Finite-constant and non-commutative operator modules: the unabsorbed multiplicity theorems of Noether–Schmeidler's 1920 theory

In 1920 E. Noether and W. Schmeidler developed the module theory of operator polynomial rings $P[\xi_1,\dots,\xi_n]$ with twisted multiplication $\xi_i a=\sigma_i(a)\xi_i+\delta_i(a)$; the paper proves a Hilbert-basis transfer (Satz III) and a decomposition theory (Satze I–X) whose multiplicity statements (Satze VIII–X) rest on the hypothesis that the coefficient field $P$ contains infinitely many constants, and it ties finite-order Restgruppen to integrals (Satze XI–XII). The problem addressed here is the analogue of the multiplicity theorems Satze VIII–X when the constants are finite — which, apart from the degenerate commutative case of a finite coefficient field, forces difference-type or mixed operators, since a nonzero derivation in characteristic $p$ has the infinite field of $p$-th powers among its constants — and when the generators do not commute.

Background

In 1920 Emmy Noether and Werner Schmeidler published "Moduln in nichtkommutativen Bereichen, insbesondere aus Differential- und Differenzenausdrücken" (Mathematische Zeitschrift 8, 1–35), prompted by a question of Edmund Landau to Noether on generalizing the product theorem for linear differential equations. The paper studies what are now called modules over operator polynomial rings. Fix a commutative coefficient field $P$ (the paper's Rationalitätsbereich) and symbols $\xi_1,\dots,\xi_n$ (differential operators $\partial/\partial x_i$, difference operators, or Frobenius-type substitutions); the operator ring $R=P[\xi_1,\dots,\xi_n]$ consists of polynomials in the $\xi_i$ with coefficients in $P$, with multiplication twisted by the rule $\xi_i a=\sigma_i(a)\xi_i+\delta_i(a)$ for $a\in P$, where $\sigma_i$ is an injective endomorphism of $P$ and $\delta_i$ is a $\sigma_i$-derivation of $P$ — an additive map satisfying $\delta_i(ab)=\sigma_i(a)\delta_i(b)+\delta_i(a)b$. For differential expressions $\sigma_i=\mathrm{id}$ and $\delta_i=\partial/\partial x_i$; for difference expressions $\sigma_i$ is a substitution and $\delta_i=0$. In modern terminology $R$ is an iterated Ore extension (skew polynomial ring), and the paper's central algebraic hypothesis, imposed from §6 on, is that the generators commute: $\xi_i\xi_j=\xi_j\xi_i$. The constants are the elements $C=\{a\in P:\xi_i a=a\xi_i\ \mathrm{for\ all}\ i\}$, i.e. the common fixed field of the $\sigma_i$ intersected with the kernels of the $\delta_i$.

The paper transfers Hilbert's basis theorem by Gordan's method: every infinite system of operator polynomials has a finite module basis (Satz III). It then develops a decomposition theory for right $R$-modules $\mathfrak M$ (systems of polynomials closed under the module operations; a module is finitely based when it has a finite basis). The divisibility language is Lasker's: a polynomial $F$ is divisible by $\mathfrak M$ when $F\in\mathfrak M$, and a module $\mathfrak N$ divides a module $\mathfrak M$ when $\mathfrak M\subseteq\mathfrak N$; the least common multiple of a system of modules is the smallest common multiple in this divisibility order. A prime module (Primmodul) is a module whose only divisors are itself and the unit module (the module of all polynomials); the paper notes this differs from the commutative usage of "prime". Two modules are coprime (teilerfremd) when their greatest common divisor is the unit module, and a system of modules is totally coprime (total teilerfremd) when the least common multiple of any subcollection is coprime to that of the rest. A module is completely reducible (vollständig reduzibel, following Loewy) when it is the least common multiple of a totally coprime system of prime modules. Two modules are of the same art (gleiche Art) precisely when their Restgruppen are isomorphic (Satz VI); in the commutative case the notion specializes to identity. The Restgruppe of $\mathfrak M$ is the system of residue classes of polynomials modulo $\mathfrak M$, with its induced addition and multiplication of a residue class by a polynomial; it is called finite (endlich) of order $m$ when some $m$ residue classes are $P$-linearly independent while any $m+1$ are linearly dependent over $P$. Additive decompositions of the Restgruppe into subgroups correspond to representations of the module as a least common multiple. Satz XI equates the order of a finite-order Restgruppe with the maximal number of linearly independent integrals of the differential system defined by a basis of $\mathfrak M$, and same-art modules correspond to birational transformations of the respective integral systems (Satz XII).

