Symmetric Triads in Lie Theory

• Symmetric triads are Lie algebra structures with two involutions that generalize symmetric pairs through a fourfold eigenspace decomposition.
• They establish a duality between non-compact pseudo-Riemannian symmetric pairs and compact symmetric triads, extending Cartan’s classical duality.
• Root-theoretic formulations and multiplicity classifications provide a foundation for understanding Hermann actions and reflective submanifolds in geometry.

In Lie theory, a symmetric triad is a natural “two-involution” generalization of the usual notion of a symmetric pair. Concretely, on the compact side it is a triple $(\mathfrak{g},\theta_1,\theta_2)$ consisting of a compact semisimple Lie algebra and two involutions; in the commutative case the involutions satisfy $\theta_1\theta_2=\theta_2\theta_1$. The central structural result is a duality between non-compact pseudo-Riemannian semisimple symmetric pairs and commutative compact semisimple symmetric triads, extending Cartan’s compact/non-compact duality for Riemannian symmetric spaces. In a complementary root-theoretic formulation, symmetric triads are encoded by triples $(\tilde{\Sigma},\Sigma,W)$, and the 2025 extension to symmetric triads with multiplicities gives a classification of the abstract objects and of the commutative compact symmetric triads they encode (Baba et al., 2020, Baba et al., 3 Jun 2025).

1. Basic definitions and algebraic setup

A non-compact semisimple symmetric pair is a pair $(\mathfrak{g}_0,\sigma)$, where $\mathfrak{g}_0$ is a non-compact real semisimple Lie algebra and $\sigma\in\mathrm{Aut}(\mathfrak{g}_0)$ is an involution. Its fixed-point subalgebra is

$\mathfrak{h}_0=\mathfrak{g}_0^\sigma,$

and the corresponding eigenspace decomposition is

$$\mathfrak{g}_0=\mathfrak{h}_0\oplus\mathfrak{q}_0,\qquad \mathfrak{q}_0=\mathfrak{g}_0^{-\sigma}.$$

If $\theta$ is a Cartan involution commuting with $\sigma$, one also has the Cartan decomposition

$\theta_1\theta_2=\theta_2\theta_1$0

When $\theta_1\theta_2=\theta_2\theta_1$1, the pair is Riemannian; otherwise it is pseudo-Riemannian (Baba et al., 2020).

A compact semisimple symmetric triad is a compact semisimple Lie algebra $\theta_1\theta_2=\theta_2\theta_1$2 equipped with two involutions $\theta_1\theta_2=\theta_2\theta_1$3. It is commutative if

$\theta_1\theta_2=\theta_2\theta_1$4

Writing

$\theta_1\theta_2=\theta_2\theta_1$5

the commuting condition yields the refined decomposition

$\theta_1\theta_2=\theta_2\theta_1$6

This fourfold eigenspace structure is the basic algebraic feature that distinguishes a triad from a symmetric pair. When $\theta_1\theta_2=\theta_2\theta_1$7, the triad $\theta_1\theta_2=\theta_2\theta_1$8 is essentially the usual compact symmetric pair (Baba et al., 2020).

Two operations are standard. The associated triad is

$\theta_1\theta_2=\theta_2\theta_1$9

and the dual triad is

$(\tilde{\Sigma},\Sigma,W)$0

These constructions are intrinsic to the two-involution formalism and recur throughout the general theory (Baba et al., 2020).

2. Duality with non-compact semisimple symmetric pairs

The duality theorem gives a bijection between equivalence classes of non-compact semisimple symmetric pairs and equivalence classes of commutative compact semisimple symmetric triads. More precisely, after choosing a Cartan involution $(\tilde{\Sigma},\Sigma,W)$1 commuting with $(\tilde{\Sigma},\Sigma,W)$2, one passes between triples $(\tilde{\Sigma},\Sigma,W)$3 and commutative compact triads $(\tilde{\Sigma},\Sigma,W)$4 by explicit constructions (Baba et al., 2020).

