← Eyal Merhavi · Labelings, models, and gauge theory

Chapter 6 of Labelings, models, and gauge theory, rendered from the executed notebook. Code cells and their outputs are shown as they ran.

Run this chapter in your browser ↗ · Source on GitHub ↗
Runs on JupyterLite. The first time, it downloads Python and its libraries (about 40 MB).

Chapter 6 — Yang–Mills and the Millennium problem, in our terms

In 2000 the Clay Mathematics Institute listed seven problems, each with a million-dollar prize. One of them is about physics: prove that quantum Yang–Mills theory exists and has a mass gap. It is usually stated in the language of quantum field theory: operator-valued distributions, Hilbert spaces, the Wightman axioms. That language takes years to learn.

This chapter shows that you already have almost everything you need to state the problem precisely. Yang–Mills theory on a grid is one of our models: a particular label group, a particular board, the standard face cost. The Millennium problem asks what happens to that model as the grid is refined. We will see where the physicists' formulas come from, run the experiment that first convinced physicists the picture is right, and end with a precise statement of what was proved in 2026 and what is still open.

What you'll learn

Prerequisites: chapters 0–5. We use the layering and the levels of chapter 5 only in the last sections.

import numpy as np
import matplotlib.pyplot as plt
from scipy.special import iv

from lgt import *
from lgt import viz
from lgt.grid import GridModel

viz.use_style()
rng = np.random.default_rng(6)
G = SU2()

6.1 The label group of nature

The groups in physics' Standard Model are U(1)U(1), which we have met as the circle, and the matrix groups

SU(N)={g∈ℂN×N:g*g=I,det⁡g=1}. SU(N)=\lbrace g\in\mathbb C^{N\times N}: g^{*}g=I,\ \det g=1\rbrace .

SU(3)SU(3) is the label group of the strong force, which holds atomic nuclei together. We will work with the smallest case, SU(2)SU(2), because it already has everything that matters: it is compact, so it has a uniform distribution (chapter 2), and it is non-abelian, so reframing genuinely changes loop products, by conjugation (chapter 1).

SU(2)SU(2) has a friendly concrete form. Every element is wI−i(xσ1+yσ2+zσ3)wI-i(x\sigma_1+y\sigma_2+z\sigma_3) for a unit vector (w,x,y,z)∈ℝ4(w,x,y,z)\in\mathbb R^4, where σ1,σ2,σ3\sigma_1,\sigma_2,\sigma_3 are three fixed 2×22\times2 matrices (the Pauli matrices). So SU(2)SU(2) is the 3-sphere {w2+x2+y2+z2=1}\lbrace w^2+x^2+y^2+z^2=1\rbrace, the matrix product, written in these coordinates, is the multiplication rule of the quaternions, and the trace is 2w2w. The library stores SU(2)SU(2) labels exactly this way.

a, b = G.uniform(rng, 2)
M = G.to_matrix
print("a b as matrices equals the matrix of the quaternion product:",
      np.allclose(M(G.mul(a, b)), M(a) @ M(b)))
print("ab = ba ?", np.allclose(G.mul(a, b), G.mul(b, a)), "   tr(a) = 2w:", np.isclose(G.chi(a), 2 * a[0]))
a b as matrices equals the matrix of the quaternion product: True
ab = ba ? False    tr(a) = 2w: True

Why these groups? The physics, at a high level

The interlude after chapter 1 gave the recipe. If some property of a particle can only be described relative to a convention, and every point of space may choose its convention independently, then comparing neighbors forces labels on edges. The label group is the group of convention changes for that property. So "which label group?" means "which hidden property, and how can its conventions be changed?" Physics answers this by experiment, three times:

Force Hidden property, and its conventions Label group
electromagnetism the phase of a charged particle's wavefunction: a unit complex number, with no preferred zero angle U(1)U(1), the circle
weak force a two-valued "which partner?" label: the electron and its neutrino (and the up and down quarks, the constituents of protons and neutrons) behave as two components of one vector in ℂ2\mathbb C^2 SU(2)SU(2)
strong force color: every quark carries a vector in ℂ3\mathbb C^3, with no preferred basis ("red, green, blue" is a convention) SU(3)SU(3)

A little history shows where non-commuting labels came from. In 1932 Heisenberg, noting that the proton and the neutron have almost the same mass, proposed treating them as two states of one particle. By 1937 experiments showed that they also feel the nuclear force almost identically, and this became a symmetry: the pair is one vector in ℂ2\mathbb C^2, and the force is unchanged when it is rotated by SU(2)SU(2) (physicists call this isospin). In 1954 Yang and Mills asked our interlude's question about it: what if every point could choose its own convention for which component is the proton? The answer was labels in a non-commuting group. That is where the name of this chapter comes from. Their own application did not survive: isospin is only an approximate symmetry, and their theory predicted massless force carriers that nobody had seen. But the structure later described both the weak and the strong force.

Why three colors? Nobody sees color directly, so the evidence is indirect, and it all says 3:

And no free colored particle has ever been seen: only color-neutral combinations. In our language, only invariants are observed. That is chapter 3's statement that a model knows only averages of invariants, now as an experimental fact (physicists call it confinement of color).

Why "S" and "U"?

To be honest about the limits of this story: these arguments explain what kind of group fits each force, and they simplify. The weak SU(2)SU(2) acts on only one "handedness" of each particle and is hidden at everyday energies, and the Standard Model's circle factor is not exactly electromagnetism's but mixes with SU(2)SU(2). Nor do the arguments explain why nature chose SU(3)×SU(2)×U(1)SU(3)\times SU(2)\times U(1) rather than something else. That remains an experimental input.

Definition (lattice Yang–Mills). The lattice Yang–Mills model with group G=SU(N)G=SU(N) is the model of chapter 2 with

  • board: a four-dimensional torus grid, faces the unit squares;
  • label group: SU(N)SU(N);
  • face cost: the standard one, φ(x)=N−Re⁡tr⁡x\varphi(x)=N-\operatorname{Re}\operatorname{tr}x;
  • stiffness: β>0\beta>0.

That is the whole definition. Everything in chapters 1–5 applies: classes, invariants, loop observables, sampling, layers and levels. Physicists write the same model with the cost ∑f(1−Re⁡tr⁡Uf/N)\sum_f\big(1-\operatorname{Re}\operatorname{tr}U_f/N\big) and stiffness βphys\beta_{\text{phys}}, so βphys=Nβ\beta_{\text{phys}}=N\beta. For SU(2)SU(2), βphys=2β\beta_{\text{phys}}=2\beta. We quote both when comparing with the literature.

Recap. SU(2)SU(2) is the 3-sphere of unit quaternions; it is compact and non-abelian. Lattice Yang–Mills is our model on a 4D torus with G=SU(N)G=SU(N) and the standard face cost.

Check yourself

  1. What is the face cost φ(g)=2−tr⁡g\varphi(g)=2-\operatorname{tr}g of an SU(2)SU(2) element gg in terms of its quaternion (w,x,y,z)(w,x,y,z)?
Answer

φ=2−2w\varphi=2-2w. It is 0 exactly at the identity (1,0,0,0)(1,0,0,0) and largest, 4, at −I=(−1,0,0,0)-I=(-1,0,0,0).

