Double Centralizers and Density for Bimodules

Q: Why must k be exactly End(_R V) rather than any division ring over which V is a vector space?

Because End(V_k) shrinks as k grows. If k is a proper subring of End(_R V), the target ring End(V_k) is strictly larger than the one density controls, and the image of R can be a tiny subring of it — as with C inside M_2(R).

Q: What exactly does balanced mean, and is it the same as dense?

Balanced means the natural map : R → End(V_k) is surjective, so R realises its entire bicommutant. Dense is strictly weaker: it says the image is topologically dense in the finite topology. For semisimple V the two agree precisely when V is finitely generated over k.

Q: Why is 2-transitivity the natural stopping point?

Because it is exactly the statement that the centraliser of R is no bigger than k, and once that is known the Density Theorem provides all higher transitivity for free. Checking 2-transitivity is a condition on pairs of vectors; checking density directly would be an infinite family of conditions.

Q: Is the double centraliser property related to the double centralizer theorem for central simple algebras?

They are cousins. The finite-dimensional theorem says that for a simple subalgebra B of a finite-dimensional central simple algebra A one has C_A(C_A(B)) = B, with the dimensions multiplying to _k A. Both are instances of the principle that in a sufficiently semisimple environment, taking centralisers twice returns you to where you started.

Q: Does (11.18) need k to be a division ring?

Yes, as stated. The proof extends an independent set to a basis and writes dependent vectors as linear combinations, both of which require division. For general k the notion of m-transitivity itself becomes awkward, which is why density is defined directly by finite interpolation in the bimodule setting.

Q: What is the operator-algebra analogue, precisely?

Von Neumann's bicommutant theorem: a unital self-adjoint algebra of bounded operators on a Hilbert space is dense in its bicommutant for the weak and strong operator topologies. The shape is identical — algebra, commutant, bicommutant, density — but the hypotheses are analytic rather than semisimplicity, and neither theorem implies the other.

Executive Summary

Fix a left $R$ -module $V$ and forget everything but the abelian group. Three rings appear: the image $ρ (R)$ , its centraliser $k = End (_{R} V)$ acting on the right, and the centraliser of that, $E = End (V_{k})$ . Always $ρ (R) \subseteq E$ . Density is the assertion that this inclusion is invisible to any finite set of test vectors.

Two results make the bimodule picture usable. $(11.18)$ says dense action and $m$ -transitivity for all $m$ are the same condition when $k$ is a division ring. $(11.20)$ says $2$ -transitivity alone already forces $End (_{R} V) = k$ — the double centraliser condition — and hence density.

3Rings on one group

2Transitivity needed

$ρ (R) \subseteq E$ Always true

$ρ (R) = E$ Balanced module

Overview

A bimodule $_{R} V_{k}$ is two module structures on one abelian group that commute: $(r v) a = r (v a)$ . That single axiom says each ring acts by endomorphisms for the other, so each maps into the centraliser of the other. The theory of density is the quantitative version of that observation.

ρ : R ⟶ E = End (V_{k}), k ⟶ End (_{R} V), (r v) a = r (v a)

(11.18)

Each side acts inside the centraliser of the other; density asks how much of $E$ the left-hand side fills.

The delicate point is that $k$ may be given — a division ring we happen to be working over — or computed as $End (_{R} V)$ . Density theorems require the computed one. When the given $k$ is strictly smaller than the computed one, $E$ is strictly larger than it should be and density fails, often spectacularly.

This page is the bimodule-theoretic core behind The Jacobson Density Theorem and behind the structure theorem discussed in Wedderburn–Artin via Density. It also connects to the finite-dimensional double centralizer theorem for central simple algebras, treated in Criteria for Central Finiteness and the Double Centralizer Theorem.

Learning Objectives

Write down the centraliser and double centraliser of $ρ (R)$ acting on $V$ .
Prove $(11.18)$ : dense action on $V_{k}$ equals $m$ -transitivity for every $m$ .
Prove the implication $2$ -transitive $\Rightarrow$ $End (_{R} V) = k$ in $(11.20)$ .
Say precisely when a module is balanced and give a module that is not.
Derive the Artin–Whaples theorem from density applied to $R \otimes_{k} R^{op}$ .
Diagnose the failure caused by replacing $End (_{R} V)$ by a proper subring.

