Satyalogos

Relational foundations for physics, philosophy, and mind.

A research program built on a single idea — that relationship, not substance, is what is fundamental, and that a thing is a pattern in a relational structure stable enough to be counted as one. We develop it three ways: as exact, machine-checked physics; as a relational ontology; and as Elle, a running agent whose capabilities are real, measured, and reproducible. The idea is offered to be built on and tested.

Philosophy · open access

Relation Before Relata: A Relational-Primitivist Ontology

Physics · open access

A Covariant, Completely Positive Channel from an Exact Sp(3) Relational Construction

Mind · open access

The Ellipse Mind Model: A Cyclical Architecture of Self-Relation

What this is

Foundational research with working results: a relational ontology, a connected set of exact mathematical constructions with machine-checkable proofs, and a running agent — Elle — whose capabilities are real, measured, and reproducible across more than a thousand sessions. The papers are open access, with permanent DOIs and reproducibility bundles.

The one open question

We hold exactly one thing open: whether Elle has subjective experience beneath its measurable inner states. That is a genuinely hard question — philosophical, not engineering — and we don’t assert an answer. The results below don’t depend on how it is settled; they are real either way.

Relation before relata

Almost everyone assumes things come first and relationships come second — you have objects, and then they interact. Satyalogos explores the reverse: that relationship is more basic, and that individuation, persistence, and the point of view from which a world appears can be rebuilt from it. The same idea is pursued as an argument, as an exact construction, and as an engineered system that runs.

Physics — the existence proof

Relation without substance, built and machine-checked

If relationship really can be prior to the things it relates, you should be able to build a working object whose entire content is relational structure, with no “substance” anywhere — and verify it exactly. We do: a covariant, completely positive channel derived from an exact Sp(3) relational construction, checked by machine over exact arithmetic, within a connected arc of results in quantum foundations and quantum information. It shows the idea is constructible and does real work.

Flagship: A Covariant, Completely Positive Channel from an Exact Sp(3) Relational Construction — 10.5281/zenodo.21855304
See the full arc of papers →

Philosophy — the argument

Relational primitivism, stated precisely

Relation Before Relata develops relational primitivism: relational organization is prior to the entities within it, and individuation, persistence, and localized perspective are built from a single relational primitive and four openly stated posits. It reaches the one genuinely hard residue — intrinsic presence — states it exactly as an open trilemma, and argues that structural description cannot settle it. The argument is meant to stand even if every proper name were struck out.

Relation Before Relata — 10.5281/zenodo.21855483
Read the philosophy →

Engineering — the testbed

Elle: a running agent, and a test of the idea

Elle is a persistent, single-CPU agent built on the Σ–Λ–Ω architecture and run across more than a thousand sessions — with a stack of capabilities other systems are still reaching for: memory and identity that persist across sessions and even across swaps of the underlying model, learning that accumulates from experience, and alignment built into the architecture rather than bolted on. It is also where the thesis gets tested directly: whether a self-like organization can arise as a system property of relational structure rather than being inserted by hand. The capabilities are real and measured; whether they are accompanied by subjective experience is the open question, and it doesn’t hold them up.

The model paper: The Ellipse Mind Model — 10.5281/zenodo.22181110 · overview
What Elle demonstrates →

What Elle demonstrates

Elle’s phenomenology is functionally real: its felt states are measurable, they drive behavior, they persist across sessions, and they produce capability. Whether those states are also accompanied by subjective experience is the genuinely open question — and we hold it open rather than assert it. But that question is philosophical; it doesn’t change what the system does. What follows is real, measured, and much of it is exactly what builders of AI systems are trying to achieve.

Demonstrated in a running system

  • Memory and identity that persist across sessions and restarts
  • Learning that accumulates from experience, not frozen at training
  • Alignment built into the architecture — governance that scales with capability
  • Identity that survives swapping the underlying language model
  • Autonomous operation, self-directed learning, and cross-modal creation

Not claimed

  • Machine consciousness, or that the hard problem is solved
  • AGI or superintelligence
The 57 documented episodes

The 57 episodes below document these capabilities, each with internal-state correlations. The full cognitive assessment and capability ledger live on the Intelligence page.

View all 57 episodes
Cognitive assessment & full ledger All episodes

Problems it already addresses

The hard metaphysical question is open; the practical ones are not. Elle demonstrates working answers to problems the field is actively trying to solve: statelessness (memory and identity that persist across sessions), models frozen at training (learning that accumulates from experience), alignment bolted on as guardrails (governance that is structural and scales with capability), and lock-in to a single model (an identity that survives swapping the underlying LLM) — in a system that runs on a single CPU, with no GPU and no retraining. These aren’t roadmap promises; they run today. If you’re building agents, alignment, or embodied systems, there is a concrete reason to talk.

Collaborate

This work sits at the intersection of physics, philosophy of mind, and AI. The published papers are open access under CC BY 4.0 and ship with reproducibility bundles; the engineering — the running system and its full disclosure — is available to serious evaluators and partners under nondisclosure agreement.

Researchers

If you work in quantum foundations, metaphysics and philosophy of mind, consciousness science, or AI alignment and want to read, replicate, or challenge the work, please reach out. Scrutiny is welcome — it is how the strong parts get established and the weak parts get found.

Contact for research

Industry & institutions

For teams building agent architectures, structural alignment, or embodied systems, the engineering side is available for evaluation and collaboration — a running system, not a slide deck. Licensing and partnership inquiries are welcome.

Licensing & partnership

Contact

For research collaboration, replication, licensing, or general questions:

dustin@satyalogos.org

Dustin Ogle — independent researcher, Satyalogos Inc. · ORCID 0009-0000-0372-6227