What If Particles Are Just Error-Correcting Codes?
An 8-qubit register, three Boolean rules, and every known fermion in the Standard Model
This is Part 4 of “Eight Easy Pieces: The Information Lattice.” In Part 1, we found that the vacuum is full of energy whose total is wrong by 10¹²¹. In Part 2, the Standard Model described only 5% of the universe with 19 unexplained parameters. In Part 3, we traced Wheeler’s “It from Bit” programme through forty years of progress — from black hole entropy to quantum error correction — and found that the one thing missing was a specific code producing specific particles. This article proposes one.
The Specification, Revisited
At the end of Article 3, we listed what any information-theoretic model of particle physics must provide. It is worth repeating these requirements, because the structure we are about to describe will be judged against them.
The model must be discrete — built from bits or qubits, not continuous fields. It must be three-dimensional. It must be error-correcting, with enough redundancy to explain why particles are stable. It must produce the correct spectrum — 45 known fermions in three generations, with the right charges, colours, and chiralities. It must include a gate that reproduces the known particle interactions. And it should have as few free parameters as possible.
These are severe constraints. Let us see how far we can get with the simplest possible starting point.
Eight Bits
Suppose a fundamental particle is represented by a register of 8 binary digits — a byte. Each bit can be 0 or 1. With 8 bits, there are 2⁸ = 256 possible combinations.
We assign each bit a label corresponding to a known quantum number of the Standard Model:
G₀ and G₁ determine the generation — which “family” the particle belongs to. The Standard Model has three generations (electron/muon/tau, up-charm-top, down-strange-bottom). Two bits can encode four combinations: (0,0), (0,1), (1,0), and (1,1). We will need one rule to trim this to three.
LQ is the lepton/quark flag — a single bit distinguishing leptons (electrons, neutrinos) from quarks (up, down, charm, strange, top, bottom).
C₀ and C₁ encode the colour charge — the property that governs the strong nuclear force. Quarks come in three “colours” (an unfortunate but entrenched terminology — it has nothing to do with visual colour). Two bits can encode three non-zero combinations: (1,0), (0,1), and (1,1). Leptons are colourless, which corresponds to (0,0).
I₃ is the isospin — distinguishing “up-type” from “down-type” within each family. Up quark versus down quark. Electron versus neutrino. A single bit toggle.
χ is the chirality — left-handed or right-handed. This is the property that makes the weak nuclear force so peculiar: it only affects left-handed particles.
W is the weak charge — whether the particle participates in weak interactions.
At this point we have done nothing clever. We have simply listed the known quantum numbers of the Standard Model and assigned one or two bits to each. Any physicist could do the same, and most would shrug. The quantum numbers are well known. Writing them as bits is just a notation.
The insight comes from what happens when you impose constraints.
Three Rules
Of the 256 possible 8-bit strings, the vast majority correspond to no known particle. An electron with colour charge, a neutrino with electric charge, a quark with no colour — none of these exist in nature. The Standard Model enforces their absence through the structure of its gauge group, SU(3) × SU(2) × U(1), which is a sophisticated piece of continuous mathematics.
We replace that continuous mathematics with three Boolean rules.
Rule 1: No Fourth Generation
G₀ · G₁ ≠ 1
The two generation bits cannot both be 1 simultaneously. This leaves three valid combinations — (0,0), (0,1), and (1,0) — corresponding to exactly three generations of matter.
The Standard Model observes three generations but does not explain why. Experiments at CERN’s Large Electron-Positron Collider showed that there are exactly three light neutrino species, but this is a measurement, not a derivation. Nothing in the gauge group SU(3) × SU(2) × U(1) forbids a fourth generation. It just doesn’t exist.
Here, the absence of a fourth generation is not a mystery. It is a one-line constraint on two bits. Whether this rule is fundamental or itself emerges from a deeper principle is a question we will return to in Article 5.
Rule 2: Chirality Locks to Weak Charge
W = χ
The weak charge bit must always equal the chirality bit. If a particle is left-handed (χ = 0), its weak charge is 0. If right-handed (χ = 1), its weak charge is 1.