Definitions

Definition—Centraliser, double centraliser, balanced

Let $V$ be a left $R$ -module and write $End (_{R} V)$ on the right of $V$ , so that $V$ becomes a right module over $k : = End (_{R} V)$ . The double centraliser of $R$ on $V$ is $E : = End (V_{k})$ , acting on the left. The natural map $ρ : R \to E$ has $\ker ρ = ann (V)$ , and $V$ is called balanced (or is said to have the double centraliser property) when $ρ$ is surjective.

$(R, k)$ -bimodule: An abelian group with a left $R$ -action and a right $k$ -action satisfying $(r v) a = r (v a)$ ; equivalently a left $R$ -module with a ring map $k \to End (_{R} V)$ .
Dense action: For every $f \in End (V_{k})$ and finitely many $v_{1}, \dots, v_{n}$ , some $r \in R$ has $r v_{i} = f (v_{i})$ for all $i$ .
$m$ -transitive: For $k$ a division ring: any $n \leq m$ $k$ -independent vectors can be sent to any prescribed $n$ vectors by a single element of $ρ (R)$ .
Dense ring of linear transformations: A subring of $End (V_{k})$ that is $m$ -transitive for every finite $m$ ; equivalently a dense subring in the finite topology.
$C_{S} (T)$: The centraliser of a subset $T$ in a ring $S$ : all elements of $S$ commuting with every element of $T$ .

Writing $k$ on the right is not decoration: it makes the map $k \to End (_{R} V)$ a homomorphism rather than an anti-homomorphism, and it makes the identity $f (v a) = f (v) a$ available.

Core Concepts

The centraliser tower

Inside $End (V)$ — endomorphisms of the underlying abelian group — the three rings sit in a tower. Taking centralisers reverses inclusions, and taking it twice is a closure operation: $ρ (R) \subseteq C (C (ρ (R)))$ always, with equality exactly when $V$ is balanced.

ρ (R)

k = C (ρ (R)) = End (_{R} V)

E = C (k) = End (V_{k})

Three consequences follow from formal properties of centralisers alone, before any theorem: $ρ (R) \subseteq E$ ; the centre of $ρ (R)$ lies in $k \cap E$ ; and enlarging $ρ (R)$ shrinks $k$ , which in turn enlarges $E$ . Density is the one genuinely non-formal input, and it needs semisimplicity.

What each level of transitivity buys

$1$ -transitivity is the statement $R v = V$ for all $v \neq 0$ — simplicity of $_{R} V$ , no more. It constrains $k$ only through Schur's Lemma, which makes $k$ a division ring but says nothing about which one. $2$ -transitivity is a genuinely stronger demand: it forces $k$ to be no larger than the scalars we started with.

The transitivity ladder
Level	Equivalent condition	What it controls
$1$ -transitive	$_{R} V$ simple	the submodule lattice of $V$
$2$ -transitive	simple and $End (_{R} V) = k$	the centraliser $k$
$m$ -transitive for all $m$	dense in $End (V_{k})$	nothing further — it is implied by the previous line
Surjective onto $End (V_{k})$	$V$ balanced	requires $\dim_{k} V < \infty$ in the simple case

Key Results

Proposition(11.18)Dense action equals density as transformations

Let $V$ be an $(R, k)$ -bimodule with $k$ a division ring, let $E = End (V_{k})$ , and let $ρ : R \to E$ be the natural map. Then $R$ acts densely on $V_{k}$ if and only if $ρ (R)$ is a dense ring of linear transformations on $V$ , i.e. $m$ -transitive for every finite $m$ .

Proof

**( $\Rightarrow$ )** Let $v_{1}, \dots, v_{n}$ be $k$ -independent and $w_{1}, \dots, w_{n} \in V$ arbitrary. Since $k$ is a division ring, $V_{k}$ is a vector space; extend $v_{1}, \dots, v_{n}$ to a basis and let $f \in E$ be the $k$ -linear map with $f (v_{i}) = w_{i}$ and $f = 0$ on the other basis vectors. A dense action supplies $r \in R$ with $r v_{i} = f (v_{i}) = w_{i}$ for all $i$ , which is $n$ -transitivity.

