Ledger–Obata Spaces Overview

• Ledger–Obata spaces are compact homogeneous spaces formed as F^m/diag(F) with tangent spaces decomposed into (m-1) irreducible Ad-invariant submodules.
• They serve as a fundamental setting for studying invariant Einstein metrics, naturally reductive geometries, and the integrability of geodesic flows.
• These spaces admit various invariant metrics whose classification leads to insight into geometric stability and scalar curvature variations.

A Ledger–Obata space is a compact homogeneous space of the form $F^m/\mathrm{diag}(F)$, where $F$ is a compact, connected simple Lie group and $\mathrm{diag}(F)$ is the diagonally embedded subgroup $\{(x,\dots,x)\mid x\in F\}$ inside the $m$-fold product $F^m$. These spaces arise as quintessential examples in the geometry of homogeneous spaces and invariant Riemannian metrics, featuring prominently in the study of invariant Einstein metrics, naturally reductive geometries, and the integrability of geodesic flows. Their algebraic and analytical structure supports explicit classification results and large families of metrics with rich geometric and dynamical properties.

1. Definition and Structural Properties

Let $F$ be a compact simple Lie group with Lie algebra $\mathfrak{f}$. For integer $m\geq 2$, the Ledger–Obata space is defined as the quotient

$$M = F^m/\mathrm{diag}(F),\quad \mathrm{diag}(F) = \{(x,\dots,x)\mid x\in F\}.$$

On the Lie algebra level, $F$0 has $F$1 ($F$2 summands), and the diagonal subalgebra is $F$3. The adjoint-invariant complement $F$4 with respect to the negative Killing form $F$5 is identified as the tangent space at the identity coset: $F$6 The dimension of $F$7 is $F$8. The isotropy representation is multiplicity-free, and $F$9 decomposes into $\mathrm{diag}(F)$0 irreducible $\mathrm{diag}(F)$1-invariant submodules, each isomorphic to $\mathrm{diag}(F)$2 (Nikolayevsky et al., 2017, Chen et al., 2016).

2. Classification of Invariant Riemannian Metrics

A $\mathrm{diag}(F)$3-invariant Riemannian metric on $\mathrm{diag}(F)$4 is determined by an $\mathrm{diag}(F)$5-invariant inner product on $\mathrm{diag}(F)$6. Every irreducible $\mathrm{diag}(F)$7-submodule in $\mathrm{diag}(F)$8 is of the form

$\mathrm{diag}(F)$9

for some $\{(x,\dots,x)\mid x\in F\}$0 with $\{(x,\dots,x)\mid x\in F\}$1. An orthonormal basis $\{(x,\dots,x)\mid x\in F\}$2 of the subspace of vectors in $\{(x,\dots,x)\mid x\in F\}$3 orthogonal to $\{(x,\dots,x)\mid x\in F\}$4 gives a decomposition

$\{(x,\dots,x)\mid x\in F\}$5

Any invariant metric is then

$\{(x,\dots,x)\mid x\in F\}$6

A uniform description is available on the standard complement $\{(x,\dots,x)\mid x\in F\}$7, with inner products parameterized by positive definite $\{(x,\dots,x)\mid x\in F\}$8 matrices $\{(x,\dots,x)\mid x\in F\}$9 via $m$0 (Nikolayevsky et al., 2017, Chen et al., 2016).

3. Naturally Reductive, Geodesic Orbit, and Einstein Metrics

Naturally Reductive Metrics: The explicit classification theorem states that a $m$1-invariant metric on $m$2 is naturally reductive if and only if one of two structures is realized:

• (a) $m$3 is an ideal, $m$4, $m$5.
• (b) There exists an $m$6-invariant quadratic form $m$7, with either all $m$8, or exactly one $m$9 with $F^m$0; the metric is $F^m$1 (Nikolayevsky et al., 2017).

Geodesic Orbit (GO) Metrics: A homogeneous Riemannian metric is geodesic orbit if every geodesic is an orbit of a one-parameter subgroup. On Ledger–Obata spaces, any GO metric must be naturally reductive [(Nikolayevsky et al., 2017), Theorem 1.3]. For $F^m$2, every invariant metric is automatically naturally reductive and GO [(Nikolayevsky et al., 2017), Prop. 2.10].