Starting from a commutative compact triad $(\tilde{\Sigma},\Sigma,W)$5, one complexifies $(\tilde{\Sigma},\Sigma,W)$6, lets $(\tilde{\Sigma},\Sigma,W)$7 be complex conjugation with respect to the compact real form, and defines the anti-linear involution $(\tilde{\Sigma},\Sigma,W)$8. Its fixed-point set is the non-compact real form

$(\tilde{\Sigma},\Sigma,W)$9

Restricting the involutions gives

$(\mathfrak{g}_0,\sigma)$0

where $(\mathfrak{g}_0,\sigma)$1 is a Cartan involution of $(\mathfrak{g}_0,\sigma)$2 commuting with $(\mathfrak{g}_0,\sigma)$3. The fixed-point algebra of $(\mathfrak{g}_0,\sigma)$4 is

$(\mathfrak{g}_0,\sigma)$5

Conversely, from $(\mathfrak{g}_0,\sigma)$6 one forms the compact real form

$(\mathfrak{g}_0,\sigma)$7

and then restricts

$(\mathfrak{g}_0,\sigma)$8

This produces a commutative compact semisimple symmetric triad. The two constructions are inverse up to equivalence, which is the content of the duality theorem (Baba et al., 2020).

In the Riemannian case, where $(\mathfrak{g}_0,\sigma)$9, the triad becomes $\mathfrak{g}_0$0, so the theory reduces to Cartan’s classical duality between non-compact and compact Riemannian symmetric spaces. The two-involution formalism therefore extends Cartan’s duality rather than replacing it.

3. Structural theory: irreducibility, types, and self-duality

The duality preserves irreducibility. On the non-compact side, irreducibility means that $\mathfrak{g}_0$1 has no non-trivial $\mathfrak{g}_0$2-invariant ideals; on the compact side, irreducibility means that $\mathfrak{g}_0$3 has no non-trivial ideals invariant under both involutions. Theorem 4.16 states that these notions correspond under the duality, and Theorem 4.15 gives a bijection between $\mathfrak{g}_0$4-invariant ideals of $\mathfrak{g}_0$5 and $\mathfrak{g}_0$6-invariant ideals of $\mathfrak{g}_0$7 (Baba et al., 2020).

For irreducible objects, the theory further refines into type correspondences. The non-compact side has the types (P-a)–(P-d), while Matsuki’s classification gives the compact-triad types (T-a)–(T-d). Theorem 4.18 establishes the precise matching

$\mathfrak{g}_0$8

This identifies the two-involution compact data as the exact compact counterpart of the pseudo-Riemannian non-compact data.

A further invariant is the class of symmetric pairs of type $\mathfrak{g}_0$9. In the non-compact setting, such a pair arises from a $\sigma\in\mathrm{Aut}(\mathfrak{g}_0)$0-grading

$\sigma\in\mathrm{Aut}(\mathfrak{g}_0)$1

with characteristic element $\sigma\in\mathrm{Aut}(\mathfrak{g}_0)$2 and a grade-reversing Cartan involution $\sigma\in\mathrm{Aut}(\mathfrak{g}_0)$3, via

$\sigma\in\mathrm{Aut}(\mathfrak{g}_0)$4

Its compact-triad characterization is especially simple. If $\sigma\in\mathrm{Aut}(\mathfrak{g}_0)$5, then type $\sigma\in\mathrm{Aut}(\mathfrak{g}_0)$6 is equivalent to inner conjugacy of the two involutions: $\sigma\in\mathrm{Aut}(\mathfrak{g}_0)$7 Corollary 4.46 then states that such pairs are self-dual (Baba et al., 2020).

