a universal definition of understanding. It observes which distinctions become necessary when an overly broad category no longer describes what is happening adequately.
The initial question — What is understanding? — becomes a methodological one: which phenomena do we merge when we use the word “understanding”?
SAME WORD ≠ SAME PHENOMENON ≠ SAME FUNCTION.
What “emergence” means here
Distinctions that arise during inquiry
The categories were not fixed in advance. They appeared when an observable difference vanished within an overly broad description.
Concrete problem
Provisional tool
Test / observation
Limit
Distinction
Counter-test
Stabilisation or revision
EMERGENCE = WHAT BECOMES NECESSARY TO DISTINGUISH DURING INQUIRY.
The emergence of an understanding
For a person or a system, certain conditions of access, relation or mobilisation can be fostered without claiming to produce understanding directly. The term has no mystical or metaphysical meaning here.
FOSTERING CONDITIONS ≠ DIRECTLY PRODUCING UNDERSTANDING.
LOCATION OF THE INTERVENTION ≠ LOCATION WHERE UNDERSTANDING EMERGES.
Seven phenomena we seek to stop conflating
These are distinct, non-exhaustive candidates: neither mandatory stages, nor a universal order, nor established natural components, nor a score.
2. Distinction
Definition. No longer merging two closely related phenomena.
Example. TRUE INFORMATION ≠ RELEVANT INFORMATION; EMOTION ≠ ACTION.
Limit. Distinguishing is not enough to understand the relation.
3. Intelligibility
Definition. Relations between several elements become intelligible.
Example. A changes B, which makes C possible.
Limit. PRESENCE OF THE ELEMENTS ≠ INTELLIGIBILITY OF THEIR RELATION.
4. Explication
Definition. A rule or organisation can be formulated.
Example. Explaining the rule already used to complete a task successfully.
Limit. EXPLICATION ≠ EVIDENCE OF THE ACTUAL MECHANISM.
5. Transfer
Definition. Using a capacity in a relevant new situation without equivalent reteaching.
Example. Reusing in a real conflict a distinction learned through a story.
Limit. One new success ≠ overall robustness; understanding can also mean not transferring beyond the domain.
6. Functional availability
Definition. Knowledge can be mobilised when the function needs it.
Example. A rule accessible in calm conditions disappears under activation.
Limit. KNOWING ≠ HAVING ACCESS ≠ MOBILISING AT THE RIGHT TIME.
7. Integration / embodiment
Definition. A distinction produces relatively lasting effects across several situations.
Example. “An emotion does not impose an action” remains available in different real contexts.
Limit. ONE CORRECT ACTION ≠ INTEGRATION; INTEGRATION ≠ IRREVERSIBLE FINAL STATE.
7 CANDIDATE PHENOMENA ≠ SCORE ≠ UNIVERSAL STAGES ≠ EXHAUSTIVE LIST.
What taxonomy tests made visible
Compression and merging trials show that grouping neighbouring phenomena too quickly can erase methodologically relevant information.
RECOGNITION ≠ DISTINCTION ≠ INTELLIGIBILITY ≠ EXPLICATION ≠ TRANSFER ≠ FUNCTIONAL AVAILABILITY ≠ INTEGRATION.
Profiles such as strong distinction / weak transfer, strong recognition / weak capacity availability, or strong explication / weak integration then become describable. They remain dissociations to test, not universal laws.
TAXONOMIC USEFULNESS ≠ ESTABLISHED ONTOLOGY.
Nine instruments for observing without reducing to a score
1. Typed relations
What relation connects the elements: supports, conditions, revises, contradicts, limits or depends on?
2. Population / referential continuity
Are we still speaking about the same object, person, capacity or information?
3. Agency
Who or what can actually produce the action, decision or transformation?
4. Local topology
Is the functional relation local, between components, or distributed across an organisation?
5. Temporality
Is what was true or necessary at T1 still so at T2?
6. Functional availability
Is the capacity merely present or genuinely available for mobilisation?
