Gilkey–Park–Sekigawa–Navarro Theorem

• The Gilkey–Park–Sekigawa–Navarro Theorem is a unique quadratic curvature identity for symmetric 2-tensor invariants on 4-dimensional Riemannian manifolds.
• It uses heat-trace asymptotics and Weyl’s invariant theory to establish the only nontrivial relation among ten canonical curvature invariants in dimension four.
• The result has significant implications for rigidity theorems, the study of critical metrics for quadratic curvature functionals, and the broader structure of Riemannian invariant theory.

The Gilkey–Park–Sekigawa–Navarro Theorem establishes the unique universal quadratic curvature identity for symmetric 2-tensor-valued invariants on 4-dimensional Riemannian manifolds. This identity, first discovered in connection with the generalized Gauss–Bonnet theorem by Berger and later re-examined by Gilkey, Park, and Sekigawa, is characterized by its universality: it holds identically on every Riemannian 4-manifold, distinguishing dimension four as exceptional with respect to second-order curvature invariants. No analogous nontrivial universal identities exist in other dimensions for tensor-valued quadratic curvature combinations. The result’s significance extends to rigidity theorems, the analysis of critical metrics for quadratic curvature functionals, and the broader structure of Riemannian invariant theory (Euh et al., 2011).

1. Statement of the Universal Curvature Identity

Let $(M^4, g)$ be a Riemannian 4-manifold. The Gilkey–Park–Sekigawa–Navarro identity is the following symmetric 2-tensor equation: $$\bigl(\lvert R\rvert^2 - 4\,\lvert \rho\rvert^2 + \tau^2\bigr) g_{ij} - R_{i}{}^{abc} R_{jabc} + 2\,\rho_{i}{}^{a} \rho_{aj} + 2\,R_{i}{}^{ab}{}_{j}\, \rho_{ab} - \tau\,\rho_{ij} = 0. \tag{1.1}$$ Here,

• $R_{abcd}$ denotes the Riemann curvature tensor,
• $\rho_{ab} = R_{a}{}^{c}{}_{bc}$ is the Ricci tensor,
• $\tau = g^{ab} \rho_{ab}$ is the scalar curvature,
• $|R|^2 = R_{abcd} R^{abcd}$ and $|\rho|^2 = \rho_{ab} \rho^{ab}$.

By introducing the notations: $$\widehat{R}_{ij} = R_{i}{}^{abc}R_{jabc}, \quad \check{\rho}_{ij} = \rho_{i}{}^{a}\rho_{aj}, \quad (L\rho)_{ij} = 2\,R_{i}{}^{ab}{}_{j} \rho_{ab}$$ the identity is succinctly rewritten as: $$(|R|^2 - 4|\rho|^2 + \tau^2)\,g - \widehat R + 2\,\check\rho + L\rho - \tau\,\rho = 0.$$ This expression exhausts all nontrivial universal quadratic curvature identities in dimension four for symmetric 2-tensors (Euh et al., 2011).

2. Universality and Dimension-Specificity

A relation among scalar invariants is termed universal if it holds identically on every Riemannian manifold of the given dimension. The theorem considers the vector space $\mathcal I_{4,4}$ of all symmetric 2-form-valued scalar invariants homogeneous of total degree four in the metric’s second derivatives (equivalently, quadratic in curvature with no covariant derivatives). This space is 10-dimensional in dimension 4.

Gilkey, Park, and Sekigawa proved, using heat-trace asymptotics and Weyl’s invariant theory, that (1.1) is the single independent universal relation in $$\bigl(\lvert R\rvert^2 - 4\,\lvert \rho\rvert^2 + \tau^2\bigr) g_{ij} - R_{i}{}^{abc} R_{jabc} + 2\,\rho_{i}{}^{a} \rho_{aj} + 2\,R_{i}{}^{ab}{}_{j}\, \rho_{ab} - \tau\,\rho_{ij} = 0. \tag{1.1}$$0 up to scaling; no other nontrivial quadratic curvature identity exists in four dimensions. For $$\bigl(\lvert R\rvert^2 - 4\,\lvert \rho\rvert^2 + \tau^2\bigr) g_{ij} - R_{i}{}^{abc} R_{jabc} + 2\,\rho_{i}{}^{a} \rho_{aj} + 2\,R_{i}{}^{ab}{}_{j}\, \rho_{ab} - \tau\,\rho_{ij} = 0. \tag{1.1}$$1, all possible linear combinations of the basis invariants are trivial—the only vanishing combination is the zero combination (Euh et al., 2011).

