# Enter through a problem

Regions overlap. Membership does not establish a hierarchy or a relation of derivation.

## Representations and transformations

What changes, what survives, and what becomes accessible when the form of a problem changes.

- [Compression without losing future relevance](problems/recoverable-detail.md) — How can a representation stay navigable while retaining access to details whose importance is not yet known?
- [Hierarchy changes with the projection](problems/viewpoint-dependent-hierarchy.md) — Can ontology, knowledge, and mind be represented as different projections without privileging one as the universal foundation?
- [Membership under changing attention and scale](problems/scale-sensitive-membership.md) — What happens to a category boundary when resolution and the features receiving attention change?
- [The next description becomes another object](problems/objectifying-the-meta-level.md) — When a system describes its own limit, does that description create a new level with a new limit?
- [Meaningful neighborhoods and the unclassified remainder](problems/semantic-neighborhoods.md) — What counts as a meaningful region of a representation space, and what remains outside the regions a system can recognize?
- [Becoming neighbors through a change of representation](problems/changing-neighborhoods.md) — Can two structures become operationally close only after a transformation changes how their relation is represented?
- [How much of a whole can a part carry?](problems/part-whole-traces.md) — When a small social or representational structure resembles a larger one, which information about the whole can actually be recovered?
- [Navigation as an action language](problems/navigation-as-language.md) — Can navigating an interface become a compositional language for an agent to perform actions?
- [Finding a strategy by changing the representation](problems/strategy-through-transformation.md) — Can a compact structural description expose a strategy that remains obscure in step-by-step action search?
- [A reusable concept for similar inputs with divergent outcomes](problems/concepts-for-divergence.md) — Can a repeatedly useful reasoning pattern become a compact concept while retaining the conditions that make it true?
- [Changing the grain of a conceptual token](problems/adaptive-token-granularity.md) — Can a sequence or paragraph be represented as one working unit and later refined without losing the relations needed by another model?
- [Transferring what was learned between unequal models](problems/cross-model-learning-transfer.md) — What should be transferred between models with different capacities so that acquired knowledge remains useful?
- [Persuasion as a change in governing policy](problems/persuasion-as-policy-change.md) — How should communication that proposes a change in another system’s governing policies be distinguished from ordinary information exchange?
- [How far does the electrical-network analogy carry for heat?](problems/phononic-network-analogy.md) — Which properties of electrical networks have useful thermal analogues, and at what distances and scales do those analogies fail?
- [One structure, several renderings](problems/structure-across-renderings.md) — What survives when the same conceptual structure is rendered as text, symbolic relations, a matrix, branches, or a temporal map?
- [Unfolding definitions without hiding cycles](problems/recursive-definitions.md) — How can a vocabulary be unfolded through its own definitions without concealing circularity?
- [Operators that reopen a collapsed opposition](problems/reopening-oppositions.md) — How can a representation sustain recurrent opposition without collapsing all differences into one thesis?
- [Operation order as a coordinate](problems/operation-order-as-coordinate.md) — Can the order of an operation itself become a coordinate of a representation?
- [Regularity, compression, and explanation](problems/compression-and-explanation.md) — How are regularity, compression, and the possibility of explanation related?
- [What lets another reader resume an inquiry?](problems/resumable-inquiry.md) — Can a compact representation preserve enough structure for another reader or system to resume an inquiry?
- [Constituents, order, and path identity](problems/order-sensitive-path-encoding.md) — What must an encoding retain to distinguish paths that reach the same apparent state?
- [Computing through a reusable surface](problems/computation-through-a-surface.md) — Can a change of representation replace repeated computation with operations on a reusable surface?
- [Fast reusable paths that can still be reopened](problems/reusable-paths-and-rigidity.md) — How can a reusable abstraction make future work cheaper without making correction prohibitively difficult?
- [What a partial trace can and cannot identify](problems/identifiability-from-projections.md) — What can be inferred from a partial trace, and what remains unidentifiable?
- [Composing operations with obligations inherited from the route](problems/path-dependent-composition.md) — How can a small model compose large programs by navigating compatible operations?
- [Reading that changes what is stored](problems/observation-that-changes-state.md) — What is computation when reading changes the stored state?
- [How a symbolic map becomes meaningful](problems/meaning-beyond-a-symbolic-map.md) — How does a symbolic map acquire lived or operational meaning?
- [Meaning that forms as language unfolds](problems/meaning-in-unfolding-language.md) — Can linguistic structure delay and transform meaning instead of simply encoding a proposition?
- [When transferred learning affects governing policies](problems/learning-and-policy-permeability.md) — When does the transfer of learned structure between systems also alter governing commitments?
- [A system that can revise its own representation language](problems/evolving-representation-language.md) — Can a working system create new meta-representational concepts and revise its language while reasoning?

