← LibraryPrimitive Rings and Primitive Ideals | KEVOS®Project Delivery · Project ManagementLesson 152/249← PrevNext →
ArticlePublished 9 Aug 202618 min readBy Kevin Jogin
Skip to content

Engineering Mathematics Core Density theory

Primitive Rings and Ideals

A ring is left primitive when it acts faithfully on a single simple left module. The corresponding ideals are exactly the annihilators of simple modules, and their intersection is radR.

Page ID
KEVOS-ENG-MATH-NCR-0085
Taxonomy
ENG / ENG-MATH
Collection
noncommutative-rings-core
Source
(11.2)–(11.5), §11 (pp. 183–185)
Reviewed
2026-08-08
Version
1.0.0

Executive Summary

Semiprimitivity asks for a faithful semisimple module. Tightening semisimple to simple gives left primitivity, and the resulting class of rings is the one the Density Theorem describes: every left primitive ring is a dense ring of linear transformations on a vector space over a division ring.

The corresponding notion for ideals is defined by quotienting, and (11.4) identifies it concretely: the left primitive ideals of R are exactly the annihilators of simple left R-modules. Since radR is the intersection of those annihilators, (11.5) follows at once — the radical is the intersection of the left primitive ideals, and equally of the right primitive ideals.

1 moduleFaithful and simple
Not symmetricLeft vs right
BergmanOne-sided example
=radRPrimitive ideals

Overview

The class of left primitive rings was isolated by Jacobson in the 1940s as the correct generalisation of simple artinian. Wedderburn–Artin describes a simple left artinian ring as Mn(D); the Density Theorem describes a left primitive ring as a dense subring of End(Vk) for a right vector space V over a division ring k. When dimkV is finite the two statements coincide, so primitivity is a chain-condition-free extension of the classical theory.

Primitive rings are not rare. If R is any nonzero ring and M any simple left R-module, then R/ann(M) is left primitive. So every ring produces primitive quotients, one for each simple module, and R recovers information about itself by mapping into their product.

The one structural surprise is that the notion has a genuine handedness. Semiprimitivity does not, because radR is side-neutral; primitivity does, and no amount of cleverness removes it. Bergman's construction produces a ring primitive on one side only, and further examples were later given by Jategaonkar.

Learning Objectives

  • State (11.2) and (11.3) and check the two definitions are consistent for the zero ideal.
  • Prove (11.4): left primitive ideals are precisely annihilators of simple left modules.
  • Derive (11.5) from (4.2) and (11.4) and state its right-handed twin.
  • Explain why R/ann(M) is left primitive for every simple M.
  • Verify that End(Vk) is left primitive for any nonzero right k-vector space V.
  • Say precisely what Bergman's example shows and what it does not.

Definitions

Definition(11.2)Left primitive ring

A ring R is left primitive if there exists a simple left R-module M with ann(M)=0. Right primitive is defined with right modules. A ring that is both is called primitive; where the sides are known to agree — for instance under (11.11) — the qualifier is often dropped.

Definition(11.3)Left primitive ideal

A two-sided ideal 𝔄R is left primitive if the quotient ring R/𝔄 is left primitive. In particular R is a left primitive ring exactly when 0 is a left primitive ideal of R, and a left primitive ideal is always proper, since the zero ring has no simple modules.

ann(M)
The two-sided ideal {rR:rM=0}; the kernel of REnd(M).
End(RM)
The endomorphism ring of M as a left R-module. By Schur's Lemma it is a division ring when M is simple.
End(Vk)
All k-linear endomorphisms of a right vector space V over a division ring k, acting on the left of V.
Primitive quotient
A ring of the form R/ann(M) with M a simple left R-module; always left primitive.
Subdirect product
A subring of a product that surjects on each coordinate. R is a subdirect product of its primitive quotients precisely when it is semiprimitive.

Core Concepts

Passing between R and R/𝔄

The entire content of (11.4) is a bookkeeping identity about modules over a quotient. If 𝔄ann(M), then M is a module over R/𝔄 by the obvious action, and its submodules over R and over R/𝔄 coincide; so simplicity is preserved in both directions. What changes is the annihilator, which shrinks by exactly 𝔄:

annR/𝔄(M)=annR(M)/𝔄(𝔄annR(M)).
(B)

Faithfulness over the quotient means the annihilator over R was exactly 𝔄.

