Why is _Z(D) D a perfect square when it is finite?

Because after extending scalars to a splitting field L Z(D), the algebra becomes M_n(L), and dimension is unchanged by base change: _Z(D) D = _L (L ⊗ D) = n^2. The integer n is called the degree of D. The argument needs finite-dimensionality; for the Weyl field there is no such invariant.

Division Rings: Basic Theory and Examples

Q: Is every division ring an algebra over a field?

Yes — over its own centre Z(D), which is always a field. The prime field (Q or F_p) also sits inside Z(D), so D is an algebra over its prime field too. What is not automatic is that D is *finite-dimensional* over Z(D); that is the centrally finite hypothesis, and it is exactly what this chapter avoids assuming.

Q: Does a division ring have to have an identity by definition?

Yes, and the definition is not weakened by the requirement: any nonzero ring without identity in which every nonzero element has an inverse relative to *some* idempotent already contains that idempotent as a two-sided identity. The genuine variants are near-fields and quasi-fields, where one distributive law is dropped; Wedderburn's theorem fails for finite near-fields.

Q: What is the difference between a division ring and a domain?

A domain has no zero divisors; a division ring additionally has inverses. Every division ring is a domain, but Z and the Weyl algebra A_1(k) are domains that are not division rings. The bridge is (13.0e): a domain that is algebraic over a central subfield *is* a division ring, and more generally an Ore domain embeds in a division ring of fractions.

Q: Can a division ring contain a proper division subring of finite index?

Only if it is finite, by (13.24). So for infinite D, every proper division subring K has [D^* : K^*] = ∈fty. Faith strengthened this: the normalizer N_D^*(K^*) itself has infinite index whenever K is proper and noncentral.

Q: Why do the theorems keep concluding commutativity?

Because noncommutativity in a division ring is extremely rigid. There are no ideals to quotient by, so any hypothesis that constrains the multiplicative structure — finiteness, nilpotence of D^*, centrality of all commutators — propagates through the whole ring and leaves commutativity as the only possibility. The rigidity is a feature: it means the hypotheses in the theorems below are genuinely sharp.

Executive Summary

A division ring is a ring with identity in which every nonzero element is invertible — a field with commutativity deleted. Wedderburn–Artin makes them unavoidable: every semisimple ring is a finite product of matrix rings $M_{n} (D)$ over division rings, so classifying semisimple rings reduces to understanding division rings, and that reduction does not simplify further.

This page fixes the notation used across the whole division-ring stream — $D^{*}$ , $Z (D)$ , $C_{D} (S)$ , additive commutators $a b - b a$ , multiplicative commutators $x^{- 1} y^{- 1} x y$ — and collects the structural facts that every later argument assumes: centralizers are division subrings, dimensions multiply along a tower, and finiteness of any kind is a severe constraint.

$0, D$ Only left ideals

$Z (D)$ Always a field

$n^{2}$

\dim_{Z (D)} D

if finite

1843First example

The deliberate policy of this chapter is to avoid assuming $\dim_{Z (D)} D < \infty$ . Almost everything below therefore applies to the badly behaved division rings as well as to the classical ones.

Overview

Division rings occupy an odd position. They are the simplest possible noncommutative rings — no proper one-sided ideals, no zero divisors, no radical — and simultaneously the hardest to classify. The finite ones are all commutative (Wedderburn); the ones algebraic over a finite field are all commutative (Jacobson); the ones algebraic over $ℝ$ are $ℝ$ , $ℂ$ or $ℍ$ (Frobenius). Outside those regimes no classification exists.

Two invariants organise everything. The centre $Z (D)$ is a field, and $D$ is an algebra over it; the centralizers $C_{D} (S)$ interpolate between $Z (D)$ and $D$ . The dichotomy between $\dim_{Z (D)} D < \infty$ (the centrally finite case, the subject of Brauer group theory) and the infinite-dimensional case runs through the entire subject; the results of this section were chosen precisely because they need no such hypothesis.