This single equation has an extraordinary consequence. The weak nuclear force — mediated by the W and Z bosons — only affects particles whose weak charge is “active.” By locking weak charge to chirality, Rule 2 ensures that only left-handed particles participate in weak interactions. Right-handed particles are automatically excluded.
In 1957, Chien-Shiung Wu demonstrated experimentally that the weak force violates parity — it distinguishes left from right. This discovery shocked the physics community. The Standard Model accommodates it by building the SU(2) gauge symmetry to act only on left-handed fermion doublets, but it does not explain why parity is violated. The asymmetry is an input, not an output.
Here, parity violation is the equation W = χ. One bit equals another. The asymmetry between left and right is as simple as a constraint can possibly be.
Rule 3: Colour Means Quark
LQ = 0 implies (C₀, C₁) = (0, 0). LQ = 1 implies (C₀, C₁) ≠ (0, 0).
If a particle is a lepton (LQ = 0), both colour bits must be zero — leptons are colourless. If a particle is a quark (LQ = 1), at least one colour bit must be non-zero — quarks must carry colour.
This separates the world into two sectors: colourless leptons (electrons, neutrinos, muons, taus) and coloured quarks (up, down, charm, strange, top, bottom). In the Standard Model, this separation is achieved by assigning quarks to the fundamental representation of SU(3) and leptons to the trivial representation. It works, but it is imposed by hand — the gauge group is chosen to produce this separation, not derived from anything deeper.
Here, the separation is a Boolean biconditional: LQ is 1 if and only if at least one colour bit is non-zero. Quarks have colour because the code says so. Leptons don’t because the code says so. The “why” is the rule; the rule is one line.
Counting the Survivors
How many of the 256 possible 8-bit strings satisfy all three rules simultaneously?
The arithmetic is straightforward, and the reader is encouraged to verify it independently.
Rule 1 eliminates every string with G₀ = G₁ = 1. That removes one-quarter of the 256 states, leaving 192.
Rule 2 forces W = χ, which means W is no longer a free bit — it is determined by χ. This halves the remaining states to 96.
Rule 3 splits the survivors into two sectors. The lepton sector (LQ = 0) forces both colour bits to zero, leaving only the I₃ and χ bits free — that gives 2 × 2 = 4 lepton states per generation. The quark sector (LQ = 1) requires at least one colour bit non-zero, giving three colour combinations (10, 01, 11) times 2 (I₃) times 2 (χ) = 12 quark states per generation.
Each generation therefore contains 4 + 12 = 16 valid states. Three generations give 3 × 16 = 48 valid codewords.
What Are They?
Here are the 16 states of the first generation (G₀ = 0, G₁ = 0):
Leptons (LQ = 0, colourless):
The left-handed electron neutrino is 00000000. Every bit is zero. It is the simplest possible codeword — the “null state” of the register. The left-handed electron is 00000100 — identical except that the isospin bit I₃ is flipped to 1. The right-handed versions have χ = W = 1: the right-handed neutrino is 00000011 and the right-handed electron is 00000111.
Quarks (LQ = 1, coloured):
A left-handed down quark in “red” (colour code 10) is 00110000. A left-handed up quark in “red” is 00110100 — the same, but with I₃ flipped to 1. Each of three colours (10, 01, 11) times two flavours (up, down) times two chiralities gives 12 quark states.
The second generation (G₀ = 0, G₁ = 1) is identical in structure but produces the muon, muon neutrino, charm, and strange. The third generation (G₀ = 1, G₁ = 0) produces the tau, tau neutrino, top, and bottom.
Of the 48 valid codewords, 45 correspond exactly to the known fermions of the Standard Model. Every particle is present. No particle is missing. No extra particle is predicted that shouldn’t be there — with three notable exceptions.
The Three Strangers
Codewords 3, 19, and 35 — one per generation — are valid under all three rules but correspond to no known Standard Model particle. They are the right-handed neutrinos: ν_eR (00000011), ν_μR (01000011), ν_τR (10000011).
