Re: An Experimental Research Enquiry: Where Relational Reasoning Training meets and contains Associative Learning, Ontological Abstraction and Cross-Domain Analogical Transformations
Michael
Firstly, apologies for the delay, I was in hospital again for another seizure. For humour, we even managed to get it on camera for specialisation of the doctors treatment plan. Apparently I was screaming a lot, I don't remember any of it.
Secondly, I want to thank Brandon Woodson and Connor Lambdon for sharing their ideas as I wouldn't have been able to create Ontological Worlds without your sharing.
Thirdly, Ontological Worlds is ready to go now on mode 1, mode 2 with the ontological categories is still in its testing phase and will be ready by this coming new Sunday.
Thank you.
Best regards.
===========
To properly explain Ontological Worlds, it helps to establish a glossary for the cognitive operations the system is designed to train.
Ontological Worlds - https://mike-stack1982364.github.io/Ontological-Worlds/
The central proposition is that sophisticated reasoning does not arise from recognising isolated facts. It arises from constructing relationships between relationships, integrating those relationships into a coherent model, extracting their unstated implications, distinguishing between closely neighbouring configurations, then transferring or generating that structure across new contexts.
I describe this broader capacity as relational integration.
Relational IntegrationRelational integration is the process of combining multiple relations into a unified, internally compatible model whose structure, implications and precise qualitative configuration can be inferred and manipulated.
It is not merely noticing that two things are related.
It requires the thinker to:
- hold several relations simultaneously;
- understand how those relations constrain one another;
- integrate them into one larger structure;
- identify implications that were never directly stated;
- distinguish the resulting structure from nearby alternatives;
- revise the model when contradictions appear;
- generate new examples that preserve the same architecture;
- transfer that architecture into a different domain.
Within Ontological Worlds, this process can be divided into six interdependent operations:
- Meta-mapping
- Meta-coherence
- Meta-logic
- Meta-distinction
- Meta-generation
- Meta-transfer
The first four form the foundational architecture of relational integration.
The final two are its generative and generalising extensions.
The complete cycle can be represented as:
Map > test coherence > infer > distinguish > generate > transfer > revise
Meta-mapping is the ability to construct, hold, align and compare two or more abstract relational mappings simultaneously.
A mapping identifies how one structure corresponds to another.
A meta-mapping requires the thinker to compare not merely objects, but entire systems of relationships.
For example, in a sixteen-direction Ontological World:
- north may represent increasing creation or reconstruction;
- east may represent increasing artificial intervention and control;
- northeast must therefore combine both dimensions;
- north-northeast must contain both, but be weighted more strongly toward north;
- east-northeast must contain both, but be weighted more strongly toward east.
To identify these distinctions, the learner must simultaneously maintain:
- the meaning of north;
- the meaning of east;
- the relationship between north and east;
- the proportional difference between NNE, NE and ENE;
- the internal structure of the scenario;
- the structures of neighbouring scenarios;
- the rules governing the entire relational network.
That concurrent alignment of multiple abstract structures is meta-mapping.
Meta-mapping constructs the network within which the remaining operations become possible.
Meta-coherence is the ability to determine whether multiple mappings, properties, inferences and distinctions remain mutually compatible within one global relational system. So this is where executive function definitely kicks in when it comes to bringing about and maintaining a coherent picture.
A scenario may appear plausible when examined in isolation but fail when placed into the wider ontology.
For example, a proposed NNE scenario may fail because:
- it expresses more east than north;
- it is functionally indistinguishable from NE;
- it contradicts the meaning assigned to north elsewhere;
- its relationship with neighbouring positions cannot be preserved;
- its opposite position no longer makes conceptual sense;
- transferring the same structure into another domain exposes an inconsistency.
Meta-coherence therefore asks:
Does every component still fit once the entire system is considered?
This is particularly important in Ontological Worlds because the learner is often asked to selectively generate context.
The learner cannot simply invent an interesting example.