The two commutator constructions are the working tools. Additive commutators detect the centre and generate $D$ ; multiplicative commutators do the same job with a multiplicative proof, and the two theories run in parallel — Cartan–Brauer–Hua is the multiplicative twin of the Lie-ideal result stated here.

Learning Objectives

State the definition of a division ring and prove the equivalence with having no nontrivial left ideals.
Use $D^{*}$ , $Z (D)$ , $C_{D} (S)$ and both commutator notations fluently.
Prove that $C_{D} (S)$ is a division subring containing $Z (D)$ .
Apply $\dim_{F} D = (\dim_{F} K) (\dim_{K} D)$ for a tower $F \subseteq K \subseteq D$ of division rings.
Give one centrally finite and one infinite-dimensional division ring with proof of the dimension claim.
Explain why finiteness — of $D$ , of a subring, of an index $[D^{*} : K^{*}]$ — forces commutativity or triviality.

Definitions

Definition(13.0)Division ring

A division ring $D$ is a ring with $1 \neq 0$ in which every nonzero element has a two-sided multiplicative inverse. Equivalently, $D \neq 0$ and $D^{*} : = D$ minus ${0}$ is a group under multiplication. A commutative division ring is a field; the terms skew field and sfield are synonyms for division ring in the older literature.

$D^{*}$: The multiplicative group of $D$ , that is $D$ with $0$ removed. Every statement about $D$ has a shadow as a statement about the group $D^{*}$ .
$Z (D)$: The centre, ${z \in D : z d = d z for all d \in D}$ . It is a field, and $D$ is a $Z (D)$ -algebra.
$C_{D} (S)$: The centralizer of a subset $S \subseteq D$ : all $d \in D$ with $d s = s d$ for every $s \in S$ . Written $C (S)$ when $D$ is clear.
Additive commutator: $a b - b a$ for $a, b \in D$ . Also written $[a, b]$ ; it is the value of the inner derivation $δ_{a}$ at $b$ up to sign.
Multiplicative commutator: $x^{- 1} y^{- 1} x y$ for $x, y \in D^{*}$ — the group commutator in $D^{*}$ . Often just called a commutator.
Division subring: A subring containing $1$ that is itself a division ring. The division subring generated by $S$ is the intersection of all those containing $S$ .
Centrally finite: $\dim_{Z (D)} D < \infty$ . Lam's chapter deliberately avoids this hypothesis; the centrally finite theory is the theory of central simple algebras.

C_{D} (S) = {d \in D : d s = s d for all s \in S}, Z (D) = C_{D} (D)

(13.0a)

The centre is the centralizer of everything; $C_{D} (\emptyset) = D$ .

Throughout, $D$ denotes a division ring, $F$ usually denotes $Z (D)$ , and ring means ring with identity.

Core Concepts

One-sided ideals, and why the radical theory is empty here

If $0 \neq a \in D$ and $𝔄$ is a left ideal containing $a$ , then $1 = a^{- 1} a \in 𝔄$ , so $𝔄 = D$ . A division ring therefore has exactly two left ideals and exactly two right ideals. It follows at once that $D$ is simple, left and right artinian, left and right noetherian, and $rad D = 0$ : every radical studied in this collection vanishes on $D$ . The interest lies entirely in the multiplicative structure.

Modules are vector spaces

Every left $D$ -module is free: a maximal $D$ -independent subset is a basis, by the same argument as over a field, and the rank is well defined because $D$ has invariant basis number. This is what licenses the phrase * $\dim_{D} V$ * for a module over a division ring and is the reason ${End}_{D} (V) ≅ M_{n} (D^{op})$ for $\dim V = n$ ; the opposite ring appears because composing endomorphisms written on one side reverses multiplication.