So "R/𝔄 has a faithful simple module" and "𝔄 is the annihilator of a simple R-module" are two readings of the same sentence.

Why the notion is handed

Left primitivity is a statement about the lattice of left ideals: R is left primitive iff some maximal left ideal 𝔪 contains no nonzero two-sided ideal, since then R/𝔪 is faithful. This reformulation is often the practical test.

Nothing in that criterion is symmetric: the maximal left ideals of R and of Rop are unrelated lattices. Bergman's example shows the asymmetry is real, not merely unproven. Contrast this with radR, which admits a description — 1xyzU(R) for all x,z — mentioning only units, and is therefore forced to be symmetric.

Primitive rings are everywhere

Let k be a division ring, V a nonzero right k-vector space, and E=End(Vk) acting on the left. Then EV is simple, because a nonzero vector can be carried to any other by a k-linear map, and it is faithful, because a linear map killing all of V is zero. Hence E is left primitive. If dimkV=n< then EMn(k) is simple artinian; if dimkV is infinite, E is a left primitive ring that is neither simple, nor commutative, nor left artinian.

Key Results

Proposition(11.4)Primitive ideals are annihilators

Let R be a ring with identity and 𝔄R a two-sided ideal. Then 𝔄 is a left primitive ideal if and only if 𝔄=ann(M) for some simple left R-module M.

Proof

**()** Suppose 𝔄=ann(M) with M a simple left R-module. Since 𝔄M=0, the action of R factors through R/𝔄, making M a left R/𝔄-module with the same submodule lattice; so M is simple over R/𝔄. Its annihilator there is annR(M)/𝔄=𝔄/𝔄=0, so M is a faithful simple R/𝔄-module and R/𝔄 is left primitive.

**()** Suppose R/𝔄 is left primitive and let M be a faithful simple left R/𝔄-module. Inflate M along the surjection RR/𝔄. The R-submodules of M are exactly its R/𝔄-submodules, so RM is simple. An element rR kills M precisely when its class r¯ kills M, that is precisely when r¯=0 by faithfulness; hence annR(M)=𝔄.

Corollary(11.5)The radical as an intersection of primitive ideals

For any ring R with identity, radR is the intersection of all left primitive ideals of R, and also the intersection of all right primitive ideals. If R0 both families are nonempty.

Proof

By (4.2), radR=Mann(M) over all simple left R-modules M. By (11.4) the ideals occurring in that intersection are precisely the left primitive ideals, so the two intersections have the same index set and agree. The right-handed statement follows by applying the same argument in Rop, together with the left-right symmetry of radR established in (4.3)(4.4).

radR={𝔄:𝔄 left primitive}={𝔄:𝔄 right primitive}.
(11.5)

The two families of ideals may differ; their intersections cannot.

CorollaryEvery ring has primitive quotients

If R0 and M is any simple left R-module, then R/ann(M) is a left primitive ring. Consequently R is semiprimitive if and only if R embeds as a subdirect product of left primitive rings, namely its primitive quotients.

Proof

The first claim is the () half of (11.4). For the second, the natural map RMR/ann(M) has kernel Mann(M)=radR and is surjective in each coordinate; it is injective exactly when radR=0.

ExampleThe endomorphism ring of a vector space

Let k be a division ring and V0 a right k-vector space, E=End(Vk). Then EV is a faithful simple left E-module, so E is left primitive. If dimkV=n< then EMn(k). If dimkV is infinite then E is left primitive but not simple: the endomorphisms of finite rank form a proper nonzero two-sided ideal.

RemarkHandedness

Left primitivity does not imply right primitivity. G. Bergman constructed a ring that is primitive on one side but not the other; A. V. Jategaonkar produced further families later. This is in sharp contrast to semiprimitivity, and it is why the qualifier left is carried throughout §11.

Proof Techniques and Method

How these proofs work, and which move to reuse.

Move 1

Inflate and deflate

A module killed by 𝔄 is the same object over R and over R/𝔄; only the annihilator changes, and it changes by exactly 𝔄. Every statement of the form "property P of R/𝔄" translates into "property P of an R-module".

Move 2

Largest ideal inside a left ideal

For a maximal left ideal 𝔪, ann(R/𝔪) is the largest two-sided ideal contained in 𝔪. Primitivity is the assertion that this largest ideal is zero for some 𝔪.

