DEV Community

Cover image for Elementary Particles Seen Through Discrete Values: Matter, Forces and Generations on the 16-Cell and the 24-Cell
MORC
MORC

Posted on

Elementary Particles Seen Through Discrete Values: Matter, Forces and Generations on the 16-Cell and the 24-Cell

Introduction

In an earlier article (in Japanese, on Zenn), I placed one generation of quarks and leptons on the eight vertices of the four-dimensional 16-cell. The point there was that a proton cannot be a face of the three-dimensional Sierpinski gasket, yet it becomes a face in four dimensions.

This article continues from there. On the same eight vertices, I ask:

  • Where do the forces (bosons) live?
  • How are the three generations arranged?
  • Where do left and right separate?
  • Where does mass appear?

Everything is counted with integers, addition and comparison only (the numbers shown inside the figures are recounted by the same procedure). Charges are written as integers in units of 1/3 (u = 2, d = −1, e = −3, ν = 0).

This article does not replace any established theory of elementary particles. It is one viewpoint: what the same world looks like when seen only through discrete values and addresses. The counts below are consistent within the geometric construction; what they mean physically is a separate question, and I mark the places where I am only offering a reading.


1. The device: eight vertices and two seats

Write the eight vertices of the 16-cell as (i, s).

  • i = 0 is the lepton axis; i = 1, 2, 3 are the axes of the three colors (red, green, blue)
  • s = +1 is the "up" type, s = −1 the "down" type

A single formula fixes the charge:

$$
Q = \frac{3s + c}{2},\qquad c = -3\ (\text{lepton}),\quad c = +1\ (\text{color})
$$

Vertex + (up type) − (down type)
Axis 0 (lepton) ν: 0 e: −3
Axes 1–3 (color) u: 2 d: −1

Each vertex has a second seat, obtained by flipping the sign of c. I place there the state that I make correspond to the antiparticle: the ν vertex also holds ē (+3), and the u vertex also holds d̄ (+1). This gives 8 vertices × 2 seats = 16 states, all with integer charge. (Charge alone does not make a state an antiparticle; this is a correspondence chosen within the device.)

The + and − ends of the same axis are antipodal and are not joined by an edge. Every other pair of vertices is joined by an edge (24 edges in total).

fig1:The eight vertices of the 16-cell as an octagon(fig1_16cell_en.png)
fig1:The eight vertices of the 16-cell as an octagon


2. Forces live in "moves"

Instead of putting bosons on vertices, I put them on moves from one vertex to another. I sorted every move between the eight vertices by how it moves and which axes it connects, and counted the change in charge.

How it moves Axes Change in charge
To the antipode (+ ↔ −) same axis ±3 only
Edge, same sign color ↔ color 0 only
Edge, same sign lepton ↔ color ±2 only
Edge, opposite sign color ↔ color ±3
Edge, opposite sign lepton ↔ color ±1, ±5
Stay — 0

fig2:Forces live in moves(r1_moves_bosons_en.gif)
fig2:Forces live in moves

From this table, the addresses of the forces can be read off.

W is the move to the antipode.
The move to the antipode changes the charge by 3 (that is, by 1) in all 8 cases (4 axes × 2 seats). It is the same 3 for u ↔ d and for ν ↔ e. The change comes only from the 3s part of the formula, so it stays 3 whatever value c takes. That W acts on leptons and quarks with the same charge comes out of the shape of the device.

The gluon count: color edges plus stays.
There are 6 directed edges that move between color axes while keeping the sign. Adding the 3 moves that stay put on a color vertex gives 9. The stay that weights all three colors equally does not distinguish colors, so subtracting it leaves 8. Every one of these changes the charge by 0. This is a match in number with the eight gluons; matching their actual algebra is a further step.

Edges joining lepton and color.
There are also 6 directed same-sign edges between the lepton axis and the color axes, and each changes the charge by ±2. It is +2 both for ν → u (the + side) and for e → d (the − side). These moves carry a fractional charge, not a multiple of 3.

Every move is a sum of two parts.
For every move, the change in charge splits into

  • a part that moves the sign (W): ±3
  • a part that moves the axis: 0 between colors, ±2 between lepton and color

On same-sign edges the sign part cancels, so the change was always the same on the + side and the − side.

The charges of one generation add up to 0.
The sign part 3s summed over the 8 vertices is 0, and the axis part c summed over the 4 axes is −3 + 1 + 1 + 1 = 0. Adding all the charges of one generation gives 0. As a control, giving the lepton c = +1 as well makes the sum of c equal to 4, and the cancellation breaks.