Towers and the transitivity of dimension

Let $F \subseteq K \subseteq D$ be division rings with $F$ contained in $Z (D)$ . Then $D$ is a left $K$ -vector space and $K$ is an $F$ -vector space, and choosing bases ${u_{i}}$ of $K$ over $F$ and ${v_{j}}$ of $D$ over $K$ makes ${u_{i} v_{j}}$ an $F$ -basis of $D$ .

\dim_{F} D = (\dim_{F} K) (\dim_{K} D)

(13.0b)

In particular, if $\dim_{F} D = n$ is finite then $\dim_{F} K$ divides $n$ for every intermediate division ring $K$ — the divisibility used in Wedderburn's Little Theorem.

The two commutators

The additive commutator $a b - b a$ measures failure of commutativity inside the additive group; the multiplicative commutator $x^{- 1} y^{- 1} x y$ does so inside $D^{*}$ . Each generates a parallel body of results. Both are trivial exactly when $D$ is commutative, and in both cases the sharper statement is available: if the commutators merely lie in the centre, $D$ is already commutative.

All commutators

= 0

All commutators central

D

commutative

Key Results

Proposition(13.0c)Ideal characterisation

Let $R$ be a ring with $1 \neq 0$ . Then $R$ is a division ring if and only if its only left ideals are $0$ and $R$ .

Proof

If $R$ is a division ring the argument above gives the ideal condition. Conversely, suppose $0$ and $R$ are the only left ideals and let $a \neq 0$ . Then $R a$ is a nonzero left ideal, so $R a = R$ and $b a = 1$ for some $b$ . Now $b \neq 0$ , so the same argument produces $c$ with $c b = 1$ . Then $c = c (b a) = (c b) a = a$ , so $a b = 1$ as well and $a$ is invertible.

Proposition(13.0d)Centralizers

For any subset $S \subseteq D$ , the centralizer $C_{D} (S)$ is a division subring of $D$ containing $Z (D)$ . Moreover $C_{D} (S) = C_{D} (T)$ where $T$ is the division subring generated by $S$ , and $S \subseteq C_{D} (C_{D} (S))$ .

Proof

Closure under subtraction and multiplication is immediate from bilinearity of the ring operations, and $1 \in C_{D} (S)$ , so $C_{D} (S)$ is a subring. If $d \in C_{D} (S)$ is nonzero and $s \in S$ , then from $d s = s d$ we get $s d^{- 1} = d^{- 1} (d s) d^{- 1} = d^{- 1} (s d) d^{- 1} = d^{- 1} s$ , so $d^{- 1} \in C_{D} (S)$ ; hence $C_{D} (S)$ is a division subring. Central elements commute with everything, so $Z (D) \subseteq C_{D} (S)$ . For the second claim, the set of elements commuting with a fixed $d$ is a division subring, so if $d$ centralises $S$ it centralises the division subring generated by $S$ ; the last inclusion is a restatement of the definition.

Proposition(13.0e)Algebraic domains are division algebras

Let $F$ be a field and $A$ an $F$ -algebra (so $F \subseteq Z (A)$ ) that is a domain and is algebraic over $F$ , meaning every element of $A$ satisfies a nonzero polynomial with coefficients in $F$ . Then $A$ is a division ring.

Proof

Fix $0 \neq d \in A$ . Since $F$ is central, $F [d]$ is a commutative subring, and it is a domain because $A$ is. It is finite-dimensional over $F$ because $d$ is algebraic. A finite-dimensional commutative domain over a field is a field: multiplication by $d$ is an injective $F$ -linear endomorphism of the finite-dimensional space $F [d]$ , hence surjective, so $d e = 1$ for some $e \in F [d]$ . Thus $d$ is invertible in $A$ .

Theorem(13.1)Wedderburn's Little Theorem

Every finite division ring is a field. Consequently a noncommutative division ring is infinite, and so is every noncommutative simple artinian ring.

