The Generative Topology of Information Space
情報空間の生成的位相論
P versus NP, the Discrete Positions of Creation, Viewing, and Evaluation, and Interface Ontology
Tanaka, Tomokazu (bigakusha-haha). 2026. The Generative Topology of Information Space: P versus NP, the Discrete Positions of Creation, Viewing, and Evaluation, and Interface Ontology. Research Monograph Series, Version 1.0. Sayama Aesthetics School / Institute of Aesthetics.
Abstract / Overview
This study develops the proposition that information has time and distance and that information space therefore exists.
It does not reduce the P versus NP problem to a loose metaphor for creativity.
In computational complexity theory, P is the class of decision problems decidable in polynomial time by a deterministic model of computation.
NP is the class of decision problems whose positive instances have certificates verifiable in polynomial time.
The P versus NP problem asks whether polynomial-time verifiability and polynomial-time decidability coincide.
The present study does not identify this question with artistic creativity.
It uses P versus NP as a rigorous reference model for a more limited structure.
Within one shared problem setting and one shared set of operational rules, proximity to verification need not coincide with proximity to construction.
Generative Topological Ontology does not use topology here as a direct synonym for mathematical topological space.
Topology names the structure of connectability generated through relations.
Information space is therefore not treated as a container holding informational items.
It is a relational structure generated through admissible operations, transformation paths, computational resources, order, irreversibility, and boundary conditions.
Informational distance is the operational difference involved in reaching one information state from another.
It can involve computational steps.
It can involve memory.
It can involve communication.
It can involve energy.
It can involve background knowledge.
It can involve institutional access and authorization.
Distance is therefore relational rather than absolute.
The same information can be near for one algorithm, language, body, institution, or history and remote for another.
Informational time is also more than external clock duration.
It includes order.
It includes dependency.
It includes repetition.
It includes waiting.
It includes memory.
It includes irreversibility.
Once an answer is known, the prior state of not knowing cannot always be fully restored.
After encountering a work, a viewer may no longer inhabit exactly the same world as before.
Information acquisition transforms its recipient.
The transformed recipient then changes the meaning and reachability of later information.
The study further distinguishes Discreteness from Dispersion.
Discreteness describes separated elements that share one coordinate system, one set of rules, one problem definition, and one scale of comparison.
Dispersion describes plural phases belonging to different bodies, scales, temporalities, causalities, values, and reality conditions without one final coordinate system that completely contains them.
This distinction is used to reconstruct the positions of creation, viewing, and evaluation.
These are not treated as essential capacities belonging to three fundamentally different subjects.
They are discrete functional positions within one generative phase.
Creation already includes viewing and evaluation.
Viewing also includes creation and evaluation.
Evaluation includes viewing and creation.
The three do not operate only as serial stages.
They function in parallel within each act.
One position can become locally foregrounded while the others remain operative.
The work is therefore an Interface.
It connects the positions of creation, viewing, and evaluation.
It exchanges them.
It redistributes them.
Evaluation is also a generative Force.
It does not merely measure a completed work.
Collection changes the social existence of a work.
Review changes it.
Pricing changes it.
Canonization changes it.
Inclusion within education changes it.
Evaluation directs visibility, recognition, preservation, imitation, citation, and future production.
It changes informational distance within the field.
The study also connects information space with Self-Externalization.
Self-Externalization is not the replication of an already completed self.
It generates paths from self toward non-self.
A work, utterance, record, or bodily Mark transforms what it carries.
The return path is asymmetrical.
The self that encounters its externalization again is no longer identical with the self before externalization.
Self-Externalization therefore has informational time.
Its return is irreversible.
Inverse Problem Art actively designs this asymmetry.
A conventional forward problem begins with conditions and laws and derives an outcome.
An inverse problem begins with outcomes or boundary conditions and infers possible generative structures.
Inverse Problem Art goes further.
A declaration can be established first.
An outline can be established first.
A place can be established first.
An institution can be established first.
A web address or exhibition condition can be established first.
These become boundary conditions from which meaning, relation, subjectivity, and Workhood emerge retrospectively.