7. Provenance and access level
Where does the information come from and how is it currently accessible?
8. Scope / domain of validity
Under which conditions is this distinction or capacity valid?
9. Variable explanatory status
Is it an observation, hypothesis, provisional explanation, conclusion or unknown?
These instruments guide observation; they do not describe nine parts of the mind.
9 INSTRUMENTS ≠ 9 DIMENSIONS OF A SCORE.
OBSERVATION INSTRUMENT ≠ COMPONENT OF THE PHENOMENON.
Studying distinctions — and what happens between them
The Aksoydan Laboratory studies not only the phenomena that “understanding” tends to merge, but also their interactions: what becomes accessible without being mobilised, what transfers without being stable, what remains functional while its historical support changes, or what appears equivalent until a perturbation reveals a difference.
Access × mobilisation
A capacity may be accessible without being mobilised at the relevant moment. ACCESS ≠ MOBILISATION, but their interaction may determine actual success.
Transfer × stability
A capacity may work elsewhere without being durably stable, or remain stable in a narrow context without transferring. TRANSFER ≠ STABILITY.
Provenance × current support
A structure needed to install an effect may no longer be its main support. HISTORICAL PROVENANCE ≠ CURRENT SUPPORT.
Output × robustness
Two systems may produce the same output today and diverge when the context changes. SAME CURRENT OUTPUT ≠ SAME ROBUSTNESS.
Correction × descendants
Upstream information changes while some downstream effects remain active. SOURCE REVISION ≠ AUTOMATIC REVISION OF THE TRAJECTORY.
Configurations rather than isolated variables
At some points, tracking one more variable no longer improves observation. It becomes more useful to ask which organisation of relations must remain available. This does not mean that all understanding is configurational or distributed.
The seven phenomena help distinguish what is observed; the nine instruments examine the relations, conditions and statuses under which these phenomena become observable. Their intersection may guide an inquiry, but the work is not about filling 63 boxes.
INTERACTION BETWEEN DIMENSIONS ≠ MERGING PHENOMENA.
INTERACTION BETWEEN DIMENSIONS ≠ SINGLE UNDERLYING MECHANISM.
ALWAYS FOLLOWING THE NEXT LINK ≠ ALWAYS UNDERSTANDING BETTER.
OBSERVABLE CONFIGURATION ≠ INDEPENDENT ENTITY.
7 × 9 = POSSIBLE OBSERVATION TOOL, NOT A SCORE OR A COMPLETE TABLE TO FILL IN.
How do we know whether a distinction is useful?
Which precise error becomes undetectable if we remove this distinction?
- Without access / mobilisation, “known but unusable” disappears.
- Without provenance, what needs revision becomes difficult to track.
- Without temporality, an old condition may still seem necessary.
- Without scope, a local result may falsely become general.
A USEFUL DISTINCTION MUST INCREASE OBSERVATIONAL CAPACITY.
MORE DISTINCTIONS ≠ BETTER OBSERVATION.
Trace, access and mobilisation
- Trace
- Information, a capacity or an organisation remains present somewhere.
- Access
- The person or system can retrieve it.
- Mobilisation
- It is actually used in the relevant function.
Example. A rule is known and retrievable in calm conditions but is not used during a real situation. Saying only that it is “present” would conceal the point of failure.
TRACE ≠ ACCESS ≠ MOBILISATION.
Installation, maintenance and current support
A structure A may enable the installation of an organisation B. Later, removing A no longer changes the effect, whereas perturbing B makes it disappear. A may therefore have mattered historically without remaining the main current functional support.
Here, functional support means what a function relies on to be installed, maintained, mobilised or expressed. It must not be confused with epistemic support — the elements that sufficiently support a proposition or conclusion.
INSTALLING ≠ MAINTAINING.
NECESSARY FOR INSTALLATION ≠ NECESSARY FOR MAINTENANCE.
HISTORICAL PROVENANCE ≠ CURRENT SUPPORT.