3. Construction and Basis of Curvature Invariants

The invariants forming the basis of $$\bigl(\lvert R\rvert^2 - 4\,\lvert \rho\rvert^2 + \tau^2\bigr) g_{ij} - R_{i}{}^{abc} R_{jabc} + 2\,\rho_{i}{}^{a} \rho_{aj} + 2\,R_{i}{}^{ab}{}_{j}\, \rho_{ab} - \tau\,\rho_{ij} = 0. \tag{1.1}$$2 are: $$\bigl(\lvert R\rvert^2 - 4\,\lvert \rho\rvert^2 + \tau^2\bigr) g_{ij} - R_{i}{}^{abc} R_{jabc} + 2\,\rho_{i}{}^{a} \rho_{aj} + 2\,R_{i}{}^{ab}{}_{j}\, \rho_{ab} - \tau\,\rho_{ij} = 0. \tag{1.1}$$3

A general universal identity would have the form: $$\bigl(\lvert R\rvert^2 - 4\,\lvert \rho\rvert^2 + \tau^2\bigr) g_{ij} - R_{i}{}^{abc} R_{jabc} + 2\,\rho_{i}{}^{a} \rho_{aj} + 2\,R_{i}{}^{ab}{}_{j}\, \rho_{ab} - \tau\,\rho_{ij} = 0. \tag{1.1}$$4 for all 4-manifolds. The Gilkey–Park–Sekigawa–Navarro Theorem asserts that up to scaling, the only nontrivial solution is given by: $$\bigl(\lvert R\rvert^2 - 4\,\lvert \rho\rvert^2 + \tau^2\bigr) g_{ij} - R_{i}{}^{abc} R_{jabc} + 2\,\rho_{i}{}^{a} \rho_{aj} + 2\,R_{i}{}^{ab}{}_{j}\, \rho_{ab} - \tau\,\rho_{ij} = 0. \tag{1.1}$$5 which precisely recovers (1.1) (Euh et al., 2011).

4. Proof Strategies and Methodological Innovations

The original proof by Gilkey, Park, and Sekigawa leveraged heat-trace asymptotics and invariant theory. An alternative approach, due to Euh, Jeong, and Park, is the universal-examples strategy:

1. Basis Introduction: Define ten canonical invariants $$\bigl(\lvert R\rvert^2 - 4\,\lvert \rho\rvert^2 + \tau^2\bigr) g_{ij} - R_{i}{}^{abc} R_{jabc} + 2\,\rho_{i}{}^{a} \rho_{aj} + 2\,R_{i}{}^{ab}{}_{j}\, \rho_{ab} - \tau\,\rho_{ij} = 0. \tag{1.1}$$6 spanning $$\bigl(\lvert R\rvert^2 - 4\,\lvert \rho\rvert^2 + \tau^2\bigr) g_{ij} - R_{i}{}^{abc} R_{jabc} + 2\,\rho_{i}{}^{a} \rho_{aj} + 2\,R_{i}{}^{ab}{}_{j}\, \rho_{ab} - \tau\,\rho_{ij} = 0. \tag{1.1}$$7.
2. A Priori Identity: Assume $$\bigl(\lvert R\rvert^2 - 4\,\lvert \rho\rvert^2 + \tau^2\bigr) g_{ij} - R_{i}{}^{abc} R_{jabc} + 2\,\rho_{i}{}^{a} \rho_{aj} + 2\,R_{i}{}^{ab}{}_{j}\, \rho_{ab} - \tau\,\rho_{ij} = 0. \tag{1.1}$$8 for all 4-manifolds.
3. Linear System Construction: Test this relation on five independent “test metrics,” specifically:
   • Products of constant-curvature surfaces,
   • Products of a 3-sphere and a line,
   • The round 4-sphere,
   • Left-invariant metrics on solvable Lie groups,
   • Non-homogeneous warped-product metrics.
4. Linear Solution: The linear system’s solution determines the coefficients $$\bigl(\lvert R\rvert^2 - 4\,\lvert \rho\rvert^2 + \tau^2\bigr) g_{ij} - R_{i}{}^{abc} R_{jabc} + 2\,\rho_{i}{}^{a} \rho_{aj} + 2\,R_{i}{}^{ab}{}_{j}\, \rho_{ab} - \tau\,\rho_{ij} = 0. \tag{1.1}$$9, revealing that the only nontrivial relation is (1.1); in particular, $R_{abcd}$0 and the other coefficients coincide with the Berger–Gilkey–Park–Sekigawa combination.

This “universal-examples” technique provides an elementary alternative to heat-kernel arguments and confirms the exceptional character of dimension four (Euh et al., 2011).

5. Geometric, Variational, and Theoretical Significance

The significance of the Gilkey–Park–Sekigawa–Navarro identity is multifold:

• As initially noted by Berger, (1.1) is the variational derivative of the Gauss–Bonnet integrand in dimension four.
• The identity is algebraically distinguished by the specific cancellations present in the four-dimensional Gauss–Bonnet theorem, with no analog in other dimensions.
• Labbi extended analogous universal curvature identities to higher even dimensions using double-form techniques, although the exceptional case of four dimensions yields a unique tensor-valued identity.
• The vanishing of the quadratic combination in (1.1) is foundational for rigidity results concerning critical points of quadratic curvature functionals and for generalized Einstein conditions in four-dimensional geometry (Euh et al., 2011).

6. Contextualization and Broader Implications

The Gilkey–Park–Sekigawa–Navarro Theorem cements the status of dimension four as unique within the algebra of Riemannian invariants, mirroring the exceptional role of the Gauss–Bonnet theorem. The result explains why searches for additional universal quadratic curvature identities—either scalar or tensor-valued—are fruitless in four dimensions beyond (1.1), and why higher dimensional analogs require fundamentally different techniques. The theorem’s conclusions underpin diverse research, including the analysis of the heat kernel, the structure of moduli spaces of Riemannian metrics, and the geometry of critical metrics. A plausible implication is the centrality of this identity for further explorations in geometric analysis on four-manifolds (Euh et al., 2011).