A work is therefore not a correct answer to a pre-given question.
It is not a certificate guaranteeing correctness.
It is an Interface that changes what can appear as a question and what can count as an answer.
This is also the limit of the analogy with P versus NP.
P versus NP concerns decision and verification within a fixed formal problem phase.
Art can transform the problem phase itself.
Creativity is therefore not merely efficient search through a given informational space.
It can open relations that were previously excluded from that space.
It can reorganize the distinction involving problem and solution.
A work should therefore not be treated as one additional informational point.
A work can alter distances among existing points.
It can open new paths.
It can bring previously disconnected experiences into a new neighborhood.
When an encounter with a work changes the meaning of past experience, changes the appearance of a familiar site, or changes the boundary involving self and other, the work has transformed the viewer's information space.
Viewing is therefore not simply the decoding of meaning installed in advance by the artist.
Work A, viewing position B, and their contact can generate new reality C.
C was not stored in completed form inside the artist.
It was not stored in completed form inside the work.
It was not stored in completed form inside the viewer.
It is generated through relation.
Web Installation provides a direct example.
The location of the work cannot be reduced to one file or one screen.
URL, server, device, communication path, viewer, temporal difference, update, and failure all participate in appearance.
Web Installation does not erase informational time and distance.
It exposes them as constituents of the work.
Artificial intelligence can also be understood as an Interface that reorganizes informational distance and neighborhoods.
Generative AI can shorten the time required to produce outputs.
This does not automatically eliminate the time required to determine whether an output is grounded in reality.
It does not guarantee that citations exist.
It does not guarantee contextual adequacy.
It does not guarantee value.
The shortening of a generative path can transfer verification costs to the user.
This structure is related to P versus NP but is not identical with it.
AI may produce quickly while leaving verification expensive.
Large-scale generation can therefore amplify the burden of evaluation.
Human judgments can then return through training, ranking, recommendation, and visibility.
An evaluation leverage structure can emerge.
Creativity in the age of AI cannot therefore be defined only by output generation.
It also requires decisions concerning generative conditions, Reality-Grounding, preserved Difference, and responsibility for the costs of verification and evaluation.
A prompt is not a container holding a completed meaning.
It is a boundary condition imposed upon a generative space.
AI generates possible outputs from that condition.
Workhood depends on whether those outputs open new relations involving self, non-self, place, institution, and reality.
The study ultimately reformulates informational Being as connectability.
To be informationally present is not merely to occupy storage.
It is to be reachable.
It is also to enable further reaching.
Information has a where.
That where is the path through which it can be reached from elsewhere and from which it can generate further relations.
Creation opens new connectability where no such path previously existed.
Full Text
This research monograph develops a Generative Topology of Information Space.
Information is often treated as abstract content without location.
A string can be copied across different media.
A formula can be copied.
An image can be copied.
An algorithm can be implemented in different physical systems.
The informational content can still be recognized as the same.
This can make information appear independent of place.
It can appear to have no distance.
It can appear to have no time.
Yet storage is not the same as reachability.
An encrypted archive can be physically present and still be remote.
An undeciphered script can be present and still be remote.
An unproved theorem can be present as a statement and still remain operationally remote.
Data can exist but remain unreachable because it is not indexed.
A text can be physically available while remaining remote because its presuppositions are unavailable.
Geographically distant information can also become operationally near through a protocol, index, algorithm, or language.
Information therefore has a spatiality that cannot be reduced to physical coordinates.
The position of information depends upon the states through which it can be reached.
It depends upon admissible operations.
It depends upon the resources required for transformation.
Generative Topological Ontology understands Being through relational connectability rather than fixed substance.
Relation precedes Being.
Topology names the structure of connectability generated through relation.
The present study extends this ontology into the informational domain.
Information space is not a container of informational items.
It is generated through reachability.
It is generated through transformation costs.
It is generated through operational rules.
It is generated through order and irreversibility.
It is generated through boundary conditions.
The central proposition is simple.
Information has time and distance.