In the classic Standard Model (before the discovery of neutrino oscillations), right-handed neutrinos were excluded by fiat. They were simply declared not to exist. After neutrino oscillations proved that neutrinos have mass — requiring some form of right-handed neutrino to generate that mass through the seesaw mechanism — the situation became murkier. Most extensions of the Standard Model now include right-handed neutrinos, but their properties are unknown and they have never been directly detected.
In this framework, the right-handed neutrinos are not added by hand. They emerge automatically from the same three rules that produce everything else. But they have a remarkable property: they interact with nothing.
Check each interaction channel. Strong force? They are colourless (C₀ = C₁ = 0). Electromagnetic? Their isospin is zero (I₃ = 0). Weak force? They are right-handed (χ = 1), so the gate (described below) doesn’t fire on them. They satisfy every rule but couple to no force. They are dynamically sterile — valid patterns that the universe’s machinery ignores.
This makes them natural dark matter candidates. They have mass (via the seesaw mechanism), they are stable (nothing can decay them), and they are invisible (they interact with nothing except gravity). Whether they constitute the 27% of the universe that is dark matter is an open experimental question — but the framework predicts their existence without any additional assumptions.
The Gate
The three rules define which states are allowed. But particles also transform — a down quark becomes an up quark, an electron becomes a neutrino. Something must govern these transitions.
In quantum computing, state transitions are performed by logic gates. The simplest non-trivial two-qubit gate is the CNOT — the Controlled NOT. It has a control bit and a target bit. If the control is in the active state, the target flips. If the control is inactive, nothing happens.
The gate governing particle transitions in this framework is a zero-controlled CNOT. Its control bit is χ (chirality) and its target bit is I₃ (isospin). It fires when χ = 0 (left-handed) and flips I₃:
When χ = 0 (left-handed): I₃ flips. Up quark becomes down quark. Electron becomes neutrino. Charm becomes strange. Top becomes bottom. Every known weak-force transition.
When χ = 1 (right-handed): nothing happens. The particle is unchanged. Right-handed particles do not participate in weak interactions.
This single gate reproduces the entire weak nuclear force. Not approximately. Not in some limit. Exactly. Every transition the weak force permits, the gate permits. Every transition the weak force forbids, the gate forbids. And it does so for the correct reason — chirality — producing parity violation as a structural consequence rather than an imposed asymmetry.
Furthermore, the gate has specific properties that reproduce the Standard Model’s conservation laws:
It never touches LQ. Therefore lepton number and baryon number are absolutely conserved. A quark cannot become a lepton. A lepton cannot become a quark. Proton decay is structurally impossible — not suppressed, not rare, but forbidden by the wiring of the gate.
It never touches C₀ or C₁. Therefore colour charge is conserved in weak interactions, exactly as observed.
It never touches G₀ or G₁. Therefore generation number is conserved at tree level — the gate does not change an electron into a muon or a down quark into a strange quark. (Generation mixing, which is observed in nature via the CKM and PMNS matrices, enters through higher-order effects of the walk operator, discussed in Article 5.)
Every conservation law of the Standard Model is a bit that the gate cannot reach.
Where Do the Bits Live?
We have a code — 8 bits, 3 rules, 1 gate, 48 valid states matching the Standard Model. But a code needs hardware. Bits need a physical substrate. Where do these 8 bits live?
In quantum computing, qubits are hosted on physical systems: superconducting circuits, trapped ions, photonic modes. Each qubit occupies a specific location in space and interacts with its neighbours through specific connections.
We need a geometry that hosts exactly 8 qubits in a symmetric arrangement, where each qubit has a well-defined neighbourhood and the adjacency structure supports distance-4 error correction (enough to detect and correct single-bit errors, explaining the extraordinary stability of particles).
There is a natural candidate: the regular octahedron.
An octahedron is the dual of a cube — or equivalently, two square-based pyramids joined at their base. It has 6 vertices, 12 edges, and exactly 8 triangular faces. One qubit per face gives exactly the 8-qubit register.
