Yang–Mills theory is the classical field theory underlying the Standard Model of particle physics: a connection with compact structure group over space-time, with curvature and action . Its quantum version with is quantum chromodynamics, the theory of the strong nuclear force. Experiment and large-scale numerical simulation leave little doubt that this quantum theory exists and has a mass gap: every excitation of the vacuum costs energy at least , which is why the strong force is short-ranged although the classical Yang–Mills equations, like Maxwell's, describe massless waves. There is no mathematical construction of the theory and no mathematical argument for the gap. The Clay Mathematics Institute made this one of its seven Millennium Prize Problems; the official statement is Arthur Jaffe and Edward Witten's Quantum Yang–Mills theory (2000, p. 6):
Yang–Mills Existence and Mass Gap. Prove that for any compact simple gauge group , a non-trivial quantum Yang–Mills theory exists on and has a mass gap . Existence includes establishing axiomatic properties at least as strong as those cited in [Streater–Wightman, Osterwalder–Schrader].
Timeline.
Jaffe–Witten do not fix a definition of "quantum Yang–Mills theory" (they write that one does not have "even a precise definition of quantum gauge theory in four dimensions"), but in §6.5 they single out Wilson's lattice approximation as the route: "One must then verify the existence of limits of appropriate expectations of gauge-invariant observables as the lattice spacing tends to zero and as the volume tends to infinity", with reflection positivity, which "holds for the Wilson approximation", supplying the Hilbert space. This mission states the problem in exactly that form, following Seiler's monograph and Chatterjee's survey. All objects live in the Lean namespace YangMills.
Gauge group. is a compact Hausdorff topological group together with a continuous injective homomorphism into the unitary group of . Such a is a compact Lie group, and every compact Lie group has such a . is a compact simple gauge group (IsCompactSimpleGaugeGroup) when it is connected, non-abelian, and every closed normal subgroup is either finite or all of . For compact connected Lie groups this is equivalent to the Lie algebra being simple; it allows a finite centre (, , , ) and excludes tori, , and all products.
Wilson's lattice gauge theory. Fix a dimension (the goal uses ). Sites are (Site d); the box of radius is (box d R). A gauge field assigns a group element to every edge of the box (Config d R G). The plaquette holonomy is (plaquette), the Wilson action is
and the expectation of an observable at inverse coupling is
where is the product over the edges of the box of the Haar probability measure of . Boundary conditions are free, and link variables of edges outside the box are the junk value .
Dyadic loops and Wilson loops. A dyadic loop (Loop d) is an axis-parallel closed lattice path together with a scale : its integer vertices stand for the physical points . Read at a finer scale (Loop.atScale), the base point is multiplied by and every step is subdivided into unit steps. The Wilson loop is , the trace of the ordered product of the link variables along (wilsonLoop, pathHol), and for a finite family of loops
is its correlation function at lattice spacing in the box of physical half-side . A family is admissible (IsAdmissible) when every loop is simple (non-empty, no vertex and no edge visited twice) and the loops are pairwise disjoint as subsets of .
Symmetries. Dyadic translations (Loop.translate), signed coordinate permutations (Loop.mapCoords, the hypercubic group), forward translation in Euclidean time by (Loop.timeTranslate), and the Osterwalder–Schrader reflection (Loop.reflect): reflect and reverse the orientation, so that for unitary .
A continuum theory is a functional on families of loops. It is a continuum limit of the Wilson theories with gauge group (IsContinuumLimit ρ β L Z W) if there are inverse couplings , boxes and multiplicative renormalization constants for each loop such that
for every admissible family . It is lattice invariant (IsLatticeInvariant) if invariant under all dyadic translations and the hypercubic group; reflection positive (IsReflectionPositive) if is real and for all admissible families lying in ; it has mass gap (HasMassGap W Δ) if for all admissible there is with
the time translation; and it has finite mass (HasFiniteMass) if it does not have mass gap for every : Jaffe–Witten's "the supremum of such is the mass , and we require ".