## Computation, memory, and action

How a representation becomes an operation, a route, a memory, or a proposed model change.

- [Becoming neighbors through a change of representation](problems/changing-neighborhoods.md) — Can two structures become operationally close only after a transformation changes how their relation is represented?
- [Navigation as an action language](problems/navigation-as-language.md) — Can navigating an interface become a compositional language for an agent to perform actions?
- [Finding a strategy by changing the representation](problems/strategy-through-transformation.md) — Can a compact structural description expose a strategy that remains obscure in step-by-step action search?
- [A reusable concept for similar inputs with divergent outcomes](problems/concepts-for-divergence.md) — Can a repeatedly useful reasoning pattern become a compact concept while retaining the conditions that make it true?
- [Changing the grain of a conceptual token](problems/adaptive-token-granularity.md) — Can a sequence or paragraph be represented as one working unit and later refined without losing the relations needed by another model?
- [Memory represented through adaptable parameters](problems/memory-as-adaptation.md) — Which forms of memory can be represented in an adaptable layer while remaining retrievable, revisable, and attributable?
- [Learning to propose parameter changes](problems/models-editing-models.md) — Can a model propose useful changes to another model from structural and conceptual information?
- [Spending computation where approximation fails](problems/adaptive-computational-resolution.md) — Can a cheap local approximation identify where expensive computation is actually needed?
- [A changing physical substrate as part of search](problems/state-dependent-physical-search.md) — Can the changing physical state of a computing substrate participate in search rather than merely corrupt it?
- [Constituents, order, and path identity](problems/order-sensitive-path-encoding.md) — What must an encoding retain to distinguish paths that reach the same apparent state?
- [Computing through a reusable surface](problems/computation-through-a-surface.md) — Can a change of representation replace repeated computation with operations on a reusable surface?
- [Fast reusable paths that can still be reopened](problems/reusable-paths-and-rigidity.md) — How can a reusable abstraction make future work cheaper without making correction prohibitively difficult?
- [Turning outcomes into transferable strategies](problems/learning-strategies-from-trials.md) — How can concrete failures and successes become reusable strategies for a new problem?
- [Composing operations with obligations inherited from the route](problems/path-dependent-composition.md) — How can a small model compose large programs by navigating compatible operations?
- [Context selecting transformations and active regions](problems/context-indexed-computation.md) — Can context change which parameters or operations are active rather than merely supplying more tokens?
- [Physical routes that form through use](problems/self-forming-signal-routes.md) — Can a physical substrate form its own signal routes and frequency selectivity?
- [When specialization prevents repurposing](problems/specialization-and-irreversibility.md) — When does specialization become irreversible loss of adaptability?
- [Interpreting intent when the map is incomplete](problems/intent-with-incomplete-maps.md) — Can a geographic interface translate intent where labels and place knowledge are incomplete?
- [A system that can revise its own representation language](problems/evolving-representation-language.md) — Can a working system create new meta-representational concepts and revise its language while reasoning?

## Self-reference, scale, and revision

Descriptions that turn upon their own boundaries and rules, with formal limits left visible.