Local or configurational support
A local support is an identifiable element that appears to carry a function. A configurational support is an organisation of several elements or relations. In a musical group, no single person necessarily carries the complete arrangement: the function also depends on their coordination.
EPISTEMIC SUPPORT ≠ FUNCTIONAL SUPPORT.
ABSENCE OF A SINGLE CENTRAL VARIABLE ≠ ABSENCE OF FUNCTIONAL SUPPORT.
DISTRIBUTED ≠ EVERYWHERE.
SUPPORT ≠ EXPRESSION.
SUPPORT ≠ COMPLETE MECHANISM.
NAMING A CONFIGURATION ≠ REIFYING AN ENTITY.
Provenance and historical descendants
Provenance means the history through which information, a conclusion or a capacity acquired its current status: source, transformations, revisions, informational descendants and present justification.
A descendant is something downstream that exists in part because earlier information, a decision, an action or a configuration helped produce it.
“Paul is probably unavailable”
“Paul is unavailable”
“Let us not assign him the case”
If the source is corrected, its descendants do not disappear automatically. Some may even become antecedents of new decisions. This calls for targeted re-examination of those whose justification or epistemic status — what the available elements can actually support — depended on the revised source.
SOURCE CORRECTION ≠ AUTOMATIC CORRECTION OF DESCENDANTS.
NOT AUTOMATICALLY CORRECTED ≠ EXEMPT FROM RE-EXAMINATION.
HISTORICAL DESCENDANT ≠ CURRENT DEPENDENCE.
HISTORICAL PROVENANCE ≠ CURRENT JUSTIFICATION.
Robustness: resistance to which variations?
A capacity is robust if it continues to work when certain conditions change. A one-off success, local stability, resistance to a given perturbation and robustness across several contexts must be distinguished.
Example. Completing a task with assistance and then without it informs one precise perturbation; it does not establish general robustness.
LOCAL SUCCESS ≠ ROBUSTNESS.
LOCAL ROBUSTNESS ≠ GENERAL ROBUSTNESS.
AUTONOMY ≠ TRANSFER ≠ ROBUSTNESS.
Adaptive resolution
Too little detail sometimes merges important phenomena. Too much detail may fragment the description until the observed function is lost.
What level of resolution makes the difference relevant to the current question visible?
Resolution too low: risk of mergingResolution too high: risk of fragmentation
Compression may remain useful if the differences it groups can be recovered when the question requires it.
MORE DETAIL ≠ BETTER OBSERVATION.
REVERSIBLE COMPRESSION ≠ IRREVERSIBLE MERGING.
Experimental future-space
Here, future-space means differences invisible in the current situation that may become observable when a relevant condition changes.
Example. Two people complete the same task successfully today. Assistance is removed: one still succeeds, the other does not. The difference may already have existed in their functional organisation before appearing in the output.
It is a tool for designing testable divergences, not a prediction of destiny, a metaphysics or a general theory.
CURRENT EQUIVALENCE ≠ EQUIVALENCE UNDER FUTURE PERTURBATION.
Same output, different organisations
Two people or systems may give the same correct answer while differing in access, provenance, supports, capacity for transfer or robustness.
SAME OUTPUT ≠ SAME ORGANISATION ≠ SAME HISTORY.
CORRECT OUTPUT ≠ ESTABLISHED UNDERSTANDING.
Topology and organisation
The same components may be connected differently; the same topology may evolve through different dynamics; the same current function may have several histories.
SAME COMPONENTS ≠ SAME FUNCTIONAL ORGANISATION.
SAME TOPOLOGY ≠ SAME DYNAMICS.
DISTRIBUTED ≠ GLOBAL.
Functional invariants
A function may remain stable while its implementation changes. Two solutions may preserve the same functional constraint without sharing the same realisation.
FUNCTIONAL INVARIANT ≠ FIXED IMPLEMENTATION.
Representation, use and test independence
Information may be present, representable or decodable without being used by the system in the function under study.
DECODABLE REPRESENTATION ≠ FUNCTIONAL USE.