The multiplicity theorems at the heart of the paper, Satze VIII–X, all carry the standing hypothesis that the constants are infinite. Satz VIII: a Restgruppe representable as a sum of finitely many (at least two) pairwise isomorphic subgroups — no irreducibility assumed — admits infinitely many distinct such representations. Satz IX: a completely reducible module, the least common multiple of a totally coprime system of prime modules, is representable as a least common multiple of prime modules in infinitely many distinct ways exactly when at least two of the prime modules are of the same art. Satz X: a module is divisible by infinitely many distinct prime modules exactly when either it admits no greatest completely reducible divisor — the least common multiple of all its prime divisors is not itself a least common multiple of finitely many prime modules — or that divisor is itself divisible by at least two distinct prime modules of the same art.

The paper is explicit about its own restrictions, and these mark four unexplored directions. First, §3 defines two-sided modules (right and left simultaneously) but restricts the theory to right modules, noting the left case is analogous. Second, the commutativity of the $\xi_i$ is imposed only in §6. Third, Satze VIII–X carry the standing hypothesis that the constants are infinite, never examined for finite constant fields. Fourth, the introduction only sketches, in a footnote, that Lasker's module-theoretic decomposition theorems can be established without the unique-factorization theorem, by methods related to those of the paper, without carrying this out.

All four directions were transformed by later mathematics, though not uniformly. Noncommutative ideal and module theory became a field: Ore's 1933 theory of non-commutative polynomials introduced the general skew polynomial rings (Ore extensions) with arbitrary endomorphism and derivation, removing the commutativity restriction at the level of basis, ideal, and factorization theory; Jacobson's pseudo-linear transformations (1935, 1937) developed the structure theory of these rings. The modern theory of iterated Ore extensions of PBW/solvable type — covering Weyl algebras (rings of differential operators with polynomial coefficients), universal enveloping algebras of Lie algebras, and quantum-type rings with relations $\xi_j\xi_i=c_{ij}\xi_i\xi_j+\text{lower terms}$ — supports one-sided and two-sided Gröbner bases (canonical generating sets that make ideal-membership algorithmic; Pesch 1998; Seiler 2009; textbook treatment in McConnell–Robson 2001), and Hilbert-basis transfers for such rings remain an active topic (Bäck–Lundström–Öinert–Richter 2026). The Lasker program was realized by Noether herself in commutative form (Idealtheorie in Ringbereichen, 1921, via chain conditions without factorization) and in noncommutative form by Fitting (primary decomposition in noncommutative rings, 1935; the connection of "same art" to elementary-divisor equivalence, 1936; carried to rings of differential operators by Kazimirskii 1964). Two-sided module theory over operator rings developed into the homological theory of rings of differential operators (Noether 1929; Hauger 1978; Björk 1979). Differential and difference algebra grew into independent fields (Ritt 1950; Cohn 1965; Levin 2008, whose chapter on difference modules is the direct descendant of the 1920 module theory), but their structure theorems presuppose an arbitrary, typically infinite, field of constants. Over finite constant fields, the operator rings themselves are heavily studied — factorization of skew polynomials over finite fields is a developed algorithmic theory (Caruso–Le Borgne 2017), the module theory of Frobenius skew polynomial rings has a Matlis-duality framework (Sharp–Yoshino 2011), and the enumeration of submodules invariant under an endomorphism is an active subject (Rossmann 2017) — yet the specific multiplicity phenomena of Satze VIII–X under a finite constant field, and for non-commuting generators, have no treatment in this literature.