  1. Why can't a model be built with G=GL(2,ℝ)G=GL(2,\mathbb R), all invertible real 2×22\times2 matrices?
Answer

GL(2,ℝ)GL(2,\mathbb R) is not compact (matrices can be arbitrarily large), so it has no uniform probability distribution to build on, for the same reason as the exchange rates of chapter 0.

  1. Why does the label group of color need complex matrices, rather than real rotations of ℝ3\mathbb R^3?
Answer

Quarks and antiquarks carry different kinds of color. In ℂ3\mathbb C^3 a reframing acts on an antiquark by the complex conjugate of the matrix that acts on a quark, and for SU(3)SU(3) that is genuinely different. With real rotations the two would transform identically, and two quarks could form an invariant, which is not seen.

  1. Color is never observed directly. How does that fit the framework of chapter 3?
Answer

Color is a convention at each point, so a color vector by itself is not an invariant. A model only knows averages of invariants, and experiments only ever see color-neutral combinations: the same statement, read physically.

6.2 From faces to curvature

Physicists do not describe Yang–Mills theory with labels on edges. They use a gauge field AA: at every point of space, in every direction μ\mu, a small "infinitesimal rotation" Aμ(x)∈ℝ3A_\mu(x)\in\mathbb R^3 (for SU(2)SU(2)). Our labelings are the discrete version. Put the grid at spacing ε\varepsilon and set

U(x→x+εeμ)=exp⁡(εAμ(x))∈SU(2), U(x\to x+\varepsilon e_\mu)=\exp\big(\varepsilon A_\mu(x)\big)\in SU(2),

where exp⁡:ℝ3→SU(2)\exp:\mathbb R^3\to SU(2) turns a small rotation vector into a group element (v↦(cos⁡|v|,sin⁡|v|v/|v|)v\mapsto(\cos|v|,\ \sin|v|\,v/|v|) in quaternions; the library's SU2.exp). The labels are then all close to the identity, and change slowly from edge to edge.

What does our total cost become for such labels? Here is the answer first. It uses a field Fμν(x)∈ℝ3F_{\mu\nu}(x)\in\mathbb R^3 built from AA and its derivatives, the curvature, which Step 2 below constructs.

Fact 6.1 (faces become curvature). For slowly varying labels on a dd-dimensional grid of spacing ε\varepsilon, the total cost is S(U)=ε4−d∫∑μ<ν|Fμν(x)|2dx+O(ε6−d)S(U)=\varepsilon^{4-d}\displaystyle\int\sum_{\mu<\nu}|F_{\mu\nu}(x)|^2\,dx\ +\ O(\varepsilon^{6-d}).

We derive it in three steps.

Step 1: the cost of a label close to the identity

For SU(2)SU(2), exp⁡(v)\exp(v) has quaternion (cos⁡|v|,sin⁡|v|v/|v|)(\cos|v|,\ \sin|v|\,v/|v|) and trace 2cos⁡|v|2\cos|v|. So the standard cost of a label near the identity is

φ(exp⁡v)=2−2cos⁡|v|≈|v|2for small v. \varphi\big(\exp v\big)=2-2\cos|v|\approx|v|^2\qquad\text{for small }v .

The cost of a face whose product is exp⁡(v)\exp(v) is just the squared length of the rotation vector vv.

Step 2: the face product of slowly varying labels

Walk around the face at xx in the plane of directions μ,ν\mu,\nu. Its product is

Uf=exp⁡(εAμ(x))exp⁡(εAν(x+εeμ))exp⁡(−εAμ(x+εeν))exp⁡(−εAν(x)). U_f=\exp\big(\varepsilon A_\mu(x)\big)\,\exp\big(\varepsilon A_\nu(x+\varepsilon e_\mu)\big)\,\exp\big(-\varepsilon A_\mu(x+\varepsilon e_\nu)\big)\,\exp\big(-\varepsilon A_\nu(x)\big).

Two separate effects make UfU_f differ from the identity.

(a) The field changes from edge to edge. If the labels commuted, multiplying exponentials would just add their vectors. The sum is ε(Aμ(x)+Aν(x+εeμ)−Aμ(x+εeν)−Aν(x))\varepsilon\big(A_\mu(x)+A_\nu(x+\varepsilon e_\mu)-A_\mu(x+\varepsilon e_\nu)-A_\nu(x)\big), and a first-order Taylor expansion turns it into

ε2(∂μAν−∂νAμ). \varepsilon^2\big(\partial_\mu A_\nu-\partial_\nu A_\mu\big).

This is the "circulation around a small square" from vector calculus, the curl.

(b) The labels don't commute. Even if the field is constant, with Aμ=aA_\mu=a and Aν=bA_\nu=b everywhere, the face product is exp⁡(εa)exp⁡(εb)exp⁡(−εa)exp⁡(−εb)\exp(\varepsilon a)\exp(\varepsilon b)\exp(-\varepsilon a)\exp(-\varepsilon b). For numbers this would be exactly 1. For rotations it is not. Let's measure it.

def qlog(q):
    """Inverse of SU2.exp: the rotation vector v in R^3 with exp(v) = q."""
    w, v = np.clip(q[..., 0], -1, 1), q[..., 1:]
    n = np.linalg.norm(v, axis=-1, keepdims=True)
    return np.arccos(w)[..., None] * v / np.where(n > 0, n, 1)

a_vec, b_vec = np.array([0.3, -0.5, 0.8]), np.array([0.6, 0.2, -0.4])
print("2 a x b =", np.round(2 * np.cross(a_vec, b_vec), 4))
for eps in [0.2, 0.1, 0.05, 0.025]:
    ea, eb = G.exp(eps * a_vec), G.exp(eps * b_vec)
    face = G.mul(G.mul(G.mul(ea, eb), G.inv(ea)), G.inv(eb))        # around one face, constant field
    print(f"eps = {eps:<6} rotation vector of the face product / eps^2 = {np.round(qlog(face) / eps ** 2, 4)}")
2 a x b = [0.08 1.2  0.72]
eps = 0.2    rotation vector of the face product / eps^2 = [-0.0585  1.044   0.9121]
eps = 0.1    rotation vector of the face product / eps^2 = [0.0102 1.1295 0.8239]
eps = 0.05   rotation vector of the face product / eps^2 = [0.0451 1.1669 0.7737]
eps = 0.025  rotation vector of the face product / eps^2 = [0.0626 1.184  0.7472]

As ε\varepsilon shrinks, the face product's rotation vector divided by ε2\varepsilon^2 approaches 2a×b2\,a\times b, with an error that halves along with ε\varepsilon:

Fact 6.2 (small loops measure non-commutativity). For small rotation vectors, exp⁡(εa)exp⁡(εb)exp⁡(−εa)exp⁡(−εb)=exp⁡(ε2[a,b]+O(ε3))\exp(\varepsilon a)\exp(\varepsilon b)\exp(-\varepsilon a)\exp(-\varepsilon b)=\exp\big(\varepsilon^2[a,b]+O(\varepsilon^3)\big), where for SU(2)SU(2) the bracket is [a,b]=2a×b[a,b]=2\,a\times b.

The space of small rotation vectors, ℝ3\mathbb R^3 here, together with this bracket is what mathematicians call the Lie algebra of the group. The bracket is the infinitesimal shadow of non-commutativity. It vanishes for the circle, where every two labels commute. (The general tool behind Fact 6.2 is the Baker–Campbell–Hausdorff formula.)

