Multiplicative Commutators in Division Rings

Q: How does (13.17) differ from the additive result (13.7) in strength?

They have the same shape but incomparable hypotheses. (13.7) assumes K is a Lie ideal — closed under x ax - xa — and needs char K ≠ 2. (13.17) assumes K^* is normal in D^* and needs nothing about the characteristic. Neither hypothesis implies the other, but in practice normality is the easier one to verify, which is why (13.18) and (13.19) go through the multiplicative version.

Q: What is the status of Herstein's conjecture?

Herstein proved it when D is countable. Lam's 1991 text records the general case as open, and this page follows that record rather than asserting a resolution. The hypothesis is a substantial weakening of (13.16): instead of demanding that commutators be central, it demands only that some power of each be.

Executive Summary

A division ring is simple, so its subobjects must be detected multiplicatively. The multiplicative commutators $x^{- 1} y^{- 1} x y$ do this. Their common centralizer is exactly $Z (D)$ ; if they are all central then $D$ is a field; and if $D$ is not a field they generate it outright — no central coefficients required.

The centrepiece is the Cartan–Brauer–Hua theorem: a proper division subring $K ⊊ D$ whose unit group is normal in $D^{*}$ must be central. Unlike its additive analogue $(13.7)$ , it holds in every characteristic. Everything follows from one identity, obtained by setting $b = a - 1$ and comparing what $a$ and $b$ do to a fixed element $c$ .

$x^{- 1} y^{- 1} x y$ The basic object

$b = a - 1$ The trick

AnyCharacteristic required

5Results, (13.15)–(13.19)

Overview

Section 13 treats commutators twice. The additive pass, $(13.4)$ through $(13.7)$ , is quick and leaves a characteristic- $2$ gap. The multiplicative pass, $(13.15)$ through $(13.19)$ , reaches the same four conclusions with no gap, at the cost of a less obvious identity.

The obstruction to copying the additive proofs is plain. There the key was $[x, x y] = x [x, y]$ : one commutator equals another divided by $x$ , so $x$ can be recovered. Multiplicatively there is no such factorisation — commutators multiply, they do not scale. The substitute exploits the fact that a division ring has an additive structure too: from $a$ one may form $b = a - 1$ , and $a$ , $b$ then differ by a central element while behaving differently under conjugation.

That single line proves $(13.15)$ and $(13.17)$ , and those two prove $(13.16)$ , $(13.18)$ and $(13.19)$ . Note that $b \neq 0$ because $a \neq 1$ , and $b$ fails to commute with $c$ precisely because $a$ does — subtracting a central element changes nothing about commutation.

Learning Objectives

Verify identities $(13.13)$ and $(13.14)$ line by line and explain why each side is nonzero.
Prove $(13.15)$ : the common centralizer of the multiplicative commutators is $Z (D)$ .
Deduce $(13.16)$ : all multiplicative commutators central implies $D$ is a field.
State the Cartan–Brauer–Hua theorem $(13.17)$ with the hypotheses $K$ normal and $K \neq D$ .
Prove $(13.18)$ : the conjugates of a noncentral element generate $D$ .
Prove $(13.19)$ and explain why no central coefficients are needed, in contrast with $(13.6)$ .

Definitions

Definition§13Multiplicative commutators and normality

For $x, y \in D^{*}$ the element $x^{- 1} y^{- 1} x y$ is a multiplicative commutator; it is common to drop the adjective and simply say commutator when the multiplicative sense is clear. For division rings $K \subseteq D$ , we say $K$ is **normal in $D$ ** if $x K x^{- 1} \subseteq K$ for every $x \in D^{*}$ — equivalently, $K^{*}$ is a normal subgroup of $D^{*}$ .

$x^{- 1} y^{- 1} x y$: The commutator of $x$ and $y$ in the group $D^{*}$ . It equals $1$ exactly when $x$ and $y$ commute.
$a^{- 1} c a c^{- 1}$: Also a multiplicative commutator — of the pair $(a, c^{- 1})$ . This is the form in which commutators appear in identity $(13.14)$ .
$K$ normal in $D$: $x K x^{- 1} \subseteq K$ for all $x \in D^{*}$ . Taking $x^{- 1}$ as well gives equality, so normality is a genuine invariance condition.
$b = a - 1$: The standard substitution in this section. It satisfies $b \in D^{*}$ whenever $a \neq 1$ , and $b$ commutes with an element exactly when $a$ does.
Conjugacy class of $d$: ${x d x^{- 1} : x \in D^{*}}$ . By $(13.18)$ its elements generate $D$ as a division ring as soon as $d \notin Z (D)$ .