SUFFICIENT REPRESENTATION ≠ SUFFICIENT EXPERIMENTAL EVIDENCE.
Adding more tasks does not automatically provide more information if they all involve the same hidden dependency.
MORE TASKS ≠ MORE INDEPENDENT CONSTRAINTS.
Same functional question, another field
Education
A person may recognise and explain a rule, complete an exercise successfully, then fail to transfer or mobilise the distinction elsewhere. This separates correct answer, explication, transfer and availability.
Programming
Copying, imitation, a local procedure and reconstruction may produce a similar artefact with different histories. FUNCTIONAL ARTEFACT ≠ ESTABLISHED UNDERSTANDING. COPYING ≠ RECONSTRUCTION.
Neuroscience
Retrieval, distributed coordination, supports, temporality, plasticity and reorganisation provide families of questions. No specific mechanism is attributed here without dedicated verification.
AI
Information present but unused, generalisation, memory, robustness, provenance, revision and fragile trajectories may be tested in comparable ways. FUNCTIONAL ANALOGY ≠ IDENTITY OF MECHANISMS.
Different corpora, one discipline of translation
Research at the Aksoydan Laboratory does not begin from a single corpus. It brings together scientific, methodological, human, historical, philosophical, spiritual, symbolic and narrative materials. These corpora do not have the same function or evidential status.
A question may arise in one corpus, then be reformulated, translated, tested or abandoned in another.
External scientific literature
Cognitive science, neuroscience, psychology, education, philosophy and epistemology, linguistics and translation, and AI research and engineering can constrain, compare, contradict, contribute empirical results and propose different segmentations.
SCIENTIFIC PARALLEL ≠ VALIDATION OF OUR WORK.
Aksoydan methodological corpora
Methodological emergence, observations, distinctions, deltas, syntheses, counter-tests, revisions, limits and protocols document how a question becomes a distinction, how it is tested and how its status changes.
Earlier corpora by Serpil Aksoydan
Comment je fonctionne vraiment is an earlier corpus of observation, modelling and formulation concerning human functioning. Some questions were subsequently decomposed, corrected or tested more strictly. EARLIER CONCEPT ≠ CURRENT LABORATORY RESULT.
Transmutation is a narrative and symbolic corpus that may provide forms, relations, metaphors or questions. SYMBOLIC FECUNDITY ≠ SCIENTIFIC VALIDITY.
Ancient knowledge, philosophies and traditions
Philosophical or spiritual traditions, religious or sacred texts and historical systems of thought may transmit categories, practices, narratives, architectures of meaning and questions about human experience. They do not replace modern experimental results or scientific literature.
Human observations and narrative corpora
Observed situations, learning, interactions, narratives, examples and trajectories may bring a question to light, provide a case or motivate a test. SINGULAR OBSERVATION ≠ GENERAL LAW. NARRATIVE ≠ ESTABLISHED CAUSAL MECHANISM.
DIFFERENT CORPORA MAY PRODUCE USEFUL QUESTIONS WITHOUT PRODUCING THE SAME KIND OF EVIDENCE.
The Laboratory does not ask a scientific text to serve the function of a spiritual text, or a spiritual text to serve the function of scientific evidence.
AGE ≠ VALIDITY. TRADITION ≠ EVIDENCE. RESONANCE ≠ VALIDATION.
SACRED TEXT ≠ EXPERIMENTAL SOURCE.
CULTURAL SURVIVAL ≠ EXPERIMENTAL ROBUSTNESS.
ORIGIN OF A QUESTION ≠ EVIDENCE FOR THE ANSWER.
TRANSLATING BETWEEN CORPORA ≠ EQUIVALENT STATUS.
Illustrative genealogy of a question
Earlier corpus / observation / literature
Question
Methodological decomposition
Test / comparison / counterexample
Revision
Current status
RESEARCH TRAJECTORY ≠ UNIVERSAL PIPELINE.
Translating between branches — and back again
A question may circulate between human experience, education, neuroscience, philosophy, AI or symbolic systems to observe what remains functionally relevant and what breaks. It may become a test in a second corpus, then return to change the original distinction.