The face-adjacency graph of the octahedron — the graph in which two faces are connected if they share an edge — is Q₃, the three-dimensional Boolean hypercube. This is a well-studied graph in combinatorics: 8 vertices, each with degree 3, and the vertices can be labelled by 3-bit binary strings (000 through 111) such that two vertices are adjacent if and only if their labels differ in exactly one bit.
Q₃ supports the [8,4,4] extended Hamming code — an error-correcting code with 8 physical bits, 4 logical (information-carrying) bits, and minimum distance 4. This is a known optimal code: no smaller code achieves the same error-correction capability. It can detect any single-bit error and correct the most common error patterns. It is, in a precise mathematical sense, the smallest code that does the job.
The octahedron is the unique convex polyhedron whose face-adjacency graph is Q₃. No other 3D solid has this property. The geometry is not chosen from a menu of options — it is forced by the requirements of the code.
Four Antipodal Pairs
An octahedron has 8 faces, which form 4 pairs of opposite (antipodal) faces. When we map the 8 code bits onto the 8 faces, a striking structure emerges in these pairings:
Pair 1: G₀ sits opposite W. Generation is paired with weak charge — the two properties that together determine how a particle decays.
Pair 2: G₁ sits opposite χ. Generation is paired with chirality — the property that determines whether the weak force acts on the particle at all.
Pair 3: LQ sits opposite I₃. The quark/lepton flag sits directly opposite the isospin bit — the bit that determines electric charge. The bit that decides what you are faces the bit that decides what charge you carry.
Pair 4: C₀ sits opposite C₁. The two colour bits occupy maximally separated faces — the two bits governing the strong force are as far apart on the octahedron as geometry allows.
The three rules operate on specific face adjacencies:
Rule 2 (W = χ) links two adjacent faces at Hamming distance 1 — faces that share an octahedral edge. The constraint connects nearby faces.
Rule 3 (LQ = C₀ ∨ C₁) links three faces forming a triangle on Q₃. The constraint spans a local neighbourhood.
The gate (χ controls I₃) connects two faces at Hamming distance 2. The control and target are separated by two steps on Q₃.
The constraints operate locally. The complementary physical roles are held at maximum distance. This is a hallmark of good error-correcting code design: the parity checks are local, but the information they protect is distributed across the full structure.
The Ninth Qubit
The [8,4] code has 4 parity-check bits built into its 8 faces. But the octahedron has one more geometrically distinguished point: its centre. A 9th qubit, sitting at the centroid of the void, carries the global parity of the other 8:
P = G₀ ⊕ G₁ ⊕ LQ ⊕ C₀ ⊕ C₁ ⊕ I₃ ⊕ χ ⊕ W
Because Rule 2 forces χ ⊕ W = 0, this simplifies to P = S ⊕ I₃, where S = G₀ ⊕ G₁ ⊕ LQ ⊕ C₀ ⊕ C₁ is the “structural root” — a fixed quantity determined by the particle’s generation and colour.
Exactly 24 of the 48 valid codewords have P = 0, and 24 have P = 1. The global parity is not a constraint — it is a dynamical variable. The 9th qubit is a genuine independent degree of freedom, extending the code from [8,4] to [9,4] with enhanced error detection.
Every time the CNOT gate fires (flipping I₃ during a weak interaction), the global parity P flips. The centre qubit records the parity of the void’s weak-interaction history — a counter, modulo 2, of how many times the particle has undergone a weak transition.
The number 9 = 8 + 1 (eight faces plus one centre) appears to be significant. The weak mixing angle — the parameter that determines the relative strengths of electromagnetism and the weak force — evaluates to sin²θ_W = 2/9 in this framework, matching the experimental value of 0.2229 ± 0.0004 to within 0.3%. The 9 is not a fitted parameter; it is the count of structural elements per void.
What This Does and Does Not Explain
It is important to be precise about what this structure achieves and where its current limits lie.