Both (13.7) and (13.17) require the subring to be proper; without that hypothesis K = D satisfies every condition and the conclusion is false.

Core Concepts

Deriving the identity

Fix $a, c \in D$ that do not commute, so $a \neq 0, 1$ and $c \neq 0$ . Set $b = a - 1$ , which is nonzero and again does not commute with $c$ . Expand, using $a = b + 1$ :

a (a^{- 1} c a - b^{- 1} c b) = c a - a b^{- 1} c b = c (b + 1) - (b + 1) b^{- 1} c b = (c b + c) - (c b + b^{- 1} c b) = c - b^{- 1} c b .

(13.13)

The right-hand side is nonzero because $b$ does not commute with $c$ .

Multiplying on the right by $c^{- 1}$ converts both conjugates into genuine commutators:

a (a^{- 1} c a c^{- 1} - b^{- 1} c b c^{- 1}) = 1 - b^{- 1} c b c^{- 1} \neq 0 .

(13.14)

$a^{- 1} c a c^{- 1}$ and $b^{- 1} c b c^{- 1}$ are multiplicative commutators; the right-hand side is $1$ minus one of them.

Two ways to read the identity

Both readings solve for $a$ , and which one you want depends on what you control.

What do you know about the two commutators appearing in the identity?

They commute with $c$ Then so do the difference and the right-hand side. Cancelling the invertible difference gives

a c = c a

— a contradiction. This is

(13.15)

They lie in a division subring $K$ Use the form

(13.13)

: if

a^{- 1} c a

b^{- 1} c b

and

c

all lie in

K

, then

a

is a quotient of two nonzero elements of

K

, so

a \in K

. This is the Cartan–Brauer–Hua theorem

(13.17)

NeitherThe identity is still true but says nothing. Its whole value lies in transporting a hypothesis about commutators back onto

a

Why $b = a - 1$ and not some other shift

Two properties are needed simultaneously: $b$ must be invertible, and $a$ must be expressible in terms of $b$ by a central correction so that $a b^{- 1} c b$ can be re-expanded. Taking $b = a - ζ$ for central $ζ \neq 0$ works equally well; $ζ = 1$ is the normalisation that keeps the algebra shortest. What is essential is that $a$ and $b$ generate the same commutation behaviour while being genuinely different elements.

Key Results

Proposition(13.15)Multiplicative commutators detect the centre

Let $D$ be a division ring. If $c \in D$ commutes with every multiplicative commutator of $D$ , then $c \in Z (D)$ .

Proof

Suppose not, so $c a \neq a c$ for some $a \in D$ . Then $c \neq 0$ , $a \neq 0$ and $a \neq 1$ , so $b : = a - 1 \in D^{*}$ , and $b$ does not commute with $c$ either. Put

d : = a^{- 1} c a c^{- 1} - b^{- 1} c b c^{- 1} .

(13.15a)

Both terms are multiplicative commutators, so by hypothesis $c$ commutes with each, hence with $d$ . By $(13.14)$ , $a d = 1 - b^{- 1} c b c^{- 1} \neq 0$ , so $d \neq 0$ ; and $c$ commutes with $1 - b^{- 1} c b c^{- 1} = a d$ as well. Therefore

(c a) d = c (a d) = (a d) c = a (d c) = a (c d) = (a c) d .

(13.15b)

Cancelling the invertible $d$ on the right gives $c a = a c$ , a contradiction. Hence $c \in Z (D)$ .

Corollary(13.16)Central commutators force commutativity

If every multiplicative commutator of a division ring $D$ lies in $Z (D)$ , then $D$ is a field.

Proof

Every $c \in D$ commutes with each central element, hence with every multiplicative commutator; $(13.15)$ then puts $c \in Z (D)$ . So $D = Z (D)$ .

Theorem(13.17)Cartan–Brauer–Hua

Let $K \subseteq D$ be division rings with $K$ normal in $D$ — that is, $x K x^{- 1} \subseteq K$ for all $x \in D^{*}$ — and suppose $K \neq D$ . Then $K \subseteq Z (D)$ .