Theorem(13.12)Frobenius

Let $D$ be a division algebra over $ℝ$ which is algebraic over $ℝ$ — finite-dimensionality is not assumed. Then $D$ is isomorphic as an $ℝ$ -algebra to $ℝ$ , $ℂ$ or $ℍ$ . The same conclusion holds over any real-closed field in place of $ℝ$ .

Theorem(13.26)Herstein: conjugacy classes are infinite

If $a$ is a noncentral element of a division ring $D$ , then $a$ has infinitely many conjugates $d a d^{- 1}$ , $d \in D^{*}$ . (Scott's refinement: the conjugacy class has the same cardinality as $D$ .)

Corollary(13.21)Nilpotence of the multiplicative group

$D^{*}$ is a nilpotent group if and only if $D$ is a field. The same holds with solvable in place of nilpotent, a considerably harder theorem of L. K. Hua.

Corollary(13.24)Finite index forces finiteness

For division rings $K ⊊ D$ , the index $[D^{*} : K^{*}]$ is finite if and only if $D$ is finite. So a proper division subring of an infinite division ring always has infinite index.

Proofs of $(13.1)$ , $(13.12)$ , $(13.21)$ and $(13.24)$ are developed on the dedicated pages of this stream; $(13.0 c)$ – $(13.0 e)$ are proved above because everything else silently uses them.

Proof Techniques and Method

How arguments about division rings actually proceed, and which move to reuse.

Move 1

Invert to force membership

To show $a$ lies in a division subring $K$ , exhibit $a$ as a quotient $u v^{- 1}$ of two elements already known to be in $K$ . Nearly every result on this stream — Lie ideals, Cartan–Brauer–Hua, generation by commutators — ends with exactly this step.

Move 2

Count in $D^{*}$

Translate a ring statement into a group statement about $D^{*}$ and use group theory: class equations, normalizers, upper central series, finite subgroups. Wedderburn's proof is the model.

Move 3

Linearise by a derivation

Replace the element $a$ by the operator $δ_{a}$ acting on $D$ as a vector space over a commutative subring, then use linear algebra — eigenvectors, minimal polynomials, the Frobenius power $δ^{p^{n}}$ .

A fourth habit is worth naming: never assume a side. Statements about left dimension and right dimension over a division subring can differ, and a division subring need not be closed under conjugation. When a result is one-sided in the source it is one-sided here.

Worked Example

Centralizers and dimensions inside $ℍ$

Let $ℍ = ℝ \oplus ℝ i \oplus ℝ j \oplus ℝ k$ with $i^{2} = j^{2} = k^{2} = - 1$ and $i j = k = - j i$ . Write a general element $α = a_{0} + a_{1} i + a_{2} j + a_{3} k$ .

**Compute $C_{ℍ} (i)$ .** From $i α - α i$ we get, using $i j = k$ , $i k = - j$ :

i α - α i = 2 a_{2} k - 2 a_{3} j,

(E.1)

so $α$ commutes with $i$ exactly when $a_{2} = a_{3} = 0$ .

Hence $C_{ℍ} (i) = ℝ \oplus ℝ i ≅ ℂ$ , a maximal subfield. Check the tower formula $(13.0 b)$ with $F = Z (ℍ) = ℝ$ and $K = C_{ℍ} (i)$ :

\dim_{ℝ} ℍ = 4, \dim_{ℝ} K = 2, \dim_{K} ℍ = 2, 4 = 2 \cdot 2 .

(E.2)

Adding a second generator collapses the centralizer: $C_{ℍ} ({i, j}) = C_{ℍ} (i) \cap C_{ℍ} (j) = (ℝ \oplus ℝ i) \cap (ℝ \oplus ℝ j) = ℝ = Z (ℍ)$ , consistent with ${i, j}$ generating $ℍ$ as a division ring.

