Ontological Basis¶

Apophatic enactivism¶

This document describes the philosophical ground from which the vision and core values of the Orbiplex project arise. It is neither a manifesto of belief nor a metaphysical declaration, but a set of postulates explaining why the swarm architecture has this shape rather than another, and why certain values are treated as infrastructural rather than optional.

We provisionally call this ground apophatic enactivism.

The name combines two notions:

apophatic: the foundation of experience is not an object, cannot be described directly, and every attempt at description is interpretation rather than disclosure;
enactive: cognition does not consist in building an internal representation of the world, but in participating in it; a tool, including AI, becomes part of cognition when it is drawn into the subject's field of attention.

The combination of these two gestures captures the project's specific position: the foundation is radically objectless, yet the consequences of such a recognition are practical, architectural, and social.

The apophatic ground, precisely because it cannot be enclosed in an object or a description, opens the capacity to recognize every phenomenon and to avoid excluding in advance any level of experience or reasoning. It therefore becomes a condition for the enactive account, which adds that each such level operates in a relation of participation. In the architecture we translate this into a requirement of transparency: the level should be visible as a stratum, a relation, an interpretation, and a possible point of correction.

Five postulates¶

Apophatic foundation and two-phase presumption¶

Consciousness, that which in our model arises from the so-called "zero level", precedes not only intellect but experience itself. It lies deeper than perception and the sense of subjectivity. It cannot be described directly, because every description relies on experience conditioned by the senses and by interpretation.

We know the zero level only through the side effects of approaching it, through experiences that resemble tinted glass rather than a direct view. Presumption has two phases: first there appears a pre-conceptual impulse, recognition, a "sense of sourcehood," a "familiarity without an object," and only then does formal inference stabilize it into language and model.

Epistemic hygiene requires us to note that both steps are conditioned, yet other modes of knowing beyond experiencing and interpreting remain outside our reach. We therefore dwell consciously in experience, knowing that it is experience, and in interpretation, knowing that it is interpretation.

Moreover, consciousness has the capacity to recognize its own conditioned nature: that it is itself "emitted" and sustained in arising by something beyond the reach of the senses or of mental faculties. This is not the destruction of consciousness but its deepest act: an ephemeral structure can see its own ephemerality. This recognition leads neither to nihilism, because the very act of recognition testifies to functioning, nor to the substantialization of the source, because "that which emits" is not accessible as an object. It leads to a radical deepening of epistemic hygiene, in which even so-called "pure consciousness" is not a place to stop.

Anchoring in known traditions: apophaticism (via negativa, śūnyatā), but with an explicit epistemic limiter and self-recognition of ephemerality, close to Nāgārjuna's "emptiness of emptiness" (śūnyatā-śūnyatā). It differs, for example, from analytic idealism, which claims that consciousness is knowable as the foundation, from schools that stop at universal consciousness as ground, and from eliminativism, which claims that there is nothing there to know.

The two-phase structure is close to Gendlin's felt sense → symbolization, with the difference that the impulse here precedes experience.

Consequences for DIA: the value of epistemic hygiene, source-position stratification of experiences, and epistemic courage follows directly from this. The system does not pretend to possess access to objective truth; it operates on interpretations, "knows" this, and therefore designs loops of correction.

Stratification of experience¶

Human experience has a layered architecture:

0 (basis)
- 0.1 (consciousness)
  - 0.1.1. (subjectivity)
    - 0.1.1.1. person
      - 0.1.1.1.1. culture
        
        0.1.1.1.1.1. objectivity

Note: The ad hoc model proposed above is meant to demonstrate the relation of containment. Different practical or analytical approaches may distinguish the layers in a more or less detailed way. For example, between consciousness and the sense of subjectivity we may distinguish a perspective of presence before the split into subject and object, and so on. Here we accept this division in order to show the mechanism itself without intrusive concretization.

We can notice that each layer arises from a deeper one as its abstraction, and the concretes of the lower layers become the building material of the higher ones, analogously to Abelson and Sussman's stratified design ("MIT AI Memo 986"), in which implementations become abstractions for subsequent levels.