4. Abstract symmetric triads and multiplicities

The root-theoretic abstraction replaces a concrete compact triad by a triple

$\sigma\in\mathrm{Aut}(\mathfrak{g}_0)$8

where $\sigma\in\mathrm{Aut}(\mathfrak{g}_0)$9 is an irreducible root system in a Euclidean space $\mathfrak{h}_0=\mathfrak{g}_0^\sigma,$0, $\mathfrak{h}_0=\mathfrak{g}_0^\sigma,$1 is another root system, and $\mathfrak{h}_0=\mathfrak{g}_0^\sigma,$2 is a nonempty subset such that $\mathfrak{h}_0=\mathfrak{g}_0^\sigma,$3, together with reflection-compatibility conditions that govern how roots move between $\mathfrak{h}_0=\mathfrak{g}_0^\sigma,$4 and $\mathfrak{h}_0=\mathfrak{g}_0^\sigma,$5. The lattice

$\mathfrak{h}_0=\mathfrak{g}_0^\sigma,$6

enters the equivalence relation $\mathfrak{h}_0=\mathfrak{g}_0^\sigma,$7, which allows a phase-twist by an element $\mathfrak{h}_0=\mathfrak{g}_0^\sigma,$8 and, in effect, can interchange the roles of the $\mathfrak{h}_0=\mathfrak{g}_0^\sigma,$9- and $$\mathfrak{g}_0=\mathfrak{h}_0\oplus\mathfrak{q}_0,\qquad \mathfrak{q}_0=\mathfrak{g}_0^{-\sigma}.$$0-parts along specified directions (Baba et al., 3 Jun 2025).

A symmetric triad with multiplicities is a quintuple

$$\mathfrak{g}_0=\mathfrak{h}_0\oplus\mathfrak{q}_0,\qquad \mathfrak{q}_0=\mathfrak{g}_0^{-\sigma}.$$1

where $$\mathfrak{g}_0=\mathfrak{h}_0\oplus\mathfrak{q}_0,\qquad \mathfrak{q}_0=\mathfrak{g}_0^{-\sigma}.$$2 are Weyl-invariant multiplicity functions, $$\mathfrak{g}_0=\mathfrak{h}_0\oplus\mathfrak{q}_0,\qquad \mathfrak{q}_0=\mathfrak{g}_0^{-\sigma}.$$3 supported on $$\mathfrak{g}_0=\mathfrak{h}_0\oplus\mathfrak{q}_0,\qquad \mathfrak{q}_0=\mathfrak{g}_0^{-\sigma}.$$4 and $$\mathfrak{g}_0=\mathfrak{h}_0\oplus\mathfrak{q}_0,\qquad \mathfrak{q}_0=\mathfrak{g}_0^{-\sigma}.$$5 supported on $$\mathfrak{g}_0=\mathfrak{h}_0\oplus\mathfrak{q}_0,\qquad \mathfrak{q}_0=\mathfrak{g}_0^{-\sigma}.$$6, with additional compatibility conditions on $$\mathfrak{g}_0=\mathfrak{h}_0\oplus\mathfrak{q}_0,\qquad \mathfrak{q}_0=\mathfrak{g}_0^{-\sigma}.$$7. Besides the isomorphism relation $$\mathfrak{g}_0=\mathfrak{h}_0\oplus\mathfrak{q}_0,\qquad \mathfrak{q}_0=\mathfrak{g}_0^{-\sigma}.$$8, the theory also uses a finer equivalence $$\mathfrak{g}_0=\mathfrak{h}_0\oplus\mathfrak{q}_0,\qquad \mathfrak{q}_0=\mathfrak{g}_0^{-\sigma}.$$9, which preserves $\theta$0, $\theta$1, $\theta$2, and the multiplicities exactly.

The 2025 theory adds type (IV) symmetric triads with multiplicities. In that case $\theta$3 is an irreducible root system, one chooses $\theta$4 with

$\theta$5

and the multiplicities come from a single root-system multiplicity function $\theta$6. Types (I)–(III) encode the case of distinct commuting involutions; type (IV) encodes the case of conjugate involutions. Theorem 3.16 classifies abstract symmetric triads with multiplicities up to $\theta$7, and Theorem 3.34 classifies the type (IV) cases. As applications, the paper gives classifications for commutative compact symmetric triads, with two types depending on the choice of equivalence relation (Baba et al., 3 Jun 2025).