Information space is the topology of connectability formed through this time and distance.
Creation does not simply add a new point to information space.
Creation generates new paths and neighborhoods.
P versus NP is introduced as a rigorous reference model.
It is not treated as a mathematical theory of artistic creativity.
P is the class of decision problems decidable in polynomial time by a deterministic model of computation.
NP is the class of decision problems whose positive instances possess certificates verifiable in polynomial time.
The open question asks whether these classes coincide.
This concerns decision problems.
It does not directly classify general discovery, artistic production, meaning, or aesthetic value.
Search problems involving witness construction require related formalisms such as FNP.
For some self-reducible problems, decision and witness search are tightly connected.
This does not justify extending the relation to every kind of creation.
P versus NP nevertheless makes one informational asymmetry visible.
The path involving a supplied certificate and its verification is not identical with the path involving the original input and the construction of an appropriate certificate or affirmative decision.
A short verification path does not automatically imply a short construction path.
Even within one shared problem phase, the neighborhood of verification can differ from the neighborhood of construction.
This provides a rigorous limit model for thinking about informational distance.
Informational distance is not the visual difference among strings.
Two statements can differ by one symbol while having radically different consequences.
Two representations can appear entirely different while remaining operationally near because an efficient translation connects them.
Distance is therefore not an intrinsic property of isolated informational items.
It depends upon operations and Interfaces.
Informational distance can involve computational steps.
It can involve memory.
It can involve communication.
It can involve energy.
It can involve background knowledge.
It can involve authorization.
It can involve institutional access.
It can involve standards of evaluation.
These dimensions are not identical.
Each participates in determining whether an information state is reachable.
The same information can therefore be near for one algorithm and far for another.
It can be near within one language and far within another.
It can be near for one body and far for another.
It can be near within one institution and inaccessible within another.
An Interface does not transparently eliminate distance.
It opens some paths.
It closes others.
It reorganizes informational neighborhoods.
Informational time is also more than clock time.
Transformation requires order.
It requires dependencies.
It can require repetition.
It can require waiting.
It can require memory.
It can be irreversible.
Once an answer is known, the prior state of not knowing cannot always be restored.
After an encounter with a work, the viewer may not return to exactly the same world.
Information acquisition transforms the recipient.
That transformed recipient changes the meaning of later information.
Informational time is therefore an irreversible generation in which the history of connection changes future connectability.
The distinction involving Discreteness and Dispersion becomes important here.
Discreteness describes elements that remain separate while sharing one coordinate system.
They share rules.
They share a problem definition.
They share a scale of comparison.
Distances among them can be measured within one system.
Classification, operation, choice, optimization, and solutions remain possible in principle.
Dispersion describes a different condition.
Elements can belong to different bodies.
They can belong to different scales.
They can belong to different temporalities.
They can belong to different causalities.
They can belong to different values and reality conditions.
No final coordinate system completely includes all of them.
Translation remains possible across dispersed phases.
A remainder can nevertheless persist.
This distinction allows creation, viewing, and evaluation to be reconstructed.
Conventional art institutions place them in different subjects.
The artist creates.
The audience receives.
Critics, markets, museums, and academic institutions evaluate.
This social division is real.
Ontologically, however, the three are not necessarily dispersed across irreducible phases.
They share a work.
They share artistic institutions.
They share language.
They share history.
They share evaluative coordinates.
They can therefore be understood as discrete positions within one generative phase.
Creation already includes viewing.
The maker observes what appears.
The maker chooses.
The maker retains.
The maker rejects.
The maker modifies.
Creation already includes evaluation.
Viewing also includes creation.
A viewer does not merely receive sensation.
The viewer composes contours.
The viewer generates meanings.
The viewer generates relations.
The viewer activates memories.
Viewing also includes evaluation.
Evaluation includes viewing and creation.
An evaluator changes visibility.
An evaluator changes circulation.
An evaluator changes preservation.
An evaluator changes citation.
An evaluator changes imitation.
An evaluator changes future production conditions.
Creation, viewing, and evaluation therefore do not operate only as three serial stages.