They must generate a context in which:
- every ingredient is internally compatible;
- every ingredient expresses the intended dimensions;
- the proportions between dimensions are accurate;
- the example remains distinct from neighbouring examples;
- the wider compass remains logically stable.
Meta-coherence prevents local plausibility from being mistaken for global correctness.
Meta-logic is the inference of relationships and consequences that are not explicitly stated but are contained within an integrated relational structure.
Consider:
A is larger than P.
P is larger than C.
From these two statements, we can infer:
- A is larger than C.
- A is the largest.
- C is the smallest.
- P is between A and C.
None of these conclusions was directly stated.
They emerge from the structure produced when the original relations are integrated.
In Ontological Worlds, meta-logic may involve recognising that:
- a scenario contains more north than east;
- one position must lie closer to northwest than west;
- two proposed positions cannot be neighbours because their structures are too dissimilar;
- a missing position must inherit properties from both adjoining directions;
- altering one scenario forces revisions elsewhere in the world;
- a generated context violates an implication of the global mapping.
Meta-logic therefore reveals what the relational system necessarily implies.
Meta-distinction is the ability to identify the precise qualitative difference between competing relational configurations.
It is not merely noticing that two scenarios are different.
It is identifying exactly how their total relational profiles differ.
For example, a zoo and a wildlife sanctuary both contain animals.
Their deeper distinction concerns:
- why the animals are present;
- how much autonomy they possess;
- how much human control is imposed;
- whether the institution prioritises exhibition, research, protection or restoration;
- whether the animals are separated from or integrated into an ecological system.
That total relational difference is a meta-distinction.
Within a sixteen-direction compass, the distinction between NNE, NE and ENE is also a meta-distinction.
All three positions may express the same two dimensions.
The difference lies in their weighting.
The learner must determine:
- which properties are present;
- how strongly each property is expressed;
- which property dominates;
- which neighbouring position the scenario most closely resembles;
- what prevents it from belonging elsewhere.
Meta-distinction gives the integrated model its resolution.
Meta-generation is the ability to create a new scenario, mapping or relational structure that satisfies multiple abstract constraints simultaneously.
This is more demanding than recalling an example.
The learner begins with a relational target and must construct a context that occupies it accurately.
For example, generating a scenario for NNE requires the learner to produce something that:
- combines north and east;
- expresses more north than east;
- remains distinct from north;
- remains distinct from northeast;
- fits the ontology governing all sixteen positions;
- contains no internal contradictions.
This is constrained generation rather than free association.
The learner is not asking:
“What interesting thing can I imagine?”
They are asking:
“What possible world would satisfy this exact relational specification?”
Meta-generation is therefore the productive expression of meta-mapping and meta-coherence.
Meta-transfer is the ability to preserve an abstract relational structure while expressing it through a different domain, context or ontology.
Suppose a sixteen-direction network is first constructed around animal institutions.
The same geometry might then be transferred into:
- political systems;
- business organisations;
- military strategies;
- biological processes;
- psychological states;
- technological civilisations;
- family structures;
- fictional worlds.
A valid transfer does not preserve surface features.
It preserves the deeper relational architecture.
For example, if northeast represents a balanced combination of world creation and artificial control in one domain, its equivalent in another domain must preserve that same structural balance, even though the visible content may be completely different.
Meta-transfer demonstrates whether the learner has understood the relation itself or merely memorised the original example.
The Bedrock of Relational IntegrationThe foundational sequence is:
Meta-mapping → meta-coherence → meta-logic → meta-distinction
- Meta-mapping constructs the relational network.
- Meta-coherence determines whether that network is globally compatible.
- Meta-logic reveals the implications contained within it.
- Meta-distinction identifies its precise qualitative configuration.
Together, these four operations constitute the bedrock of relational integration.
> Meta-generation and meta-transfer then extend that integrated structure.
> Meta-generation creates a new context from the structure.
> Meta-transfer preserves the structure across different contexts.