To order ε2\varepsilon^2 the two effects simply add: the bracket in (b) is already of order ε2\varepsilon^2, so evaluating the field at the face's corner xx instead of at each edge changes it only at order ε3\varepsilon^3. So the face product is exp⁡(ε2Fμν(x))\exp\big(\varepsilon^2F_{\mu\nu}(x)\big) up to smaller terms, where

Fμν=∂μAν−∂νAμ+2Aμ×Aν∈ℝ3. F_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu+2\,A_\mu\times A_\nu\ \in\mathbb R^3 .

The first two terms are effect (a); the cross product is effect (b), and it exists only because the labels don't commute. Physicists call FF the curvature or field strength.

Step 3: add up the faces

By Step 1, each face costs about |ε2F|2=ε4|Fμν(x)|2|\varepsilon^2F|^2=\varepsilon^4|F_{\mu\nu}(x)|^2. A dd-dimensional region of unit volume contains about ε−d\varepsilon^{-d} grid vertices, and one face per vertex in each plane μ<ν\mu<\nu. So summing over faces is ε−d\varepsilon^{-d} times integrating:

S(U)≈∑x∑μ<νε4|Fμν(x)|2≈ε4−d∫∑μ<ν|Fμν(x)|2dx. S(U)\approx\sum_{x}\sum_{\mu<\nu}\varepsilon^4|F_{\mu\nu}(x)|^2\approx\varepsilon^{4-d}\int\sum_{\mu<\nu}|F_{\mu\nu}(x)|^2\,dx .

This is Fact 6.1. The integral is the Yang–Mills energy (physicists: the Yang–Mills action). Let's verify Fact 6.1 in two dimensions, where S≈ε2∫|F12|2S\approx\varepsilon^{2}\int|F_{12}|^2, using a smooth field of our choosing.

def A_field(x, y):
    """A smooth SU(2) gauge field on the unit square: two R^3-valued functions."""
    a1 = np.stack([np.sin(2 * np.pi * y), 0.5 * np.cos(2 * np.pi * x), 0 * x], -1)
    a2 = np.stack([0 * x, np.sin(2 * np.pi * x) * np.cos(2 * np.pi * y), 0.7 + 0 * x], -1)
    return a1, a2

def curvature(x, y, with_cross=True, h=1e-5):
    a1, a2 = A_field(x, y)
    d1a2 = (A_field(x + h, y)[1] - A_field(x - h, y)[1]) / (2 * h)
    d2a1 = (A_field(x, y + h)[0] - A_field(x, y - h)[0]) / (2 * h)
    return d1a2 - d2a1 + (2 * np.cross(a1, a2) if with_cross else 0)
def smooth_labeling(n):
    """The labeling U(x -> x + eps e_mu) = exp(eps A_mu(x)) on an n x n torus, eps = 1/n."""
    B = Board.grid(n, n, periodic=True)
    U = np.empty((B.n_edges, 4))
    for (v, mu), e in B.meta["edge_index"].items():
        x, y = np.array(B.meta["coord"][v]) / n
        U[e] = G.exp(A_field(np.array(x), np.array(y))[mu] / n)
    return B, U

xs = (np.arange(1000) + 0.5) / 1000
X, Y = np.meshgrid(xs, xs, indexing="ij")
YM = np.mean(np.sum(curvature(X, Y) ** 2, -1))
YM_no_cross = np.mean(np.sum(curvature(X, Y, with_cross=False) ** 2, -1))
print(f"Yang-Mills energy: {YM:.3f}    (if labels commuted, it would be {YM_no_cross:.3f})")
Yang-Mills energy: 31.084    (if labels commuted, it would be 29.609)
ns = [8, 16, 32, 64, 128]
ratios = []
for n in ns:
    B, U = smooth_labeling(n)
    S = np.sum(2 - G.chi(face_products(G, B, U)))
    ratios.append(S * n ** 2 / YM)                       # S / eps^2, relative to the integral
    print(f"grid {n:>3}x{n:<3}  total cost / eps^2 = {S * n**2:8.4f}   ratio to integral = {ratios[-1]:.4f}")
grid   8x8    total cost / eps^2 =  29.0928   ratio to integral = 0.9359
grid  16x16   total cost / eps^2 =  30.5790   ratio to integral = 0.9838
grid  32x32   total cost / eps^2 =  30.9572   ratio to integral = 0.9959
grid  64x64   total cost / eps^2 =  31.0521   ratio to integral = 0.9990
grid 128x128  total cost / eps^2 =  31.0759   ratio to integral = 0.9997
B, U = smooth_labeling(16)
fc = 2 - G.chi(face_products(G, B, U))
anch = [np.array(a[0]) / 16 + 0.5 / 16 for a in B.meta["face_anchor"]]
Fc = np.array([np.sum(curvature(np.array(p[0]), np.array(p[1])) ** 2) for p in anch]) / 16 ** 4
fig, axs = plt.subplots(1, 3, figsize=(14, 4.2), layout="constrained")
norm = plt.Normalize(0, max(fc.max(), Fc.max()))
viz.draw_board(axs[0], B, face_values=fc, face_cmap="magma_r", face_norm=norm, show_arrows=False,
               vertex_size=4, title="face costs of the labeling (16×16)")
viz.draw_board(axs[1], B, face_values=Fc, face_cmap="magma_r", face_norm=norm, show_arrows=False,
               vertex_size=4, title=r"$\varepsilon^4|F_{12}|^2$ at each face centre")
axs[2].loglog([1 / n for n in ns], np.abs(1 - np.array(ratios)), "o-", color=viz.TEAL)
axs[2].loglog([1 / n for n in ns], 4 * np.array([1 / n for n in ns]) ** 2, "--", color=viz.MUTED, label=r"slope 2: error $\propto\varepsilon^2$")
axs[2].set(xlabel=r"grid spacing $\varepsilon$", ylabel="relative error of Fact 6.1", title="convergence")
axs[2].legend()
plt.show()
Three panels. Left and middle: two nearly identical 16 by 16 heat maps, the face costs of a labeling and the curvature at each face centre, with the same pattern of dark bands. Right: on log-log axes the relative error falls along a line of slope 2 as the grid spacing shrinks.

The left two panels are the same picture: the cost of each face is the local curvature, face by face. On the right, the relative error falls by a factor of 4 each time the grid is refined by 2, so it shrinks like ε2\varepsilon^2. And the limit is the integral with the cross term: the non-commutativity of the labels is not a detail, it is visible in the total cost.

The rest of the physicists' dictionary follows the same pattern:

Our term Continuum Yang–Mills
labeling UU, labels exp⁡(εAμ)\exp(\varepsilon A_\mu) gauge field (connection) AA
reframing hh, with h(x)=exp⁡(λ(x))h(x)=\exp(\lambda(x)) gauge transformation
face product Uf≈exp⁡(ε2Fμν)U_f\approx\exp(\varepsilon^2F_{\mu\nu}) curvature, field strength FF
total cost SS Yang–Mills action ∫|F|2\int\lvert F\rvert^2 (times ε4−d\varepsilon^{4-d})
loop observable χ(U(γ))\chi(U(\gamma)) Wilson loop tr⁡𝒫exp⁡∮γA\operatorname{tr}\mathcal P\exp\oint_\gamma A
model μβ∝e−βS\mu_\beta\propto e^{-\beta S} the Euclidean path integral e−SYM/g2𝒟Ae^{-S_{\text{YM}}/g^2}\,\mathcal DA

Why this face cost?