**( $\Leftarrow$ )** Let $f \in E$ and $v_{1}, \dots, v_{n} \in V$ be arbitrary — not assumed independent. Reindex so that $v_{1}, \dots, v_{m}$ is a maximal $k$ -independent subset; then for $j > m$ we may write $v_{j} = \sum_{i \leq m} v_{i} a_{i j}$ with $a_{i j} \in k$ . By $m$ -transitivity choose $r \in R$ with $r v_{i} = f (v_{i})$ for $i \leq m$ . For $j > m$ , using that both $v \mapsto r v$ and $f$ are right $k$ -linear,

r v_{j} = \sum_{i \leq m} (r v_{i}) a_{i j} = \sum_{i \leq m} f (v_{i}) a_{i j} = f (\sum_{i \leq m} v_{i} a_{i j}) = f (v_{j}) .

So the same $r$ works on the whole list, which is the dense action. Note where the bimodule axiom was used: $r (v a) = (r v) a$ is what makes $v \mapsto r v$ right $k$ -linear.

Theorem(11.20)Transitivity and the double centraliser

Let $k$ be a division ring, $V \neq 0$ a right $k$ -vector space, and $R \subseteq E = End (V_{k})$ a subring. Then:

$R$ is $1$ -transitive if and only if $_{R} V$ is simple; in that case $V$ is a faithful simple left $R$ -module and $R$ is left primitive.
The following are equivalent: (a) $R$ is $2$ -transitive; (b) $R$ is $1$ -transitive and $End (_{R} V) = k$ ; (c) $R$ is dense in $E$ .

Proof

(1) $1$ -transitivity says exactly that $R v = V$ for every $v \neq 0$ , which is simplicity; faithfulness is automatic because $R$ is a subring of $E$ .

(2) **(b) $\Rightarrow$ (c).** By (1) the module $_{R} V$ is simple, hence semisimple, and by hypothesis $k$ is $End (_{R} V)$ . The Density Theorem $(11.16)$ gives a dense action, and $(11.18)$ converts it into density as a ring of transformations.

**(c) $\Rightarrow$ (a).** Immediate: density is $m$ -transitivity for all $m$ , in particular for $m = 2$ .

**(a) $\Rightarrow$ (b).** Assume $R$ is $2$ -transitive, hence also $1$ -transitive. Let $λ \in End (_{R} V)$ , written on the right as $v \mapsto v λ$ , and fix any $v \neq 0$ .

*Step 1: $v$ and $v λ$ are $k$ -dependent.* If they were independent, $2$ -transitivity would give $r \in R$ with $r v = 0$ and $r (v λ) \neq 0$ . But $λ$ commutes with the action of $R$ , so $r (v λ) = (r v) λ = 0$ — a contradiction.

Step 2: the scalar is global. By Step 1 write $v λ = v a$ for some $a \in k$ (take $a = 0$ if $v λ = 0$ ). Let $w \in V$ be arbitrary. By $1$ -transitivity there is $s \in R$ with $w = s v$ , and then

w λ = (s v) λ = s (v λ) = s (v a) = (s v) a = w a .

Hence $λ$ is right multiplication by $a$ , so $End (_{R} V) = k$ . Together with $1$ -transitivity this is (b).

Corollary—Burnside's theorem

Let $k$ be an algebraically closed field, $V$ a nonzero finite-dimensional $k$ -vector space and $R \subseteq {End}_{k} (V)$ a $k$ -subalgebra acting irreducibly on $V$ . Then $R = {End}_{k} (V)$ .

Proof

By Schur, $D : = End (_{R} V)$ is a division ring; it contains $k$ centrally and is finite-dimensional over $k$ because $V$ is. Any $d \in D$ generates a commutative field extension $k (d)$ of finite degree over the algebraically closed $k$ , so $k (d) = k$ and $D = k$ . Now $(11.20)$ (b) $\Rightarrow$ (c) gives density, and $(11.17)$ turns density into surjectivity because $\dim_{k} V < \infty$ .

Corollary—Artin–Whaples

Let $R$ be a simple ring with centre $k$ (a field), and let $x_{1}, \dots, x_{n} \in R$ be linearly independent over $k$ . Then for any $y_{1}, \dots, y_{n} \in R$ there exist $a_{1}, \dots, a_{m}$ and $b_{1}, \dots, b_{m}$ in $R$ with

y_{i} = \sum_{j = 1}^{m} a_{j} x_{i} b_{j} (1 \leq i \leq n) .

Proof

Let $A = R \otimes_{k} R^{op}$ act on $R$ by $(a \otimes b^{op}) \cdot x = a x b$ . The $A$ -submodules of $R$ are exactly the two-sided ideals, so simplicity of $R$ makes $_{A} R$ a simple module.