Problem Statement

Fix the setting of the 1920 paper: $P$ a commutative field (the Rationalitätsbereich); $R=P[\xi_1,\dots,\xi_n]$ the operator polynomial ring with multiplication twisted by $\xi_i a=\sigma_i(a)\xi_i+\delta_i(a)$ for $a\in P$, where $\sigma_i$ is an injective endomorphism of $P$ and $\delta_i$ is a $\sigma_i$-derivation (an additive map with $\delta_i(ab)=\sigma_i(a)\delta_i(b)+\delta_i(a)b$); generators commuting, $\xi_i\xi_j=\xi_j\xi_i$; constants $C=\{a\in P:\sigma_i(a)=a\ \mathrm{and}\ \delta_i(a)=0\ \mathrm{for\ all}\ i\}$. Modules are finitely based right $R$-modules (finitely generated modules of polynomials); the Restgruppe of a module is the system of residue classes of polynomials modulo it, and it is finite of order $m$ when some $m$ residue classes are $P$-linearly independent while any $m+1$ are linearly dependent. Satze VIII–X of Noether–Schmeidler show that when $C$ is infinite the decomposition theory has built-in infinite multiplicity: a Restgruppe representable as a sum of finitely many (at least two) pairwise isomorphic subgroups admits infinitely many distinct such representations (Satz VIII); a completely reducible module — the least common multiple of a totally coprime system of prime modules — is representable as a least common multiple of prime modules in infinitely many distinct ways exactly when at least two of the prime modules are of the same art, i.e. have isomorphic Restgruppen (Satz IX); and a module is divisible by infinitely many distinct prime modules exactly when either it admits no greatest completely reducible divisor — the least common multiple of all its prime divisors is not itself a least common multiple of finitely many prime modules — or that divisor is divisible by at least two distinct prime modules of the same art (Satz X).

Part 1 (primary open core — finite constants). Determine the analogues of Satze VIII–X when $C$ is finite. The basic instances are operator rings with difference-type generators: e.g. $P=\mathbb{F}_q(t)$ with $\xi$ acting by the Frobenius substitution $t\mapsto t^q$ (so $C=\{a\in P:a^q=a\}=\mathbb{F}_q$), or by a mixed law $\xi a=a^q\xi+\delta(a)$ with a Frobenius-derivation $\delta$ (constants contained in $\mathbb{F}_q$). Note that the pure differential case ($\sigma_i=\mathrm{id}$, $\delta_i\neq 0$) admits no finite-constant instances: in characteristic $p$ the constants of a nonzero derivation contain the field of $p$-th powers $P^p$ (since $\delta(a^p)=p\,a^{p-1}\delta(a)=0$), and in characteristic $0$ they contain the prime field, so they are always infinite; hence $C$ finite with all $\sigma_i=\mathrm{id}$ forces $P$ finite and all $\delta_i=0$, where $R$ degenerates to the commutative polynomial ring over the finite field $P$ and the module theory is Lasker's commutative theory.

(a) Finiteness: for a finitely based right $R$-module whose Restgruppe has finite order, is the number of distinct decompositions into pairwise isomorphic summands finite? Is the number of prime-module divisors finite? Prove finiteness theorems with explicit bounds, or refute them.

(b) Parametrization: where finiteness holds, give an explicit, decidable description of the parameter set of decompositions and of prime-module divisors.

(c) Multiplicity failure: alternatively, exhibit a finitely based module over an operator ring with finite constant field, whose Restgruppe has finite order, that nevertheless admits infinitely many pairwise distinct decompositions into isomorphic summands or infinitely many prime-module divisors — showing that the infinite-constant hypothesis of Satze VIII–X is essential to the statements and not an artifact of the diagonal-matrix proof.