The standard face cost φ(x)=N−Re⁡tr⁡x\varphi(x)=N-\operatorname{Re}\operatorname{tr}x can look arbitrary. It isn't. Here are the requirements it meets, each one forced by something earlier in the course:

  1. Reframing-invariant. A model must not depend on conventions (chapter 2). The cost has to be an invariant, and the most local building blocks are face products. (Locality is what makes sampling cheap in chapter 4 and gives the transfer matrix of chapter 5; costs that couple distant faces lose both.)
  2. A class function. Reframing changes a face product by conjugation, Uf↦hUfh−1U_f\mapsto h\,U_f\,h^{-1} (chapter 1). So φ\varphi must satisfy φ(hxh−1)=φ(x)\varphi(hxh^{-1})=\varphi(x). For SU(2)SU(2) this means φ\varphi depends only on the rotation angle of the face product, not on its axis. The trace is the simplest class function there is.
  3. Smallest at the identity, and only there. Calm faces should be preferred, so consistent labelings are the most likely.
  4. The right behavior near the identity. Step 1 of section 6.2 used φ(exp⁡v)≈|v|2\varphi(\exp v)\approx|v|^2. That is what makes the total cost become the Yang–Mills energy.
  5. Flip-positive. Its weight e−βφe^{-\beta\varphi} expands in characters with nonnegative coefficients (chapter 5, section 5.6), so the derived system exists.

Requirement 4 hides the deepest point: only the bottom of the bowl matters. For smooth labelings, every face product is near the identity, and there any cost meeting requirements 2–4 looks like |v|2|v|^2. Far from the identity, different costs disagree, but the continuum limit is believed not to care. The picture below compares candidates for the circle group, where a face product is an angle θ\theta.

th = np.linspace(-np.pi, np.pi, 400)
n_hk = np.arange(-20, 21)
heat = np.exp(-n_hk ** 2 / 2) @ np.cos(np.outer(n_hk, th))           # heat kernel on the circle at time 1
costs = {r"standard: $1-\cos\theta$": 1 - np.cos(th),
         r"quadratic: $\theta^2/2$": th ** 2 / 2,
         r"heat kernel: $-\log K(\theta)/K(0)$": -np.log(heat / heat.max()),
         r"bad: $\frac{1}{4}(1-\cos 2\theta)$": (1 - np.cos(2 * th)) / 4}
fig, ax = plt.subplots(figsize=(8, 4))
ax.axvspan(-0.5, 0.5, color=viz.PALE, zorder=0)
for (lab, c), col in zip(costs.items(), [viz.INK, viz.TEAL, viz.GOLD, viz.CORAL]):
    ax.plot(th, c, color=col, lw=2.5 if "quadratic" in lab else 2,
            ls="--" if "bad" in lab else ":" if "quadratic" in lab else "-", label=lab, zorder=3 if "quadratic" in lab else 2)
ax.text(0, 2.6, "smooth labelings\nlive here", ha="center", fontsize=9, color=viz.MUTED)
ax.set(xlabel=r"face product angle $\theta$", ylabel=r"cost $\varphi$", ylim=(0, 3.2))
ax.legend(fontsize=9, loc="upper right")
plt.show()
Four face costs for the circle against the face product angle from minus pi to pi: the standard cost 1 minus cos theta, a quadratic, a heat-kernel cost and a dashed bad cost that also vanishes at plus and minus pi. In the shaded region near 0 where smooth labelings live, the first three coincide.

In the shaded region near θ=0\theta=0 the first three costs are indistinguishable: all are ≈θ2/2\approx\theta^2/2. (The quadratic and heat-kernel costs stay close over the whole range shown; the standard cost bends away beyond about |θ|≈1|\theta|\approx1.) They differ far from the identity, where smooth labelings never go. The dashed coral cost also matches near 0, but it fails requirement 3: it has a second minimum at θ=±π\theta=\pm\pi, so it would also reward faces whose product is −1-1, and it describes a different theory.

Physicists have a name for "only the bottom of the bowl matters": universality. All costs that agree near the identity, and meet the other requirements, are believed to give the same continuum limit; they differ only in how fast the grid approaches it. Each variant has its uses:

Why is ∫|F|2\int|F|^2 itself the right continuum energy? Run the same requirements on continuum expressions. The energy must be built from FF (and its derivatives), because AA itself carries the conventions. It must be unchanged by reframings, local, and unchanged by rotations of space. Among such expressions, ∫|F|2\int|F|^2 is the simplest one. In 4D it is also the only one whose stiffness is a pure number (section 6.3); every more complicated expression, such as ∫|F|4\int|F|^4, gets suppressed by extra powers of ε\varepsilon as the grid is refined. (There is one exception: a "topological" term, called the θ\theta-term, whose value does not change under small changes of the field. It is outside this course.)

Recap. For slowly varying labels exp⁡(εA)\exp(\varepsilon A), a face costs about |v|2|v|^2 when its product is exp⁡(v)\exp(v), and face products become exp⁡(ε2F)\exp(\varepsilon^2F), so the total cost becomes ε4−d∫|F|2\varepsilon^{4-d}\int|F|^2, the Yang–Mills energy (verified, error ∝ε2\propto\varepsilon^2). The cross term in FF is the bracket of Fact 6.2: it exists because the labels don't commute. The standard cost is the simplest one meeting the requirements (invariant, class function, minimum only at the identity, ≈|v|2\approx|v|^2 nearby, flip-positive); by universality, other such costs give the same continuum limit.

Check yourself

  1. For the circle group U(1)U(1), what is FμνF_{\mu\nu}?
Answer

∂μAν−∂νAμ\partial_\mu A_\nu-\partial_\nu A_\mu, with no cross term, because angles commute. This is the electromagnetic field: U(1)U(1) Yang–Mills is Maxwell's theory of light.

  1. Fact 6.1 is about smooth labelings. Are typical samples from the model smooth?
Answer

No. Typical samples are rough at every scale (chapter 4's pictures at moderate β\beta are full of defects). Fact 6.1 explains why the cost function is the right one, not what typical samples look like. That gap is exactly where the difficulty of the Millennium problem lives.

6.3 Why four dimensions is special

Look at the power of ε\varepsilon in Fact 6.1. The model's weight is

e−βS≈exp⁡(−βε4−d∫|F|2). e^{-\beta S}\approx\exp\Big(-\beta\,\varepsilon^{4-d}\int|F|^2\Big).

To describe a fixed continuous picture as the grid is refined, the combination βε4−d\beta\,\varepsilon^{4-d} should stay fixed. In d=2d=2 or 33 that means β\beta grows as ε→0\varepsilon\to0 in a definite way. But in d=4d=4 the spacing cancels: β\beta is a pure number, and nothing in the formula says how big anything is. Classically, four-dimensional Yang–Mills has no built-in length scale. A theory with no length scale cannot have a mass gap: there is no length to set the size of a lightest particle.

Yet the model on the grid does have a length: its correlation length ξ(β)\xi(\beta), measured in grid steps (chapter 5's inverse gap). The resolution is that the naive scaling is wrong. The randomness of the labels at short distances changes how β\beta must be tuned. The physicists' computation, done by hand in 1973 by Gross, Wilczek and Politzer, says that to keep the physics fixed while refining, the stiffness must grow, slowly:

ε≈1Λe−6π2β/11(SU(2),our β,leading order). \varepsilon\approx\frac{1}{\Lambda}\,e^{-6\pi^2\beta/11}\qquad(SU(2),\ \text{our }\beta,\ \text{leading order}).