Einstein Metrics: Any invariant Einstein metric is a critical point of the scalar curvature functional under the constraint $F^m$3 for the parameter matrix $F^m$4. For $F^m$5, the classification shows exactly three invariant Einstein metrics up to isometry and scaling:

• The standard (Killing) metric;
• The Jensen metric;
• A third non-standard naturally reductive metric (Chen et al., 2016). The number of invariant Einstein metrics on $F^m$6 satisfies a lower bound equal to the number of integer partitions of $F^m$7, growing super-exponentially with $F^m$8 (Chen et al., 2016).

4. Geodesic Flows and Integrability

For $F^m$9, the geodesic flow of the normal (standard) metric is integrable:

• The cotangent bundle can be reduced via left-trivialization and symplectic reduction, identifying $F$0 with a subspace $F$1.
• $F$2-invariant metrics correspond to positive definite operators on $F$3. The geodesic flow Hamiltonian for the normal metric is $F$4.
• Invariant functions separate into the pull-back of invariants on $F$5 and the ring of $F$6-invariant polynomials on $F$7.
• By modifying the argument shift method, a maximal commutative subalgebra of invariant polynomials is constructed to establish Liouville integrability of the geodesic flow, both for the normal metric and Nikonorov's invariant Einstein metrics when $F$8 is simple (Jovanovic, 2010).

The key result is that the number of independent commuting integrals matches half the dimension of the phase space, and explicit polynomial first integrals, constructed via flag subalgebras and argument shifts, exist for all $F$9 (Jovanovic, 2010).

5. Reducibility, Decomposition, and Isometry Groups

Ledger–Obata spaces $\mathfrak{f}$0 are reducible as Riemannian products if and only if $\mathfrak{f}$1 decomposes into a sum of $\mathfrak{f}$2-invariant submodules, each corresponding to a lower-dimensional Ledger–Obata factor $\mathfrak{f}$3. The decomposition respects $\mathfrak{f}$4. The reducibility criterion is verified using the holonomy algebra associated to extended endomorphisms acting in block-diagonal form, corresponding to certain combinatorial structures (trees on $\mathfrak{f}$5 vertices) (Nikolayevsky et al., 2017).

An irreducible Ledger–Obata space has full connected isometry group $\mathfrak{f}$6, acting by left multiplication on each factor [(Nikolayevsky et al., 2017), Cor. 1.6]. Geodesic orbit property thus reduces to a product of factors, each with a naturally reductive metric [(Nikolayevsky et al., 2017), Cor. 1.7].

6. Stability and Scalar Curvature Functional

The standard invariant metric on every Ledger–Obata space $\mathfrak{f}$7 is Einstein with Einstein constant $\mathfrak{f}$8 (Gutiérrez et al., 2023). The second variation of the scalar curvature under volume-preserving $\mathfrak{f}$9-invariant deformations is calculated via the Lichnerowicz operator. For $m\geq 2$0, the standard metric is unstable: the coindex (number of independent directions of positive second variation) satisfies

$m\geq 2$1

showing that the instability grows linearly in $m\geq 2$2 (Gutiérrez et al., 2023). This instability holds for the family of standard (Killing) metrics, and the computation applies to $m\geq 2$3 acting on $m\geq 2$4.

7. Examples, Growth of Einstein Metrics, and Topological Properties

For $m\geq 2$5, the Ledger–Obata space $m\geq 2$6 is a symmetric space with the bi-invariant metric (Nikolayevsky et al., 2017, Chen et al., 2016). For $m\geq 2$7, all invariant metrics are naturally reductive and geodesic orbit (Nikolayevsky et al., 2017). For $m\geq 2$8, there are three isometry classes of invariant Einstein metrics: standard, Jensen, and a new non-standard naturally reductive metric (Chen et al., 2016). For higher $m\geq 2$9, the number of non-isometric invariant Einstein metrics grows at least as fast as the number of integer partitions, with further non-product type critical points implying even faster asymptotic growth (Chen et al., 2016). All such metrics are homogeneous, of positive Ricci curvature, but uniqueness fails for $$M = F^m/\mathrm{diag}(F),\quad \mathrm{diag}(F) = \{(x,\dots,x)\mid x\in F\}.$$0. Topologically, $$M = F^m/\mathrm{diag}(F),\quad \mathrm{diag}(F) = \{(x,\dots,x)\mid x\in F\}.$$1 is diffeomorphic to $$M = F^m/\mathrm{diag}(F),\quad \mathrm{diag}(F) = \{(x,\dots,x)\mid x\in F\}.$$2, and the fundamental group is determined by the projection covering $$M = F^m/\mathrm{diag}(F),\quad \mathrm{diag}(F) = \{(x,\dots,x)\mid x\in F\}.$$3 (Chen et al., 2016).