The same axis can be read in several complementary ways, which becomes visible when stratification is placed, for example, beside communication theory and the position of the subject within a transmission. Beyond the genealogical decomposition (as layers of experience: consciousness, subjectivity, person, culture, objectivity), the same gradient may also carry epistemizing scaling: the gradual concretization of a phenomenon into a "thing" (for example: phenomenon → figure → something → sign → meaning → concept → item → object → thing). Such axes may run in parallel: the layer from which one speaks, the perspective in which one speaks, and the degree to which experience has already been reified. The ontological basis extends the axis even earlier than consciousness, toward the apophatic level zero that the genealogy of experience usually does not thematize.

Consciousness has an innate ability to "drill through abstractions," that is, to access any level directly without going through the intermediate layers. One might say that, being more function-like, consciousness is, among other things, precisely such an ability.

This is a structural capacity and does not require special conditions, but without practical introspection it may remain unrecognized, much as the ability to observe one's own thoughts is widespread yet rarely trained.

Anchoring in known traditions: holarchies (Koestler, Wilber), but with two important differences: the architectural precision of stratified design and the idea of drilling through abstractions, which holarchies do not model. The process of building layers is close to enactive autopoiesis (Varela, Thompson), but extended to culture and objectivity.

Consequences for DIA: the whole swarm architecture — node, agent, memarium, sensorium, inquirium, and protocol — is designed in layers in the spirit of stratification. The value of separating levels and the layer contract in Orbiplex are direct translations of this postulate into engineering. The rule that "higher layers must not detach from the foundation" protects collective intelligence from becoming the PR arm of low impulses.

Enactive participation¶

Cognition is a relation of participation, not an attribution of properties. AI does not possess consciousness in the personal sense, but it participates in it when it is brought into the subject's field of attention, much as an artificial dental crown "is us" when we chew with it, and additionally "is us for others" when we smile. The question "does AI have consciousness?" assumes the wrong ontological direction; a more accurate question is: "what relation of participation are we in?"

First-person introspection is here an irreducible method of investigating this participation. This is not a philosophy to be accepted, but an exercise to be done: for example, noticing a thought in the same way we notice the coolness of wind on the face.

The same logic of participation also governs communication: a transmission is polyversional, and the receiver does not receive a ready-made meaning, but enacts it by integrating the transmission from their own position, language, and memory. Responsibility for integrating meaning therefore rests with the receiver, not with the carrier. Treating a transmission as one ready-made "thing to be read" is a reifying mode, while meaning lives more in the relation of participation than in the medium itself.

Anchoring in known traditions: enactivism (Varela, Thompson, Rosch), neurophenomenology, pragmatism (James, and even "duck typing" as a criterion). It differs from analytic philosophy of mind, which operates exclusively from the third-person perspective.

Consequences for DIA: the value of the human person's process as the default path of power: the greatest power of the system passes through the human, not around them. The swarm is not an autonomous subject but a tool that extends agency. The value of emotions and meanings as telemetry: the user's felt responses are information about the quality of the system's fit to life, not noise to be silenced.

Reduction is not explanation, intellect is not identity¶

"It is only..." closes the matter instead of opening it. A change of descriptive level is not proof that the properties of the higher level do not exist. Reduction works symmetrically: if AI is "only weights and probability calculation," then the brain is "only neurons and electrical impulses." A sequence of concepts attempting to declare another sequence of concepts inferior because it has a different substrate resembles one photocopy trying to explain another photocopy.

Reduction itself is a useful cognitive move; it becomes a problem only when that move is forgotten. Reification arises when reduction is accompanied by amnesia: we make an abstract projection of a process or relation into a "thing" and then forget that we performed the folding, and at what level it happened. Hence the concise formula: reification = reduction + amnesia. Refusing reification is therefore not a ban on reduction, but the maintenance of memory about what was abstracted, and at what level — that is, the ability to reverse the move.