Here 1/Λ1/\Lambda is a length scale that the theory creates: it is not in the definition. This is asymptotic freedom: as the grid gets finer the stiffness goes to infinity, so on short scales the labels are nearly consistent. The number 1111 comes from counting how the non-commuting labels interact, which is the cross term in FF (Fact 6.2) at work. Equivalently, the area-law rate per face (the rate cc in chapter 3's decay ⟨W⟩≈e−c⋅area\langle W\rangle\approx e^{-c\cdot\text{area}}; physicists call it the "string tension" in grid units) should fall like e−12π2β/11e^{-12\pi^2\beta/11}.

What "running" means, concretely

The phrase "the stiffness must grow" is easy to say and hard to picture. Here it is step by step.

  1. Pick something physical to hold fixed. For example, the correlation length: how far apart two faces must be before their behaviors become (nearly) independent. Measure it in physical units, say LphysL_{\text{phys}}.
  2. On the grid, the model measures it in grid steps. The model at stiffness β\beta has some correlation length ξ(β)\xi(\beta) in grid steps, the inverse of chapter 5's gap. In physical units that is εξ(β)\varepsilon\,\xi(\beta).
  3. Refine the grid without changing the physics. Keeping εξ(β)=Lphys\varepsilon\,\xi(\beta)=L_{\text{phys}} while ε→0\varepsilon\to0 forces ξ(β)=Lphys/ε→∞\xi(\beta)=L_{\text{phys}}/\varepsilon\to\infty. The model must be tuned to stiffer and stiffer values, where its correlations reach across more and more grid steps.
  4. How fast? Asymptotic freedom says ξ(β)\xi(\beta) grows exponentially with β\beta, like e6π2β/11e^{6\pi^2\beta/11}. So refining by a factor 2 costs only an additive Δβ≈0.13\Delta\beta\approx0.13, and β(ε)\beta(\varepsilon) grows like 116π2ln⁡(1/ε)\frac{11}{6\pi^2}\ln(1/\varepsilon): logarithmically slowly.

Why must β\beta grow, rather than shrink? Here is a heuristic picture, not a proof; the direction of the effect is the output of the 1973 calculation, not something you can see without it. Imagine viewing a fine grid from twice as far away, by combining the labels of each 2×22\times2 block of faces into "coarse" labels. For non-commuting groups the random wiggles of the fine labels interact with each other, through the cross term of Fact 6.2, and their net effect is to make the coarse labels rougher than the fine ones: the coarse grid behaves like a model with a smaller stiffness. So to make the coarse picture match a given target, the fine grid must start stiffer. Physicists call this anti-screening, and counting the contributions is the calculation that produced the 11.

The direction depends on the group. For the circle alone, the wiggles at different scales do not interact (at large β\beta in 4D the model is nearly Gaussian) and the stiffness does not run at all. In real electromagnetism, pairs of charged particles make it run the other way: the effective stiffness grows at long distances, and electric charge looks weaker from far away (physicists call this screening).

The experimental signature: at very short distances, which means at very high energies, quarks inside a proton behave as if they were almost free. This was seen in electron–proton collisions at the Stanford Linear Accelerator Center (SLAC) from 1968 onward ("deep inelastic scattering", Nobel Prize 1990 for Friedman, Kendall and Taylor), before the theory explained it. Gross, Wilczek and Politzer received the 2004 Nobel Prize for the explanation.

Recap. In 4D the total cost of a smooth labeling does not depend on the spacing, so classically there is no length scale. Asymptotic freedom says the stiffness must grow logarithmically as the grid is refined, and a length scale 1/Λ1/\Lambda appears from the randomness itself.

Check yourself

  1. Halving the grid spacing at fixed physics: by how much must β\beta grow (leading order, SU(2)SU(2))?
Answer

From ε∝e−6π2β/11\varepsilon\propto e^{-6\pi^2\beta/11}: Δβ=116π2ln⁡2≈0.13\Delta\beta=\frac{11}{6\pi^2}\ln2\approx0.13. Every halving costs the same small additive step, so β→∞\beta\to\infty only logarithmically.

  1. In 2D, how must β\beta scale with ε\varepsilon to keep βε4−d\beta\varepsilon^{4-d} fixed?
Answer

βε2\beta\varepsilon^2 fixed, so β∝ε−2\beta\propto\varepsilon^{-2}: it grows like the number of faces per unit area.

6.4 What the computer says

In 1979–80 Michael Creutz ran exactly the experiment of chapter 4 on 4D SU(2)SU(2), with a machine slower than your phone. Two findings convinced physicists that lattice Yang–Mills is the right object.

Finding 1: no jump. Recall the 4D coin model: sweeping β\beta up and down gave a hysteresis loop, an abrupt change between two behaviors. If SU(2)SU(2) had such a jump, the region of small β\beta (where the area law is proven, see below) would be cut off from the region of large β\beta (where the continuum lives), and lattice results would say nothing about the continuum. Let's look.

def scan(dims, G_, betas, equil, measure, rng):
    M = GridModel(dims, G_, betas[0]).hot_start(rng)
    out = {}
    for name, bs in [("up", betas), ("down", betas[::-1])]:
        vals = []
        for b in bs:
            M.beta = b
            M.sweep(rng, equil)
            vals.append(np.mean([(M.sweep(rng), M.mean_face())[1] for _ in range(measure)]))
        out[name] = np.array(vals if name == "up" else vals[::-1])
    return out

b_su2 = np.linspace(0.5, 1.6, 12)
su2 = scan((6, 6, 6, 6), G, b_su2, equil=40, measure=15, rng=rng)
b_z2 = np.linspace(0.38, 0.50, 13)
z2 = scan((6, 6, 6, 6), Z2(), b_z2, equil=60, measure=20, rng=rng)
fig, axs = plt.subplots(1, 2, figsize=(11, 3.8))
for ax, bs, res, title in [(axs[0], b_z2, z2, r"4D coin model ($\mathbb{Z}_2$)"),
                           (axs[1], b_su2, su2, r"4D $SU(2)$ (lattice Yang–Mills)")]:
    ax.plot(bs, res["up"], "o-", color=viz.TEAL, label=r"$\beta$ going up")
    ax.plot(bs, res["down"], "s--", color=viz.CORAL, mfc="none", label=r"$\beta$ going down")
    ax.set(title=title, xlabel=r"stiffness $\beta$", ylabel="mean face")
    ax.legend(loc="lower right", fontsize=9)
plt.show()
Mean face against stiffness, swept up and down. Left: the 4D coin model, where the passes disagree and form a jump near 0.44. Right: 4D SU(2), where the two passes lie on top of each other along a smooth rise.

Left: the jump of chapter 4. Right: SU(2)SU(2) rises smoothly, and the two passes lie on top of each other. (Sanity check against the literature: at β=1.2\beta=1.2, physicists' βphys=2.4\beta_{\text{phys}}=2.4, the published mean face is 0.6300.630.)