Part 2 (secondary residue — non-commuting generators). Take up the same questions in the paper's general setting, dropping Part 1's finite-constant restriction: the coefficient field $P$ and its constants are unrestricted, so that differential, difference, and mixed operators all remain available and the integral correspondences of Satze XI–XII keep their objects. Relax only the commutativity of the generators to PBW-type (Poincaré–Birkhoff–Witt type) relations $\xi_j\xi_i=c_{ij}\xi_i\xi_j+\text{lower-order terms}$, $c_{ij}\in P^{\times}$ (iterated Ore extensions of solvable type, where the finiteness theorem Satz III survives), and determine which of Satze VIII–X and of the Restgruppe–integral correspondences of Satze XI–XII remain true, fail, or require modification. The basic instances required are the two-generator $q$-commuting rings $R=P[\xi_1,\xi_2]$ with $\xi_2\xi_1=q\,\xi_1\xi_2$, $q\in P^{\times}$, over a coefficient field $P$ of the paper's general kind.

Part 1 concerns the difference case ($\sigma_i\neq\mathrm{id}$, $\delta_i=0$), the mixed cases ($\sigma_i\neq\mathrm{id}$ with nonzero $\sigma_i$-derivations), and the degenerate commutative case ($P$ finite, all $\sigma_i=\mathrm{id}$, all $\delta_i=0$), for both finiteness questions of (a) — for decompositions into isomorphic summands and for prime-module divisors — over operator rings with finite constant field. Part 2 concerns, for the basic instances of two generators with $q$-commutation, which of Satze VIII–X and of the Restgruppe–integral correspondences of Satze XI–XII remain true, fail, or require modification.