The conjugacy class of $i$

For a unit quaternion $u$ , the map $α \mapsto u α u^{- 1}$ preserves the real part and the norm, so every conjugate of $i$ is a pure quaternion of norm $1$ . Conversely every pure unit $v = b_{1} i + b_{2} j + b_{3} k$ with $b_{1}^{2} + b_{2}^{2} + b_{3}^{2} = 1$ satisfies $v^{2} = - 1$ , and one checks that $u = (1 + v i^{- 1})$ , when nonzero, conjugates $i$ to $v$ ; the exceptional case $v = - i$ is handled by $u = j$ . The class of $i$ is therefore the whole unit $2$ -sphere of pure quaternions — uncountably infinite, exactly as $(13.26)$ predicts.

An infinite-dimensional example

Let $K = ℚ (t)$ and let $σ$ be the automorphism with $σ (t) = 2 t$ , of infinite order. In the skew Laurent series division ring $D = K ((x; σ))$ , where $x c = σ (c) x$ for $c \in K$ , a series $\sum a_{i} x^{i}$ is central iff $a_{i} c = a_{i} σ^{i} (c)$ for all $c \in K$ and all $i$ . Since $σ$ has infinite order this forces $a_{i} = 0$ for $i \neq 0$ , and $a_{0}$ fixed by $σ$ , so $Z (D) = ℚ$ and $\dim_{Z (D)} D = \infty$ : $D$ is not centrally finite.

Comparison and Classification

Division rings people actually use
Division ring	Centre	$\dim$ over centre	How it is built
A field $k$	$k$	$1$	Trivial case
$ℍ$	$ℝ$	$4$	Hamilton's relations; norm form anisotropic
Quaternion algebra $(\frac{a, b}{k})$	$k$	$4$	Division iff the norm form is anisotropic over $k$
Cyclic algebra $(K / k, σ, a)$	$k$	$n^{2}$	Crossed product of a cyclic extension
$K ((x; σ))$ , $σ$ of order $n$	$Fix (σ) ((x^{n}))$	$n^{2}$	Skew Laurent series
$K ((x; σ))$ , $σ$ of infinite order	$Fix (σ)$	infinite	Skew Laurent series
Weyl field $Frac (A_{1} (k))$ , $char k = 0$	$k$	infinite	Ore localisation of the Weyl algebra
Mal'cev–Neumann series over an ordered group	varies	usually infinite	Formal series with well-ordered support

Which finiteness hypothesis buys which conclusion
	$D$ finite	$D$ algebraic over a finite field	$D$ algebraic over $ℝ$	$D$ centrally finite	No hypothesis
$D$ is commutative	yes	yes	no	no	no
$D$ classified up to isomorphism	yes	yes	yes	partial	no
$\dim_{Z (D)} D$ is a perfect square	yes	partial	yes	yes	no
Finite subgroups of $D^{*}$ are cyclic	yes	yes	no	partial	partial
Every element is algebraic over $Z (D)$	yes	yes	yes	yes	no

Which finiteness hypothesis buys which conclusion

Part in row 3 for the algebraic-over-a-finite-field column records that such a $D$ is commutative, so the dimension is $1$ only when $D = Z (D)$ ; the square condition is then vacuous rather than informative.

Relationship Map

Division rings sit at the bottom of the structural hierarchy of rings, and every containment below is strict.

Ringsno hypotheses

Semiprimitive

rad R = 0

Semisimplefinite product of

M_{n_{i}} (D_{i})

Simple artinian

R ≅ M_{n} (D)

D

a division ring

Division rings

n = 1

Fieldscommutative; contains every finite division ring

Inside a fixed $D$ , the subobject lattice is organised by centralizers.

$D$ — the whole division ring
- Maximal subfields
  - $C_{D} (a)$ for $a$ generating a maximal commutative subring
  - in the centrally finite case, of degree $n$ over $Z (D)$ where $\dim_{Z (D)} D = n^{2}$
- Centralizers $C_{D} (S)$
  - division subrings containing $Z (D)$
  - $\dim_{Z (D)} C_{D} (S)$ divides $\dim_{Z (D)} D$ when finite
- $Z (D)$
  - a field
  - the centralizer of everything
  - $C_{D} (C_{D} (Z (D))) = Z (D)$

D

finite

D

commutative

D^{*}

cyclic if finite

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

Schur endomorphism rings

${End}_{R} (M)$ for simple $M$ is a division ring; whether it equals the ground field is exactly the question of absolute irreducibility, and drives the theory of splitting fields and Schur indices.