For every compact simple gauge group there exist , , and a functional on families of loops in such that
This is YangMills.existence_and_mass_gap. It is open; a proof would resolve the Millennium Prize Problem in the lattice formulation of Jaffe–Witten §6.5.
lattice_reflection_positivity; Osterwalder–Seiler 1978): for every box, every and every family of loops in , is real and .strong_coupling_clustering; Osterwalder–Seiler 1978): there is such that for and every scale , the infinite-volume limits of all loop correlation functions exist and connected correlations decay in the time separation (lattice units) like with .strong_coupling_area_law; Wilson 1974, Osterwalder–Seiler 1978): for there are and with for every rectangle of plaquettes inside the box, uniformly in the box.two_dimensional_rectangle_law; Migdal 1975, Gross–Witten 1980): for and every , the expectation of a rectangular Wilson loop enclosing plaquettes equals for i.i.d. with law .A bridging lemma, wilsonLoop_gaugeTransform, states the gauge invariance of Wilson loops.
The result itself. Existence would give the first non-trivial relativistic quantum field theory in four dimensions satisfying the Osterwalder–Schrader axioms; Jaffe–Witten: "at present we do not know any non-trivial relativistic field theory that satisfies the Wightman (or any other reasonable) axioms in four dimensions." The mass gap explains the short range of the strong force and implies clustering of correlations, which is what would allow the theory to be transplanted from to general four-manifolds and connect with Donaldson theory and its descendants. Both parts together would put the predictions of lattice QCD on a mathematical footing.
Formalizing it. Nothing in this mission has a machine-checked proof, and Mathlib contains no lattice gauge theory, no cluster expansions, no Osterwalder–Schrader reconstruction and no representation theory of compact Lie groups beyond the Haar measure. Milestones 1 and 4 are finite-dimensional statements about integrals over compact groups whose proofs are known and short; milestone 1 is exactly the property Jaffe–Witten call "a major advantage" of Wilson's approximation. Milestones 2 and 3 are the strong-coupling cluster expansion, a substantial but fully rigorous piece of mathematical physics. The goal fixes in Lean a precise version of what the prize asks for, so that partial results are formalized against a common statement.
Every known method fails between the two regimes the milestones control. The strong-coupling expansion (milestones 2–3) converges only for and produces a correlation length of order one lattice unit, so along with bounded the physical mass tends to infinity and the limit is ultralocal, with ; HasFiniteMass rules this out. Asymptotic freedom predicts that the continuum limit requires (Chatterjee: "one needs to take while sending the lattice spacing to zero, to obtain a nontrivial correlation function"), and at large the Wilson measure concentrates near flat connections, where the analysis needs gauge fixing on phase cells, renormalization, and control of large fields; Bałaban's renormalization group achieves ultraviolet stability but gives no infrared control and no gap. Jaffe–Witten: "no present ideas point the direction to establish the existence of a mass gap that is uniform in the volume", and footnote 2 excludes weak-existence (compactness) arguments unless the properties of the limit are established. The renormalization constants are forced by the perimeter divergence of Wilson loops (raw expectations of smooth loops are expected to vanish in the limit), but they cannot manufacture a gap: connected correlations are rescaled, never created.
The formalization adds a second difficulty: even the two-dimensional milestone needs the change of variables from link to plaquette variables on a planar lattice and Haar invariance under the resulting maps, and the strong-coupling milestones need polymer/cluster expansions over the plaquettes of a growing box with bounds uniform in the volume.
[Group G] [TopologicalSpace G] [IsTopologicalGroup G] [CompactSpace G] [T2Space G] [MeasurableSpace G] [BorelSpace G]; the Haar probability measure is Mathlib's Measure.haarMeasure ⊤. The representation is a monoid homomorphism into Matrix.unitaryGroup (Fin N) ℂ, continuous and injective in the goal. Any faithful is allowed; the Wilson action uses without normalization, so absorbs the choice.γ.scale ≤ k, and limits are unaffected.HasFiniteMass.No trivializing formalization. Choosing (independent links) or (flat connections), or any bounded , yields limits whose connected correlations vanish for all large time separations, hence have mass gap for every ; HasFiniteMass excludes them. The constants multiply each loop separately and cannot create time dependence in connected correlations.