They operate in parallel within each act.
One position can become locally foregrounded.
The others remain operative.
The work is therefore an Interface connecting these positions.
It exchanges them.
It redistributes them.
The question is not only who the creator is.
A more fundamental question concerns what generates the positions of creation, viewing, and evaluation.
Another question concerns what temporarily occupies those positions.
Human beings can occupy them.
Works can occupy them.
Exhibition spaces can affect them.
Languages can affect them.
Algorithms can affect them.
Markets can affect them.
Preservation systems can affect them.
Artificial intelligence can affect them.
Evaluation must also be distinguished from formal verification.
A computational verifier checks whether a certificate satisfies conditions already fixed by a formal system.
Artistic evaluation often changes the conditions of existence of what it evaluates.
Collection changes a work.
Review changes a work.
Pricing changes its social reality.
Canonization changes its position.
Educational inclusion changes its future.
Evaluation is therefore not merely a neutral measure applied after Being is complete.
Evaluation is a generative Force.
It directs actions and choices.
Proof can stabilize this Force institutionally.
Value can sustain it through time.
Evaluation generates further evaluation.
Proof generates demands for further proof.
Value generates further value.
An evaluation leverage structure can emerge.
The positions of creation, viewing, and evaluation are therefore neither fixed nor equal.
Institutions distribute speech unevenly.
They distribute capital unevenly.
They distribute visibility unevenly.
They distribute capacities of preservation unevenly.
Evaluation changes informational distance.
It can make one work searchable.
It can make another effectively invisible.
It can authorize one expression as a legitimate problem.
It can dismiss another as noise.
These actions reorganize neighborhoods and boundaries within information space.
The critique is therefore not simply directed at the existence of artists, audiences, and evaluators as different social roles.
It is directed more fundamentally at the naturalization of discrete positions as fixed identities and authorities.
Art does not need to erase all positions.
It can reveal how positions are generated.
It can reveal how they are exchanged.
It can reveal how they acquire power.
It can make them reconfigurable.
The theory of information space also clarifies Self-Externalization.
Self-Externalization is not a secondary expression of an already completed interior self.
The self comes into being through relation with what is other than itself.
At the same time, it cannot be completely reduced to others, institutions, language, evaluation, works, or technology.
From the perspective of information space, Self-Externalization is the formation of paths involving self and non-self.
A work is not a transparent copy of the self.
An utterance is not a transparent copy.
A record is not a transparent copy.
A bodily Mark is not a transparent copy.
Externalization produces transformation.
The return path becomes asymmetrical.
The self that encounters what it externalized is no longer identical with the self before externalization.
Self-Externalization therefore has informational time.
Its return is irreversible.
Inverse Problem Art actively designs this asymmetry.
A conventional forward problem begins with initial conditions and laws.
An outcome is obtained through them.
An inverse problem begins with observed outcomes or boundary conditions.
Possible generative structures are inferred.
Inverse Problem Art does more than reconstruct a missing cause.
It can establish a declaration first.
It can establish an outline first.
It can establish a place first.
It can establish an institution first.
It can establish a web address first.
It can establish an exhibition condition first.
These become boundary conditions.
Meaning can emerge retrospectively.
Relations can emerge retrospectively.
Subjectivity can emerge retrospectively.
Workhood can emerge retrospectively.
The work is therefore not the correct answer to a pre-given question.
It is not a certificate guaranteeing correctness.
It is an Interface that changes what can appear as a question.
It also changes what can count as an answer.
This indicates the limit of the analogy with P versus NP.
P versus NP concerns decision and verification within a shared formal problem phase.
Inverse Problem Art asks how the problem phase itself is generated.
It asks who defines its boundary.
It asks which realities are allowed to appear as problems.
Creativity is therefore not merely efficient search through a fixed search space.
Creativity can open relations previously excluded from that space.
It can reorganize the distinction involving problem and solution.
A work should not be treated merely as one informational item added to a preexisting space.
A work can change distances among existing points.
It can open new paths.
It can bring previously disconnected experiences into a new neighborhood.