Geometry and robotics

Quaternion kinematics

Unit quaternions parametrise $S O (3)$ two-to-one. That $ℍ$ is a division ring is what makes normalisation and inversion of rotations numerically stable, and the conjugacy computation above is the rotation action itself.

Coding theory

Space-time codes

Cyclic division algebras over number fields supply space-time block codes for multi-antenna channels: the division property is precisely the non-vanishing determinant condition that guarantees full diversity.

Symbolic computation

Ore localisation

The Weyl algebra and skew polynomial rings are Ore domains, so they embed in division rings of fractions. Computer algebra systems use this to solve linear systems over differential and difference operators.

Control theory

Time-varying systems

Linear time-varying systems are modules over skew polynomial rings; passing to the division ring of fractions gives well-defined notions of rank, transfer function and controllability index.

Number theory

Brauer groups

Centrally finite division algebras over a field $k$ are the elements of $Br (k)$ , and local-global principles for them are a pillar of class field theory.

The honest summary: outside quaternions and cyclic algebras, division rings are infrastructure. They are what you land on when you take a simple module and ask for its endomorphisms, and their theory is consumed rather than exhibited.

Standards and Notation

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

Preferred

D

for a division ring;

D^{*}

for its multiplicative group

Variants

\dot{D}

D^{\times}

for the multiplicative group; skew field and sfield for the object

Centre

Z (D)

; some authors write

Cent (D)

K (D)

Commutators

[a, b] = a b - b a

additively;

(x, y)

[x, y]

for

x^{- 1} y^{- 1} x y

multiplicatively — always say which

Degree

\deg D = n

where

\dim_{Z (D)} D = n^{2}

, only for centrally finite

D

MarkupPresentation MathML per ISO/IEC 40314; symbol conventions per ISO 80000-2

GAP / MagmaQuaternionAlgebra, IsDivisionRing, CyclicAlgebra

SageQuaternionAlgebra(K,a,b), A.is_division_algebra()

Computational Notes

Computational notes cover algorithms, cost and library behaviour rather than manufacturing process.

Deciding the division property. For a quaternion algebra over a number field this is a finite computation: check that the norm form is anisotropic at every place, which by Hilbert reciprocity reduces to finitely many Hilbert symbols.
Structure constants. A finite-dimensional algebra given by structure constants is tested for the division property by computing its radical, applying Wedderburn–Artin to the semisimple quotient, and checking that a single matrix block of size $1$ results — polynomial time over a finite field or a number field.
**Arithmetic in $ℍ$ .** Multiplication costs $16$ real multiplications naively, $8$ with a Karatsuba-style scheme; inversion is one conjugation and a division by the norm, so it is numerically well conditioned away from $0$ .
Ore localisation. Solving linear systems over $k [x; σ, δ]$ requires the Ore condition and a noncommutative Gaussian elimination; degree growth, not the operation count, dominates the cost.
Undecidability. There is no algorithm that decides whether a finitely presented ring is a division ring — the word problem already fails to be decidable for finitely presented rings.

Failure Modes and Common Mistakes

Do not assume $\dim_{Z (D)} D$ is finite; the theorems on this stream were chosen because they do not need it.
Do not assume every maximal subfield has the same degree unless $D$ is centrally finite — in the infinite-dimensional case maximal subfields can be wildly different sizes.
Do not read ${End}_{D} (V) ≅ M_{n} (D)$ without checking sides; for a left $D$ -space acting on the left, the opposite ring $D^{op}$ appears.
Do not expect $D^{*}$ to be abelian-by-anything: $(13.21)$ says nilpotence already forces commutativity.