5. Geometric realization: compact groups, Hermann actions, and root data

For a compact connected semisimple Lie group $\theta$8 with involutions $\theta$9, let $\sigma$0 be the identity component of $\sigma$1, and write the Lie algebra decomposition

$\sigma$2

If the triad is commutative, choose a maximal abelian subspace

$\sigma$3

Hermann’s theorem gives

$\sigma$4

so the action of $\sigma$5 on $\sigma$6 is hyperpolar with flat, totally geodesic section $\sigma$7. This is the geometric setting in which compact symmetric triads arise naturally (Baba et al., 3 Jun 2025).

The corresponding root data are extracted from the restricted root spaces

$\sigma$8

which decompose according to the $\sigma$9-eigenspaces of $\theta_1\theta_2=\theta_2\theta_1$00: $\theta_1\theta_2=\theta_2\theta_1$01 Then

$\theta_1\theta_2=\theta_2\theta_1$02

and the multiplicities are

$\theta_1\theta_2=\theta_2\theta_1$03

This produces the symmetric triad with multiplicities of $\theta_1\theta_2=\theta_2\theta_1$04.

The same data can be described from double Satake diagrams. Proposition 4.9 expresses $\theta_1\theta_2=\theta_2\theta_1$05, $\theta_1\theta_2=\theta_2\theta_1$06, and the multiplicities in terms of the projections of compact, noncompact, and complex roots: $\theta_1\theta_2=\theta_2\theta_1$07 with multiplicity formulas

$\theta_1\theta_2=\theta_2\theta_1$08

$\theta_1\theta_2=\theta_2\theta_1$09

These formulas show that symmetric triads compress the local geometry of Hermann actions into restricted-root combinatorics. The resulting classification also furnishes an alternative route to the classification of reflective submanifolds and clarifies the compact side of the generalized compact/non-compact duality (Baba et al., 3 Jun 2025).

6. Terminological scope and distinct usages

A recurrent source of confusion is terminological. In Lie theory, “symmetric triad” has the precise two-involution and root-theoretic meanings just described. In other areas, the same expression is used for different structures.

In the Macdonald/DIM setting, “triad” refers to an embedding of symmetric polynomials, Baker–Akhiezer polynomials, and a Noumi–Shiraishi-type power series into a common framework; the “symmetric triad” denotes the symmetric-polynomial corner of that three-cornered structure (Mironov et al., 10 Mar 2025). In the theory of tridiagonal algebras, the $\theta_1\theta_2=\theta_2\theta_1$10-symmetric tridiagonal algebra is described as a threefold enlargement of the usual tridiagonal algebra, with a symmetric triad of node operators and a symmetric triad of edge operators acting on $\theta_1\theta_2=\theta_2\theta_1$11 (Terwilliger, 2024). In rational homotopy theory, a Lie model of the triangle with $\theta_1\theta_2=\theta_2\theta_1$12-symmetry is presented as a symmetric triad of three vertices, three edges, and a 2-cell encoded in a complete DGLA or cdgl (Buijs et al., 2018, Griniasty et al., 2018). In general relativity, the phrase appears in connection with a triad formalism for twist-free axisymmetric spacetimes, where a three-dimensional frame is adapted to the reduced Einstein equations (Brink et al., 2013).

This suggests a terminological polysemy rather than a single cross-disciplinary formalism. Within Lie theory, however, the term has a stable technical content: a symmetric triad is the compact two-involution object, or its abstract root-theoretic shadow, that organizes the structure of pseudo-Riemannian symmetric pairs, Hermann actions, reflective submanifolds, and their classification.