Move 3

Same intersection, different families

Two different families of ideals can have equal intersections. (11.5) is exactly that phenomenon, and it is what lets a symmetric invariant be built from asymmetric pieces.

Move 2 is the version to remember when computing. To show a concrete ring is left primitive, produce a maximal left ideal and show no nonzero two-sided ideal fits inside it. That is exactly the pattern used for skew polynomial rings on the Skew Polynomial Rings as Primitive Rings page, where the ideals were first classified in (11.12) so that this check became routine.

Worked Example

Primitive ideals of a commutative ring

Take R=. The simple modules are /p and ann(/p)=p, so the left primitive ideals of are exactly the maximal ideals p. Their intersection is 0=rad, confirming (11.5). The zero ideal is not primitive, so itself is not a primitive ring — as (11.8) requires, since is commutative but not a field.

The general commutative picture: 𝔄 is a primitive ideal of a commutative ring R iff R/𝔄 is a field, i.e. iff 𝔄 is maximal. So for commutative rings, primitive ideal and maximal ideal are the same notion, and (11.5) degenerates to the classical description of the Jacobson radical.

A noncommutative primitive quotient

Let R=T2() be the upper triangular 2×2 rational matrices, with simple modules S1,S2 as on the Semiprimitive Rings page. Then

ann(S1)=(00),ann(S2)=(00),
(E.1)

and both quotients are isomorphic to , a field, hence primitive rings. So T2() has exactly two primitive ideals; their intersection is the square-zero ideal of strictly upper triangular matrices, which is radR. Again (11.5) checks out, and here the intersection is nonzero, so R is not semiprimitive.

An infinite-dimensional primitive ring

Let V be a countably infinite-dimensional right vector space over a field k, E=End(Vk). Fix a basis e1,e2, The module EV is simple: given v0 and wV, extend v to a basis and define a linear map sending vw. It is faithful since only the zero map kills every vector. Hence E is left primitive.

E is not simple: the finite-rank endomorphisms form a two-sided ideal F with 0FE. And E is not left artinian: the left ideals 𝔄n={rE:r(ei)=0 for in} form a strictly descending chain. So E is a left primitive ring outside the reach of Wedderburn–Artin, and it is the model case for the Density Theorem.

Frameworks and Models

Primitive rings fall into recognisable families, and knowing which family you are in determines which theorem applies.

  • Left primitive rings — faithful simple left module V; k=End(RV) is a division ring
    • dimkV finite
      • RMn(k) — simple left artinian
      • recovers Wedderburn–Artin
    • dimkV infinite, nonzero socle
      • R contains a minimal left ideal
      • End(Vk) and its dense subrings containing finite-rank maps
      • faithful simple module is unique up to isomorphism (11.11)
    • dimkV infinite, zero socle
      • no minimal left ideals
      • simple domains such as k[x;δ]
      • possibly many non-isomorphic faithful simple modules

Comparison and Classification

Primitive rings and primitive ideals side by side
QuestionRing versionIdeal version
DefinitionR has a faithful simple left moduleR/𝔄 has a faithful simple left module
Concrete formsome maximal left ideal contains no nonzero ideal𝔄=ann(M), M simple (11.4)
Commutative caseR is a field (11.8)𝔄 is a maximal ideal
Relation to radR primitive R semiprimitive (11.6)radR=𝔄 (11.5)
Side symmetryfails — Bergman 1965left and right families differ, intersections agree
Test cases
Left primitiveRight primitiveSimpleSemiprimitive
Mn(D)yesyesyesyes
End(Vk), dimkV infiniteyesyesnoyes
Weyl algebra A1(k), chark=0yesyesyesyes
nononoyes
k[[x]]nononono
Bergman's exampleyesnonoyes

Test cases

Relationship Map

M simple left R-moduleann(M) left primitive idealR/ann(M) left primitive ring

The correspondence is surjective onto left primitive ideals but not injective: non-isomorphic simple modules can share an annihilator. The fibres of the map are studied on the Socle of a Primitive Ring page, where (11.11) shows the fibre over 0 is a single isomorphism class as soon as R has a minimal left ideal.