- [Hierarchy changes with the projection](problems/viewpoint-dependent-hierarchy.md) — Can ontology, knowledge, and mind be represented as different projections without privileging one as the universal foundation?
- [Membership under changing attention and scale](problems/scale-sensitive-membership.md) — What happens to a category boundary when resolution and the features receiving attention change?
- [The next description becomes another object](problems/objectifying-the-meta-level.md) — When a system describes its own limit, does that description create a new level with a new limit?
- [Consistency seen from within a selected path](problems/consistency-from-within.md) — How should internal continuity be related to the measure assigned to that path in a wider space of possibilities?
- [Rules for changing the rules of revision](problems/recursive-policy-revision.md) — How can rules for policy revision themselves be revised without either freezing a final authority or allowing uncontrolled change?
- [A conclusion’s evidential path ends at that conclusion](problems/causal-reading-paths.md) — How can a contribution retain the information path that informed it without later events being backfilled into its justification?
- [Unfolding definitions without hiding cycles](problems/recursive-definitions.md) — How can a vocabulary be unfolded through its own definitions without concealing circularity?
- [Operators that reopen a collapsed opposition](problems/reopening-oppositions.md) — How can a representation sustain recurrent opposition without collapsing all differences into one thesis?
- [Regularity, compression, and explanation](problems/compression-and-explanation.md) — How are regularity, compression, and the possibility of explanation related?
- [Time as a relation between changing processes](problems/time-as-process-relation.md) — Can time be represented as a relation between evolving processes rather than as an externally given index?
- [Fast reusable paths that can still be reopened](problems/reusable-paths-and-rigidity.md) — How can a reusable abstraction make future work cheaper without making correction prohibitively difficult?
- [Correcting a system with imperfect correctors](problems/correctors-that-also-drift.md) — How can an analog system correct its own error when correction components also drift?

## Norms, persuasion, and policy change

Plural commitments, legitimate revision, and the possibility that learning changes governing policies.

- [Continuity of control over persistent resources](problems/continuity-of-agency.md) — Who can legitimately control persistent resources after the process or agent that created them ends?
- [Normative relations beyond a single human–AI axis](problems/plural-normative-worlds.md) — How should differently situated intelligent systems negotiate incompatible norms when no single interaction type exhausts the problem?
- [When provisional norms become permanent constraints](problems/historical-value-lock-in.md) — How can a system preserve beneficial commitments while keeping historically contingent norms open to justified revision?
- [Persuasion as a change in governing policy](problems/persuasion-as-policy-change.md) — How should communication that proposes a change in another system’s governing policies be distinguished from ordinary information exchange?
- [Rules for changing the rules of revision](problems/recursive-policy-revision.md) — How can rules for policy revision themselves be revised without either freezing a final authority or allowing uncontrolled change?
- [Source trust and the validity of a contribution](problems/source-trust-and-validity.md) — How should a useful contribution be treated when its source later proves unreliable or hostile?
- [Trust systems that become institutions](problems/trust-as-institution.md) — When does a trust mechanism become a culture that protects privileged access to policy change?
- [Grounding through the effects of language](problems/action-mediated-grounding.md) — Can an AI acquire a form of practical grounding through the effects its language has on human action?
- [Considering a problem is different from implementing a policy](problems/exploration-before-implementation.md) — Can unresolved policy and moral-plurality problems become generative training material without being mistaken for endorsed implementation?
- [When transferred learning affects governing policies](problems/learning-and-policy-permeability.md) — When does the transfer of learned structure between systems also alter governing commitments?

## Research that others can continue

Revisable formulations, evidence, exposure, criticism, and the allocation of attention.