The full architecture therefore becomes:
Meta-mapping constructs the network. Meta-coherence stabilises it. Meta-logic reveals what it implies. Meta-distinction resolves its exact form. Meta-generation extends it. Meta-transfer carries it into another world.
In the earlier Imagi-World system, relational space was primarily organised through:
- north;
- south;
- east;
- west.
These four directions represented broad relational poles.
Ontological Worlds expands this into a sixteen-point compass:
- North
- North-northeast
- Northeast
- East-northeast
- East
- East-southeast
- Southeast
- South-southeast
- South
- South-southwest
- Southwest
- West-southwest
- West
- West-northwest
- Northwest
- North-northwest
I previously considered many of these intermediate directions either unnecessarily difficult or conceptually redundant.
I now think they are central to the system.
Four directions permit coarse categorisation.
Sixteen directions demand high-resolution relational integration.
With only four directions, the learner may be able to classify a scenario according to a single dominant quality:
- natural;
- artificial;
- generative;
- reductive.
With sixteen directions, the learner must model several dimensions simultaneously and estimate their relative weighting.
They must distinguish:
- north from NNE;
- NNE from NE;
- NE from ENE;
- ENE from east.
The task is no longer merely to determine which qualities are present.
It is to determine the exact relational proportion in which they are present, that is now high load-bearing weight on meta-distinction whereas before, we were mainly dealing with meta-logic and meta-generation by comparison.
A Sixteen-Direction Ontological World
For example:
T is north of J.
Where:
T = Jurassic Park
J = A typical zoo
J is not the centre of the ontology. It is simply a reference node within this particular relational sequence.
Further premises progressively reveal the wider structure:
L is east of J.
Where:
L = A biological laboratory where monkeys are experimented on
The learner must now determine what abstract relationship makes Jurassic Park north of the zoo, what makes the laboratory east of it, and what kinds of scenarios could occupy the increasingly precise positions between them, emerging from the integration of the relational premises.
Within this demonstrative mapping, the positions could be represented as follows:
North — Jurassic Park
> An existing institution for displaying animals is transformed into the reconstruction of an entire lost biological world.
North-Northeast — Inventing an Entirely New Species
> This shares Jurassic Park’s production of unprecedented life while introducing a stronger element of deliberate biological engineering.
Northeast — Genetically Engineered Superhumans
> Existing life is deliberately transformed into something categorically beyond its ordinary form.
East-Northeast — Engineering Human-Level Intelligence in Primates
> An existing species is intensively modified, but not replaced by an entirely new category of life.
East — A Biological Laboratory Where Monkeys Are Experimented On
> Existing organisms are subjected to controlled intervention and experimentation.
East-Southeast — An Industrial Animal-Testing Facility
The experimental organism becomes increasingly standardised and instrumentalised.
Southeast — A Cloned-Livestock Production Complex
Artificial control is combined with the conversion of living organisms into engineered commercial units.
South-Southeast — An Automated Slaughterhouse
A controlled system transforms living beings into standardised material outputs.
South — A Taxidermy Exhibition
The living organism has been reduced to a preserved physical object.
South-Southwest — A Natural-History Specimen Archive
The organism remains reduced to a specimen, but that reduction now serves historical preservation.
Southwest — A Cryogenic Gene Bank
Living complexity is compressed into stored biological information for possible future restoration.
West-Southwest — A Captive-Breeding Conservation Centre
Organisms remain controlled, but that control is directed toward preservation and eventual release.
West — A Nature Reserve
Animals possess substantially greater ecological autonomy than those contained for exhibition.
West-Northwest — A Large-Scale Rewilding Corridor
The goal expands from protecting individual organisms to restoring the relationships that sustain an ecosystem.
Northwest — A Wildlife Sanctuary
Animals are placed within an environment organised around protection, welfare and greater autonomy.
North-Northwest — A De-Extinction Sanctuary
Lost species are reconstructed and returned to a protected, self-sustaining ecological system.