TRANSLATING A QUESTION ≠ TRANSFERRING A MECHANISM.
SAME WORD ACROSS TWO BRANCHES ≠ SAME OBJECT.
SAME APPARENT FUNCTION ≠ SAME SUPPORT.
A DISTINCTION MAY CROSS FIELDS WITHOUT LICENSING A CROSS-FIELD ONTOLOGY.
INTER-CORPUS TRANSLATION = TOOL FOR CONSTRAINT AND REVISION, NOT EVIDENCE OF UNITY.
The value of this intersection does not lie in adding several forms of expertise. It lies in what each prevents the others from concluding too quickly. This discipline of translation is also connected to Serpil Aksoydan’s interdisciplinary path.
FLEXIBILITY IN WHAT MAY OPEN A QUESTION.
RIGOUR IN WHAT MAY BECOME A CONCLUSION.
Verified scientific crossings
These publications provide external constraints, not a decorative bibliography. They may meet some Laboratory distinctions, but may also complicate them, segment them differently or require new counter-tests.
The Aksoydan Laboratory studies distinctions and their interactions. Scientific crossings are especially useful when they show that the same apparent capacity may depend on several organisations, that one dimension may change without the others, or that different frameworks segment the same container term differently.
INTERACTION ≠ MERGING.
DIVERGENT SEGMENTATION ≠ FAILURE.
CONVERGENCE ≠ IDENTICAL SEGMENTATION.
Memory, trace and retrieval
What the literature examines. Yang, He and Cai distinguish transient reactivation from more lasting reorganisation after a reminder stimulus; its context also changes behavioural consequences. Kerrén, Michelmann and Doeller link successful retrieval to an expansion of cortical representation preceded by a brief hippocampal oscillation — a ripple.
Crossing. These results make it useful to separate maintenance, active manifestation, reactivation, transformation and functional consequence. Retrieval does not necessarily resemble replaying an identical copy.
What differs or remains open. They do not establish that trace, access and mobilisation correspond to three distinct biological mechanisms. The status of “activity-silent” working-memory states remains debated: Oberauer and Awh test the alternative of retrieval from episodic memory, while Foster, Vogel and Awh defend a distinction between active and offline representations.
What this changes for our tests. Compare two retrievals yielding the same answer, then examine whether their later transformations and transfers diverge.
SAME EXPERIMENTAL PHENOMENON ≠ SINGLE THEORETICAL INTERPRETATION.
Working memory ≠ understanding as a whole.
Yang, C., He, X. & Cai, Y. “Reactivating and reorganizing activity-silent working memory: two distinct mechanisms underlying pinging the brain”. Cerebral Cortex, 2025. DOI: 10.1093/cercor/bhae494.
Oberauer, K. & Awh, E. “Is There an Activity-silent Working Memory?”. Journal of Cognitive Neuroscience, 2022. DOI: 10.1162/jocn_a_01917.
Foster, J. J., Vogel, E. K. & Awh, E. “Working Memory as Persistent Neural Activity”. The Oxford Handbook of Human Memory, 2024. DOI: 10.1093/oxfordhb/9780190917982.013.13.
Kerrén, C., Michelmann, S. & Doeller, C. F. “Hippocampal ripples initiate cortical dimensionality expansion for memory retrieval”. Nature Communications, 2026. DOI: 10.1038/s41467-026-75345-6.
Support, topology and stability
What the literature examines. Human intracranial recordings link the co-occurrence of ripples between distant regions to neuronal coordination that varies with memory load. In mice and artificial recurrent networks, Osako and colleagues observe specialised subspaces reused for new content. In macaques, Abbaspoor, Aljishi and Hoffman describe traces becoming more stable with age while also observing broader reorganisation.
Crossing. Distributed support, local specialisation and reorganisation may coexist. Stability does not entail immobility, and a functional invariant does not fix its content.
