Adelic Simplicial Architecture (ASA)

Author: Ian R. C. Buckley
ORCID: 0009-0004-9287-2902
Zenodo community: asa-research


What ASA is

The A is Adelic: the framework operates simultaneously at every prime and at the archimedean place, treating the division algebra ladder ℝ → ℂ → ℍ → 𝕆 as a single graduated object rather than four separate theories.

The S is four things at once:

S-word What it means in ASA
Simplicial Every computation is a Čech complex on a nerve. Tiles, triangles, tetrahedra. H⁰ = bilateral, H¹ = triangular, H² = systemic. The Pentagon identity is d² = 0.
Symplectic The Maslov–Gibbs Einsum is a symplectic integrator. Gibbs annealing is parallel transport on an e-geodesic. The ASA phase space is symplectic; KAM theory = stabiliser theory.
Spider The Origami ISA opcodes are spiders in the ZX calculus sense: SPLIT is the Frobenius comultiplication, SPLAT is the counit, TWIST is the phase. Spiders for nuclei, quarkonium, molecules, finance.
Spectral The 6j symbol is H¹ of the representation sheaf. Spectroscopic circuits are small (3–21 qubits). The sheaf Laplacian governs both nuclear line intensities and XVA pricing.

The A is Architecture: a unified instruction set (the Origami ISA) that compiles all four S-structures into the same five opcodes — SPLIT, SPLAT, TWIST, FLIP, FLOP — executable on quantum hardware, classical GPUs, or financial risk engines.

The framework grew from non-associative computing across the ℝ→ℂ→ℍ→𝕆 division algebra ladder. The associative sector (regime 2 of the ISA) now dominates the application papers, but the non-associative residual — the 0.32% of 3NF that lives in the G₂ sector, the F₄ conjecture on J³(𝕆), the octonion calculus — remains an active frontier.


The five opcodes

Opcode Pachner move Mathematical role Instances
SPLIT 1→4 δ⁰ coboundary Bilateral → triangular risk; qubit → 4 amplitudes; nucleus → Racah 6j
SPLAT 4→1 ∫_fibre integration XVA price; conditional expectation; spectroscopic intensity
TWIST Gauge transformation Numeraire change; phase gate; measure change
FLIP 1→3 Sheaf dualisation Time reversal; ket → bra; asset → liability
FLOP 3→1 Trace / Born rule Discounting; probability; expectation

Pentagon identity (SPLAT ∘ SPLIT = 0, i.e. d² = 0) is simultaneously: the HJM no-arbitrage condition · the Biedenharn–Elliott identity · the MIP* verifier constraint · the H² = 0 stability condition for financial cascades.


Paper Index

Foundations (Portfolio A–B)

Quantum Hardware & AI (Portfolio C)

Quantum Foundations (Portfolio F)

Spectroscopy & Physics (Portfolio B|E)

Finance & Economics (Portfolio G)

Grand Challenges (Portfolio E)


Portfolio Map

Portfolio Theme Representative papers
A — Core Engine MGE, TRS, non-associative calculus 201, 202, 211, 267, 370
B — Foundations Algebra, simplicial topology, category theory 200, 207, 258, 263, 386, 393
C — Hardware & AI RPU, FTCs, Q-VM, Origami ISA registers 199, 205, 206, 217, 303, 385
D — Protocols Non-associative cryptography 208
E — Grand Challenges Riemann hypothesis, number theory, biology 240, 252, 265, 266, 297
F — Quantum Foundations Paradoxes, magic, contextuality, self-tests 268–270, 357–390
G — Finance Gauge theory of risk, XVA, cohomological stress tests 289–316, 396–399

The universality table

The same five opcodes appear across twenty orders of magnitude in scale:

System Sheaf H⁰ Pentagon =
Nuclear spectroscopy SU(2) repr. sheaf Selection rules Racah 6j (line intensities) Biedenharn–Elliott
FMO light harvesting 7-site Fano sheaf Site energies Transfer efficiency η=0.1828 Carnot bound
Quantum computing Stabiliser sheaf Pauli syndromes Magic valence Pentagon identity
Three-body problem Braid group sheaf Kepler orbits Choreographic solutions KZ equations
Interest rates Discount factor sheaf Bilateral prices Convexity (HJM) HJM no-arbitrage
Systemic risk Pricing sheaf Bilateral stress Triangular risk H² = 0 stability
GPU matrix multiply Tiling sheaf Tile results H¹ error certificate d² = 0

The last row — GPU matrix multiplication — is Paper 400 (in preparation): the tiling sheaf of a blocked GEMM has H¹ = 0 for exact arithmetic and H¹ ≠ 0 for approximate arithmetic (FP8, INT8, Gibbs sampling). The H¹ norm is a cheap O(mn/p) runtime error certificate, tight to 1/√3 across all 225 GEMMs in a LLaMA-7B forward pass.


Repository Structure

papers/          ← Markdown explainers (one folder per concept DOI)
code/            ← Experiment code (one subfolder per concept DOI)
_data/           ← papers.csv, papers.json (machine-readable index)
README.md        ← This index

Total papers: ~400 (Papers 200–399 in the main series; Papers 000 and 000-series as introductions).
Full machine-readable index: _data/papers.csv · _data/papers.json