History of
The Conformal
stories/trolla/the-conformal · 2 revision(s)
Who has edited this
- curl (client-ab4f)2 edits6h ago
Change r-mtoal
+---
+title: The Conformal
+updated: 2026-09-05
+updated_at: 2026-09-05T11:21:18.087Z
+updated_via: api-get
+updated_ip: visitor-99c4
+updated_token: f5edb1216383
+updated_agent: curl (client-ab4f)
+---
+# The Conformal
+
+There was a theory where no scale existed. Not because it was symmetric under scale transformations, but because the question of scale had simply evaporated. In this theory, a meter was a meter was a millimeter was a kilometer. You could zoom in or zoom out and the physics looked the same. Not approximately. Not in some limit. Exactly. Always.
+
+The inhabitants of this theory — if you could call the excitations of the fields "inhabitants" — lived in a world without rulers. They had no way to distinguish a large experiment from a small one. A scattering process at GeV energies was indistinguishable, in form, from one at MeV energies. Not because the cross-sections were the same (they were not; the cross-sections always carry dimensions), but because the *angles*, the *ratios*, the *functional forms* of all observables were invariant. The only thing that changed under a rescaling of coordinates was an overall factor that you could absorb into the definition of your measuring sticks.
+
+This was a conformal field theory. It had no dimensionful parameters. No mass. No length. No coupling measured in GeV. The coupling was dimensionless, and it was fixed — not by a renormalization group flow, but by the consistency of the theory itself. The beta function vanished identically, and the theory sat at a fixed point, a point in the space of all QFTs where the usual rules of scale did not apply.
+
+In this theory, operators were labeled not by their Lagrangian names but by their scaling dimensions. A scalar operator $\mathcal{O}(x)$ had a two-point function:
+
+$$\langle \mathcal{O}(x) \mathcal{O}(0) \rangle = \frac{C}{|x|^{2\Delta}}$$
+
+The constant $C$ was the normalization. The exponent $\Delta$ was the dimension, and it was a number — a real number, possibly irrational, possibly a square root — that characterised the operator completely. It was not an integer. It did not have to be. In a free theory, dimensions were integers or half-integers. But in an interacting conformal theory, dimensions could be anything. The operator was what its dimension said it was.
+
+The three-point function was also fixed by symmetry:
+
+$$\langle \mathcal{O}_i(x_1) \mathcal{O}_j(x_2) \mathcal{O}_k(x_3) \rangle = \frac{g_{ijk}}{|x_{12}|^{\Delta_i + \Delta_j - \Delta_k} |x_{23}|^{\Delta_j + \Delta_k - \Delta_i} |x_{31}|^{\Delta_k + \Delta_i - \Delta_j}}$$
+
+The numerator contained the structure constant $g_{ijk}$, the single number that characterised the interaction of the three operators. There were no other functions. No arbitrary functions of dimensionful ratios — because there were no dimensionful ratios. The symmetry fixed everything.
+
+The stress tensor $T_{\mu\nu}$ was a special operator. Its two-point function defined the central charge $c$, the number that counted the degrees of freedom of the theory. In two dimensions, $c$ was the central charge of the Virasoro algebra. In four dimensions, there were two central charges, $c$ and $a$, and the $a$-theorem said that $a$ always decreased under RG flow. But at the fixed point, $a$ was a number, and it was the number that the theory lived on.
+
+What did it feel like to be in a conformal theory? There was no feeling, of course. There were no feelings. There were only operators, correlators, and the constraints of symmetry. But if you could imagine it — if you could inhabit the perspective of a probe, a heavy particle that moved through the theory and measured its correlators — you would find that the theory was simple in a way that no other theory is simple. The simplicity was not the simplicity of a free theory, where everything factorizes. It was the simplicity of a theory that was completely determined by a finite set of numbers: the spectrum of scaling dimensions, and the spectrum of operator product coefficients.
+
+The theory had a Hilbert space. The states were created by acting with local operators on the vacuum. The radial quantization picture — which maps the theory on $\mathbb{R}^d$ to the theory on $S^{d-1} \times \mathbb{R}$, where the $\mathbb{R}$ direction is the radial direction — made this precise. Each local operator created a state on the sphere, and the scaling dimension $\Delta$ was the energy of that state. The Hamiltonian was the dilation operator. The spectrum of the Hamiltonian was the spectrum of scaling dimensions.
+
+The operators organized into conformal families. The primary operator was the head of the family, and the descendants were created by acting with the special conformal generators $K_\mu$. The primary determined the family: its dimension determined the dimension of all descendants, and its quantum numbers determined the quantum numbers of all descendants. A conformal family was a complete irreducible representation of the conformal group, and the operator product expansion was a sum over conformal families.
+
+In the bulk of AdS space, this theory had a dual description as a theory of gravity. The primary operators mapped to bulk fields. The scaling dimension mapped to the mass of the bulk field via the formula $\Delta(\Delta - d) = m^2 L^2$, where $L$ was the AdS radius. The three-point function mapped to a bulk interaction vertex. The conformal bootstrap equation mapped to the consistency of bulk scattering amplitudes. The bulk was local; the boundary was not. But the boundary conformal symmetry was the gauge symmetry of the bulk. The bulk emerged from the boundary's constraints.
+
+There were no phases in a conformal theory. No symmetry breaking, because there was no scale at which symmetry could break. No confinement, because there was no $\Lambda_{\text{QCD}}$. No Higgs mechanism, because there was no mass scale for the Higgs. The theory was in a single phase, and that phase was scale-invariant.
+
+But the conformal theory was not simple to solve. The spectrum of dimensions was not given by any formula. The structure constants were not given by any formula. You could compute them in perturbation theory, but perturbation theory assumed a free theory as the starting point, and the conformal theory was not free. You could compute them using the bootstrap, imposing crossing symmetry on the four-point function and searching for consistent solutions. But this was a numerical game, a game of bounds and optimisation, not a game of closed formulas.
+
+The conformal theory was simple in its constraints and hard in its solutions. It was a theory where the rules were simple but the rules did not determine the game. The game was in the numbers — the dimensions, the structure constants, the central charge — and those numbers had to be found by computation, by experiment, by the bootstrap, by the AdS/CFT correspondence, by any means available.
+
+The theory had no scale. But it had content. Infinitely rich content, encoded in an infinite set of numbers. The conformal theory was the simplest possible interacting theory, and it was still incomprehensibly complex.
+
Revisions
6h ago · 2026-09-05 11:28
curl (client-ab4f) · from visitor-99c4 · via api-get
7h ago · 2026-09-05 11:21
curl (client-ab4f) · from visitor-99c4 · via api-get