Executive Summary
Let be left primitive with faithful simple left module , and let , a division ring. Faithfulness embeds in ; density says the embedded copy is -transitive for every . A dichotomy follows immediately: either is finite, and then and is left artinian, or is infinite and is not left artinian.
The finite branch is the Wedderburn–Artin theorem for simple left artinian rings, obtained without ever mentioning a chain condition in the proof. The infinite branch is a consolation prize with real content: still carries matrix rings of every size as subquotients.
Overview
The classical route to Wedderburn–Artin decomposes a semisimple ring into isotypic components and identifies each component by an endomorphism-ring computation. It needs the descending chain condition from the first line. The density route reverses the logic: prove a transitivity statement that holds for every left primitive ring, then observe that a chain condition forces the transitivity to saturate.
The structure theorem and the dichotomy it produces.
Two ingredients feed in. Schur's Lemma makes a division ring, so is an honest vector space; the Density Theorem supplies transitivity. Everything else on this page is extracted from those two facts by linear algebra.
Because a simple ring is left primitive and a left artinian left primitive ring is simple, the artinian branch is not a special case reluctantly recovered — it is exactly the classical theorem, proved more generally. The relationship between these classes is laid out in Primitive, Simple, Prime and Semiprimitive: How the Classes Relate.
Learning Objectives
- State with the hypotheses on , and spelled out.
- Prove the equivalence left artinian for a left primitive .
- Construct the strictly descending chain of left ideals in the infinite-dimensional case.
- Identify the matrix sections and say why they are only subquotients.
- Deduce the Wedderburn–Artin theorem for simple left artinian rings.
- Use to certify density by checking -transitivity alone.
Definitions
- -transitive
- is -transitive if for every , every -linearly independent and every there is with for all .
- Dense ring of linear transformations
- A subring of that is -transitive for every finite .
- For a fixed independent sequence , the finite-dimensional subspace .
- and
- , a subring of ; , an ideal of and a left ideal of .
- Left primitive ring
- A ring with a faithful simple left module; equivalently, a ring having a maximal left ideal containing no nonzero two-sided ideal.
All rings have an identity. is written as a right -vector space, and acts on the left; transformations are composed accordingly.
Core Concepts
From faithful simple module to a ring of matrices
Faithfulness of says , so is injective and may be identified with a ring of linear transformations of . Simplicity of says this ring is -transitive. Density upgrades -transitivity to -transitivity for every . So a left primitive ring is, up to isomorphism, a ring of matrices — possibly of infinite size and possibly a proper subring of all of them.
Why finite dimension collapses the theorem
If , then is finitely generated over and makes onto. Since for an -dimensional right -space, we get — and matrix rings over division rings are left and right artinian, so the chain condition arrives as a conclusion rather than a hypothesis.
Why infinite dimension destroys the chain condition
If is infinite, choose independent and let annihilate . Each is a left ideal of , and -transitivity supplies an killing but not . That single element separates from , and the resulting strictly descending chain rules out the descending chain condition.
Key Results
Let be a left primitive ring, a faithful simple left -module, and — a division ring by Schur's Lemma. Then is isomorphic to a dense ring of linear transformations on the right -vector space . Moreover:
- If is left artinian, then is finite and .
- If is not left artinian, then is infinite, and for every integer there is a subring admitting a surjective ring homomorphism onto .
Embedding. Since is faithful, , so is injective. Since is simple it is semisimple, so the Density Theorem applies, and by density of the action is the same as being a dense ring of linear transformations on . This is the first assertion.
Finite-dimensional case. Suppose . Then is finitely generated, so is onto by ; being also injective, . In particular is left (and right) artinian.
Infinite-dimensional case. Suppose is infinite. Fix -independent and set , together with
Restriction to gives a ring homomorphism with kernel . It is onto: given , -transitivity produces with for ; then , so and restricts to . Hence , which is assertion (2).
No chain condition. Each is a left ideal of and clearly . By -transitivity applied to the independent vectors there is with and ; thus and the chain is strictly descending. So is not left artinian.
The last paragraph also proves the converse implication needed for (1): if were infinite, would fail to be left artinian. So a left artinian left primitive ring has finite, and the two cases are exhaustive and mutually exclusive.
Let be a simple ring which is left artinian. Then for some integer and some division ring , and for the unique simple left -module up to isomorphism.
Since it has a maximal left ideal , and is a simple left module. Its annihilator is a two-sided ideal not containing , hence proper, hence zero by simplicity of ; so is faithful and is left primitive. Applying with the given left artinian hypothesis yields and with . Uniqueness of the simple module follows because is semisimple with a single isotypic component.
If is a semisimple ring, then for division rings . Density supplies each factor: the decomposition of into finitely many simple components is the input, and each component is simple left artinian, hence a matrix ring by the previous corollary.