Infrastructure needed and welcome contributions. Haar-measure change of variables on finite products of compact groups (gauge fixing, plaquette variables); Schur orthogonality and characters of compact groups; polymer and cluster expansions; the transfer-matrix/Osterwalder–Schrader reconstruction from reflection-positive lattice data; convergence of Wilson loop expectations in two dimensions. Theorems in the YangMills namespace, decompositions of the milestones, and Mathlib lemmas about Haar measures on compact groups are all welcome. The Wightman-axiom (Minkowski) side, general 4-manifolds, and theories with matter fields are out of scope.
namespace YangMills
theorem existence_and_mass_gap {G : Type*} [Group G] [TopologicalSpace G] [IsTopologicalGroup G]
[CompactSpace G] [T2Space G] [MeasurableSpace G] [BorelSpace G]
(hG : IsCompactSimpleGaugeGroup G) {N : ℕ} (ρ : G →* Matrix.unitaryGroup (Fin N) ℂ)
(hρ : Continuous ρ) (hρ' : Function.Injective ρ) :
∃ (β : ℕ → ℝ) (L : ℕ → ℕ) (Z : ℕ → Loop 4 → ℝ) (W : List (Loop 4) → ℂ),
IsContinuumLimit ρ β L Z W ∧ IsLatticeInvariant W ∧ IsReflectionPositive W ∧
(∃ Δ : ℝ, 0 < Δ ∧ HasMassGap W Δ) ∧ HasFiniteMass W := by sorry
end YangMillsThis is the Clay Millennium Prize Problem of Jaffe and Witten (p. 6): "Prove that for any compact simple gauge group , a non-trivial quantum Yang–Mills theory exists on and has a mass gap . Existence includes establishing axiomatic properties at least as strong as those cited in [Streater–Wightman, Osterwalder–Schrader]", stated in the form Jaffe–Witten describe in §6.5: as the existence of the continuum and infinite-volume limit of Wilson's lattice gauge theory with the properties of a reflection-positive Euclidean theory with a mass gap.
Let be a compact Hausdorff group which is a compact simple gauge group (connected, non-abelian, every closed normal subgroup finite or all of ), and let be a continuous injective homomorphism (so is a compact Lie group with simple Lie algebra). Then there exist
such that:
where is the Wilson lattice gauge theory expectation at lattice spacing in the box of physical half-side with action and free boundary conditions; 2. (Euclidean invariance, lattice part) is invariant under all dyadic translations and all signed coordinate permutations of ; 3. (Reflection positivity) for all and admissible families of loops with strictly positive time coordinates, is real and , the time reflection (with orientation reversal); 4. (Mass gap) there is such that for all admissible there is with for all dyadic times at which is admissible, the forward time translation; 5. (Non-triviality: ) property 4 does not hold for every , i.e. the mass is finite.
Under Osterwalder–Schrader reconstruction, property 3 yields the Hilbert space, vacuum and positive Hamiltonian of the theory, and property 4 is equivalent to having no spectrum in ; property 5 is Jaffe–Witten's requirement "we require " and excludes the degenerate limits (independent links, flat connections, bounded couplings) whose connected correlations vanish at all large time separations. This is an open problem; a proof would settle the Millennium Prize Problem in this formulation.
Formalization Note The theory is defined through Wilson-loop correlation functions of axis-parallel loops with dyadic vertices (YangMills.Loop 4), following Jaffe–Witten §6.5, Seiler and Chatterjee; smeared field operators and full Euclidean invariance are not encoded. The infinite-volume limit is taken jointly with the continuum limit along (weaker than requiring the infinite-volume limit at each fixed spacing). The renormalization constants are multiplicative and per loop, as expected from the perimeter divergence of Wilson loops; they cannot create time dependence in connected correlations. Nothing is asserted about the values of on non-admissible families, about uniqueness of the theory, about the rate at which , or about matter fields.
No submissions yet on this mission's goal. Log in and be the first to prove it.