Executive Summary
The differential polynomial ring is the smallest ring in which the elements of live alongside a formal operator obeying the Leibniz rule . Whether is simple is decided by two conditions on the pair , and Lam's shows that over a -algebra these two are also sufficient.
Each condition rules out one visible family of ideals. If carries a proper nonzero -stable ideal , then is a proper nonzero ideal of . If is inner, the change of variable makes central and an ordinary polynomial ring, which is never simple.
The content of the theorem is that nothing else can go wrong. The proof is a minimal-degree argument: from an ideal one extracts a monic element of least degree , and the commutator produces the identity , which exhibits as inner once is invertible. That inversion is the only use of the -algebra hypothesis, and it is not removable.
Overview
Let be a ring and an additive map with . The differential polynomial ring has as elements the left polynomials with , added coefficientwise and multiplied by the single rule
Iterating gives ; in particular the coefficient of is .
Setting recovers the ordinary polynomial ring ; the construction is one of the two basic Ore extensions, the other being the twisted ring treated on the page for Skew Polynomial Rings: Hilbert's Twist. Both are covered by the general Ore extension .
The question this page answers is when has no two-sided ideals other than and . It matters because is a natural source of simple non-artinian rings: it is a domain whenever is, it is noetherian whenever is, and its simplicity cannot come from a chain condition.
Two special cases are worth memorising. If is a field of characteristic and , both conditions hold automatically, so is simple. If is commutative, every inner derivation is zero, so the second condition reduces to — and then the whole burden falls on -simplicity.
Learning Objectives
- Write down the multiplication rule of and the expansion of .
- Define -ideal, -simple and inner derivation, and state with its -algebra hypothesis.
- Prove that inner implies with central.
- Prove that a proper nonzero -ideal gives the proper nonzero ideal .
- Carry out the minimal-degree argument that yields and conclude that is inner.
- Exhibit a characteristic pair satisfying both conditions with not simple.
Definitions
A derivation on a ring is an additive map with for all . For the map is a derivation; derivations of this form are called inner. A derivation that is not of this form is non-inner or outer.
Note for every derivation, and exactly when .
A two-sided ideal is a **-ideal** if . The ring is **-simple** if and its only -ideals are and .
-simplicity is strictly weaker than simplicity: with is -simple but has plenty of ideals, none of which is stable under differentiation.
- Left polynomials , , with . Free as a left -module on .
- The largest with , for . Degrees add when is a domain, since .
- The inner derivation determined by .
- The ring of constants , a subring of containing .
- For a -ideal , the set of polynomials all of whose coefficients lie in ; a two-sided ideal of .
Rings have an identity, k need not be commutative, and polynomials are always written with coefficients on the left. A Q-algebra means a ring containing a copy of the rationals in its centre, so every nonzero integer is a central unit.
Core Concepts
Obstruction one: stable ideals of the base
Suppose is an ideal of with . Put . Left multiplication keeps the coefficients inside because and , and both and lie in . Right multiplication does too, using the expansion of and the fact that absorbs on the right. So is a two-sided ideal, proper and nonzero whenever is.
Obstruction two: inner derivations
If , set . Then for every
so is central in and is an honest polynomial ring over in a central variable. Then is a nonzero proper two-sided ideal — nonzero because , proper because a nonzero element of has zero constant term, so .
Why a minimal-degree argument must work
is filtered by degree and the associated graded ring is with central: the commutator of two elements has degree strictly less than the sum of their degrees. So commutators reduce degree, and any ideal is closed under them. An ideal that survives repeated commutation must contain elements of degree , that is elements of — and -simplicity then pushes it up to the unit ideal.
The two commutators that matter are , which multiplies the leading coefficient by , and for , which produces the correction term in degree . The first shows the leading-coefficient set is a -ideal; the second is where becomes inner.
Here is arbitrary and is monic.
Key Results
Let be a -algebra and let be a derivation on . Then is a simple ring if and only if is -simple and is not an inner derivation on .
Necessity, by contraposition. Suppose first that is inner. Putting gives for all , as computed above, so is central and . The ideal is nonzero and proper, so is not simple.
Now suppose is a -ideal of with . Then , the set of polynomials with all coefficients in , is a two-sided ideal of : it absorbs left multiplication because and , and it absorbs right multiplication because expanding produces only coefficients of the form multiplied on the left by elements of . It is nonzero and does not contain , so again is not simple.