Proof

Step 1. Let $a \in D ∖ K$ and $c \in K$ ; we claim $a c = c a$ . If not, then $c \neq 0$ , $a \neq 1$ , and $b : = a - 1 \in D^{*}$ . By normality, $a^{- 1} c a$ and $b^{- 1} c b$ lie in $K^{*}$ , and $c \in K^{*}$ , so both sides of $(13.13)$ involve only elements of $K$ : the factor $a^{- 1} c a - b^{- 1} c b$ lies in $K$ , and the right-hand side $c - b^{- 1} c b$ lies in $K$ and is nonzero. Since $a$ times the first equals the second and the first is therefore nonzero,

a = (c - b^{- 1} c b) (a^{- 1} c a - b^{- 1} c b)^{- 1} \in K,

(13.17a)

contradicting $a \notin K$ . So every element outside $K$ centralises $K$ .

Step 2. Fix $c \in K$ and let $c' \in K^{*}$ . Choose $a \in D ∖ K$ , possible since $K \neq D$ . Then $a c' \notin K$ , for otherwise $a = (a c') (c')^{- 1} \in K$ . By Step 1 both $a$ and $a c'$ commute with $c$ , and hence so does $a^{- 1}$ ; therefore $c' = a^{- 1} (a c')$ commutes with $c$ . Thus $c$ commutes with all of $K$ and, by Step 1, with all of $D ∖ K$ — so $c \in Z (D)$ .

This is the same Step 2 as in the additive result $(13.7)$ ; only Step 1 differs, and only Step 1 needed the characteristic hypothesis there. A fuller discussion, including Faith's strengthening, is given in The Cartan–Brauer–Hua Theorem.

Corollary(13.18)Conjugates of a noncentral element generate

Let $D$ be a division ring and $d \in D ∖ Z (D)$ . Then $D$ is generated as a division ring by the set of all conjugates $x d x^{- 1}$ , $x \in D^{*}$ .

Proof

Let $K$ be the division subring generated by all conjugates of $d$ . For $x \in D^{*}$ and $y \in D^{*}$ we have $x^{- 1} (y d y^{- 1}) x = (x^{- 1} y) d (x^{- 1} y)^{- 1}$ , again a conjugate of $d$ ; so $x^{- 1} K x$ contains every conjugate of $d$ and hence contains $K$ . Applying this with $x^{- 1}$ in place of $x$ gives $x K x^{- 1} \subseteq K$ for all $x$ , so $K$ is normal in $D$ . Since $d \in K$ and $d \notin Z (D)$ , we have $K not \subseteq Z (D)$ , so $(13.17)$ forces $K = D$ .

Corollary(13.19)Commutators generate a noncommutative division ring

A noncommutative division ring $D$ is generated as a division ring by its multiplicative commutators alone.

Proof

Let $K$ be the division subring generated by all multiplicative commutators. The set of commutators is carried to itself by every ring automorphism of $D$ , so $K$ is invariant under every automorphism, in particular under every inner automorphism; hence $K$ is normal in $D$ . Since $D$ is not commutative, $(13.16)$ says some multiplicative commutator is noncentral, so $K not \subseteq Z (D)$ . By $(13.17)$ , $K = D$ .

Conjecture—Herstein's conjecture

Suppose that for all $a, b \in D^{*}$ there is a positive integer $n = n (a, b)$ with $(a b a^{- 1} b^{- 1})^{n} \in Z (D)$ . Is $D$ necessarily a field? Herstein settled the case of countable $D$ ; Lam records the general case as open.

Proof Techniques and Method

How these proofs work, and which move to reuse.

Every proof above is the same three-line manoeuvre. It is worth naming the parts.

Shift by one. Replace $a$ by the pair $(a, b)$ with $b = a - 1$ . Both are invertible, both fail to commute with $c$ , and their difference is central — so a single expansion relates their conjugation actions.
**Express $a$ as a quotient.** Identity $(13.13)$ writes $a$ as (something built from conjugates of $c$ ) divided by (something built from conjugates of $c$ ). Whatever contains those conjugates contains $a$ .
Choose the container. For $(13.15)$ the container is the centralizer of $c$ ; for $(13.17)$ it is $K$ . The theorem you get is determined by which subobject you feed the identity.
Upgrade outside to inside. Step 2 — multiply an outside element by a unit of $K$ to get another outside element, then divide — promotes "everything outside centralises $K$ " to " $K$ is central". It is characteristic-free and is shared with $(13.7)$ .

Assume $a$ and $c$ do not commuteThis makes

a \neq 0, 1

and

c \neq 0

, so

b = a - 1

is a unit.

Write down (13.13)

a (a^{- 1} c a - b^{- 1} c b) = c - b^{- 1} c b

, with the right-hand side nonzero.

Divide by $c$ on the right if you want commutatorsGives

(13.14)

, whose two subtracted terms are genuine multiplicative commutators.

