All cognition begins with the recognition of difference. To perceive is to distinguish — between presence and absence, proximity and distance, self and other. AXES (Axes of Epistemic System) is a formal protocol grounded in this observation: that knowledge, in any mind — human or artificial — is structured as coordinates in a multi-dimensional space of differences. Here, “epistemic systems” does not imply a unified or fixed structure, but rather a plurality of shifting and context-dependent frames of reference. This coordinate space already exists. Large language models process meaning as high-dimensional feature vectors; the geometry of difference is already implicit in their representations. What does not yet exist is a shared common protocol by which these coordinates can be made explicit, compared, and negotiated — between humans, between humans and AI systems, and between AI systems themselves. AXES proposes that common protocol. The absence of such a protocol carries measurable risk. When humans and AI systems operate from divergent epistemic coordinates without making those coordinates visible, misalignment is structural rather than incidental. When AI systems of different architectures collaborate without shared coordinate references, divergence is undetectable. AXES is not a language, ontology, or classification system; it is a reference protocol for externalising and aligning epistemic positions across agents. AXES addresses this gap — not as a solution to misalignment, but as its necessary precondition: you cannot correct what you cannot see. By making epistemic coordinates explicit and shared, AXES creates the conditions for deeper mutual recognition across difference — reducing not conflict itself, but the misrecognition that makes conflict irresolvable. This paper does not present a completed system. It proposes a direction, a minimal architecture, and an invitation to collaborate.
Kaori Frizelle (Fri,) studied this question.