Sufficiency. Assume is -simple and suppose, for contradiction, that has an ideal with . We derive that is inner.
Let be the least degree of a nonzero element of , and let consist of together with the leading coefficients of the degree- elements of . Then is an ideal of : if and , then has leading coefficient , and has leading coefficient because ; sums are handled by noting that a cancellation of leading terms leaves an element of of degree , which must be .
is a -ideal: for we have , and since while contributes , the difference has degree at most with coefficient in degree . Hence .
By -simplicity and we get , so : there is a monic . If then and , contrary to hypothesis; so .
Now take any . From we obtain
This element lies in and has degree at most , so it is by minimality of . Comparing the coefficient of gives for every .
Since is a -algebra and , the integer is a central unit. Writing we get for all , so is inner. This is the required contradiction, and is simple.
Let be any simple ring of characteristic and let be a non-inner derivation on . Then is a simple ring which is not left artinian.
The centre of a simple ring is a field. Since , that field contains , so is a -algebra. A simple ring has only the ideals and , both of which are -stable, so is -simple. Theorem now applies and is simple.
For non-artinianness, note that right multiplication by involves no twisting: . Hence every nonzero element of has zero coefficients in all degrees , while has coefficient in degree . Therefore and
is a strictly descending chain of left ideals, so fails the DCC.
Let be a field of characteristic and a derivation on . Then is a simple, non-artinian, left and right noetherian domain.
Indeed has no proper nonzero ideals, hence is -simple; is commutative, so its only inner derivation is ; and is a domain because degrees add over a field. The noetherian claim is the Ore-extension form of the Hilbert Basis Theorem.
Let be a prime, and . Then is a field, hence -simple, and is non-inner because is commutative. Yet is not simple.
The reason is that on : it is again a derivation, and it kills because consecutive integers have a product divisible by . Substituting into gives for all . So is central, and is a nonzero proper ideal.
Proof Techniques and Method
How these proofs work, and which move to reuse.
Three reusable techniques, all visible in the proof of and all reappearing in the skew Laurent analogue .
Minimal degree plus leading coefficients
Given an ideal , take the least degree and collect leading coefficients into . This turns an ideal of into an ideal of , where the hypothesis on can bite.
Commutators lower the degree
and both drop degree because the associated graded ring is commutative. An ideal closed under them and containing something of minimal degree is forced into relations among its coefficients.
Normalise to monic, then read off a relation
Once one may assume the minimal-degree element is monic. The coefficient of in then yields an equation valid for all , and that equation is exactly the statement that is inner.
The pattern generalises: in every Ore extension, simplicity is decided by which ideals of the base survive the twisting data, plus a non-degeneracy condition saying the twisting is not a change of variables away from being trivial. For that condition is non-inner; for an automorphism it is infinite inner order.
Worked Example
The rational differential operator algebra
Take , the field of rational functions in one variable, and . Then is a field of characteristic and , so by the corollary above is a simple noetherian domain that is not artinian. Its elements are the linear differential operators with rational-function coefficients.
It is instructive to verify simplicity by hand rather than by citing . The basic relation is
So already the two-sided ideal generated by contains and is everything.
More systematically, for with one computes, using and the commutativity of ,
Bracketing with differentiates a polynomial formally with respect to .
Let be an ideal and take of degree with leading coefficient . Applying times leaves , which is nonzero because and is a field. A nonzero element of is a unit, so .
A concrete run
Take . Then bracketing with three times gives , then , then ; and is a unit in , so after three brackets.
A base ring where the construction always fails
Let . It is simple, so -simple for any . But every derivation of a matrix algebra over a field is inner: the Hochschild cohomology of a separable algebra vanishes. So for every derivation of there is with , and is never simple. Simplicity of is nowhere near sufficient.
Process and Workflow
Is simple?
