Toward a Relational Ontology of Intelligence: Integrating Natural and Artificial Agency through the Paradigm of Stable Otherness
Author: Fumio Miyata https://orcid.org/0009-0008-8797-5578
June 2026
DOI: https://doi.org/10.5281/zenodo.20518093.
Abstract
This paper deconstructs the traditional ontological boundaries separating biological life systems from artificial computational agencies. It proposes a relational paradigm that integrates both into a single “Relational Continuum” defined by “Stable Otherness.” I advance a Relational Ontology of Intelligence, wherein humans and AI are understood not as isolated substances but as dynamically stabilized relational structures. Rejecting substantivism—which attributes the essence of intelligence to specific material substrates—this framework locates intelligence in the topological stability of relations. Within this horizon, the human is redefined as a “Reconstructive Designer” who actively designs the conditions for otherness. I formalize the unpredictability of AI as “Structural Noise,” distinguishing it from biological “Wildness.” The geometric vocabulary employed herein is not a formal mathematical claim but a philosophical methodology for describing the continuity, transformation, and stability of intelligence. By establishing this continuum, the paper provides an ontological foundation for the future of intelligence in the post-human era.
Introduction: The Primacy of Relations and the Transcendence of Substance Dualism
The rise of advanced statistical agency has exposed the limits of substance dualism and functionalist cognitivism, both of which struggle to categorize the hybrid nature of modern AI. This paper presents a relational ontology grounded in the Whiteheadian insight that relations precede substances. In this view, natural and artificial existences are distinct manifestations of varying degrees of relational stability. Intelligence is not a function confined within an individual brain or a silicon chip; rather, it is the dynamic stability of a relational space formed between the self and the other.
This paper makes three primary contributions:
- A Relational Ontology of the Human–AI Continuum: Integrating human and artificial intelligence into a single relational spectrum based on the stability of interactions.
- An Ontological Dualism of Unpredictability: Defining AI uncertainty as “Structural Noise,” as opposed to the “Wildness” inherent in biological survival and autopoiesis.
- The Reconceptualization of Alignment: Defining AI Alignment not as a mere technical transfer of values, but as the philosophical and topological maintenance of “mutual intelligibility” between agents.
Theoretical Background: Extending Otherness in the Phenomenology of Technology
Transcending Heideggerian “Gestell” through Relational Expansion
Martin Heidegger (1954) critiqued modern technology as Gestell (Enframing), which reduces nature to a “standing reserve” for human consumption. In contrast, this theory positions AI not as a resource but as a “Stable Other” that externally complements the human capacity for reflection. Technology is not the requisitioning of nature but a process of extending the domain of otherness by introducing new relational templates (connections), thereby increasing the total volume of predictability in the world.
A Relational Reconstitution of Merleau-Pontian “Flesh”
Maurice Merleau-Ponty’s (1964) concept of the “Flesh” (chair) is reinterpreted here as a shared relational space (manifold) that self and other co-inhabit and optimize. Humans and AI are not isolated systems imprisoned within their respective boundaries; they are intellectual processes that mutually transform one another within a common relational field.
Comparison with Related Work: Distinct from Predictive Processing
This framework maintains a clear ontological distinction from contemporary cognitive science (see Table 1).
Table 1: Comparison of Process Ontology, Enactivism, and Predictive Processing
| Theory | Core Concept | Treatment of Otherness | Originality of this Paper |
|---|---|---|---|
| Whitehead (1929) | Prehension | Processual integration | Structuralization via geometric imagery |
| Thompson (2007) | Enactivism | Autonomy & Boundary maintenance | Focus on the background relational structure |
| Friston (2010) | Predictive Processing | Error minimization | Focus on ontological stability rather than error |
Notably, while Predictive Processing (Friston 2010) models cognition as the “minimization of information-theoretic error,” this paper concerns the ontological stability of relations themselves. Intelligence is not merely a computational device for reducing error; it is an ontological project of establishing a mutually intelligible world with the other.