Historical Notes and Lessons Learned

1843Hamilton constructs $ℍ$ Commutativity is abandoned deliberately in order to obtain a four-dimensional real division algebra; the first noncommutative division ring in print.
1877FrobeniusClassifies the finite-dimensional real division algebras: only $ℝ$ , $ℂ$ and $ℍ$ . The statement in $(13.12)$ needs only algebraicity.
1903–1905Moore, then WedderburnMoore classifies finite fields; two years later Wedderburn proves that no other finite division rings exist. Witt's cyclotomic proof follows in 1931.
1929–1932Brauer, Hasse, Noether, AlbertCentrally finite division algebras over number fields are shown to be cyclic; the Brauer group becomes a computable object.
1949–1950Cartan, Brauer, HuaIndependently prove that a conjugation-invariant proper division subring is central — the multiplicative analogue of the Lie-ideal result of this section.
1955AmitsurDetermines exactly which finite groups embed in the multiplicative group of a division ring, closing the question raised by the quaternion and binary tetrahedral groups in $ℍ^{*}$ .

The lesson that repeats: every finiteness hypothesis one can impose on a division ring — finite order, finite subgroup, finite index, algebraic over a finite or real-closed field — collapses the object drastically. Progress in the subject has come from learning which hypotheses can be dropped, not from adding more.

Quick Reference

DefinitionEvery nonzero element invertible;

D^{*}

is a group

IdealsOnly

0

and

D

, on either side

Radical

rad D = 0

;

D

is simple artinian

Centre

Z (D)

is a field;

D

is a

Z (D)

-algebra

Centralizer

C_{D} (S)

is a division subring containing

Z (D)

Towers

\dim_{F} D = (\dim_{F} K) (\dim_{K} D)

for

F \subseteq K \subseteq D

F

central

Finite caseFinite

\Rightarrow

commutative

(13.1)

ConjugacyNoncentral elements have infinitely many conjugates

(13.26)

Group side

D^{*}

nilpotent

iff

D

a field

(13.21)

Notation used throughout the division-ring stream
Symbol	Meaning	First appearance
$D$	a division ring	§13
$D^{*}$	multiplicative group of $D$	§13
$Z (D)$	centre of $D$	§13
$C_{D} (S)$ , $C (S)$	centralizer of a subset	§13
$a b - b a$	additive commutator	(13.4)
$x^{- 1} y^{- 1} x y$	multiplicative commutator	(13.15)
$δ_{a}$	inner derivation $x \mapsto a x - x a$	(13.7)
$F (S)$	division subring generated by a field $F$ and $S$	(13.10)

Frequently Asked Questions

Is every division ring an algebra over a field?

Yes — over its own centre $Z (D)$ , which is always a field. The prime field ( $ℚ$ or $𝔽_{p}$ ) also sits inside $Z (D)$ , so $D$ is an algebra over its prime field too. What is not automatic is that $D$ is finite-dimensional over $Z (D)$ ; that is the centrally finite hypothesis, and it is exactly what this chapter avoids assuming.

Why is

\dim_{Z (D)} D

a perfect square when it is finite?

Because after extending scalars to a splitting field $L \supseteq Z (D)$ , the algebra becomes $M_{n} (L)$ , and dimension is unchanged by base change: $\dim_{Z (D)} D = \dim_{L} (L \otimes D) = n^{2}$ . The integer $n$ is called the degree of $D$ . The argument needs finite-dimensionality; for the Weyl field there is no such invariant.

Does a division ring have to have an identity by definition?

Yes, and the definition is not weakened by the requirement: any nonzero ring without identity in which every nonzero element has an inverse relative to some idempotent already contains that idempotent as a two-sided identity. The genuine variants are near-fields and quasi-fields, where one distributive law is dropped; Wedderburn's theorem fails for finite near-fields.