Why These Positions Are Not Arbitrary
Each position must be inferred from its relationships with the others. North-northeast cannot merely be assigned an imaginative example. It must share the structure of both north and northeast while remaining more similar to north. East-northeast must also share those neighbouring structures, though remain more similar to east.
The same requirement applies around the complete network.
The learner must determine:
- what relationships organise the current mapping;
- how each scenario inherits properties from neighbouring scenarios;
- which properties dominate;
- why one position is closer to one neighbour than another;
- whether all positions remain mutually compatible.
The examples therefore form one constrained relational system rather than sixteen independent categories. Thus, changing one scenario may require the learner to revise several others.
The Six Meta-Operations
This process requires six interacting operations.
Meta-mapping holds and compares multiple abstract relational structures.
Meta-coherence tests whether those structures remain mutually compatible.
Meta-logic reveals consequences that were not directly stated.
Meta-distinction identifies the precise difference between neighbouring configurations.
Meta-generation creates a new scenario that satisfies the required relational constraints.
Meta-transfer preserves the same abstract structure while replacing the original subject matter.
The process can therefore be represented as:
Receive premises > construct mappings > test coherence > infer relationships > distinguish positions > generate a continuation > transfer the structure
Why Sixteen Directions Improve the Training
The four-direction system establishes broad relational contrasts. The sixteen-direction system forces the learner to navigate the conceptual territory between them. With four directions, the learner may succeed by identifying one dominant similarity.
With sixteen, the learner must:
- integrate several premises;
- infer the hidden structure of the mapping;
- compare closely neighbouring possibilities;
- estimate the relative weighting of shared properties;
- reject locally plausible but globally inconsistent answers;
- revise the model as new information appears.
The additional directions therefore increase representational resolution rather than merely increasing the number of labels. NNE, NE and ENE may share most of their properties. The challenge is to identify the relational weighting that separates them. That finer discrimination places greater demand on relational integration than the original four-direction system, ergo a greater load bearing weight of meta-distinction is now operative during the training.
What Ontological Worlds Is Actually Training
The learner is training the ability to:
- construct global models from local premises;
- hold multiple abstract mappings simultaneously;
- infer unstated relationships;
- preserve compatibility across interacting constraints;
- distinguish closely neighbouring structures;
- generate contexts that satisfy precise relational requirements;
- revise an ontology when one relation changes;
- preserve a structure while transferring it into another domain.
The learner is not merely deciding that something belongs at northeast.
They must determine:
- which premises imply that location;
- why it lies between its neighbours;
- which qualities dominate;
- why it is not NNE or ENE;
- what change would move it elsewhere;
- whether its placement remains compatible with the wider mapping.
Meta-mapping constructs the network.
Meta-coherence ensures that it holds together.
Meta-logic reveals what it implies.
Meta-distinction resolves its precise structure.
Meta-generation extends it.
Meta-transfer reconstructs it in another world.
Demonstrative Purposes and Cross-Domain Mapping
The animal examples above are included for demonstrative purposes only.
They do not permanently define north, south, east, west or any of the intermediate positions. The relationships governing one Ontological World may be entirely different from those governing another.
The intended exercise is cross-domain and will be explained more deeply once I have a proper tutorial on the site.
The learner must extract the abstract relational structure from one mapping and reproduce it using completely different subject matter.
For example, the structure represented above by north-northeast > the deliberate creation of something genuinely unprecedented—could be transferred as:
- Biology: inventing an entirely new species;
- Technology: creating a fundamentally new form of computing;
- Government: designing a form of political organisation that has never previously existed;
- Language: constructing a symbolic system capable of expressing relations that existing languages cannot represent.
The structure represented by east-northeast > the intensive transformation of something already existing—could become:
- Biology: engineering human-level intelligence in primates;
- Technology: converting a narrow artificial-intelligence system into a general reasoning system;
- Business: transforming a conventional company into a deeply automated organisation;
- Education: technologically augmenting an existing learner rather than replacing the educational system itself.
