On Information Physics: From State Distinguishability to Causal Accumulation
Human cognitive architecture mirrors the mathematics of this note because the human brain is an embeddedphysical transducer forged under the invariant laws of the environment it evolved to predict. Human cognition is anembodied physical subsystem of nature, and it has no Markovian closure; a physical universe containing an evolved,non-Markovian observer cannot itself be fundamentally Markovian. Markovian dynamics is therefore confined totwo limits, forgetful and faithful; conservation and the Landauer bound leave a closed system only the forgetful one,and physical memory is lossy, neither faithful to every history nor forgetful of all of them. We begin by defininginformation as the distinguishability of physical states and causality as change in physical state, including its rate ofchange. The fundamental theorem of calculus connects instantaneous change to accumulated change, but endpointdisplacement alone does not retain a weighted causal history. Causal histories compose, motivating a path-categorydescription; comparison of contributions at different spacetime points additionally requires transport. The de Rhamcomplex places endpoint dependence, topological path dependence, and curvature on three tiers, and situates thehistory functional outside all of them. We then construct the source algebra, which yields two finite-dimensionalconstructions: positive biquaternionic closure with Born-rule trace readout, and Lorentzian determinant geometryon the same matrix carrier. The real forms of that determinant separate the Hermitian closure slice, which carriesMinkowski four-vectors, from the anti-Hermitian generator slice, on whose traceless part the determinant is theKilling form of su(2). Dyads of closure vectors give rank-two tensors that satisfy the weak, null and dominant energyconditions by positivity alone, and these conditions survive causal accumulation automatically for interaction depthαfrac ≥ 1. The proposed Causal Accumulation Law (CAL) then combines these ingredients in a past-supportedintegral. For an action weighted by a positive function, the gradient of the weight adds a friction-type term to theEuler–Lagrange equations; we derive this and propose its identification with accumulated causal weight. A closingsketch states the cognitive-evolutionary hypothesis that motivates the program and the evidence that supports it.This note develops the foundational mathematical structure from which the full field dynamics are intended to bemapped. The construction of those dynamics is an active research effort, not a result asserted here; further fieldequations, dynamical regimes, and gravitational extensions fall beyond the present scope.Keywords: Non-Markovian dynamics Causal Accumulation Law (CAL) Biquaternions Born rule Landauer principle Energy conditions (weak, null, dominant) Killing form, su(2) de Rham complex Path category Lorentzian determinant geometry Fractional calculus Information as state distinguishability Memory kernels Lossy physical memory Dissipative Lagrangian mechanics / weighted action principle Parallel transport and holonomy Embodied cognition Predictive processing Evolutionary epistemology Foundations of physics Physics of information
Authors
- Samuel L Leizerman (ORCID: https://orcid.org/0009-0000-0133-2291)
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-12
- DOI
- https://doi.org/10.5281/zenodo.24182669
- Primary Topic
- Relativity and Gravitational Theory
- Type
- preprint