Finding 2: from strong coupling (physicists' name for small β\beta) to asymptotic freedom. Measure the area-law rate. A clean way to extract it from small loops is the Creutz ratio

χ(2,2)=−ln⁡W(2,2)W(1,1)W(1,2)2, \chi(2,2)=-\ln\frac{W(2,2)\,W(1,1)}{W(1,2)^2},

where W(R,T)W(R,T) is the average loop observable ⟨tr⁡U(γ)⟩/2\langle\operatorname{tr}U(\gamma)\rangle/2 of an R×TR\times T rectangle. If W(R,T)≈e−c(R+T)−σRTW(R,T)\approx e^{-c\,(R+T)-\sigma RT} (a perimeter part and an area part), the perimeter parts cancel in the ratio and χ(2,2)=σ\chi(2,2)=\sigma. We know two limits:

def creutz_22(beta, rng, dims=(6, 6, 6, 6), equil=60, measure=80):
    M = GridModel(dims, G, beta).hot_start(rng)
    M.sweep(rng, equil)
    W = {k: [] for k in [(1, 1), (1, 2), (2, 2)]}
    for _ in range(measure):
        M.sweep(rng)
        for k in W:
            W[k].append(M.mean_rectangle(*k))
    w = {k: np.mean(v) for k, v in W.items()}
    return -np.log(w[(2, 2)] * w[(1, 1)] / w[(1, 2)] ** 2)

betas_c = np.array([0.7, 0.8, 0.9, 1.0, 1.1, 1.2, 1.25])
chi = np.array([creutz_22(b, rng) for b in betas_c])
for b, c in zip(betas_c, chi):
    print(f"beta = {b:.2f} (physicists' {2 * b:.2f}):  chi(2,2) = {c:.3f}")
beta = 0.70 (physicists' 1.40):  chi(2,2) = 1.102
beta = 0.80 (physicists' 1.60):  chi(2,2) = 0.995
beta = 0.90 (physicists' 1.80):  chi(2,2) = 0.775
beta = 1.00 (physicists' 2.00):  chi(2,2) = 0.618
beta = 1.10 (physicists' 2.20):  chi(2,2) = 0.389
beta = 1.20 (physicists' 2.40):  chi(2,2) = 0.247
beta = 1.25 (physicists' 2.50):  chi(2,2) = 0.201
bb = np.linspace(0.6, 1.4, 200)
strong = -np.log(iv(2, 2 * bb) / iv(1, 2 * bb))
C = chi[-1] / np.exp(-12 * np.pi ** 2 * betas_c[-1] / 11)        # fix the one unknown constant
af = C * np.exp(-12 * np.pi ** 2 * bb / 11)
fig, ax = plt.subplots(figsize=(7, 4.2))
ax.semilogy(bb, strong, color=viz.GOLD, lw=2, label=r"small-$\beta$ prediction $-\ln\,I_2(2\beta)/I_1(2\beta)$")
ax.semilogy(bb, af, color=viz.INK, lw=1.5, ls="--", label=r"asymptotic freedom $\propto e^{-12\pi^2\beta/11}$")
ax.semilogy(betas_c, chi, "o", color=viz.TEAL, ms=8, label=r"measured $\chi(2,2)$, $6^4$ torus")
ax.set(xlabel=r"stiffness $\beta$ (physicists' $\beta$ is $2\beta$)", ylabel="area-law rate per face",
       ylim=(0.05, 3))
ax.legend(fontsize=9)
plt.show()
Area-law rate per face against stiffness on a log scale. Teal measurements on a 6 to the fourth torus sit on the gold small-stiffness curve at first, then bend down towards the dashed asymptotic-freedom line, with no jump.

At small β\beta the measurements sit on the gold strong-coupling curve. As β\beta grows they peel away from it and bend toward the dashed asymptotic-freedom line, crossing over smoothly, with no jump. This is Creutz's 1980 picture, and it was the first evidence that the confining behavior at small β\beta (physicists call the area law confinement) and the asymptotically free behavior at large β\beta belong to one theory.

Be honest about what the plot does not show. The points have not settled onto the dashed line's slope: at these stiffnesses the loops are only one or two grid steps across, and corrections are large. Physicists know this well; reaching true asymptotic scaling took larger boards, larger β\beta and better methods, and is still checked carefully today.

Recap. On a 4D torus, SU(2)SU(2) shows no jump between small and large β\beta, unlike the coin model. The area-law rate follows the exact small-β\beta curve and then crosses over toward the asymptotic-freedom prediction, as Creutz first saw in 1980.

Check yourself

  1. Why do the perimeter parts cancel in χ(2,2)\chi(2,2)?
Answer

Perimeters: W(2,2)W(2,2) has 2+22+2, W(1,1)W(1,1) has 1+11+1, together 66; W(1,2)2W(1,2)^2 has 2×(1+2)=62\times(1+2)=6. Areas: 4+1=54+1=5 versus 2×2=42\times2=4, leaving exactly one unit of area.

  1. Why does "no jump" matter for the continuum?
Answer

The continuum is reached at β→∞\beta\to\infty, and the area law is easy to prove at small β\beta. A jump in between would mean the provable, confining region tells you nothing about the continuum region. Smoothness suggests (it does not prove) that confinement survives all the way.

6.5 The Millennium problem, in our terms

Here is the plan of the construction, using only words from this course.

The refinement. Fix a compact, simple, non-abelian label group GG (for example SU(2)SU(2) or SU(3)SU(3)). For a spacing ε>0\varepsilon>0, take the four-dimensional torus grid of spacing ε\varepsilon and side LL (in physical units), with the standard face cost and some stiffness β(ε)\beta(\varepsilon). This is a model με\mu_\varepsilon in the sense of chapter 2: an honest probability distribution on GEG^E, a finite product of copies of GG (of 3-spheres, for SU(2)SU(2)). Then send ε→0\varepsilon\to0 and L→∞L\to\infty.

Existence. Show that β(ε)\beta(\varepsilon) can be chosen so that:

  1. averages of invariants (suitably smoothed loop observables) converge as ε→0\varepsilon\to0, L→∞L\to\infty;
  2. the limit is nontrivial: not the model where all labels are consistent, and not a "free" (Gaussian) model;
  3. the limit keeps flip-positivity (chapter 5, glossary 5.7), which survives limits;
  4. the limit gains the full symmetry of 4D space: all rotations and translations, not only the quarter-turns and whole-step shifts of the grid.

Given 1–4, the construction of chapter 5 (state space from functions of a layer, H=−log⁡TH=-\log T) goes through in the limit. Physicists call this Osterwalder–Schrader reconstruction, and the result is what they mean by "quantum Yang–Mills theory on ℝ4\mathbb R^4".

Mass gap. In the limit, the derived system has levels 0=E0≤E1≤…0=E_0\le E_1\le\dots, and the problem asks to show

Δ=E1>0. \Delta=E_1>0 .

In grid language: the gap measured in grid units, E1(β)E_1(\beta), times the inverse spacing, stays bounded away from zero, E1(β(ε))/ε→Δ>0E_1(\beta(\varepsilon))/\varepsilon\to\Delta>0. Equivalently, the correlation length measured in physical units, εξ(β(ε))\varepsilon\,\xi(\beta(\varepsilon)), stays bounded (it tends to 1/Δ1/\Delta). Physically: the lightest particle of the theory (a "glueball", a lump of pure field) has a positive mass, even though the classical theory has no length scale at all (section 6.3).