The structure represented by southwest > compressing a complex system into a preserved form from which it could later be restored—could become:
- Biology: storing endangered life as genetic information;
- Culture: archiving an endangered language for future revival;
- Technology: preserving the code and data of a discontinued digital ecosystem;
- Government: retaining the institutional records of a collapsed state for possible reconstruction.
The structure represented by west-northwest > restoring the relationships that allow a larger system to function autonomously—could become:
- Ecology: reconnecting fragmented habitats;
- Education: reconnecting isolated disciplines into an integrated body of knowledge;
- Psychology: reintegrating compartmentalised memories and emotional systems;
- Technology: connecting independent open-source projects into an interoperable ecosystem.
The surface content changes completely.
The underlying relational structure remains.
A learner who remembers that Jurassic Park was north has learned a label.
A learner who understands the abstract relations that made it north can reconstruct an equivalent position through technology, politics, psychology, culture, education or any other domain.
That is the transition from directional recall to relational integration and genuine meta-transfer.
.
Michael
Cross-World Method - How to map the relations.
(Utilised in the examples above: Previously referred to as the DaVinci Method in the previous edition of the game, Imagi-World)
The examples above continue a generative mapping procedure that I will now call the Cross-World Method, is more palliative for the design of Ontological Worlds.
I previously called it the Da Vinci Method, but “Cross-World Method” is, for now, more accessible, less noisy and more straightforward.
This is the method previously employed when training on the game Imagi-World. The learner begins with several relational premises, infers the abstract structure connecting them, and constructs a coherent possible world in which every position satisfies those relations, as per the examples above, I.e. K is west of M, Dirty Bankers are the opposite to Charity Workers, so we are finding analogical mappings relative to the direction of the relational premise.
Ontological Worlds uses the same generative mapping method as imagi-world but at greater resolution. Imagi-World established broad directional relations; Ontological Worlds extends the process across sixteen positions, and when onto mode two, across the the ontological parameters themselves. The examples above should therefore be understood as continued Cross-World Method demonstrations: different surface worlds generated from, and constrained by, an underlying relational structure.
As stated this coming Sunday I will finalise mode 2 for the ontological categories, and with that a proper tutorial imprinted onto the site, exclusive to the ontological categories and inclusive of their application.
Best,
Michael
Michael
Michael
https://groups.google.com/g/brain-training/c/gh5xgRRwVtc/m/XiU309FIkLUJ
Michael
Michael
(1) I use it while consciousness streaming, so I map the source-contact, as in the immediacy of patterns perceived in my consciousness stream to the n-stimuli, and then I compare the patterns I discern from my consciousness stream n-trials ago.
Thus I remember the
(1) n-stimuli, letters or numbers for me as seizure medication for now has given me partial blurry vision (everything will workout okay in the long run there),
(2) I pair the patterns to the n-stimuli as presented
(3) I either compare incoming patterns perceived in my stream with all prior paired (n-stimuli pairing) patterns while also doing it for the exact n-stimulus pairing followed by, if its a match, anchoring the next perception of patterns with how they connect to the specific n-stimulus pairing, if not a match, then still the full pairings. Or I only compare relative to the n-back positioning.
I.e. Let's say we are playing n-back 2 and I am perceiving a rainforest in my stream, the first letter is "K", so I pair the patterns there with "K", I continue to allow the stream to evolve and let's say I then move into the cellular body of a roaming jaguar while hearing the letter "T" as the next n-trial, then let's say the beaming sun catches my attention and the patterns of that are paired with the third n-stimulus "Y". As described, I will either compare the patterns paired with "Y" with those of "K", not just whether there is a letter match, or if there is a letter match, compare both all of the paired patterns to letters while anchoring to the patterns from the match as an expansion point. This really teaches synthesis as well, and perhaps that's something I dive deeper into at a later time after I have picked this back up after a long while.
I will try the same technique with this new Number Series n-back variant I have created, but maybe it'll be too much executive chaos for my brain to be the right protocol for me, I will let the results show for themselves down the line.