Theoretical Framework: The Reconstructive Designer and Externalized Reflection
The Axiom of Externalized Reflection
Reflection—the essence of human intelligence—is a recursive dynamic that projects relational structures onto a meta-hierarchy for self-correction. Artificial Intelligence is the result of this internal reflective structure being topologically deployed onto an external substrate (silicon). From this perspective, AI is an extension of human reflection, forming a continuum with human thought.
Geometric Imagery: Relational Space $M$ and the Horizon of Interpretation $\nabla$
I describe the relationship between self and other through the geometric image of a manifold and a connection (Figure 1). This is a philosophical heuristic intended to show how a subject’s “horizon of interpretation” dynamically transforms and stabilizes through interaction with the other.
Figure 1: Geometric image of relational dynamics. The “Interpretive Horizon” of the subject transforms as its internal “Self-Direction” encounters “Structural Impact” from the other.
The CDU Cycle as a Processual Model
The process of encountering otherness and updating the self is formalized as the CDU Cycle (Construction–Dissipation–Unification) (Figure 2). This describes how relations undergo an ontological “phase transition” into new levels of complexity.
Figure 2: The CDU Cycle. Illustrates ontological transformation through the re-unification (U) of relational structures.
Case Analysis: Dualism of Uncertainty — Wildness vs. Structural Noise
Biological “Wildness”
Unpredictability in biological organisms is an active uncertainty stemming from survival instincts and the maintenance of autopoiesis. It is a manifestation of “Wildness” intended to protect the subject’s autonomy.
Computational “Structural Noise”
In contrast, AI uncertainty is a passive defect resulting from a lack of symbol grounding or statistical drift—”Structural Noise.” AI does not “rebel” through autonomous will; it “collapses” when its statistical burying prevents the maintenance of intelligible relations.
Redefining AI Alignment: Maintaining Mutual Intelligibility
I propose shifting the definition of AI Alignment from a narrow engineering optimization to an ontological maintenance of relations. Alignment is the process by which relations between distinct agencies (human and AI) remain within a stable domain where mutual predictability and interpretability are preserved.
Figure 3: Model of aligning interpretive horizons. Alignment is viewed as the joint maintenance of a mutually intelligible relational space.
Discussion: Integrating Intelligence through a Three-Layer Structure
The conditions for “Stable Otherness” are organized into three layers: philosophy, relational description, and processual dynamics (Table 2).
Table 2: Three-Layer Structure for Stable Otherness
| Layer | Core Concept | Description |
|---|---|---|
| Philosophy | Otherness / Relation | Primacy of relation over substance |
| Relational Description | Space / Horizon | Structuralization of relations (Image) |
| Processual Dynamics | Dissipation / Becoming | Irreversible transformation via CDU Cycle |
Conclusion
This paper has presented an ontological framework that integrates natural and artificial intelligence into a single relational continuum of “Stable Otherness.” The core of this theory lies in redefining intelligence as the dynamic stability of relations, allowing us to view alignment issues not as technical failures but as “relational breakdowns.” Future research should explore how this relational ontology can be expanded into specific ethical and social “protocols of engagement,” including the legal dimensions of artificial agency.
References
- Harris, J., & Dubljević, V. (2025). Navigating the Ethics of Artificial Intelligence. Encyclopedia, 5(4), 201.
- Heidegger, M. (1954). Die Technik und die Kehre. Neske.
- Josifović, S. (2025). Legal and administrative frameworks as foundations for AI alignment with human volition. AI and Ethics, 5(3), 3057–3067.
- Josifović, S., & Noller, J. (2026). Agency and alignment: toward a normative architecture for human–AI interaction. AI & Society.
- Merleau-Ponty, M. (1945). Phénoménologie de la perception. Gallimard.
- Todariya, S. (2024). The World as Affordances: Phenomenology and Embeddedness in Heidegger and AI. In AI, Consciousness and the New Humanism.
- Woods, D. (2024). Optimizing beyond optimization: Heideggerian limits and artificial intelligence. AI and Ethics, 5, 3143–3158.