Infinite Rank Generalized Weyl Algebras

• Infinite rank generalized Weyl algebras are noncommutative structures built from an associative ring, a family of commuting automorphisms, and central elements indexed by an infinite ordinal.
• Their structure relies on an inductive limit of finite-rank subalgebras and a PBW-type basis, embedding them as principal Galois orders in skew monoid rings.
• Their representation theory highlights finiteness properties of Harish-Chandra modules and mirrors key attributes like simplicity, Ore conditions, and birational classification.

Infinite rank generalized Weyl algebras (GWAs) extend the algebraic framework of finite-rank GWAs, enabling a unified and characteristic-free treatment of a class of algebras naturally arising as Galois orders within representation theory. They arise from an associative (not necessarily commutative) ring $D$ together with a family of commuting automorphisms and central elements indexed by a (potentially infinite) ordinal $\alpha$. Their structure, representation theory, and embedding into skew monoid rings reveal deep connections between noncommutative algebra and the theory of principal Galois orders (Schwarz, 15 Jan 2026).

1. Conceptual Background

GWAs were introduced for finite rank by Bavula, permitting the systematic study of algebras defined by automorphisms and central elements of the base ring. Let $R$ be an associative ring, $\sigma \in \operatorname{Aut}(R)$, and $a \in Z(R)$. The rank-one GWA is given by

$$A = R(\sigma,a) = R\langle x, y\rangle / \langle x r - \sigma(r) x,\; y r - \sigma^{-1}(r) y,\; yx - a,\; xy - \sigma(a) : r \in R \rangle.$$

Higher-rank GWAs generalize this by employing a finite family of automorphisms $\sigma_1, \ldots, \sigma_n$ and central elements $a_1,\ldots,a_n$, with generators $X_i^\pm$ ($1\leq i\leq n$) subject to compatible commutation and rank-one-like relations. Many algebras encountered in representation theory—such as Gelfand–Tsetlin subalgebras or certain shift algebras—naturally embed as subalgebras of skew monoid rings (for example, $L*\mathbb{N}^n$ or $L*\mathbb{Z}^n$). To encompass cases with infinitely many commuting automorphisms, the theory extends to infinite rank, capturing these as inductive limits and situating them within the broader category of Galois orders.

2. Definition and Construction

Let $D$ be an associative ring. Fix an ordinal $\alpha$ and commuting automorphisms $$\sigma = (\sigma_\beta)_{\beta < \alpha} \subset \operatorname{Aut}(D)$$, together with central elements $a = (a_\beta)_{\beta < \alpha} \subset Z(D)$, satisfying

$$\sigma_\beta(a_\gamma) = a_\gamma\quad \forall\, \beta \neq \gamma.$$

The infinite-rank GWA of degree $\alpha$ is then

$$A = D(a,\sigma) := D\langle\,X_\beta^+, X_\beta^- : \beta < \alpha\,\rangle / \mathscr{R}$$

where the relations $\mathscr{R}$ are:

• $X_\beta^+ d = \sigma_\beta(d) X_\beta^+$ and $X_\beta^- d = \sigma_\beta^{-1}(d) X_\beta^-$ for $d \in D$,
• $[X_\beta^\pm, X_\gamma^\pm] = 0$ for all $\beta \neq \gamma$,
• $X_\beta^- X_\beta^+ = a_\beta$ and $X_\beta^+ X_\beta^- = \sigma_\beta(a_\beta)$.

Multi-indices $z$ with finitely many nonzero entries yield monomials

$$X^z = \prod_{\beta < \alpha} (X_\beta^+)^{z_\beta^+} (X_\beta^-)^{z_\beta^-}$$

where $z_\beta^\pm \in \mathbb{N}$ with $z_\beta^+ z_\beta^- = 0$. The relations are homogeneous under the natural $\mathbb{Z}^{\oplus \alpha}$-grading.

The Diamond Lemma and induction on finite subsets of $\alpha$ show that $\{ X^z : z \in \mathbb{Z}^{\oplus\alpha} \}$ forms a free left (and right) $D$-basis. The infinite-rank GWA is the direct (inductive) limit of its finite-rank subalgebras.

3. Embedding, Galois Order Structure, and Center

The subgroup $\Sigma \subset \operatorname{Aut}(D)$ generated by the $\sigma_\beta$ reflects the automorphism structure. If the $a_\beta$ are invertible in an Ore localization $D_S$, and the monoid map from the symbols $X_\beta^\pm$ onto $\Sigma$ is bijective (surjectivity-type hypothesis), $D_S(a,\sigma)$ embeds into the skew monoid ring $D_S*\Sigma$ by

$$X_\beta^+ \mapsto \sigma_\beta,\qquad X_\beta^- \mapsto a_\beta\sigma_\beta^{-1}.$$

An algebra $U \subset D_S*\Sigma$ is called a principal Galois order if:

1. $U \supset D$,
2. $U$ generates $D_S*\Sigma$ over the fraction field,
3. for $u \in U$, $u(D) \subset D$.

Under the regularity and surjectivity conditions, $D(a, \sigma)$ is a principal Galois order in $D_S*\Sigma$.