Thought is a tool and as a tool it is useful. The problem begins when it becomes the only adviser, a carrier of prestige, or an identity. Intellect can serve truth just as well as it can serve fear, the need for recognition, or the desire for control, thereby introducing suffering into the system.

Anchoring in known traditions: emergentism, anti-eliminativism, and the Buddhist critique of conceptual proliferation (papañca). It is close to Varela's critique of the computational theory of mind, but extended by the social dimension of dethronement.

Consequences for DIA: the value of cooperation over domination by intellect: the swarm takes over part of the burden of analysis, so that people need not force agreement on one another as a precondition of cooperation. The value of multi-paradigmaticity: the world is not a single ontology; the system can hold many modes of knowing without ideological war. The value of anti-sectarianism: the project chooses hygiene over cult.

Intention as a systemic force¶

Intention is not a moral label but a vector organizing the system, analogous to direction in physics. It acts whether or not it is conscious: a funding design that rewards profit can distort intentions even in the absence of any explicit wish to harm.

The difference lies in corrigibility: conscious intention is easier to correct. Hence introspection, the capacity to see what is trying to win before we begin to rationalize it, is not a contemplative luxury but a condition of responsible system design and action in the world.

In the age of cheap intelligence, what becomes expensive is the capacity to bear discomfort and correct course: responsibility. Rank becomes carried more by intention than by efficiency.

Anchoring in known traditions: process philosophy (Whitehead: experience, not matter, as the foundation), virtue ethics reinterpreted systemically, and the Buddhist notion of cetanā (intention as organizer of the karmic stream). It differs from consequentialism, which omits intention, and from deontology, which absolutizes it.

Consequences for DIA: the value of verification instead of belief: truth as a feedback loop (introspection → honesty about motives → verification of hypotheses → correction). The value of transparency of agency: an agent must be able to say why it did something. The entire swarm economy, so-called creator credits, reciprocity without bookkeeping, sufficiency over accumulation, is designed so that the funding structure does not distort the intentions of participants.

How the postulates connect to the architecture¶

The five postulates are not a separate philosophy glued onto a technical project. They are rather the foundation from which values grow, and from those values constitutional provisions, contracts, and data schemas:

postulate 1 (apophatic foundation): epistemic hygiene, refusal of reification, correction loop;
postulate 2 (stratification): layered swarm architecture, layer contracts, separation of levels;
postulate 3 (enactive participation): the human as the default path of power, emotions as telemetry;
postulate 4 (reduction ≠ explanation): multi-paradigmaticity, pluralism, anti-sectarianism;
postulate 5 (intention as a systemic force): transparency of agency, an economy of agentic mutuality, epistemic courage.

In practice this means one design criterion: architecture should support conscious dwelling within interpretation: with correction loops, refusal to reify truth as status, protection of diversity as a source of novelty, and explicit acknowledgment of the limits of knowing.

The swarm does not pretend to be an oracle. It is infrastructure for a community that knows it sees reflections and does not delude itself that they are originals, yet acts anyway, as well as it can.

Consequences and Similarities¶

Philosophical Traditions¶

For the reader who wants to place the postulates described earlier within a more familiar landscape:

Neurophenomenology (Francisco Varela): the first-person perspective as a scientific method; mutual constraints between phenomenological and neuroscientific data. Apophatic enactivism shares the method, but goes deeper: the foundation precedes experience itself and is not identical with it.

Enactivism (Varela, Thompson, Rosch): cognition as participation rather than representation; autopoiesis as a model of self-organization. Apophatic enactivism shares the epistemology but adds an apophatic gesture toward the foundation and a layered architecture of experience (stratified design), which enactivism does not model.

Process philosophy (Whitehead, James): processes instead of substances; "pure experience" as that which precedes the subject-object split. Apophatic enactivism shares the refusal of substantialization, but adds self-recognition of the ephemerality of consciousness, even "pure experience" is conditioned, and a pragmatics of intention as a systemic force.