References

1. Noether, E. and Schmeidler, W. (1920). Moduln in nichtkommutativen Bereichen, insbesondere aus Differential- und Differenzenausdrücken. Mathematische Zeitschrift 8(1–2), 1–35. DOI 10.1007/BF01212856. https://doi.org/10.1007/BF01212856
2. Lasker, E. (1905). Zur Theorie der Moduln und Ideale. Mathematische Annalen 60, 20–116. DOI 10.1007/BF01447495. https://doi.org/10.1007/BF01447495
3. Noether, E. (1921). Idealtheorie in Ringbereichen. Mathematische Annalen 83, 24–66. DOI 10.1007/BF01464225. https://doi.org/10.1007/BF01464225
4. Noether, E. (1929). Hyperkomplexe Größen und Darstellungstheorie. Mathematische Zeitschrift 30, 641–692. DOI 10.1007/BF01187794. https://doi.org/10.1007/BF01187794
5. Krull, W. (1927). Algebraische Erweiterungen kommutativer hyperkomplexer Systeme. Mathematische Annalen 97, 473–489. DOI 10.1007/BF01447878. https://doi.org/10.1007/BF01447878
6. Ore, Ø. (1933). Theory of Non-Commutative Polynomials. The Annals of Mathematics (2) 34(3), 480–508. DOI 10.2307/1968173. https://doi.org/10.2307/1968173
7. Jacobson, N. (1935). On Pseudo-Linear Transformations. Proceedings of the National Academy of Sciences 21(12), 667–670. DOI 10.1073/pnas.21.12.667. https://doi.org/10.1073/pnas.21.12.667
8. Jacobson, N. (1937). Pseudo-Linear Transformations. The Annals of Mathematics (2) 38(2), 484–507. DOI 10.2307/1968565. https://doi.org/10.2307/1968565
9. Fitting, H. (1935). Primärkomponentenzerlegung in nichtkommutativen Ringen. Mathematische Annalen 111, 19–40. DOI 10.1007/BF01472198. https://doi.org/10.1007/BF01472198
10. Fitting, H. (1936). Über den Zusammenhang zwischen dem Begriff der Gleichartigkeit zweier Ideale und dem Äquivalenzbegriff der Elementarteilertheorie. Mathematische Annalen 112, 572–582. DOI 10.1007/BF01565430. https://doi.org/10.1007/BF01565430
11. Kazimirskii, P. S. (1964). Theorem on elementary divisors for a ring of differential operators [Russian original title: Теорема об элементарных делителях для кольца дифференциальных операторов]. Ukrainian Mathematical Journal 16(3), 309–318. DOI 10.1007/BF02538008. https://doi.org/10.1007/BF02538008
12. Ritt, J. F. (1950). Differential Algebra. American Mathematical Society Colloquium Publications 33, American Mathematical Society, viii+184 pp. ISBN 0-8218-3205-0. https://www.ams.org/online_bks/coll33/
13. Cohn, R. M. (1965). Difference Algebra. Interscience Tracts in Pure and Applied Mathematics 17, Interscience Publishers / John Wiley and Sons, xiv+355 pp.
14. Levin, F. (1956). On ideals in multidifferential polynomial rings. Mathematische Annalen 132, 289–309. DOI 10.1007/BF01360044. https://doi.org/10.1007/BF01360044
15. Cohn, P. M. (1969). Factorization in general rings and strictly cyclic modules. Journal für die reine und angewandte Mathematik 239–240, 185–200. DOI 10.1515/crll.1969.239-240.185. https://doi.org/10.1515/crll.1969.239-240.185
16. Hauger, G. (1978). Injektive Moduln über Ringen linearer Differentialoperatoren. Monatshefte für Mathematik 86, 189–201. DOI 10.1007/BF01659719. https://doi.org/10.1007/BF01659719
17. Björk, J.-E. (1979). Rings of Differential Operators. North-Holland Mathematics Library 21, North-Holland Publishing Company, xvii+374 pp.
18. McConnell, J. C. and Robson, J. C. (2001). Noncommutative Noetherian Rings, revised edition. Graduate Studies in Mathematics 30, American Mathematical Society, xx+636 pp. ISBN 0-8218-2169-5. https://www.ams.org/books/gsm/030/
19. Pesch, M. (1998). Two-sided Gröbner Bases in Iterated Ore Extensions. In: Symbolic Rewriting Techniques. Birkhäuser, Basel, 225–243. ISBN 978-3-0348-9779-2. DOI 10.1007/978-3-0348-8800-4_11. https://doi.org/10.1007/978-3-0348-8800-4_11
20. Seiler, W. M. (2009). A combinatorial approach to involution and δ-regularity I: involutive bases in polynomial algebras of solvable type. Applicable Algebra in Engineering, Communication and Computing 20, 207–259. DOI 10.1007/s00200-009-0098-0. https://doi.org/10.1007/s00200-009-0098-0
21. Levin, A. B. (2008). Difference Algebra. Algebra and Applications 8, Springer, xi+519 pp. ISBN 978-1-4020-6946-8. DOI 10.1007/978-1-4020-6947-5. https://link.springer.com/book/10.1007/978-1-4020-6947-5
22. Caruso, X. and Le Borgne, J. (2017). A new faster algorithm for factoring skew polynomials over finite fields. Journal of Symbolic Computation 79, 411–443. DOI 10.1016/j.jsc.2016.02.016. https://doi.org/10.1016/j.jsc.2016.02.016
23. Rossmann, T. (2017). Enumerating submodules invariant under an endomorphism. Mathematische Annalen 368, 391–417. DOI 10.1007/s00208-016-1499-6. https://doi.org/10.1007/s00208-016-1499-6
24. Sharp, R. Y. and Yoshino, Y. (2011). Right and left modules over the Frobenius skew polynomial ring in the F-finite case. Mathematical Proceedings of the Cambridge Philosophical Society 150(3), 419–438. DOI 10.1017/S0305004110000617. https://doi.org/10.1017/S0305004110000617
25. Bäck, P., Lundström, P., Öinert, J. and Richter, J. (2026). Hilbert's Basis Theorem for Poisson Ore Extensions. Mathematische Nachrichten (online first). DOI 10.1002/mana.70248. https://doi.org/10.1002/mana.70248