What differs or remains open. Observed coordination ≠ complete causality. Intracranial recordings concern a clinical population; mouse ≠ human; macaque ≠ general human understanding. The stability study does not directly test “installation ≠ maintenance”.
What this changes for our tests. Ask which level — component, module or network — is necessary for the function, and whether the same correct output depends on different supports.
DISTRIBUTED ≠ HOMOGENEOUS ≠ EVERYWHERE.
DISTRIBUTED SUPPORT ≠ ABSENCE OF SPECIALISATION.
STABILITY ≠ IMMOBILITY.
Verzhbinsky, I. A. et al. “Cross-region neuron co-firing mediated by ripple oscillations supports distributed working memory representations”. Nature Neuroscience, 2026. DOI: 10.1038/s41593-026-02403-z.
Osako, Y. et al. “Reusable modular architecture enables flexible cognitive operations in the mouse brain and artificial recurrent networks”. Nature Neuroscience, 2026. DOI: 10.1038/s41593-026-02410-0.
Abbaspoor, S., Aljishi, A. & Hoffman, K. L. “Experience reorganizes content-specific memory traces in macaques”. Nature Neuroscience, 2026. DOI: 10.1038/s41593-026-02357-2.
Explication and transfer
What the literature examines. In physics tasks, structured self-explanation prompts improve explanation quality, error correction and near transfer, but not the reported far transfer. A second-order meta-analysis frequently finds near transfer, while far transfer becomes weak or null after controlling for certain biases. A 2025 school-based study nevertheless observes a small far-transfer effect for metacognitive regulation.
Crossing. Explication does not guarantee transfer; transfer is not all-or-nothing, and its domain of validity is part of the result.
What differs or remains open. Self-explanation in physics does not exhaust our notion of explication; these studies’ near/far measures do not exhaust our notion of transfer.
What this changes for our tests. Ask whether an explicated capacity without far transfer may nevertheless be robust in its local domain, and systematically vary the distance and nature of tasks.
EXPLICATION ≠ GUARANTEED TRANSFER.
ABSENCE OF FAR TRANSFER ≠ ABSENCE OF UNDERSTANDING.
PRESENCE OF TRANSFER ≠ UNDERSTANDING AS A WHOLE.
Zhang, Q. & Fiorella, L. “Effects of self-explaining feedback on learning from problem-solving errors”. Contemporary Educational Psychology, 2024. DOI: 10.1016/j.cedpsych.2024.102326.
Sala, G. et al. “Near and Far Transfer in Cognitive Training: A Second-Order Meta-Analysis”. Collabra: Psychology, 2019. DOI: 10.1525/collabra.203.
Wirth, J. et al. “Far Transfer of Metacognitive Regulation: From Cognitive Learning Strategy Use to Mental Effort Regulation”. Educational Psychology Review, 2025. DOI: 10.1007/s10648-024-09983-x.
Other philosophical segmentations of understanding
What the literature examines. Belkoniene presents grasping — grasping an explanation — as knowing how to use that explanation to account for a phenomenon. Strevens instead analyses it as a recognition capacity.
Crossing. Other frameworks may group under “understanding” capacities that we separate here in order to test them.
What differs or remains open. This difference in segmentation is not an anomaly to eliminate by definition: it is itself a research question. Philosophy ≠ experimental data.
What this changes for our tests. Examine whether recognition, intelligibility, explication or use of an explanation genuinely dissociate across tasks, rather than defending the taxonomy by definition.
AKSOYDAN SEGMENTATION ≠ ESTABLISHED NATURAL SEGMENTATION.
ANOTHER TAXONOMY ≠ AN ERROR TO CORRECT.
Belkoniene, M. “Grasping in Understanding”. The British Journal for the Philosophy of Science, 2023. DOI: 10.1086/714816.
Strevens, M. “Grasp and Scientific Understanding: A Recognition Account”. Philosophical Studies, 2024. DOI: 10.1007/s11098-024-02121-x.