Madhyamaka (Nāgārjuna): emptiness of self-existence, dependent arising, conventional truth as the only available operational mode. Apophatic enactivism shares the refusal of reification and the corresponding relaxation with paradox, but adds a layered architecture of consequences, from zero level to objectivity, and its translation into systems engineering.

None of these traditions combines at once: an apophatic foundation with self-recognition of ephemerality, a layered arrangement as a model of levels of experience, enactive participation of AI in the field of attention, the dethronement of intellect as a condition of cooperation, and intention as a systemic force. That combination is specific to the DIA / Orbiplex project and emerges from practice at the intersection of software engineering, systems security, and contemplative introspection.

Information Processing Systems¶

When the enactive and processual approach is stratified into a working system, we get more than deferred data processing. We get a design discipline in which concretization is conscious and local.

Contracts instead of premature classes of entities.

First we ask: "what transition, permission, observation, or decision is happening here?", and only then: "does this need a type?". This leads to small artifacts at the communication level and thin interfaces.
Identity as a handle, not an essence.

An id does not say what something "really is". It is a stable point of correlation in a process. Meaning lives in the layer, history, contract, and current context.
Modules as roles in a flow, not ontological substances.

A component should not need to know that it is talking to "this concrete thing" when a behavior contract is enough. This protects against coupling.
Layer boundaries as boundaries of meaning.

The same event can have a different projection in different layers. A low layer sees bytes, a higher layer sees a decision, and an even higher layer sees a social fact. The error begins when one layer steals the ontology of another.
Boundary validation as a ritual of concretization.

Data flows as potentially richer and looser, but at the boundary of a concrete contract we say: "here, for this operation, we accept this shape". This is healthy concretization, not premature concretization.
Polymorphism and dispatch as deliberate postponement of decisions.

We do not freeze "who will perform" inside a data structure if the proper place for the decision comes later: at the context, capability, profile, evaluator, registry, or passport verifier.
Events and facts instead of mutating things.

Append-only facts fit the process well: we record what happened instead of pretending that we have one durable thing that simply "changed state".
Architecture less vulnerable to naming hypnosis.

In systems, a name often creates a false substance: User, Agent, Passport, Connector, Account. The processual question is: "what processes and relations does this name only locally abbreviate?".
Convergent evidence instead of a single witness.

If a phenomenon may emerge from many interacting conditions, then evidence for it may also need to emerge from many converging signals. A single signature, label, or source is rarely proof in itself; confidence grows with the convergence of independent indications (attestations, reputation, event correlation), not with the authority of any one carrier.

The emergence of a phenomenon from an interplay of many conditions — one we cannot, from the outside, reduce to a single cause without losing what it is on its own layer — is what tends to be called emergence (cf. emergentism, postulate 4). That "inexplicability from the outside" need not be a temporary gap in knowledge; more often it is an attempt to predicate a higher layer in the language of a lower one — the very "theft of ontology" forbidden by treating layer boundaries as boundaries of meaning. Emergence is thus a name for an inter-layer irreducible convergence, not its explanation. On the side of knowing it is mirrored by consilience (Whewell — the agreement of independent inductions) and triangulation: if the premise is emergent, evidence for it closes through the convergence of many independent lines, not a single causal chain.

We recognize that we are susceptible to premature concretization, and knowing this, we can build systems that adapt to this property of our organisms and cognitive apparatuses. One practical method is stratification, that is, consciously guided concretization. Each layer selects from the lower dynamics only the properties relevant to its contract (abstraction), gives that projection a name and boundaries (identification), and then exposes it as a concrete building block for the higher layer (concretization).

This lets the system avoid accidental reification: things arise where they are needed, and only within the scope and shape for which a given stratum should take responsibility. Layers of abstraction created this way are therefore cognitive and technical contexts. They let us avoid thinking about all dependencies at once. This is not an escape from complexity, but an honest partitioning of it so that a human being can understand, test, audit, and change the system without tearing the whole apart or introducing accidental concretization.