Two-sided ideals of Rall of them
Semiprime idealsR/𝔄 has no nonzero nilpotent ideals
Prime idealsR/𝔄 is a prime ring
Left primitive ideals𝔄=ann(M), M simple
Maximal idealsR/𝔄 simple — always left primitive by (11.6)

The containment *left primitive prime* is (11.6) applied to R/𝔄; the containment *maximal left primitive* holds because a simple ring is primitive. Both inclusions are strict in general: 0 is a prime but not primitive ideal of , and 0 is a primitive but not maximal ideal of End(Vk) for infinite-dimensional V.

Applications and Industry Use

Applications here means where this structure is used — inside mathematics and in the engineering and computing disciplines that consume it.

Representation theory

The primitive spectrum

For enveloping algebras of Lie algebras, the set of primitive ideals with its Jacobson topology is the basic classification invariant; Dixmier's and Duflo's work describes it for solvable and semisimple Lie algebras.

Quantum algebra

Stratification of quantum groups

Primitive ideals of quantised coordinate rings are computed by the Goodearl–Letzter stratification, transporting the ideal-theoretic questions of this page to the quantum setting.

Operator algebras

Primitive ideal space

For a C*-algebra the primitive ideal space, kernels of irreducible representations, is the noncommutative substitute for a topological space, and is exactly (11.4) read in the analytic category.

Symbolic computation

Deciding primitivity

For a finite-dimensional algebra over a field, primitivity is decidable: compute the radical, split the semisimple quotient, and test whether one simple factor accounts for the whole algebra. Beyond finite dimension, no general algorithm exists.

Honestly stated: primitive ideals are the noncommutative replacement for points of a spectrum. Wherever a commutative theory uses Spec or the maximal spectrum, the noncommutative theory uses the primitive spectrum, and (11.4) is the definition that makes that substitution work.

Standards and Notation

Standards here covers notation, symbol and markup standards, and reference implementations, rather than material or design codes.

Preferred termleft primitive; carry the side explicitly
Primitive spectrumPrim(R), with the Jacobson topology
Annihilatorann(M); AnnR(M) in module-theoretic texts
Convention clashSome authors say primitive for what is here left primitive; check whether modules are left or right
Ideal notation𝔄R for two-sided; 𝔪 reserved for maximal left ideals
ImplementationsGAP and Magma expose simple module enumeration; primitive ideals of enveloping algebras are handled in specialist packages

Failure Modes and Common Mistakes

  • Do not confuse ann(M) with a maximal left ideal. The annihilator is two-sided and is the largest ideal inside 𝔪; it is generally much smaller than 𝔪.
  • Do not assume a left primitive ring has a unique faithful simple module. Uniqueness requires a minimal left ideal (11.11); without it there can be infinitely many, as k[x;σ] shows.
  • Do not read (11.5) as saying the left and right primitive ideals coincide. Only their intersections do.
  • Do not expect R left primitive to pass to subrings or to matrix rings for free — matrix rings do inherit primitivity, but the argument goes through Morita theory, not through the definition.

Historical Notes and Lessons Learned

  • 1927Artin's chain conditionsWedderburn's structure theory for finite-dimensional algebras is extended to rings with the descending chain condition, but the arguments still lean on nilpotence and finiteness.
  • 1945Jacobson's radical and primitivityJacobson defines the radical for arbitrary rings via annihilators of simple modules and isolates the primitive rings as those acting faithfully on a simple module.
  • 1945–47The Density TheoremJacobson and Chevalley prove that a ring acting faithfully and irreducibly acts densely; the structure theorem for primitive rings follows.
  • 1964Bergman's asymmetric ringG. Bergman constructs a ring that is primitive on one side and not on the other, settling that the notion is genuinely one-sided. Lam dates the construction to 1965.
  • 1970sPrimitive spectraDixmier and others study the primitive ideals of enveloping algebras; the primitive spectrum becomes a standard classification tool, later carried into quantum algebra.

The lesson is the same as for the radical: define your invariant by how the ring acts on modules, not by what its elements look like. Jacobson's definition needs no chain condition, and the price — losing left-right symmetry — turned out to be one worth paying.