3. The three generations live inside the 24-cell

The 24 vertices of the 24-cell split exactly into three 16-cells that share no vertex. (Doubling the coordinates: the 8 points ±2eᵢ, the 8 points (±1,±1,±1,±1) with an even number of minus signs, and the 8 with an odd number.) I place one generation in each of the three.

fig3:The 24-cell split into three 16-cells(fig3_24cell_en.png)
fig3:The 24-cell split into three 16-cells

Three points with the same name form a triangle.
I fixed the arrangement of generation 1 and tried every arrangement of generations 2 and 3 (384 × 384 = 147,456; 384 = 2⁴ × 4! is the order of the symmetry group of the 16-cell). Of these, 168 arrangements make "the three points with the same name (for example u, c, t) form a triangle of the 24-cell, for all eight names". As a control, asking for squared distance 8 instead of a triangle gives 0. What the number 168 means is not explained yet; it is simply the result of the exhaustive search.

Moving by a single rotation: 12 on the left, 12 on the right.
Narrowing further to "each generation is reached from the previous one by multiplying all vertices by a single quaternion", 12 arrangements remain for multiplication from the left and 12 for multiplication from the right. The left 12 and the right 12 do not overlap at all. The way generations move also comes in a left family and a right family.


4. Left and right: the residue separates them

The two 1s of Fibonacci are not the same

For neighboring pairs (F(n+1), F(n)) of the Fibonacci sequence, compute

$$
F(n+1)^2 - F(n+1)F(n) - F(n)^2
$$

For every n from 1 to 100 it alternates −1, +1, −1, +1, … Already at the first pair (1, 1) it sits on the −1 side. Step by step the pair hops between two hyperbolas and never settles symmetrically. In B13 I call this leftover the residue (zansa).

As a control, the Lucas numbers starting from 2, 1 give −5, +5, … The residue is ±1 only when the sequence starts from 1, 1.

In four dimensions, left and right are the order of multiplication

The eight vertices of the 16-cell can be read directly as ±1, ±i, ±j, ±k (axis 0 = 1, color axes = i, j, k). Over all 8 × 8 products I compared the result of multiplying from the left with the result of multiplying from the right.

fig4:Multiply from the left, or from the right(r2_left_right_en.gif)
fig4:Multiply from the left, or from the right

Compared set Equal Differ only by sign Other
16-cell (8 points) 40 24 0
Control: 24-cell (24 points) 168 24 384
Control: plane, ±1, ±i 16 0 0

In the 16-cell, the difference between left and right can always be written as a sign (±) alone — the same form as the two-dimensional residue ±1. In the plane the order of multiplication does not matter, so no difference ever appears.

The left–right difference is the W difference

Take the 8 vertices v, the same 8 quaternions as multipliers m, and the 2 seats: 8 × 8 × 2 = 128 cases. For each, compare the change in charge of the left move v → mv with that of the right move v → vm. In 48 of the 128 cases they differ, and whenever they differ the gap is exactly ±3 — the W difference. (Counting mismatches of the products themselves gives the same 48: the 24 pairs that differ by sign land on antipodes, and that shifts the charge by 3 in both seats.)

They differ only when the multiplier is a color axis (i, j, k). Multiplying by the lepton axis (±1) never separates them. Within the device, the distinction between left and right lives on the side of W and color. This is a correspondence inside the model, not a derivation of the Standard Model W boson.


5. The string of tetrahedra closes in four dimensions

A string of regular tetrahedra glued face to face twists by an amount that never divides evenly in three dimensions, so it never closes. In four dimensions it closed in two ways.

Inside the 16-cell: strings of 8 cells.
Following the 16 cells (tetrahedra) of the 16-cell face to face, every string closed after 8 cells (24 such strings). There are 12 ways to split the 16 cells into two disjoint strings. At each step of a string, one of the four vertices is replaced by its antipode on the same axis. Each step of the string is the same move as W.

Inside the 24-cell: a 6-step string across the generations.
I kept multiplying vertices from the left by a quaternion u whose real part is +1/2 (with doubled coordinates, the 8 points (1,±1,±1,±1)). For every such u and every starting vertex, the string

  • closes after 6 steps,
  • runs along edges of the 24-cell at all 6 steps, and
  • always reaches the antipode at step 3.

It passes through generation 1 → 2 → 3 → 1 → 2 → 3.

As a control, a u with real part −1/2 closes after 3 steps but never runs along an edge.

fig5:A string that closes in 4D (r3_strings_en.gif)
fig5:A string that closes in 4D

Starting from u (red), the charges read 2, 2, 2, −1, −1, −1. One lap through the three generations turns it into d across a W; the second lap brings it back to u. One lap flips the sign, two laps restore it — the same form as the Fibonacci residue.