Representation, function and causal level
What the literature examines. Zheng, Lu and Woolgar stress that not all information decodable from frontoparietal activity is necessarily related to behaviour. In mechanistic interpretability of language models, Mueller and colleagues compare several possible causal mediators — neurons, directions, attention heads or other units — according to the research objective.
Crossing. A detectable representation is not automatically used; the appropriate level of causal description depends on the question.
What differs or remains open. Neural decoding ≠ general definition of representation. AI model ≠ human brain. This work does not validate our adaptive resolution.
What this changes for our tests. Ask whether decodable but causally unused information should count as functionally available, and which distinctions survive a change in resolution.
DECODABLE REPRESENTATION ≠ FUNCTIONAL USE.
APPROPRIATE LEVEL OF DESCRIPTION ≠ FIXED UNIVERSAL UNIT.
Zheng, Y., Lu, R. & Woolgar, A. “Radical flexibility of neural representation in frontoparietal cortex and the challenge of linking it to behaviour”. Current Opinion in Behavioral Sciences, 2024. DOI: 10.1016/j.cobeha.2024.101392.
Mueller, A. et al. “The Quest for the Right Mediator: Surveying Mechanistic Interpretability for NLP Through the Lens of Causal Mediation Analysis”. Computational Linguistics, 2026. DOI: 10.1162/coli.a.572.
What these crossings change for our tests
They strengthen the case for testing separately presence and manifestation, explication and transfer, stability and immobility, local support and distributed organisation, representation and causal use, output and internal organisation.
They also complicate our categories: silent states remain debated, recognition and grasping may be grouped differently, distribution and modularity may coexist, and the scope of transfer or causal level depends on the question.
- Can an explicated capacity without far transfer be locally robust?
- Does the same correct output depend on different supports?
- Which distinctions survive a change in resolution?
- Is decodable but causally unused information functionally available?
- Which phenomena in our taxonomy become redundant across tasks?
SCIENTIFIC PARALLEL ≠ VALIDATION OF OUR WORK.
None of this work is evidence for the 7 × 9.
Current scope
What this work can do
- Better distinguish phenomena merged under “understanding”.
- Design discriminating tests and test dissociations.
- Compare trajectories, not only outputs.
- Observe provenance, temporality, support and availability.
- Track what changes under perturbation.
- Identify when more detail helps or degrades observation.
- Translate certain distinctions into other fields.
What it cannot establish
- A universal definition or general theory of understanding.
- A global score, seven mandatory stages or nine natural dimensions.
- A complete causal mechanism.
- Consciousness in an AI system or a Human / AI identity.
- Universal generalisation of local results.
7 × 9 ≠ SCORE.
USEFUL TAXONOMY ≠ ESTABLISHED ONTOLOGY.
OBSERVED FUNCTION ≠ ESTABLISHED CONSCIOUSNESS.
How to read the status of propositions
StrengthenedLocal / boundedOpenNot justified at this stage???
These documentary statuses indicate how propositions are presented on this page: resistance within the current framework, limited scope, active question, insufficient support or unknown information. They determine neither their functional status within an operation nor their ontological status — what the object would actually be — and are not assigned arbitrarily to every proposition.
EPISTEMIC STATUS ≠ FUNCTIONAL STATUS ≠ ONTOLOGICAL STATUS ≠ DOCUMENTARY STATUS.
PUBLICATION ≠ SCIENTIFIC VALIDATION.
Research that can correct itself
Transfer
More generalisation ≠ better understanding.
Resolution
More detail ≠ better observation.
Support
Cause of installation ≠ current support.
Output
Same answer ≠ same history.
These revisions do not close the framework: they show why its own distinctions must remain exposed to counter-tests.
Open questions
What remains to be tested
- Beyond what level of dissociation is it still useful to speak of “understanding”?
- Which distinctions are genuinely independent?
- Which instruments become redundant?
- Which level of resolution actually maximises observation?
- Which current differences become visible only after perturbation?
- When does a distributed organisation become the relevant functional support?
No answer is presupposed by publishing these questions.
Continue exploring
Return to the general map or examine the concepts and fields that extend this research.