An encounter with a work can give past experience another meaning.
It can make a familiar place appear as another reality.
It can transform the boundary involving self and other.
In such cases, the work has not merely transmitted information.
It has transformed the information space of the viewer.
Viewing is therefore not the decoding of a meaning installed in advance by the artist.
Work A and viewing position B can generate new reality C.
C was not stored in completed form inside the artist.
It was not stored in completed form inside the work.
It was not stored in completed form inside the viewer.
C is a new Phase generated through contact.
This is consistent with the basic propositions of Generative Topological Ontology.
Relation precedes Being.
Generation occurs through relation.
Creation is therefore a local operation that opens what was not yet visible into visibility.
Viewing is a local operation that assumes and regenerates that opening as another reality.
Evaluation is a local operation directing how the connection persists, becomes shared, and becomes institutionalized.
All three participate in the topological transformation of information space.
Web Installation provides a concrete model of this structure.
The internet is not treated as a virtual container.
It is a field in which works and subjects appear through access, connection, delay, movement, reading, display, and recognition.
The location of the work cannot be reduced to one file.
It cannot be reduced to one screen.
URL matters.
Server matters.
Device matters.
Communication path matters.
Viewer matters.
Temporal difference matters.
Update matters.
Failure matters.
Web Installation does not eliminate informational time and distance.
It exposes them as conditions of appearance.
Artificial intelligence can also be interpreted as an Interface reorganizing informational distance and informational neighborhoods.
Generative AI does not simply retrieve intact items.
It composes outputs through learned statistical relations.
Rapid output generation does not guarantee Reality-Grounding.
It does not guarantee factual accuracy.
It does not guarantee valid citations.
It does not guarantee contextual adequacy.
It does not guarantee value.
A shortened generative path can transfer the cost of verification to the user.
This produces an asymmetry related to, but distinct from, P versus NP.
AI can shorten production time while leaving verification costly.
Large-scale generation can increase the burden of evaluation.
Human judgments can then return through training.
They can return through ranking.
They can return through recommendation.
They can return through visibility systems.
An evaluation leverage structure can emerge.
Creativity in the age of AI therefore cannot be defined only by whether an output can be generated.
Generative conditions matter.
Reality-Grounding matters.
Preserved Difference matters.
Responsibility for verification matters.
Responsibility for evaluation matters.
A prompt is not a command containing completed meaning.
It is a boundary condition imposed upon a generative space.
AI generates possible outputs under that condition.
Workhood depends upon whether an output opens new relations involving self, non-self, site, institution, and reality.
The argument can be summarized through several propositions.
Informational Being consists not only in storage but in the capacity to be connected, transformed, decoded, and made operative.
Informational distance is the difference in operations, resources, knowledge, and institutional conditions required to reach one state in relation to another.
Informational time includes order, dependency, history, and irreversibility.
Informational nearness and remoteness depend upon operations and Interfaces.
Efficient verification does not automatically guarantee efficient construction or decision.
Creation, viewing, and evaluation are discrete positions within one generative phase.
Evaluation generates Being by directing visibility, recognition, preservation, imitation, and future creation.
A work is an Interface connecting, exchanging, and redistributing the positions of creation, viewing, and evaluation.
Creation does not merely add a point to information space.
It generates paths.
It generates neighborhoods.
It generates boundaries.
It generates problem definitions.
To be is not merely to occupy a fixed location.
To be is to possess connectability.
A being can reach others.
It can also enable others to reach new forms of Being.
Information therefore has a where.
That where is the path through which information can be reached by another state.
It is also the position through which further relations can be generated.
Creation opens new connectability in a region that previously had no such path.
The complete text is available as a PDF below.
Keywords
Generative Studies; Generative Topological Ontology; Information Space; P versus NP; Computational Complexity; Informational Time; Informational Distance; Connectability; Creation; Viewing; Evaluation; Discreteness; Dispersion; Self-Externalization; Inverse Problem Art; Workhood; Interface Ontology; Reality-Grounding; Web Installation; Artificial Intelligence; Generative AI