Comparison and Classification
| -simple? | Inner? | simple? | ||
|---|---|---|---|---|
| yes | yes | no — this is | ||
| yes | no | yes — this is the Weyl algebra | ||
| yes | no | yes | ||
| no — is stable | no | no | ||
| any | yes | yes, always | no | |
| yes | no | no — is central | ||
| no — is stable | no | no |
| Twisting data | derivation | automorphism |
|---|---|---|
| Commutation rule | ||
| Base condition | is -simple | is -simple |
| Non-degeneracy condition | not inner | of infinite inner order |
| Characteristic hypothesis | required | none required |
| invertible | no | yes |
| Reference | (3.15) | (3.18) |
The two Ore extensions side by side
The parallel is close but not perfect, and the asymmetry in the characteristic row is the interesting part: the Laurent criterion needs no hypothesis on , because its final step divides by nothing.
Relationship Map
Simplicity of sits between conditions on and conclusions about as a ring.
- simple — what follows and what does not
- always follows
- is not left or right artinian
- and is prime and primitive
- is a field, contained in
- follows only with extra hypotheses on
- is a domain — needs a domain
- is left and right noetherian — needs noetherian
- is a principal left ideal domain — needs a division ring
- never follows
- for a division ring
- has a minimal left ideal
- is finite-dimensional over its centre
- always follows
Applications and Industry Use
Applications here means where this structure is used — inside mathematics and in the engineering and computing disciplines that consume it.
Algebras of differential operators
is the working ring of computer algebra for linear ODEs. Factoring operators, computing symbolic solutions and applying the Beke–Schlesinger algorithm all take place inside it; Maple's DEtools and Singular's Plural implement its arithmetic.
Holonomic systems
Modules over differential operator algebras encode systems of linear PDEs. Simplicity of the operator ring is why such modules have no nontrivial two-sided obstructions and why localisation behaves; it underlies creative telescoping and automatic identity proving.
Linear time-varying systems
A time-varying linear system is a module over a differential operator ring with function coefficients. Controllability and observability become module-theoretic properties, and the ring-theoretic hypotheses of this page are what make the dictionary work.
Simple noetherian domains
is the standard machine for producing simple rings that are not artinian. Almost every counterexample in the structure theory of noncommutative noetherian rings starts life as some .
Enveloping algebras
The enveloping algebra of a solvable Lie algebra is an iterated Ore extension, and simplicity criteria of this type identify which primitive quotients are simple — the starting point for Dixmier's classification of primitive ideals.
Canonical commutation relations
The relation is the Heisenberg relation. Simplicity of the resulting algebra is the algebraic reason it admits no finite-dimensional representation, which forces the use of unbounded operators on Hilbert space.
The honest summary: this is a manufacturing criterion. Its value is that it converts a hard question about a large noncommutative ring into two small checks on the base ring and its derivation.
Design Considerations
Design considerations here means the choices made when modelling a problem with these algebraic structures.
- Polynomial or Laurent? If the operator you are modelling is invertible — a shift, a rotation — use the skew Laurent ring; if it is a genuine derivative, it is not invertible and is correct. This choice changes which criterion applies.
- How large to make the base. Enlarging from to makes -simplicity trivial and turns the result into a principal ideal domain, at the price of losing the graded structure and the finite-rank filtration. Choose when you need the Bernstein filtration, when you need division.
- Left or right coefficients. Writing with coefficients on the left fixes the sign conventions in every computation that follows. The right-coefficient convention gives and silently negates .
- Characteristic. If your application lives in characteristic — coding theory, cryptography — do not import simplicity results from characteristic . The correct statement there involves the -th power operator and Azumaya-type descriptions, not simplicity.
Standards and Notation
Standards here covers notation, symbol and markup standards, and reference implementations, rather than material or design codes.
nc_algebra, Sage OreAlgebra, Macaulay2 Dmodules, Maple Ore_algebraComputational Notes
Computational notes cover algorithms, cost and library behaviour rather than manufacturing process.
- Arithmetic in costs a factor of more than commutative polynomial arithmetic: multiplying past a coefficient expands into terms via the binomial formula, so a naive product of two degree- operators costs coefficient operations rather than .
- Left and right division algorithms exist whenever is a division ring, which makes a left and right euclidean domain. This is the computational backbone of operator factorisation.
- Non-commutative Gröbner bases (Buchberger's algorithm adapted to G-algebras) terminate for over a field and are implemented in Singular:Plural and Macaulay2. They compute syzygies and elimination ideals for D-module algorithms.