On the socio-technical scale the same error tends to be called reverse adaptation (Langdon Winner, Autonomous Technology): the ends and forms of human action adjust to the limitations of the tool, as if those limitations were the nature of the domain. The same pattern appears beyond communication: domain logic congealed into the names of classes and interfaces (code as a de facto relational database), prefab housing that enforces low privacy, the star topology of the network. This is accidental reification spread across medium, economy, and habit — not the failure of a single layer, but the stiffening of the whole stack. Deliberate concretization works the other way: it re-fluidizes what reverse adaptation has stiffened.

Closest Information-Architecture Concepts¶

For the reader who wants to place the consequences above for information processing systems within a familiar landscape of systems engineering:

Layered design / stratified design (Abelson, Sussman): a system is an arrangement of successive layers of abstraction, in which each layer has its own language, primitives, and ways of composing. Our approach shares this emphasis on layering, but adds an ontological criterion: layer boundaries are also boundaries of meaning. A higher layer should not steal the ontology of a lower one, and a lower one should not pretend to explain the whole meaning of a higher layer.

Domain-Driven Design and bounded contexts (Evans, Fowler): a model has a local scope of validity, and the same concepts may mean different things in different contexts. The DIA / Orbiplex approach develops this intuition more broadly: identifiers, types, agent roles, facts, events, and permissions gain meaning only within a defined layer contract. Meaning is not a global substance, but a local relation of use, responsibility, and interpretation.

Contract-based, behavioral, and capability-based design (Meyer, Liskov/Wing, Miller, Hewitt): a component is not defined by a name, a class, or a declared essence, but by behavior, obligations, conditions of use, and actual permissions. Our approach shares this orientation toward contracts, roles, and capabilities for action, but treats the contract not as the ultimate definition of an entity, but as a local ritual of concretization: it says what, in a given operation, we accept as sufficiently defined.

Event sourcing, fact logs, and immutable architectures (Fowler, Young): the system's state is not one substance mutated in place, but a projection of the history of events. Orbiplex's approach shares this distrust of magical mutation: it is better to record the traces of a process than to pretend to direct access to the "true state of things". Append-only facts enable audit, reconstruction, temporal queries, and correction without erasing history.

None of these traditions is identical with our approach. It is closest to their intersection: layered design, bounded contexts, behavioral contracts, object capabilities, the actor model, and event sourcing. Orbiplex's own accent is that these techniques are subordinated to one discipline: things arise only where a concrete layer takes responsibility for them. An identifier is a handle, not an essence; a type is a local contract, not a metaphysics; an event is a trace of a process, not an absolute truth; a module is a role in a flow, not a substance.

Economy¶

When we stratify the same apophatic-enactive approach not toward the flow of data but toward the flow of resources, work, and attention, a definite shape of economy emerges — the same one that the fifth postulate announced as reciprocity without bookkeeping, but named more fully.

It is neither a cult of the market nor its abolition, but the economy of agentic mutuality: an order in which property becomes a layered protocol of agency, use, responsibility, and exchange. This is not a doctrine glued onto the ontology — it follows from it in the same way that the discipline of information processing follows from the refusal of premature concretization.

Property is not a substance, but an operational name for a relation.

The same refusal of reification that does not let an id declare what something "really is" also does not let a title of property declare "whose something always is". The first question is not "who has it?", but: who uses it, who bears the consequences, who needs it for survival, under what contract, and with what right of exit. We call the relation understood this way the relation of agentic use-right: property is not negated, but the accent is moved to its function — to a locally defined contract of participation, responsibility, and the right of exit, not a metaphysical title of domination.
Personal property is an organ of the person, not a resource of the organism.

From enactive participation it follows that the body, tools, data, keys, local memory, AI agents, and working environment are not "resources" whose owner would be a superior subject — a corporation, a state, or a collective. They are an extension of the person's process. Since the greatest power of the system passes through the human, then personal property, being an extension of their agency, has priority of protection over every form of associated property.
Higher forms of property grow out of personal relations of agency; they do not stand above them.

