Hodge theory uses the Riemannian metric to select a canonical representative from each de Rham cohomology class — the unique harmonic form (satisfying Δω = 0, where Δ = dδ + δd is the Hodge Laplacian). The Hodge decomposition Ωᵏ(M) = ℋᵏ ⊕ im(d) ⊕ im(δ) splits the space of k-forms into harmonic, exact, and co-exact components. This converts a topological question (cohomology) into an analytic one (solutions of an elliptic PDE), providing both computational power and deep structural results.
De Rham cohomology defines cohomology classes as equivalence classes of closed forms modulo exact forms. A cohomology class [ω] contains infinitely many representatives (ω, ω + dα, ω + dβ, ...). Hodge theory uses the Riemannian metric to select a unique "best" representative — the one that minimizes the L² norm ‖ω‖² = ∫_M ω ∧ *ω within its cohomology class. This minimal representative is the harmonic form: the unique element satisfying Δω = 0.
The machinery requires the Hodge star operator * : Ωᵏ(M) → Ωⁿ⁻ᵏ(M), which depends on the metric and orientation. Using *, you define the codifferential δ = ±*d* : Ωᵏ → Ωᵏ⁻¹, which is the formal adjoint of d with respect to the L² inner product: (dα, β) = (α, δβ). The Hodge Laplacian Δ = dδ + δd is a second-order elliptic differential operator — the Riemannian generalization of the ordinary Laplacian. On functions, Δf = -div(grad f). A form is harmonic if Δω = 0, which (on a compact manifold) is equivalent to being both closed (dω = 0) and co-closed (δω = 0).
The Hodge decomposition theorem (for compact oriented Riemannian manifolds) states: Ωᵏ(M) = ℋᵏ(M) ⊕ d(Ωᵏ⁻¹) ⊕ δ(Ωᵏ⁺¹), where the three summands are mutually L²-orthogonal and ℋᵏ is the finite-dimensional space of harmonic k-forms. Since harmonic forms are closed and represent unique cohomology classes, this gives an isomorphism ℋᵏ(M) ≅ Hᵏ_dR(M). The proof uses the theory of elliptic operators: the Laplacian Δ is self-adjoint and elliptic, so by Fredholm theory, its kernel (harmonic forms) is finite-dimensional and complements its image.
Hodge theory has far-reaching consequences. The Betti numbers bₖ = dim ℋᵏ = dim Hᵏ are now computable by solving the eigenvalue problem Δω = λω — they are the multiplicities of the zero eigenvalue. The full spectrum of Δ (the eigenvalues) contains rich geometric information beyond Betti numbers — this is spectral geometry ("Can you hear the shape of a drum?"). On Kähler manifolds (including all smooth projective algebraic varieties), Hodge theory produces the Hodge decomposition of complex cohomology into (p,q)-types, which is a cornerstone of algebraic geometry. In physics, harmonic forms represent ground states of quantum mechanical systems, and the Hodge-de Rham operator d + δ appears in supersymmetric quantum mechanics.
No topics depend on this one yet.