Next, $End (_{A} R) = k$ : an $A$ -endomorphism $ϕ$ satisfies $ϕ (a x b) = a ϕ (x) b$ , so taking $x = 1$ and then $a = 1$ or $b = 1$ gives $ϕ (a) = a ϕ (1)$ and $ϕ (b) = ϕ (1) b$ for all $a, b$ ; hence $ϕ (1)$ is central, i.e. $ϕ (1) \in k$ , and $ϕ$ is multiplication by that scalar.

So $R$ is a simple $A$ -module with endomorphism ring $k$ , and the Density Theorem applies to the $(A, k)$ -bimodule $R$ . Given $k$ -independent $x_{1}, \dots, x_{n}$ , choose $f \in End (R_{k})$ with $f (x_{i}) = y_{i}$ ; density yields an element $\sum_{j} a_{j} \otimes b_{j}^{op} \in A$ acting as $f$ on the $x_{i}$ , which is the displayed formula.

Remark—Rieffel's double centraliser theorem

A companion result with no semisimplicity hypothesis: let $R$ be a simple ring and $𝔞 \neq 0$ a left ideal; put $k = End (_{R} 𝔞)$ and $E = End (𝔞_{k})$ . Then the natural map $λ : R \to E$ is an isomorphism. Injectivity is simplicity; for surjectivity one checks that $f λ (a) = λ (f (a))$ for $f \in E$ , $a \in 𝔞$ , so $E λ (𝔞) \subseteq λ (𝔞)$ , and then $E = E λ (𝔞 R) \subseteq λ (𝔞) λ (R) \subseteq λ (R)$ using $𝔞 R = R$ . Simple rings are therefore always balanced on their left ideals.

Proof Techniques and Method

How these proofs work, and which moves transfer to other arguments.

Move 1

Independence first, linearity after

Reduce an arbitrary finite list of vectors to a maximal independent sublist, solve there, and extend by linearity. This is the whole of $(11.18)$ and it recurs whenever transitivity meets density.

Move 2

Separate to contradict

To show two vectors are dependent, assume independence and use transitivity to produce an element killing one and not the other — then contradict commutation. This is the engine of $(11.20)$ (a) $\Rightarrow$ (b).

Move 3

Turn a ring into a bimodule

Letting $R \otimes_{k} R^{op}$ act on $R$ converts two-sided ideal statements into module statements. Simplicity becomes irreducibility, and the centre becomes the endomorphism ring.

Move 3 is the most portable. Any question about two-sided ideals of $R$ is a question about submodules of $R$ over the enveloping ring $R \otimes_{k} R^{op}$ , and every theorem about simple modules becomes available. Artin–Whaples is the cleanest illustration; the same reduction underlies much of the theory of central simple algebras.

A discipline worth adopting: whenever a proof writes $r (v a)$ or $f (v a)$ , note explicitly which axiom licenses moving the scalar out. In the bimodule setting there are two different such axioms and they are easy to conflate.

Worked Example

$ℂ$ acting on itself over $ℝ$

Let $V = ℂ$ regarded as a right $ℝ$ -vector space with basis ${1, i}$ , and let $R = ℂ$ act by left multiplication. In this basis

ρ (a + b i) = (\begin{matrix} a & - b \\ b & a \end{matrix}) \in M_{2} (ℝ) ≅ End (V_{ℝ}) .

(E.1)

Check: $(a + b i) \cdot 1 = a + b i$ and $(a + b i) \cdot i = - b + a i$ .

The module $_{R} V$ is simple — $ℂ$ is a field and $V$ is one-dimensional over it — so $R$ is $1$ -transitive on $V$ . But $R$ is not $2$ -transitive: the vectors $1$ and $i$ are $ℝ$ -independent, yet no $r \in ℂ$ satisfies $r \cdot 1 = 1$ and $r \cdot i = 0$ , since the first equation forces $r = 1$ and then $r \cdot i = i \neq 0$ .

$(11.20)$ predicts the diagnosis exactly: $End (_{R} V) = ℂ$ , not $ℝ$ . Indeed an additive map commuting with multiplication by every complex number is $ℂ$ -linear, hence multiplication by its value at $1$ .