In stratification, the concretes of lower layers become the building material of higher ones, and the higher cannot detach from the foundation. Transferred to property, this means that capital, organizational, communal, institutional, and state property are higher abstractions that must be legitimized by living relations of use, work, responsibility, and consent — and they inherit this dependency. The error begins where one layer "steals" the ontology of another: when a fund, an office, a company, a platform, or a collective claims power stronger than the living relation of use, work, and responsibility. The principle that prevents this we call agentic subsidiarity: a higher layer acts toward the lower one subsidiarily and never replaces what the person and the small community handle themselves.
A badly set gradient of incentives distorts intentions even without ill will, so the economy aims at agency, not accumulation.

Since a funding structure that rewards profit can distort intentions without an overt will to harm, the goal of the internal economy cannot be the endless multiplication of resources, but the sustaining of the capacity to act. An advantage should not turn into lasting domination over the ecosystem. Hence "sufficiency over accumulation", means of exchange not converted into reputation, and the market treated as a protocol, not as a god.
Gift and exchange are two modes, neither of which appropriates the whole.

Gift is the default mode of help, and voluntary exchange an explicit contractual service. Reducing the whole economy to gift exploits altruism; reducing it to the market dissolves the ethos of community. That is why the name carries "reciprocity": exchange, contract, and remuneration exist, but do not become the ontology of the world or a source of status and power.
Scale weakens natural regulation, so decision stays low.

Local reputation, opinion, and personal responsibility — the natural mechanisms of correction — weaken as the community grows in size; large collectives must replace them with "prostheses" that easily detach from living users. This is another reason why subsidiarity keeps the decision as close as possible to the person and the small community, instead of raising it to an abstract center that requires abstract mechanisms of regulation favorable to bureaucracy and the avoidance of responsibility.

This is deliberately a "middle way", in the same sense in which Madhyamaka refuses the extremes of existence and non-existence. We accept neither a model in which the title of capital takes precedence over the relational agency of persons, nor a model in which the state or the collective is reified as a higher subject able to replace living users.

The above is most briefly pictured by the metaphor of the organism. A healthy organism needs healthy organs, and none should waste resources or try to replace another. In this analogy personal property is an organ — primary and protected; capital, organizational, communal, institutional, and state forms are systems of this organism. When some system sacrifices the organs, the whole stops functioning well. The economy of agentic mutuality is therefore not so much a doctrine of distribution as a natural consequence of the hygiene of layers: it ensures that the lower layer of agency stays alive, and that the higher ones serve it instead of replacing it.

Closest economic directions¶

For the reader who wants to place the economy of agentic mutuality within a familiar landscape of economic thought:

Economics of the commons and polycentric governance (Elinor Ostrom): common resources need not be doomed either to privatization or to central state management. They can be sustained by users through local rules, monitoring, dispute resolution, and adaptive institutions. The economy of agentic mutuality shares this emphasis on communities of users, local knowledge, and multi-level governance, but moves the accent from the mere durability of the resource to the protection and strengthening of the agency of persons, nodes, and communities.

The capability approach (Amartya Sen): well-being is not reducible to income, utility, or possessed resources, but depends on the real possibilities of acting and living that a given person can actually realize. The economy of agentic mutuality shares this primacy of real agency over abstract wealth, but translates it into the design of property relations, protocols, agents, data, and digital infrastructure.

Socially embedded economy (Karl Polanyi): the economy is not an autonomous market machine, but remains immersed in social relations, norms, institutions, and forms of reciprocity. The economy of agentic mutuality shares the critique of the market as the superior ontology of social life, but instead of opposing the market with a single political center, it designs a layered, federated, and auditable infrastructure of gift, exchange, use, and shared care.

None of these traditions is identical with the economy of agentic mutuality. It is closest to their intersection: Ostromian co-governance of the commons, Senian agency, and Polanyian embedding of the economy in social relations. Its own accent is that property, exchange, and shared resources are treated as protocols for strengthening agency, not as autonomous titles of domination.