Invariant theory for non-reductive actions: extensions of Hilbert and Schwarz theorems

Technical Summary: Invariant Theory for Non-Reductive Actions

Problem Statement
Classical invariant theory, grounded in the work of Hilbert, Weyl, and Schwarz, establishes a robust correspondence between algebraic (polynomial) and analytic (smooth) invariants for compact and reductive Lie groups. Specifically, the Hilbert-Weyl theorem guarantees the finite generation of polynomial invariant rings, while the Schwarz theorem ensures that the algebra of smooth invariants is generated by these polynomial invariants. However, this correspondence is known to break down for non-reductive groups. The paper investigates the specific nature of this divergence in two distinct non-compact, non-reductive regimes:

1. Discrete subgroups of the Lorentz group $O(n, 1)$ acting on Minkowski space $\mathbb{R}^{n,1}$.
2. Cocompact actions of discrete groups on $\mathbb{R}^n$ (or smooth manifolds).

The central question is whether the classical theorems (Hilbert-Weyl and Schwarz) persist, fail, or transform in these settings, and how the relationship between algebraic and smooth structures shifts when compactness and reductivity are lost.

Methodology
The paper employs a combination of algebraic invariant theory, differential geometry, and topological analysis of orbit spaces.

• Algebraic Analysis: The author utilizes the block-diagonal structure of Lorentz group elements to decompose actions into orthogonal (compact) and hyperbolic (non-compact) components. This allows the application of product action lemmas to determine Hilbert bases for polynomial invariants.
• Geometric and Topological Analysis: The study of smooth invariants relies on analyzing the topology of the orbit space $\mathbb{R}^{n,1}/\Gamma$. The author examines properties such as orbit separation, the properness of the Hilbert map, and the Hausdorff nature of the quotient.
• Construction of Counterexamples: To disprove the Schwarz property in the Lorentz regime, the paper constructs specific smooth invariant functions that cannot be expressed as functions of the polynomial invariants, utilizing light-cone coordinates to exploit the disconnected nature of orbits.
• Manifold Theory: For cocompact actions, the paper leverages the fact that the quotient space is a compact smooth manifold. It applies the Whitney Embedding Theorem and the Tubular Neighborhood Theorem to construct smooth generators for the invariant algebra, bypassing the polynomial category entirely.

Key Contributions and Results

1. The Lorentz Regime: Polynomial Generation vs. Smooth Failure

• Theorem A (Polynomial Generation): For a discrete subgroup $\Gamma \subset O(n, 1)$ generated by a matrix conjugate to a block form $L_{\alpha, \beta}$ (combining a Euclidean rotation $R_\alpha$ and a hyperbolic boost $H_\beta$), the ring of polynomial invariants $P(\mathbb{R}^{n,1})^\Gamma$ is finitely generated. The Hilbert basis is the union of the basis for the Euclidean part and the quadratic form $x^2 - y^2$ from the hyperbolic part. This provides a class of non-reductive groups where Hilbert's 14th problem has a positive answer.
• Proposition 4.1 (Failure of Schwarz Theorem): Despite finite generation of polynomials, the classical Schwarz theorem fails. The algebra of smooth invariants $E(\mathbb{R}^{n,1})^\Gamma$ is not generated by the polynomial invariants.
  • Mechanism: The failure arises because the polynomial invariants fail to separate orbits. In Minkowski space, the level sets of the invariant quadratic form $x^2 - y^2 = c$ (for $c > 0$) consist of two disconnected branches (future and past light cones). The group action preserves these branches, but the polynomial invariant maps both to the same value. Consequently, one can construct smooth functions that distinguish between these branches (e.g., functions non-zero only on the right branch) which are invariant but not functions of the polynomial generator.
  • Obstructions: The paper identifies three specific obstructions to the Schwarz theorem in this setting: (i) insufficiency of polynomial invariants to separate orbits, (ii) failure of orbit separation leading to a non-Hausdorff quotient, and (iii) the absence of an averaging operator (Haar measure) and the non-properness of the Hilbert map.

2. The Cocompact Regime: Polynomial Triviality vs. Smooth Generation

• Theorem B (Polynomial Collapse): For a discrete group $\Gamma$ acting freely, properly, and cocompactly on $\mathbb{R}^n$, the algebra of polynomial invariants is trivial (i.e., consists only of constants).
  • Reasoning: Any non-constant polynomial on $\mathbb{R}^n$ is unbounded. However, a cocompact action implies the quotient space is compact, forcing any invariant function to be bounded. Thus, no non-constant polynomial invariants exist.
• Theorem C (Smooth Generation): In contrast to the algebraic collapse, the algebra of smooth invariants $E(M)^\Gamma$ is finitely generated and structurally rich.
  • Result: There exists a topological isomorphism between the smooth invariants on the manifold $M$ and the smooth functions on the compact quotient $M/\Gamma$. Furthermore, there exist a finite set of smooth invariants $\sigma_1, \dots, \sigma_m$ such that every smooth invariant is a smooth function of these generators. This establishes a "Schwarz-type" theorem for the smooth category, even though the polynomial category is degenerate.

Synthesis and Classification
The paper synthesizes these findings into a fourfold classification of invariant-theoretic regimes (Table 1):

1. Compact Groups: Polynomial invariants are finitely generated; smooth invariants are generated by polynomials (Hilbert-Weyl + Schwarz).
2. Reductive Groups: Polynomial invariants are finitely generated; smooth invariants are generated by polynomials (Luna).
3. Cocompact Actions: Polynomial invariants are trivial; smooth invariants are finitely generated by smooth functions (Theorems B & C).
4. Discrete Lorentz Actions: Polynomial invariants are finitely generated; smooth invariants are not generated by polynomials (Theorem A + Proposition 4.1).

Significance and Claims
The paper claims to delineate the boundaries of the Hilbert-Weyl and Schwarz theorems, demonstrating that the alignment of algebraic and analytic descriptions of symmetry is not a generic feature of invariant theory but relies heavily on specific structural properties:

• Properness and Compactness: The validity of the Schwarz property (smooth generation by polynomials) is shown to depend on the properness of the action and the ability to separate orbits, which fails in the Lorentz case due to topological pathologies (non-Hausdorff quotients).
• Decoupling of Algebra and Analysis: The results show that the failure of one aspect (e.g., polynomial generation in the cocompact case, or smooth generation by polynomials in the Lorentz case) does not imply a total lack of structure. Instead, the structure may reside in the geometry of the quotient manifold (cocompact case) or in the algebraic properties of the orbit space (Lorentz case).
• Extension of Hilbert's 14th Problem: The work contributes to extending Hilbert's 14th problem to the smooth category, suggesting that the existence of a categorical quotient in the sense of Geometric Invariant Theory (GIT) does not guarantee a well-behaved analytic quotient when properness is lost.

The paper concludes that the "Schwarz property" is not a consequence of finite generation alone but requires mechanisms like averaging operators or proper actions, which are absent in the studied non-reductive regimes. This framework provides a basis for investigating other non-reductive settings and understanding the limits of equivariant symmetry reduction.