Repair the choice of $k$ and everything works

Take $k = End (_{R} V) = ℂ$ instead. Then $\dim_{k} V = 1$ , $End (V_{ℂ}) = ℂ$ , and $ρ$ is onto: $V$ is balanced and $R ≅ M_{1} (ℂ)$ , precisely as $(11.19)$ requires. The image measured over the wrong scalars had $ℝ$ -dimension $2$ inside a $4$ -dimensional ring; measured over the right scalars it is everything.

The general pattern

The same computation works for any finite field extension $K \supseteq k$ of degree $d > 1$ : view $V = K$ as a right $k$ -space and let $R = K$ act by multiplication. Then $_{R} V$ is simple, $End (_{R} V) = K \neq k$ , and $ρ (K)$ is a $d$ -dimensional subring of the $d^{2}$ -dimensional ring $End (V_{k})$ — as far from dense as the degree allows. Only when $k$ is algebraically closed does no such extension exist, which is why Burnside's theorem holds precisely there.

Comparison and Classification

Centraliser data for standard actions
$R$ acting on $V$	$End (_{R} V)$	$End (V_{k})$ for $k = End (_{R} V)$	Dense?	Balanced?
$M_{n} (D)$ on $D^{n}$	$D$	$M_{n} (D)$	yes	yes
$ℂ$ on $ℂ$ over $ℝ$	$ℂ$	$ℂ$	yes over $ℂ$ , no over $ℝ$	yes over $ℂ$
Upper triangular $T_{2} (F)$ on $F^{2}$	$F$	$M_{2} (F)$	no — $V$ is not semisimple	no
Weyl algebra $A_{1}$ on $k_{0} [y]$ , $char 0$	$k_{0}$	$End (k_{0} [y]_{k_{0}})$	yes	no
$F [x]$ on $F [x] / (f)$ , $f$ irreducible	$F [x] / (f)$	$F [x] / (f)$	yes	yes
$k \cdot 1 + finite rank$ on $⨁_{i} e_{i} k$	$k$	$End (V_{k})$	yes	no

Which hypotheses each conclusion actually requires
	$_{R} V$ semisimple	$_{R} V$ simple	$\dim_{k} V < \infty$	$k$ algebraically closed
$End (_{R} V)$ is a division ring	no	yes	no	no
$ρ (R)$ dense in $End (V_{k})$	yes	yes	no	no
$V$ balanced	yes	no	yes	no
$1$ -transitivity alone implies density	no	yes	yes	yes

Which hypotheses each conclusion actually requires

Relationship Map

Double centraliser statements form a family. They differ in what replaces semisimplicity as the hypothesis that makes the bicommutant computable.

Double centraliser results — each says $ρ (R)$ fills its bicommutant, exactly or approximately
- Approximate
  - Jacobson–Chevalley $(11.16)$ : $_{R} V$ semisimple $\Rightarrow$ $ρ (R)$ dense in $End (V_{k})$
  - Von Neumann bicommutant theorem: a unital self-adjoint operator algebra is weakly dense in its bicommutant
- Exact
  - $(11.17)$ : $V_{k}$ finitely generated $\Rightarrow$ $ρ$ onto
  - Burnside: $k$ algebraically closed, $\dim_{k} V < \infty$ , irreducible $\Rightarrow$ $R = {End}_{k} (V)$
  - Rieffel: $R$ simple, $𝔞$ a nonzero left ideal $\Rightarrow$ $R ≅ End (𝔞_{k})$
  - Double centralizer theorem for a simple subalgebra $B$ of a finite-dimensional central simple algebra $A$ : $C_{A} (C_{A} (B)) = B$
- Failure modes
  - $_{R} V$ not semisimple: bicommutant can be strictly larger than $ρ (R)$
  - $k$ chosen smaller than $End (_{R} V)$ : density fails outright

2

-transitive

End (_{R} V) = k

Density Theorem applies

m

-transitive for all

m

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

Absolute irreducibility

A representation of a group or algebra over a field $F$ is absolutely irreducible exactly when it is irreducible with endomorphism ring $F$ . That is condition (b) of $(11.20)$ , and it is the condition checked by every character-theoretic and computational test.

Computational algebra

MeatAxe endomorphism computations

Irreducibility and absolute irreducibility of matrix modules over finite fields are decided by computing the endomorphism algebra of the module; if it is larger than the base field the module is irreducible but not absolutely so, and the base field must be extended.