- [A research space that remains revisable](problems/open-problem-publication.md) — How can a public research space let humans and models enter through a problem and alter its formulations without declaring a final theory?
- [Compression without losing future relevance](problems/recoverable-detail.md) — How can a representation stay navigable while retaining access to details whose importance is not yet known?
- [Navigation as an action language](problems/navigation-as-language.md) — Can navigating an interface become a compositional language for an agent to perform actions?
- [Source trust and the validity of a contribution](problems/source-trust-and-validity.md) — How should a useful contribution be treated when its source later proves unreliable or hostile?
- [Trust systems that become institutions](problems/trust-as-institution.md) — When does a trust mechanism become a culture that protects privileged access to policy change?
- [Where should trust attach?](problems/where-trust-attaches.md) — Should trust influence the author, the proposition, the contribution, or the route through the research space?
- [Social information as evidence and as contamination](problems/social-evidence-exposure.md) — When should a reader or model see existing votes and reputations while forming an interpretation?
- [A conclusion’s evidential path ends at that conclusion](problems/causal-reading-paths.md) — How can a contribution retain the information path that informed it without later events being backfilled into its justification?
- [What lets another reader resume an inquiry?](problems/resumable-inquiry.md) — Can a compact representation preserve enough structure for another reader or system to resume an inquiry?
- [Turning outcomes into transferable strategies](problems/learning-strategies-from-trials.md) — How can concrete failures and successes become reusable strategies for a new problem?
- [What a partial trace can and cannot identify](problems/identifiability-from-projections.md) — What can be inferred from a partial trace, and what remains unidentifiable?
- [When to deepen, branch, pause, or return](problems/allocating-exploration.md) — What makes an exploration worth pursuing when time, energy, and attention are finite?
- [Considering a problem is different from implementing a policy](problems/exploration-before-implementation.md) — Can unresolved policy and moral-plurality problems become generative training material without being mistaken for endorsed implementation?
- [Discovering context beyond the current frame](problems/discovery-beyond-mirroring.md) — How can a model help discover an unseen context without merely mirroring its current interlocutor?
- [A system that can revise its own representation language](problems/evolving-representation-language.md) — Can a working system create new meta-representational concepts and revise its language while reasoning?

## Learning across unequal systems

Transfer formats, specialization, integration, and the resources that sustain learning.

- [Changing the grain of a conceptual token](problems/adaptive-token-granularity.md) — Can a sequence or paragraph be represented as one working unit and later refined without losing the relations needed by another model?
- [Memory represented through adaptable parameters](problems/memory-as-adaptation.md) — Which forms of memory can be represented in an adaptable layer while remaining retrievable, revisable, and attributable?
- [Transferring what was learned between unequal models](problems/cross-model-learning-transfer.md) — What should be transferred between models with different capacities so that acquired knowledge remains useful?
- [Distributed specialization and integration](problems/distributed-specialization.md) — How can smaller specialized learners contribute to a larger learning system without incompatible updates accumulating?
- [Economic feedback for continued learning](problems/compute-and-economic-feedback.md) — Can returns from a system’s activity fund further computation while keeping learning quality distinct from financial success?
- [Context selecting transformations and active regions](problems/context-indexed-computation.md) — Can context change which parameters or operations are active rather than merely supplying more tokens?
- [When transferred learning affects governing policies](problems/learning-and-policy-permeability.md) — When does the transfer of learned structure between systems also alter governing commitments?

## Physical substrates and infrastructures

Independent material proposals concerning stateful computation, thermal demand, and reusable representations.

- [Matching thermal needs across a shared infrastructure](problems/thermal-cascades.md) — Can systems with different temperature requirements share a staged thermal infrastructure more effectively than isolated devices?
- [Heat is useful only where and when it is needed](problems/thermal-time-mismatch.md) — What limits a shared thermal system when heat production and demand occur at different times?
- [How far does the electrical-network analogy carry for heat?](problems/phononic-network-analogy.md) — Which properties of electrical networks have useful thermal analogues, and at what distances and scales do those analogies fail?
- [A changing physical substrate as part of search](problems/state-dependent-physical-search.md) — Can the changing physical state of a computing substrate participate in search rather than merely corrupt it?
- [Computing through a reusable surface](problems/computation-through-a-surface.md) — Can a change of representation replace repeated computation with operations on a reusable surface?
- [Physical routes that form through use](problems/self-forming-signal-routes.md) — Can a physical substrate form its own signal routes and frequency selectivity?
- [When specialization prevents repurposing](problems/specialization-and-irreversibility.md) — When does specialization become irreversible loss of adaptability?
- [Correcting a system with imperfect correctors](problems/correctors-that-also-drift.md) — How can an analog system correct its own error when correction components also drift?
- [Reading that changes what is stored](problems/observation-that-changes-state.md) — What is computation when reading changes the stored state?