6. Three generations that differ only in orientation

Every string is a regular hexagon

In the 3D figure the hexagon looks distorted, but in four dimensions every string is a regular hexagon. Checking with integers, for every string (8 choices of u × 8 vertices):

  • squared distances are 4 between neighbors, 12 one apart, and 16 opposite (the 1 : 3 : 4 of a regular hexagon)
  • for any three consecutive vertices, "middle = previous + next" (an addition satisfied only by a flat regular hexagon centered at the origin)

The strings make 16 regular hexagons in total, which equals the number found by an exhaustive search for regular hexagons in the 24-cell (16). The strings pass through exactly all of them. Each hexagon takes 2 vertices (an antipodal pair) from each of the three 16-cells.

fig6:A string that closes in 4D, rotating in 3D(r3_strings_3d_en.gif)
fig6:A string that closes in 4D, rotating in 3D

The shape changes with the direction of the shadow

The three 16-cells are a triple of identical shapes in four dimensions. The only difference is which direction in four dimensions each one faces.

Cast a shadow into three dimensions, however, and they separate. Projecting along the direction (1, 1, 1, 1):

Generation Vertices in 4D Shadow in 3D
1 ±2eᵢ a cube (8 separate vertices)
2 (±1,±1,±1,±1), even number of − an octahedron, plus 2 points ±(1,1,1,1) collapsed at the center
3 odd number of − the same cube as generation 1, same size

Only the generation aligned with the direction of the shadow collapses into a different shape.

From here on it is a reading. The three generations, like the electron, muon and tau, share charge and color yet behave differently. Perhaps they are the same thing in four dimensions, seen from three dimensions in different orientations. This article does not address which shadow corresponds to which mass.


7. Where mass might sit

This section is a reading, not a counted result.

Two different geometries sit close together here and should be kept apart: the light cone of spacetime (causal structure), and the mass shell in energy–momentum space. In energy–momentum space (natural units, c = 1), a particle of mass m satisfies E² − p² = m². For m = 0 this is the cone E = |p|, a sharp V; for m > 0 it is a hyperbola, the V with its corner rounded off. Taking the two coordinates along that cone, a = E − p and b = E + p,

$$
a \times b = m^2
$$

A hyperbolic rotation (a change of velocity) leaves this product unchanged. Mass 0 means a or b is 0, that is, the point lies on the cone. Mass is "how much the corner of the cone is rounded", and "the product that does not change under rotation". Plotting E against p (p on the horizontal axis, E on the vertical), the curve E(p) = √(p² + m²) bends at p = 0 with curvature 1/m, the same as the parabola m + p²/2m of non-relativistic kinetic energy.

Applied to the moves of the 16-cell: the massless gluon count runs along edges, while only the massive W moves to the antipode, through the center, off the edges. Whether the reading "a move off the edges ⇔ rounding the corner ⇔ having mass" holds up is one of the next things to count.


8. What is not drawn yet

  • Fixing the generations: even restricted to a single rotation, 12 placements of the lepton axis remain.
  • Separating the photon and the Z: both are on the side of "moves that stay", and I have not decided how to tell them apart.
  • The cell string (8 steps) cannot be made by multiplying by a single quaternion; it is a different rotation from the generation string (6 steps). How the two strings combine is open.
  • Directions of the shadow: I have not yet counted how many kinds of shadow the three generations fall into from directions other than (1, 1, 1, 1).

9. Connections to established language

The counting above is closed inside the device, but in established language it corresponds to the following.

  • Treating the lepton as a fourth color and lining up four axes is the Pati–Salam arrangement. The charge formula has the form Q = T₃ + (B−L)/2. Extending to four axes, the moves between axes number 12 + 4 − 1 = 15, and the ±2 moves joining lepton and color correspond to leptoquarks (charge 2/3).
  • The 8 for gluons is 3 × 3 − 1.
  • The split of the 24-cell into three 16-cells is, on the quaternion side, triality.
  • The split of four-dimensional rotations into left and right multiplication is written SO(4) ≅ SU(2) × SU(2)/±1.
  • The Fibonacci residue ±1 is Cassini's identity; the string of tetrahedra is the Boerdijk–Coxeter helix.

Conclusion

With one generation on the eight vertices and two seats of the 16-cell, and the forces on the moves between vertices, W becomes the move to the antipode (change ±3), the gluon count becomes the color edges plus stays (8), and the moves joining lepton and color become ±2. Every change in charge splits into "a part that moves the sign" and "a part that moves the axis", and the charges of one generation add up to 0.

The three generations sit in the three 16-cells inside the 24-cell. Left and right appeared not as a left and right of shape, but as the order of multiplying quaternions. Their difference has the same sign (±) form as the residue of Fibonacci's 1, 1; in charge it is the ±3 of W, and it lives on the side of color. The string of tetrahedra, which never closes in three dimensions, closes in four dimensions as a regular hexagon that flips after one lap and returns after two. Each hexagon takes two points from each of the three generations, and in four dimensions the three generations do not differ in shape. They differ only in orientation, and only when cast as shadows into three dimensions do they become separate shapes — a cube, or a collapsed octahedron.

Matter on vertices, forces on moves, generations on the triple of 16-cells, left and right in the order of multiplication. Seen through discrete values and addresses alone, this is how far a picture of the elementary particles can be drawn.

Top comments (0)