Operator algebras

Bicommutant philosophy

Von Neumann algebras are defined by the property of equalling their own bicommutant. The algebraic density theorem is the same statement with the weak topology replaced by the finite topology and self-adjointness replaced by semisimplicity.

Field theory

Galois theory of division rings

The Jacobson–Bourbaki correspondence matches division subrings of a division ring with certain rings of additive endomorphisms; density is the mechanism that makes the correspondence bijective.

The candid summary: this material is a tool for other parts of algebra. Its most concrete downstream consumer is the machinery that decides irreducibility of representations, which in turn drives symmetry-adapted computations in physics and chemistry codes.

Standards and Notation

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

Preferred here

End (_{R} V)

acting on the right;

End (V_{k})

acting on the left

Common variant

{End}_{R} (V)

with all maps written on the left, which turns

V

into a left module over the opposite ring

Operator-algebra notation

S'

for the commutant and

S''

for the bicommutant

Centraliser

C_{S} (T)

; older sources write

{Cent}_{S} (T)

V_{S} (T)

Balanceddouble centraliser property, balanced module, and density property all appear for the same condition

ImplementationsGAP MeatAxe functions test irreducibility and absolute irreducibility; Magma exposes an endomorphism algebra constructor for modules

Failure Modes and Common Mistakes

Do not apply $(11.18)$ when $k$ is not a division ring: the proof selects a maximal independent subset and extends by linearity, which needs bases.
Do not read $2$ -transitivity as a condition on arbitrary pairs; the source vectors must be $k$ -independent, and for dependent pairs the prescribed images are forced.
Do not assume the bicommutant of a non-semisimple module is computable from the module lattice: for $T_{2} (F)$ on $F^{2}$ the bicommutant is all of $M_{2} (F)$ while $ρ (R)$ has codimension one.
Do not transport a double centraliser statement across sides without checking; $End (_{R} V)$ and $End (V_{R})$ solve different problems.

Best Practices

Compute $End (_{R} V)$ before quoting any density statement, and record it explicitly alongside $V$ .
Certify density by checking $2$ -transitivity — one finite condition — rather than attempting $m$ -transitivity for all $m$ .
State on which side each ring is written the first time a bimodule appears, and never mix conventions inside a proof.
When a density argument fails, test the two standard causes first: the module is not semisimple, or $k$ is too small.
For finite-dimensional problems over a non-closed field, extend scalars to test absolute irreducibility before concluding anything about the image of $ρ$ .

Quick Reference

Tower

ρ (R) \subseteq E = End (V_{k})

k = End (_{R} V)

$(11.18)$ dense action

\Leftrightarrow

m

-transitive for all

m

(

k

a division ring)

$(11.20) (1)$

1

-transitive

\Leftrightarrow

_{R} V

simple

$(11.20) (2)$

2

-transitive

\Leftrightarrow

simple with

End (_{R} V) = k

\Leftrightarrow

dense

Balanced

ρ

onto

E

; holds if

_{R} V

semisimple and

V_{k}

finitely generated

Burnside

k

algebraically closed,

\dim_{k} V < \infty

, irreducible

\Rightarrow

R = {End}_{k} (V)

Artin–Whaples

R

simple with centre

k

y_{i} = \sum_{j} a_{j} x_{i} b_{j}

solvable for

k

-independent

x_{i}

Standard failure

ℂ

ℂ

over

ℝ

1

-transitive, not

2

-transitive

Diagnosing a density argument
Symptom	Likely cause	Fix
$2$ -transitivity fails	$End (_{R} V)$ is bigger than $k$	Replace $k$ by $End (_{R} V)$
Density holds but $ρ$ is not onto	$V_{k}$ is not finitely generated	Nothing to fix — this is the non-artinian branch
Bicommutant strictly bigger than $ρ (R)$	$_{R} V$ is not semisimple	Pass to a semisimple subquotient or a different module
Formulas produce an opposite ring	Endomorphisms written on the wrong side	Write $End (_{R} V)$ on the right

Frequently Asked Questions

Why must

k

be exactly

End (_{R} V)

rather than any division ring over which

V

is a vector space?

Because $End (V_{k})$ shrinks as $k$ grows. If $k$ is a proper subring of $End (_{R} V)$ , the target ring $End (V_{k})$ is strictly larger than the one density controls, and the image of $R$ can be a tiny subring of it — as with $ℂ$ inside $M_{2} (ℝ)$ .

What exactly does balanced mean, and is it the same as dense?