What it does:
It reproduces the complete fermion spectrum of the Standard Model from three Boolean constraints on an 8-bit register — 45 active particles plus 3 sterile neutrinos, with no free parameters adjusted to match the observed spectrum. Every quantum number (charge, colour, chirality, isospin, generation) emerges from specific bit positions. Every conservation law (baryon number, lepton number, colour, generation at tree level) corresponds to a bit the gate cannot reach. Parity violation is a one-line equation. The absence of a fourth generation is a one-line inequality. Three dark matter candidates appear without being added by hand.
What it does not (yet) explain:
It does not explain particle masses. The electron and the muon have identical bit patterns except for their generation bits, but the muon is 207 times heavier. The mass hierarchy requires dynamics — the propagation of these patterns through space and their interaction with the vacuum — which is the subject of Article 5.
It does not explain the three rules themselves. Are R1, R2, and R3 fundamental axioms, or do they emerge from something deeper? We will argue in Article 5 that they arise from spontaneous symmetry breaking during the crystallisation of the vacuum, but this remains a conjecture under active investigation.
It does not explain gravity. The octahedral void has a tensor excitation mode (called E_g) with the right quantum numbers for a graviton, but the gravitational coupling vertex has not been computed. This is an open calculation, not a solved problem.
And it does not, by itself, connect to the vacuum. The code describes particles. The vacuum — the “empty” space between particles, full of the mysterious energy we explored in Article 1 — requires the code to be embedded in a three-dimensional lattice. That embedding is the subject of the next article.
The Deeper Question
A sceptical reader might reasonably ask: isn’t this just numerology? You found a way to arrange 8 bits so they match the Standard Model. So what? Maybe someone could do the same with 7 bits or 9 bits or 12, if they tried hard enough.
This objection is worth taking seriously. Here is why we believe the [8,4] code is not arbitrary.
First, the structure is constrained, not fitted. We did not search through thousands of possible rules to find three that work. The rules are the three simplest Boolean constraints on 8 bits that produce a non-trivial error-correcting code. Changing any rule — weakening R1 to allow a fourth generation, breaking R2 to decouple chirality from weak charge, modifying R3 to allow coloured leptons — produces a spectrum that immediately contradicts observation. The code is brittle in the right way: it breaks if you touch it, which is the hallmark of a structure that is doing real work rather than being flexible enough to fit anything.
Second, the code has predictive content beyond what was used to construct it. The sterile right-handed neutrinos were not inputs — they are outputs. The absolute stability of the proton was not an input — it is a consequence of the gate’s wiring. The weak mixing angle (2/9) was not an input — it follows from the count of structural elements. If any of these predictions is falsified (a fourth generation is found, the proton decays, sin²θ_W deviates from 2/9 at tree level), the framework is dead. That vulnerability is what separates a model from numerology.
Third, and most importantly, the code lives on a unique geometry. Among all vertex-transitive graphs that can be realised as face-adjacency graphs of convex 3D polyhedra, contain the 4-cycles required for distance-4 error correction, and support a non-trivial code — Q₃ on the regular octahedron is the unique minimum. There is no other polyhedron that does the job. The tetrahedron has too few faces (4). The cube’s face graph has the wrong distance properties. The dodecahedron and icosahedron have far too many faces for a minimal code. The Petersen graph — a famously optimal graph in combinatorics — fails on two independent counts: it cannot be realised as the face graph of any convex solid, and it contains no 4-cycles, precluding distance-4 error correction.
The octahedron is not selected from a menu. It is the only item on the menu.
Coming Next
Article 5: “The Shape That Builds Itself” — How identical qubits, governed only by energy minimisation and the monogamy of entanglement, spontaneously crystallise into octahedral voids, tile three-dimensional space, and produce the lattice whose vacuum fluctuations resolve the 10¹²¹ catastrophe.
The complete 48-codeword table for all three generations, with gate action and conservation law analysis, is available as a companion PDF on Zenodo.
The supporting research papers:
Dave Elliman is the founder of Neuro-Symbolic Ltd and was a Professor of Computer Science at the University of Nottingham, he has since had a successful research career in industry. His research spans information theory, neuro-symbolic AI, and quantum information.
The title of this series nods to Richard Feynman’s “Six Easy Pieces” (1995). Feynman needed six. The octahedron needs eight.