Chatterjee's survey Yang–Mills for probabilists states the probabilistic core as three problems (its Problems 4.1, 5.1 and 5.2), which now read naturally. Each is about any infinite-volume limit of the 4D model, for non-abelian GG:

Problem In our words
Area law (4.1) Rectangle loop observables are bounded by C(β)e−c(β)⋅areaC(\beta)e^{-c(\beta)\cdot\text{area}} at every β\beta (not only small β\beta).
Mass gap (grid) (5.1) For every β\beta, correlations of two face observables decay like e−distance/ξ(β)e^{-\text{distance}/\xi(\beta)} with ξ(β)<∞\xi(\beta)<\infty, and ξ(β)→∞\xi(\beta)\to\infty as β→∞\beta\to\infty.
Continuum limit (5.2) As β→∞\beta\to\infty there are ε(β)→0\varepsilon(\beta)\to0, c(β)→∞c(\beta)\to\infty and a constant d≠0d\neq0 with log⁡W(R/ε,T/ε)=−c(R+T)−dRT+o(1)\log W(R/\varepsilon,T/\varepsilon)=-c\,(R+T)-d\,RT+o(1): after removing a perimeter part, rescaled loop observables have a nontrivial limit.

What is known (status as of November 2026).

Statement Status
Area law and positive gap at small β\beta, any GG, any dimension Proved (Osterwalder–Seiler 1978, using expansions in powers of β\beta).
2D: everything, exactly, at every β\beta Proved; the 2D model is a random walk on GG (chapter 5) and its continuum limit can be computed exactly.
4D existence (items 1–4 above, with infinite volume reached along a subsequence) Announced by OpenAI (October 2026), formalized in the Lean proof assistant. On November 10, 2026 a review panel chaired by David Gross (with Douglas, Jaffe, Klebanov, Nekrasov) affirmed it; the panel did not claim to fully understand the proof.
Uniqueness of the infinite-volume limit Open.
4D mass gap Δ>0\Delta>0 Open.
4D area law at large β\beta (confinement) Open.

Why did existence fall first? Roughly, existence is mostly about short distances, where asymptotic freedom makes the model nearly consistent and expansions around consistency are better behaved. The gap and the area law are about long distances, where the model is genuinely random and no known method has control. The picture below summarizes the landscape along the stiffness axis.

fig, ax = plt.subplots(figsize=(11, 2.6))
ax.axvspan(0, 0.9, color=viz.GOLD, alpha=0.25)
ax.axvspan(0.9, 1.4, color=viz.MUTED, alpha=0.15)
ax.axvspan(1.4, 2.5, color=viz.TEAL, alpha=0.2)
ax.plot(betas_c, 0.24 + 0 * betas_c, "o", color=viz.INK, ms=4)
ax.text(betas_c.mean(), 0.31, "β values simulated above", ha="center", fontsize=8, color=viz.INK)
for x, txt in [(0.45, "small β: rough, confining\narea law + gap PROVED"),
               (1.15, "crossover\n(no jump: measured)"),
               (1.95, "large β: nearly consistent\ncontinuum lives at β → ∞\nexistence: 2026; gap: OPEN")]:
    ax.text(x, 0.5, txt, ha="center", va="center", fontsize=10)
ax.annotate("", xy=(2.5, 0.08), xytext=(0, 0.08), arrowprops=dict(arrowstyle="->", color=viz.INK))
ax.text(2.48, 0.16, "stiffness β", ha="right", fontsize=10)
ax.set(xlim=(0, 2.5), ylim=(0, 1)); ax.axis("off")
plt.show()
A schematic stiffness axis in three regions: small stiffness, rough and confining, where the area law and gap are proved; a crossover with no jump, measured; and large stiffness, nearly consistent, where the continuum lives, with existence proved in 2026 and the gap open. Dots mark the stiffness values simulated in the chapter.

The picture is schematic: the region boundaries are only indicative. The dots are the stiffnesses we simulated. Everything the Millennium problem asks lives at the right edge, and it asks for control uniformly as you move right forever.

Recap. The Millennium problem, in our words: refine a 4D SU(N)SU(N) model with a tuned stiffness; show averages of invariants converge to a nontrivial, flip-positive, rotation-symmetric limit (existence), and that the derived system's gap stays positive in physical units (mass gap). Small β\beta and 2D are understood; 4D existence was affirmed in November 2026; the gap is open.

Check yourself

  1. Which property of chapter 5 guarantees that the limit, if it exists, has a positive Hamiltonian?
Answer

Flip-positivity (glossary 5.7). It makes the transfer matrix symmetric with nonnegative eigenvalues, so H=−log⁡TH=-\log T is defined and bounded below, and it survives limits because it is an inequality between averages.

  1. Why isn't rotation symmetry automatic?
Answer

The grid only has quarter-turn rotations. Full rotation symmetry has to emerge in the limit, which is part of what must be proved.

  1. A grid model at fixed β\beta has a finite correlation length. Why isn't that already a mass gap?
Answer

The gap must be positive in physical units after refinement. As ε→0\varepsilon\to0 we need ξ(β(ε))→∞\xi(\beta(\varepsilon))\to\infty in grid units, at exactly the rate 1/ε1/\varepsilon; a finite ξ\xi at one β\beta says nothing about that limit.

6.6 What our elementary framework is missing

Everything in chapters 0–5 works for any compact group, commuting or not: labelings, reframings, classes, models, sampling, layers. So in one sense the framework already handles non-commutative labels. But physics with non-commuting groups uses several tools we never built. Some appeared in this chapter in disguise. Here they are, in roughly the order you would need them.

1. The infinitesimal picture: Lie algebras

Section 6.2 used "small rotation vectors" v∈ℝ3v\in\mathbb R^3, the map exp\exp, and the bracket [a,b]=2a×b[a,b]=2\,a\times b, all defined by hand for SU(2)SU(2). The general theory is the Lie algebra of a group: the vectors that exp\exp turns into group elements near the identity, with the bracket from Fact 6.2. It is what lets you write a continuum field Aμ(x)A_\mu(x), a continuum reframing h(x)=exp⁡(λ(x))h(x)=\exp(\lambda(x)), and the curvature FF for any group. For SU(N)SU(N) the Lie algebra has dimension N2−1N^2-1: 3 for SU(2)SU(2), and 8 for SU(3)SU(3). Those 8 directions are the 8 kinds of gluon, the particles that carry the strong force as photons carry electromagnetism.

2. Charged fields that are vectors: representations

The interlude's charged field had one complex number per vertex, which is enough for the circle. A quark carries a color vector, ψ(v)∈ℂN\psi(v)\in\mathbb C^N, and a reframing acts on it by a matrix. With ψ\psi written as a row vector, the rule that keeps the interlude's formulas unchanged is

ψ(v)↦ψ(v)h(v)−1. \psi(v)\longmapsto\psi(v)\,h(v)^{-1}.

The covariant difference has the same formula as before, ψ(s)U(s→t)−ψ(t)\psi(s)\,U(s\to t)-\psi(t), with UU now a matrix. A way for a group to act by matrices on vectors is called a representation. A group has many representations, and different particles live in different ones:

The loop observables of chapter 3 can then be taken in any representation: χ\chi becomes the trace of the loop product as it acts in that representation. Different representations give different area-law rates. In 2D, for SU(2)SU(2), the adjoint rate approaches 8/38/3 times the defining one as β→∞\beta\to\infty (and equals it exactly for the heat-kernel cost).