What is the difference between a division ring and a domain?

A domain has no zero divisors; a division ring additionally has inverses. Every division ring is a domain, but $ℤ$ and the Weyl algebra $A_{1} (k)$ are domains that are not division rings. The bridge is $(13.0 e)$ : a domain that is algebraic over a central subfield is a division ring, and more generally an Ore domain embeds in a division ring of fractions.

Can a division ring contain a proper division subring of finite index?

Only if it is finite, by $(13.24)$ . So for infinite $D$ , every proper division subring $K$ has $[D^{*} : K^{*}] = \infty$ . Faith strengthened this: the normalizer $N_{D^{*}} (K^{*})$ itself has infinite index whenever $K$ is proper and noncentral.

Why do the theorems keep concluding commutativity?

Because noncommutativity in a division ring is extremely rigid. There are no ideals to quotient by, so any hypothesis that constrains the multiplicative structure — finiteness, nilpotence of $D^{*}$ , centrality of all commutators — propagates through the whole ring and leaves commutativity as the only possibility. The rigidity is a feature: it means the hypotheses in the theorems below are genuinely sharp.

References

T. Y. Lam, A First Course in Noncommutative Rings, Graduate Texts in Mathematics 131, Springer-Verlag, 1991, §13 (pp. 213–226).
T. Y. Lam, A First Course in Noncommutative Rings, §14, for multiplicative commutators, cyclic algebras and the Cartan–Brauer–Hua circle of ideas.
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 3.
P. M. Cohn, Skew Fields: Theory of General Division Rings, Encyclopedia of Mathematics and its Applications 57, Cambridge University Press, 1995.
L. H. Rowen, Ring Theory, Volume I, Academic Press, 1988, Chapter 3.

AI Suggested Questions

Construct explicitly a division ring that is infinite-dimensional over its centre and verify the centre computation.
Prove that a left and right Ore domain embeds in a division ring of fractions, and check that the Weyl algebra satisfies the Ore condition.
Give Cohn's example of a division ring extension where the left and right dimensions differ.
How does the degree of a centrally finite division algebra relate to the orders of elements in the Brauer group?
Explain why ${End}_{D} (V) ≅ M_{n} (D^{op})$ rather than $M_{n} (D)$ , with a careful choice of side conventions.
Which finite groups embed in $ℍ^{*}$ , and how does that list relate to the finite subgroups of $S U (2)$ ?
Compare the maximal subfields of a quaternion division algebra over $ℚ$ with those of $ℍ$ .

01Executive Summary

02Overview

03Learning Objectives

04Definitions

05Core Concepts

One-sided ideals, and why the radical theory is empty here

Modules are vector spaces

Towers and the transitivity of dimension

The two commutators

06Key Results

07Proof Techniques and Method

Invert to force membership

Count in D∗

Linearise by a derivation

08Worked Example

Centralizers and dimensions inside ℍ

The conjugacy class of i

An infinite-dimensional example

09Comparison and Classification

10Relationship Map

11Applications and Industry Use

Schur endomorphism rings

Quaternion kinematics

Space-time codes

Ore localisation

Time-varying systems

Brauer groups

12Standards and Notation

13Computational Notes

14Failure Modes and Common Mistakes

15Historical Notes and Lessons Learned

16Quick Reference

17Frequently Asked Questions

18Related KEVOS Topics

19References

20AI Suggested Questions

Continue learning

Executive Summary

Overview

Learning Objectives

Definitions

Core Concepts

Key Results

Proof Techniques and Method

Count in $D^{*}$

Worked Example

Centralizers and dimensions inside $ℍ$

The conjugacy class of $i$

Comparison and Classification

Relationship Map

Applications and Industry Use

Standards and Notation

Computational Notes

Failure Modes and Common Mistakes

Historical Notes and Lessons Learned

Quick Reference

Frequently Asked Questions

Related KEVOS Topics

References

AI Suggested Questions