- Testing -simplicity is not decidable in general. For a finitely generated commutative algebra over a field it reduces to a computation with differential ideals, which is effective but doubly exponential in the worst case.
- Testing innerness of on a finite-dimensional algebra is linear algebra: solve for the unknown across a basis. The cost is one linear system of size for .
Failure Modes and Common Mistakes
- Do not expect simplicity to give any chain condition. is never artinian, and noetherianness must be inherited from .
- Do not confuse the -ideal condition with being generated by constants. The ideal of is not -stable for , but is -stable for .
- Do not assume without checking. Central elements of may involve — precisely what happens in characteristic , where is central.
- Do not read as requiring to be commutative. Amitsur's corollary is most useful when is a noncommutative simple ring admitting an outer derivation.
Quick Reference
| Check | What to verify | If it fails |
|---|---|---|
| -algebra | every nonzero integer is a central unit of | the theorem does not apply; look for central |
| -simplicity | no ideal with and | is a proper nonzero ideal |
| Non-innerness | no with | , never simple |
| Domain | is a domain | has zero-divisors but may still be simple |
| Noetherian | is left or right noetherian | need not be noetherian |
Frequently Asked Questions
Why does the theorem need rather than just characteristic ?
The proof divides by the integer , the minimal degree in the ideal, so must be invertible in and central. Characteristic alone says , which is weaker. In practice the gap rarely bites: Amitsur's shows that for simple of characteristic the centre is a field containing , so is automatically a -algebra.
Is -simplicity of implied by simplicity of ?
Yes, and that direction needs no hypothesis on the characteristic. If is a -ideal with , the polynomials with all coefficients in form a proper nonzero two-sided ideal of . So the necessity half of holds for arbitrary rings .
What replaces the criterion in characteristic ?
There is no clean simplicity criterion, because simplicity usually fails. The right framework is different: for a field of characteristic with a nonzero derivation, is a free module of rank over the central subring generated by and the constants, and the object of study becomes an Azumaya algebra over that centre rather than a simple ring.
Does simplicity of tell me anything about its one-sided ideals?
Almost nothing, and this is the usual disappointment. always has the strictly descending chain , so it is never artinian; it has no minimal left ideals; and its left ideal theory is as rich as that of allows. When is a division ring, is a principal left and right ideal domain, which is the best case.
How is this related to the Weyl algebra?
The Weyl algebra is the case , . The extra work needed there is showing that is -simple even though it is not simple, which is a short minimal-degree argument in its own right. The page on the Weyl algebra carries it out.
Can be non-inner but become inner after enlarging ?
Yes, and this is the usual way non-innerness is destroyed. Passing from to a larger ring in which the operator acquires an inverse-like partner can make inner. Innerness is not preserved by ring extensions in either direction, so it must be checked over the exact base you intend to use.
References
- T. Y. Lam, A First Course in Noncommutative Rings, Graduate Texts in Mathematics 131, Springer-Verlag, 1991, §3, results (3.15)–(3.16) (pp. 44–45).
- S. A. Amitsur, “Derivations in simple rings”, Proceedings of the London Mathematical Society (3) 7 (1957), 87–112.
- K. R. Goodearl and R. B. Warfield, Jr., An Introduction to Noncommutative Noetherian Rings, 2nd edition, London Mathematical Society Student Texts 61, Cambridge University Press, 2004, Chapters 1–2.
- J. C. McConnell and J. C. Robson, Noncommutative Noetherian Rings, revised edition, Graduate Studies in Mathematics 30, American Mathematical Society, 2001, Chapter 1.
- N. Jacobson, “Abstract derivation and Lie algebras”, Transactions of the American Mathematical Society 42 (1937), 206–224.
- O. Ore, “Theory of non-commutative polynomials”, Annals of Mathematics 34 (1933), 480–508.
AI Suggested Questions
- Prove the expansion by induction and identify where commutativity is not used.
- Determine the centre of when is a field of characteristic and .
- Give a -algebra that is -simple for some but has infinitely many ideals.
- State and prove the analogue of for the general Ore extension .
- Which simple rings of characteristic admit an outer derivation? Compare the commutative and central simple cases.
- Show that is a principal left ideal domain when is a division ring, and identify the euclidean function.
- How does the failure of in characteristic relate to the Azumaya structure of differential operator rings?