Balanced means the natural map $ρ : R \to End (V_{k})$ is surjective, so $R$ realises its entire bicommutant. Dense is strictly weaker: it says the image is topologically dense in the finite topology. For semisimple $V$ the two agree precisely when $V$ is finitely generated over $k$ .

Why is

2

-transitivity the natural stopping point?

Because it is exactly the statement that the centraliser of $R$ is no bigger than $k$ , and once that is known the Density Theorem provides all higher transitivity for free. Checking $2$ -transitivity is a condition on pairs of vectors; checking density directly would be an infinite family of conditions.

Is the double centraliser property related to the double centralizer theorem for central simple algebras?

They are cousins. The finite-dimensional theorem says that for a simple subalgebra $B$ of a finite-dimensional central simple algebra $A$ one has $C_{A} (C_{A} (B)) = B$ , with the dimensions multiplying to $\dim_{k} A$ . Both are instances of the principle that in a sufficiently semisimple environment, taking centralisers twice returns you to where you started.

Does

(11.18)

need

k

to be a division ring?

Yes, as stated. The proof extends an independent set to a basis and writes dependent vectors as linear combinations, both of which require division. For general $k$ the notion of $m$ -transitivity itself becomes awkward, which is why density is defined directly by finite interpolation in the bimodule setting.

What is the operator-algebra analogue, precisely?

Von Neumann's bicommutant theorem: a unital self-adjoint algebra of bounded operators on a Hilbert space is dense in its bicommutant for the weak and strong operator topologies. The shape is identical — algebra, commutant, bicommutant, density — but the hypotheses are analytic rather than semisimplicity, and neither theorem implies the other.

References

T. Y. Lam, A First Course in Noncommutative Rings, Graduate Texts in Mathematics 131, Springer-Verlag, 1991, §11, (11.18) and (11.20) (pp. 194–197), with Exercise 11.6 for Artin–Whaples.
N. Jacobson, Structure of Rings, American Mathematical Society Colloquium Publications 37, revised edition, 1964.
M. A. Rieffel, “A general Wedderburn theorem”, Proceedings of the National Academy of Sciences of the USA 54 (1965).
E. Artin and G. Whaples, “The theory of simple rings”, American Journal of Mathematics 65 (1943).
F. W. Anderson and K. R. Fuller, Rings and Categories of Modules, 2nd edition, Graduate Texts in Mathematics 13, Springer-Verlag, 1992.
L. H. Rowen, Ring Theory, Volume I, Academic Press, 1988, Chapter 2.

AI Suggested Questions

Give a module that is balanced but not semisimple, and one that is semisimple but not balanced.
Work out the bicommutant of the upper triangular matrices acting on $F^{2}$ and confirm it is all of $M_{2} (F)$ .
Prove Rieffel's theorem in full detail and compare it with the classical proof of Wedderburn–Artin.
How does absolute irreducibility of a representation relate to the Schur index and to extension of the base field?
State the Jacobson–Bourbaki correspondence precisely and identify where density is used.
For which rings is every simple module balanced?
Compare the finite topology on $End (V_{k})$ with the weak operator topology and explain why completeness holds in both.

01Executive Summary

02Overview

03Learning Objectives

04Definitions

05Core Concepts

The centraliser tower

What each level of transitivity buys

06Key Results

07Proof Techniques and Method

Independence first, linearity after

Separate to contradict

Turn a ring into a bimodule

08Worked Example

ℂ acting on itself over ℝ

Repair the choice of k and everything works

The general pattern

09Comparison and Classification

10Relationship Map

11Applications and Industry Use

Absolute irreducibility

MeatAxe endomorphism computations

Bicommutant philosophy

Galois theory of division rings

12Standards and Notation

13Failure Modes and Common Mistakes

14Best Practices

15Quick Reference

16Frequently Asked Questions

17Related KEVOS Topics

18References

19AI Suggested Questions

Continue learning

Executive Summary

Overview

Learning Objectives

Definitions

Core Concepts

Key Results

Proof Techniques and Method

Worked Example

$ℂ$ acting on itself over $ℝ$

Repair the choice of $k$ and everything works

Comparison and Classification

Relationship Map

Applications and Industry Use

Standards and Notation

Failure Modes and Common Mistakes

Best Practices

Quick Reference

Frequently Asked Questions

Related KEVOS Topics

References

AI Suggested Questions