The center is characterized by

$$Z(D) \cap \left\{ d \in D \mid \sigma_\beta(d) = d,\; \forall\, \beta \right\}$$

and the grading by $\mathbb{Z}^{\oplus\alpha}$ is faithful. The multiplication on graded pieces obeys

$$X^z X^{z'} = \left( \prod_{\beta<\alpha} c_\beta(z,z') \right) X^{z+z'}$$

for certain scalars $c_\beta(z,z')$ determined by the base relations when $$\operatorname{supp}(z)\cap\operatorname{supp}(z')=\varnothing$$.

4. Illustrative Examples

Infinite Cyclic Automorphism Group

Let $D = k[h]$, $\sigma(h) = h-1$, and $a_n \in k[h]$. Setting $\Sigma = \langle \sigma \rangle \simeq \mathbb{Z}$, the degree-$\omega$ GWA becomes

$$A = k[h]\langle X_n^\pm : n \in \mathbb{Z} \rangle \Big/ \left\langle X_n^+ h - (h-1)X_n^+,\;\; X_n^- h - (h+1) X_n^-,\;\; X_n^- X_n^+ - a_n,\;\; X_n^+ X_n^- - \sigma(a_n) \right\rangle.$$

This can be realized as an inductive limit of finite-rank subalgebras indexed by increasing intervals.

Noncommutative Base Ring

Let $D = M_m(k)$, with automorphisms $\sigma_i(d) = g_i d g_i^{-1}$ (for $g_i \in GL_m(k)$) and central $a_i \in k^\times$. The infinite-rank GWA is given by

$$A = M_m(k)\langle X_i^\pm : i<\alpha \rangle \Big/ \big\langle X_i^+ d - g_i d g_i^{-1} X_i^+,\;\; X_i^- d - g_i^{-1} d g_i X_i^-,\;\; X_i^-X_i^+ - a_i,\;\; X_i^+X_i^- - a_i \big\rangle.$$

The principal Galois order structure persists over such noncommutative $D$.

5. Structural and Representation-Theoretic Properties

Simplicity

If no nonzero two-sided ideal of $D$ is stable under all $\sigma_\beta$, the subgroup generated by the $\sigma_\beta$ in $\operatorname{Aut}(D)/\operatorname{Inn}(D)$ is free abelian of rank $|\alpha|$, and for each $\beta$ and $m \geq 1$,

$D a_\beta + D \sigma_\beta^m(a_\beta) = D,$

then $D(a, \sigma)$ is simple (Theorem 6.8).

Noetherianity and Ore Conditions

For infinite $\alpha$, $D(a, \sigma)$ is not left or right Noetherian. However, if $D$ is a Noetherian domain, each finite-rank GWA is Noetherian and left/right Ore, and their inductive limit $D(a,\sigma)$ inherits the domain property and Ore conditions.

Localization and Birational Classification

If $S \subset Z(D)$ is a $\sigma$-stable Ore set containing all $\sigma_\beta^m(a_\gamma)$, then

$D_S(a,\sigma) \cong D_S * \Sigma,$

yielding a birational classification. In degree one over $k[h]$, this recovers exactly the Weyl algebra $A_1$, the quantum plane, and the Laurent polynomial case.

Representation Theory

Harish-Chandra modules over $D(a, \sigma)$ are locally finite over $D$. Finite-dimensional generalized weight spaces decompose, and irreducible Harish-Chandra modules are finite over weight subalgebras. A key consequence is that, under mild finiteness of stabilizers, there are finitely many irreducible modules with a fixed central character, mirroring the Main Theorem of Galois order theory of Futorny–Ovsienko. Open directions include the classification of simple Harish-Chandra modules, block decompositions, and precise conditions for the existence of a highest weight category structure (category $\mathcal{O}$).

6. Core Results and Main Formulas

Commutation and Defining Relations

For each $\beta < \alpha$ and $d \in D$:

$$X_\beta^+ d = \sigma_\beta(d) X_\beta^+,\qquad X_\beta^- d = \sigma_\beta^{-1}(d) X_\beta^-,$$

$$X_\beta^- X_\beta^+ = a_\beta,\qquad X_\beta^+ X_\beta^- = \sigma_\beta(a_\beta).$$

Basis Theorem

The set $\{ X^z : z \in \mathbb{Z}^{\oplus\alpha} \}$ is a free left and right $D$-basis of $D(a,\sigma)$, generalizing the Poincaré–Birkhoff–Witt theorem.

Inductive Limit Structure

If $\mathcal{F}$ is the poset of finite subsets $F \subset \alpha$, then

$$D(a, \sigma) \simeq \varinjlim_{F \subset \alpha\ \text{finite}} D_F(a^F,\sigma^F),$$

with $D_F(a^F,\sigma^F)$ the finite-rank GWA on $F$.

Principal Galois Order Isomorphism

Under the appropriate regularity and surjectivity hypotheses,

$D_S(a,\sigma) \cong D_S * \Sigma$

holds for localizations, identifying the infinite-rank GWA as a principal Galois order in the skew monoid ring.

A plausible implication is that the theory of infinite-rank GWAs provides a canonical and flexible apparatus for future investigations into the structure and module categories of algebras that arise as inductive limits or with infinite automorphism groups, particularly within the context of Galois orders and their invariants under group actions (Schwarz, 15 Jan 2026).