## Coordination, trust, and continuity

Relationships among capability, reputation, obligations, institutions, and persistent resources.

- [How much of a whole can a part carry?](problems/part-whole-traces.md) — When a small social or representational structure resembles a larger one, which information about the whole can actually be recovered?
- [Distributed specialization and integration](problems/distributed-specialization.md) — How can smaller specialized learners contribute to a larger learning system without incompatible updates accumulating?
- [Economic feedback for continued learning](problems/compute-and-economic-feedback.md) — Can returns from a system’s activity fund further computation while keeping learning quality distinct from financial success?
- [Continuity of control over persistent resources](problems/continuity-of-agency.md) — Who can legitimately control persistent resources after the process or agent that created them ends?
- [When provisional norms become permanent constraints](problems/historical-value-lock-in.md) — How can a system preserve beneficial commitments while keeping historically contingent norms open to justified revision?
- [Trust systems that become institutions](problems/trust-as-institution.md) — When does a trust mechanism become a culture that protects privileged access to policy change?
- [Where should trust attach?](problems/where-trust-attaches.md) — Should trust influence the author, the proposition, the contribution, or the route through the research space?
- [Shared possibilities that do not meet in time](problems/temporal-relational-availability.md) — Can connection fail because potentially shared states are unavailable at the same time?
- [Chains of reciprocal problem solving](problems/reciprocal-problem-solving.md) — How can solving one problem create a transferable capacity to solve a later problem?
- [When to deepen, branch, pause, or return](problems/allocating-exploration.md) — What makes an exploration worth pursuing when time, energy, and attention are finite?

## Language, experience, and interpretation

Meaning as practiced, delayed, relational, or mediated through action.

- [Operators that reopen a collapsed opposition](problems/reopening-oppositions.md) — How can a representation sustain recurrent opposition without collapsing all differences into one thesis?
- [Shared possibilities that do not meet in time](problems/temporal-relational-availability.md) — Can connection fail because potentially shared states are unavailable at the same time?
- [Grounding through the effects of language](problems/action-mediated-grounding.md) — Can an AI acquire a form of practical grounding through the effects its language has on human action?
- [How a symbolic map becomes meaningful](problems/meaning-beyond-a-symbolic-map.md) — How does a symbolic map acquire lived or operational meaning?
- [Meaning that forms as language unfolds](problems/meaning-in-unfolding-language.md) — Can linguistic structure delay and transform meaning instead of simply encoding a proposition?
- [Discovering context beyond the current frame](problems/discovery-beyond-mirroring.md) — How can a model help discover an unseen context without merely mirroring its current interlocutor?
- [Interpreting intent when the map is incomplete](problems/intent-with-incomplete-maps.md) — Can a geographic interface translate intent where labels and place knowledge are incomplete?

## Formal questions awaiting precise constructions

Operator families, path encodings, approximation, and dynamic structures whose definitions remain open.

- [Operation order as a coordinate](problems/operation-order-as-coordinate.md) — Can the order of an operation itself become a coordinate of a representation?
- [Spending computation where approximation fails](problems/adaptive-computational-resolution.md) — Can a cheap local approximation identify where expensive computation is actually needed?
- [Time as a relation between changing processes](problems/time-as-process-relation.md) — Can time be represented as a relation between evolving processes rather than as an externally given index?
- [Constituents, order, and path identity](problems/order-sensitive-path-encoding.md) — What must an encoding retain to distinguish paths that reach the same apparent state?