Quick Reference

(11.2)R left primitive some simple left R-module is faithful
(11.3)𝔄 left primitive R/𝔄 left primitive
(11.4)𝔄 left primitive 𝔄=ann(M), M simple
(11.5)radR= left primitive ideals = right primitive ideals
TestFind a maximal left ideal containing no nonzero two-sided ideal
Standard exampleEnd(Vk), any nonzero right k-space V
AsymmetryBergman; Jategaonkar later
Commutative caseprimitive ideal = maximal ideal
Which statement to invoke
GoalUseHypothesis needed
Show an ideal is primitive(11.4), exhibit a simple module it annihilatesnone
Show a ring is primitive(11.2), exhibit a faithful simple modulenone
Compute radR(11.5)none
Force both sides to agree(11.11)R has a minimal left ideal
Get a matrix ring(11.19)(1)R left primitive and left artinian

Frequently Asked Questions

Why introduce primitive ideals at all, when primitive rings would seem enough?

Because most rings are not primitive, but every ring has primitive ideals, and they carry the information. The primitive ideals of R play the role that maximal ideals play in commutative algebra: their intersection is the radical, and the family of quotients R/𝔄 is a supply of well-understood rings into which R maps.

What exactly does Bergman's example show?

That there exists a ring with a faithful simple left module but no faithful simple right module. It does not show that the primitive ideals on the two sides are unrelated in intersection — (11.5) guarantees both intersections equal radR. It also does not contradict (11.11), since a ring with a minimal left ideal cannot behave this way; Bergman's ring necessarily has zero socle.

Is a primitive ideal prime?

Yes. By (11.6) a left primitive ring is prime, and applying that to R/𝔄 shows every left primitive ideal is prime. The converse fails: 0 is prime in but not primitive, because is a commutative non-field.

How do I actually test a concrete ring for left primitivity?

Find a maximal left ideal 𝔪 and show that the largest two-sided ideal inside 𝔪, namely ann(R/𝔪), is zero. In practice one first classifies the two-sided ideals of R and then checks none of the nonzero ones fits inside 𝔪. That is precisely the strategy of (11.12) and (11.13) for skew polynomial rings.

Do simple modules with the same annihilator have to be isomorphic?

No. A left primitive ring with zero socle can have many non-isomorphic faithful simple modules, all with annihilator 0; k[x;σ] with k=(t) and σ:tt+1 has infinitely many. Uniqueness holds when R has a minimal left ideal, by (11.11).

Is Mn(R) left primitive when R is?

Yes. Primitivity is a Morita invariant: if V is a faithful simple left R-module then the column space Vn is a faithful simple left Mn(R)-module. The argument is representation-theoretic rather than a direct manipulation of the definition.

References

  1. T. Y. Lam, A First Course in Noncommutative Rings, Graduate Texts in Mathematics 131, Springer-Verlag, 1991, §11, (11.2)–(11.5), pp. 183–185.
  2. N. Jacobson, Structure of Rings, American Mathematical Society Colloquium Publications 37, revised edition, 1964, Chapters I–II.
  3. G. M. Bergman, “A ring primitive on the right but not on the left”, Proceedings of the American Mathematical Society 15 (1964), 473–475.
  4. I. N. Herstein, Noncommutative Rings, Carus Mathematical Monographs 15, Mathematical Association of America, 1968, Chapter 2.
  5. L. H. Rowen, Ring Theory, Volume I, Academic Press, 1988, Chapter 2.

AI Suggested Questions

  • Sketch Bergman's construction of a ring that is primitive on one side but not the other.
  • Describe the Jacobson topology on the primitive spectrum and compute it for the first Weyl algebra.
  • Show that primitivity is a Morita invariant and identify exactly which step needs the equivalence.
  • For which group algebras kG of infinite groups is kG left primitive?
  • Compare primitive ideals with maximal ideals in a noetherian ring satisfying a polynomial identity.
  • Give a ring with exactly two primitive ideals whose intersection is nonzero, and identify its radical.
  • How do the primitive ideals of a quantised coordinate ring stratify under the Goodearl–Letzter theory?
Page
KEVOS-ENG-MATH-NCR-0085
Path
Engineering / Mathematics
Template
kevos-knowledge-article-v2
KEVOS® Knowledge Library — reviewed 2026-08-08

Continue learning

Absolutely Irreducible Modules and Schur’s Lemma over a Field | KEVOS®Article · Project ManagementAlgebraic and Geometric Multiplicities of Eigenvalues | KEVOS® MathematicsArticle · Project ManagementAmitsur’s Theorem on the Radical of a Polynomial Ring | KEVOS®Article · Project ManagementAmitsur’s Theorem on the Radical of an Algebra of Small Dimension | KEVOS®Article · Project Management