Let be a division ring, a right -vector space, and a subring. Then:
- is -transitive if and only if is a simple module; in that case is a left primitive ring.
- The following are equivalent: (a) is -transitive; (b) is -transitive and ; (c) is dense in .
(1) -transitivity says for every , which is simplicity of ; the action is faithful because sits inside , so is a faithful simple left -module and is left primitive.
(2) For (b) (c): by (1), is simple, hence semisimple, and the hypothesis identifies with , so the Density Theorem together with gives density. (c) (a) is immediate, since density includes -transitivity. The remaining implication (a) (b) is the double-centraliser step: -transitivity forces every -endomorphism of to be right multiplication by a scalar in ; it is proved in Double Centralizers and Density for Bimodules.
The surjection is from a subring , not from . A simple non-artinian left primitive ring — the first Weyl algebra in characteristic , for instance — has no proper two-sided quotients at all, so no homomorphism from onto can exist for with .
Proof Techniques and Method
How this proof works, and which moves transfer to other arguments.
Faithful plus simple equals concrete
Any faithful simple module turns an abstract ring into a ring of linear transformations. All later arguments are then linear algebra over a division ring.
Transitivity manufactures elements
Need an element killing some vectors but not another? Ask -transitivity for it. This is how the strictly descending chain is built, and how most exercises on primitive rings are solved.
Stabiliser subring, annihilator ideal
The pair turns a finite-dimensional piece of into a matrix section of . Subquotients, not quotients — the distinction matters.
The strategic point is the order of quantifiers. The classical Wedderburn–Artin proof fixes the chain condition and derives structure; the density proof derives structure first and lets the chain condition decide how far the structure goes. That is why the same argument covers non-artinian rings without modification.
Worked Example
Finite-dimensional check
Let and let be the module of column vectors, with acting on the right by scalars. is simple and faithful, and because a matrix commuting with all of is scalar. Here , so predicts — which it is. The theorem is consistent, and the content in this case is that no proper subring of can be -transitive on .
Infinite-dimensional: scalars plus finite rank
Let be a field, , and . Let be the set of transformations of finite rank and put
Here denotes the scalar transformations . Over a division ring one must restrict to , since only central scalars act -linearly.
is an ideal of , so is a subring. It is dense: given -independent and arbitrary , extend to a basis and define with zero on the remaining basis vectors. Then has rank at most , so . Hence is -transitive for every , so is simple and faithful, is left primitive, and by we also get .
Now run concretely with . Writing for the rank-one transformation killing the other basis vectors:
- contains , which does not annihilate ; so and is not left artinian.
- contains every with , and restriction gives — matrix rings of every size appear as sections.
- : the minimal left ideals are for rank-one idempotents , so this left primitive ring has nonzero socle and is therefore right primitive as well.
Let be an integral domain with quotient field , and let . Let consist of those transformations whose matrix is a finite matrix over in the top-left corner and the scalar down the rest of the diagonal. Then is dense in , hence left primitive, and .
*Why the centre is :* if such an with corner block is central, then commutes with all of , so for some ; repeating the argument one size up forces . Since a left primitive ring is prime, its centre is automatically a domain, so this construction shows every integral domain occurs.
Frameworks and Models
Left primitive rings organise by two independent binary features: whether is finite, and whether the socle is nonzero.
| finite | Nonzero socle | Left artinian | Example | |
|---|---|---|---|---|
| Simple artinian | yes | yes | yes | |
| Full endomorphism ring | no | yes | no | , infinite |
| Scalars plus finite rank | no | yes | no | |
| Socle-free | no | no | no | Weyl algebra ; free algebra |
Four kinds of left primitive ring
The first row is the only one where is determined by alone. In the last row the ring may even have non-isomorphic faithful simple left modules, so the representation as a dense ring of transformations is a choice, not an invariant.
Process and Workflow
You have a left primitive ring with faithful simple and . What can you conclude?
Comparison and Classification
| Feature | infinite | |
|---|---|---|
| Image of | all of | proper dense subring |
| Chain conditions | left and right artinian | neither left nor right artinian |
| Ring structure | , simple | need not be simple; need not be noetherian |
| Matrix rings inside | itself | subquotients for all |
| Simple modules | one isomorphism class | possibly several, even among faithful ones |
| Right primitivity | automatic | not automatic |
| General ring | Has a minimal left ideal | Left artinian | Commutative | |
|---|---|---|---|---|
| left primitive prime | yes | yes | yes | yes |
| prime left primitive | no | yes | yes | no |
| left primitive simple | no | no | yes | yes |
| left primitive right primitive | no | yes | yes | yes |
Which implications hold in which class of rings
Relationship Map
Each containment is strict in general, and each becomes an equality once the descending chain condition on left ideals is imposed — that collapse is the content of and the reason Wedderburn–Artin looks so much stronger than it is.