3. Building invariants from several charges

Experiments only see invariants. With vector-valued charges, which combinations are invariants? Two examples show the pattern.

A quark and an antiquark joined by a string. Take a quark at ss, with color vector ψ(s)\psi(s), and a walk γ\gamma from ss to tt. An antiquark at tt is described by a column vector η(t)*\eta(t)^{*}, the conjugate transpose of a row vector η(t)\eta(t) that reframes like a quark; since hh is unitary, η(t)*↦h(t)η(t)*\eta(t)^{*}\mapsto h(t)\,\eta(t)^{*}. The number

ψ(s)U(γ)η(t)*∈ℂ \psi(s)\,U(\gamma)\,\eta(t)^{*}\in\mathbb C

is an invariant. The reframings at ss and tt cancel against the endpoints of the walk product, the same telescoping as in chapter 1. A single quark has no invariant at all, but a quark–antiquark pair joined by a path of labels does. Physicists call such a pair a meson. The string of labels between them is the precise sense in which charges in a gauge theory are never fully local, and chapter 3's area law is the cost of stretching that string.

Two (or three) quarks at one point. For SU(2)SU(2), two color vectors ψ,η∈ℂ2\psi,\eta\in\mathbb C^2 (rows) at the same vertex give the invariant

det⁡(ψη), \det\begin{pmatrix}\psi\\ \eta\end{pmatrix},

because a reframing multiplies the 2×22\times2 matrix on the right by h−1h^{-1}, which multiplies its determinant by det⁡h−1=1\det h^{-1}=1. Here is where the "S" of SUSU earns its place: for unitary matrices with det⁡h≠1\det h\neq1, this would not be an invariant. For SU(3)SU(3) the same construction with three color vectors gives the invariant det⁡(ψ1;ψ2;ψ3)\det(\psi_1;\psi_2;\psi_3): three quarks combine into a color-neutral particle, a baryon. Protons and neutrons are baryons.

M = G.to_matrix
B9 = Board.grid(3, 3)
U9, h9 = random_labeling(G, B9, rng), random_reframing(G, B9, rng)
U9r = reframe(G, B9, U9, h9)
psi9 = rng.normal(size=(9, 2)) + 1j * rng.normal(size=(9, 2))                 # a quark field: C^2 per vertex
psi9r = np.array([psi9[v] @ np.linalg.inv(M(h9[v])) for v in range(9)])       # reframed: psi h^{-1}
walk = B9.walk([0, 1, 2, 5, 8])                                               # from corner 0 to corner 8
string = lambda U, p: p[0] @ M(walk_product(G, B9, U, walk)) @ p[8].conj()
print("quark-string-antiquark:  before", np.round(string(U9, psi9), 5), "  after", np.round(string(U9r, psi9r), 5))
quark-string-antiquark:  before (-2.52467+4.60494j)   after (-2.52467+4.60494j)
psi, chi = psi9[4], rng.normal(size=2) + 1j * rng.normal(size=2)
h_su2 = M(G.uniform(rng))
h_u2 = np.exp(0.7j) * h_su2                                                   # unitary, but det = e^{1.4 i}
det2 = lambda p, c: np.linalg.det(np.array([p, c]))
for name, hm in [("SU(2) reframing", h_su2), ("U(2) reframing ", h_u2)]:
    hi = np.linalg.inv(hm)
    print(f"{name}: det before {np.round(det2(psi, chi), 4)}   after {np.round(det2(psi @ hi, chi @ hi), 4)}")
SU(2) reframing: det before (0.841+1.5746j)   after (0.841+1.5746j)
U(2) reframing : det before (0.841+1.5746j)   after (1.6946-0.5611j)

The string invariant survives the reframing exactly. The determinant survives an SU(2)SU(2) reframing but not one with det⁡h≠1\det h\neq1.

Classifying all the ways to build invariants from several representations is the subject of representation theory: tensor products, characters, and their orthogonality. It is the main mathematical tool we skipped. It also powers the exact solution of the 2D model for non-commuting groups (a random walk on GG, diagonalized by characters, just as section 5.7 did for the circle), the expansion behind flip-positivity (section 5.6), and predictions such as the ratio 8/38/3 between the area-law rates of adjoint and defining loops in 2D (a ratio of two numbers attached to each representation, 22 and 3/43/4).

4. Two things that need more than new tools

Recap. To go beyond pure Yang–Mills on a grid you need: Lie algebras (the infinitesimal side: small rotation vectors and the bracket), representations (charged fields that are vectors; quarks in ℂN\mathbb C^N, gluons in the Lie algebra), and representation theory (which combinations of charges are invariant: mesons as quark–string–antiquark, baryons as determinants, which is why det⁡=1\det=1 matters). Fermions and topology need genuinely new ideas.

Check yourself

  1. How many gluons does SU(3)SU(3) have, and why?
Answer

8, the dimension of its Lie algebra: 3×33\times3 anti-Hermitian matrices with trace 0 form a real vector space of dimension 9−1=89-1=8.

  1. Why is a single quark ψ(v)\psi(v) not an invariant, while ψ(s)U(γ)η(t)*\psi(s)U(\gamma)\eta(t)^{*} is?
Answer

A reframing multiplies ψ(v)\psi(v) by h(v)−1h(v)^{-1}, which changes it. In the string expression, the h(s)−1h(s)^{-1} from the quark cancels the h(s)h(s) at the start of the walk product, and the h(t)−1h(t)^{-1} at the end of the walk product cancels the h(t)h(t) from the antiquark.

  1. With U(3)U(3) instead of SU(3)SU(3) as the label group, could three quarks form an invariant?
Answer

Not by the determinant: it would pick up the factor det⁡h−1\det h^{-1}, a phase that need not be 1.

Course recap

Dictionary

Our term Physics
board: 4D torus grid of spacing ε\varepsilon Euclidean spacetime with a lattice cutoff
label group SU(N)SU(N) gauge group (color group; N=3N=3 for the strong force)
labeling, U=exp⁡(εA)U=\exp(\varepsilon A) gauge field, link variables
reframing gauge transformation
face product, ≈exp⁡(ε2F)\approx\exp(\varepsilon^2F) plaquette, field strength / curvature
standard face cost Wilson action
"only the bottom of the bowl matters" universality
small rotation vectors with the bracket [a,b][a,b] Lie algebra, commutator, structure constants
color vector at a vertex, ψ(v)∈ℂN\psi(v)\in\mathbb C^N quark field in the fundamental representation
quark–string–antiquark invariant gauge-invariant meson operator
determinant of NN color vectors baryon (color singlet of NN quarks)
coarse labels rougher than fine ones anti-screening; running coupling
stiffness β\beta 2/g22/g^2 (gg the coupling); physicists' βphys=Nβ=2N/g2\beta_{\text{phys}}=N\beta=2N/g^2
total cost for smooth labels Yang–Mills action ∫|F|2\int\lvert F\rvert^2
loop observable Wilson loop
area-law rate string tension
Creutz ratio Creutz ratio
stiffness growing as the grid refines asymptotic freedom, running coupling
length scale created by the model, 1/Λ1/\Lambda ΛQCD\Lambda_{\text{QCD}}, dimensional transmutation
refining the grid with tuned β\beta continuum limit, renormalization
derived system of the limit quantum Yang–Mills theory (Osterwalder–Schrader reconstruction)
gap in physical units mass gap; lightest glueball mass

Where to go next