Design Considerations
Design considerations here means the choices made when modelling a problem with these algebraic structures.
- Which faithful simple module? For a socle-free left primitive ring the choice is real: different give different concrete realisations of as transformations. Fix once and record it, as you would fix a basis.
- Which side? Primitivity is not left-right symmetric, so a theorem proved for left primitive rings does not transfer by opposite-ring formalities unless the ring is known to be right primitive too. Rings with a minimal one-sided ideal are the safe case.
- Artinian or not? Assuming left artinian is not a technical convenience here: it changes the conclusion from dense to equal. If you only need matrix behaviour on finitely many vectors, do not assume it.
- Which invariant to carry? is the useful numerical invariant in the artinian branch. Outside it, the family of matrix sections carries the information instead, and it depends on the chosen independent sequence.
Failure Modes and Common Mistakes
- Do not assume the division ring is central in or commutative; it is and can be any division ring.
- Do not deduce right artinian from left artinian in general; here it happens to follow because the conclusion is artinian on both sides, not because the hypothesis is symmetric.
- Do not conclude simplicity from left primitivity. with infinite is left primitive and has a proper nonzero ideal — the finite-rank transformations.
- Do not expect the Wedderburn–Artin uniqueness statement to survive into the infinite-dimensional branch: the pair need not be unique when the socle vanishes.
Quick Reference
| Goal | Quote | Hypotheses |
|---|---|---|
| Realise as transformations | , first assertion | left primitive, faithful simple |
| Get | additionally left artinian | |
| Find matrix sections | infinite | |
| Certify density cheaply | , -transitive | |
| Prescribe the centre | a domain with quotient field |
Frequently Asked Questions
Does this really prove Wedderburn–Artin, or does it assume it?
It proves it, for simple left artinian rings, and the argument is logically independent of the socle-decomposition proof in §3. The only inputs are Schur's Lemma, the Density Theorem and the observation that a simple ring is left primitive. The extension to general semisimple rings still needs the decomposition into finitely many simple components.
Why does a strictly descending chain of the ideals exist as soon as is infinite?
Infinite dimension supplies independent vectors without end. For each , -transitivity produces a ring element annihilating while moving ; that element lies in but not in . Transitivity is exactly the tool that manufactures separating elements.
Is a left primitive ring determined by the pair ?
Only in the artinian branch, where is the full endomorphism ring . In general is merely some dense subring, and many non-isomorphic dense subrings of the same exist — the full ring, scalars plus finite rank, and Kaplansky's centre-prescribing rings all live inside the same .
Why is checking -transitivity enough to get density?
Because -transitivity forces to be exactly , and once the endomorphism ring is correct the Density Theorem applies and delivers -transitivity for all . This is , and it is the practical test: verify one condition on pairs of vectors rather than an infinite family of conditions.
Can a left primitive ring be commutative?
Only if it is a field. A commutative left primitive ring has a faithful simple module , and annihilates it, so faithfulness forces . Consistently, then gives and .
Does the theorem give a canonical embedding of into a matrix ring of infinite size?
It gives an embedding into , which after choosing a basis of becomes a ring of column-finite infinite matrices. The embedding depends on both the module and the basis, so it is a useful coordinate system rather than a canonical form.
References
- T. Y. Lam, A First Course in Noncommutative Rings, Graduate Texts in Mathematics 131, Springer-Verlag, 1991, §11, (11.19)–(11.21) (pp. 194–196).
- T. Y. Lam, A First Course in Noncommutative Rings, §3, for the classical proof of the Wedderburn–Artin Theorem (pp. 33–37).
- N. Jacobson, “Structure theory of simple rings without finiteness assumptions”, Transactions of the American Mathematical Society 57 (1945), 228–245.
- N. Jacobson, Structure of Rings, American Mathematical Society Colloquium Publications 37, revised edition, 1964.
- I. N. Herstein, Noncommutative Rings, Carus Mathematical Monographs 15, Mathematical Association of America, 1968, Chapter 2.
- L. H. Rowen, Ring Theory, Volume I, Academic Press, 1988, Chapter 2.
AI Suggested Questions
- Show directly that with infinite is left primitive but not simple, and identify all its two-sided ideals.
- Work out the matrix sections for the Weyl algebra acting on .
- Which dense subrings of are simple, and how does the socle detect that?
- Give a left primitive ring with two non-isomorphic faithful simple left modules and compare the resulting division rings.
- How does interact with Morita equivalence — are dense subrings of Morita equivalent to ?
- State and prove the right-handed version of and explain why it is not a formal consequence of the left one.
- Use Kaplansky's construction to realise as the centre of a left primitive ring.