Apply the hypothesisEither both commutators centralise

c

, or both lie in

K

Solve for $a$ and contradictCancel the nonzero difference; conclude

a c = c a

a \in K

, contrary to assumption.

Run Step 2Promote the centralising statement from

D ∖ K

to all of

D

Worked Example

A noncentral commutator in the quaternions

Work in $D = ℍ$ , with $Z (ℍ) = ℝ$ . The obvious commutator is central and therefore useless: $i^{- 1} j^{- 1} i j = (- i) (- j) (i j) = (i j) (i j) = k^{2} = - 1 \in ℝ$ . By $(13.16)$ a noncentral commutator must exist, since $ℍ$ is not a field. Take $x = i$ and $y = 1 + j$ , so $y^{- 1} = \frac{1}{2} (1 - j)$ :

x^{- 1} y^{- 1} x y = (- i) \cdot \frac{1}{2} (1 - j) \cdot i \cdot (1 + j) = \frac{1}{2} (- i + k) \cdot i \cdot (1 + j) = \frac{1}{2} (1 + j) (1 + j) = j .

(E.1)

Using $k i = j$ and $j^{2} = - 1$ ; the commutator is $j$ , which is not real.

By the symmetry of $ℍ$ under the automorphisms permuting $i, j, k$ , the elements $i$ and $k$ are also multiplicative commutators. Hence the division subring they generate contains $ℝ$ , $i$ , $j$ and $k$ — that is, all of $ℍ$ . This is $(13.19)$ verified by hand, and note that no central coefficients had to be adjoined: they came for free from $j^{2} = - 1$ .

Testing Cartan–Brauer–Hua

Let $K = ℂ = ℝ + ℝ i$ , a proper division subring of $ℍ$ that is not contained in $Z (ℍ) = ℝ$ . $(13.17)$ predicts $K^{*}$ is not normal in $ℍ^{*}$ . Take $x = \frac{1}{\sqrt{2}} (1 + j)$ , a unit quaternion with $x^{- 1} = \frac{1}{\sqrt{2}} (1 - j)$ :

x i x^{- 1} = \frac{1}{2} (1 + j) i (1 - j) = \frac{1}{2} (i - k) (1 - j) = \frac{1}{2} (i - k - k - i) = - k \notin ℂ,

(E.2)

Using $i j = k$ , $k j = - i$ . Conjugation by $x$ is the quarter-turn about the $j$ -axis.

So normality fails, exactly as the theorem requires. Note that some conjugations do preserve $ℂ$ — for instance $j i j^{- 1} = - i$ — so checking a single conjugation is never enough.

Testing (13.18)

Take $d = i$ , which is noncentral. Its conjugacy class in $ℍ^{*}$ is the set of pure quaternions of norm $1$ — the whole unit sphere in $ℝ i \oplus ℝ j \oplus ℝ k$ , as $(E . 2)$ already indicates. That set contains $i$ , $j$ and $k$ , so the division subring it generates is $ℍ$ , confirming $(13.18)$ .

Comparison and Classification

The additive and multiplicative theories side by side
Question	Additive answer	Multiplicative answer
Common centralizer of all commutators	$Z (D)$ , by $(13.4)$	$Z (D)$ , by $(13.15)$
All commutators central	$D$ is a field, $(13.5)$	$D$ is a field, $(13.16)$
Do commutators generate $D$ ?	Yes, over $Z (D)$ , by $(13.6)$	Yes, outright, by $(13.19)$
Invariant proper division subring	Lie ideal, central if $char \neq 2$ , $(13.7)$	Normal, central in all characteristics, $(13.17)$
Driving identity	$[x, x y] = x [x, y]$	$a (a^{- 1} c a - b^{- 1} c b) = c - b^{- 1} c b$
Extra structure used	Multiplication only	Addition, via $b = a - 1$

The last row is the resolution of the apparent paradox. The multiplicative theory looks purely group-theoretic but is not: forming $a - 1$ is an additive operation, and it is exactly what a normal subgroup of an abstract group cannot do. Cartan–Brauer–Hua is a theorem about division rings, not about groups.

Do these results survive weakening the hypotheses?
	Simple rings	Matrix rings over $D$	Division rings
(13.15) commutators detect the centre	partial	partial	yes
(13.17) Cartan–Brauer–Hua as stated	no	partial	yes
(13.18) conjugates generate	partial	partial	yes
Inverses available for the identity	no	no	yes

Do these results survive weakening the hypotheses?

The "no" in the second row is not a gap in the literature but a genuine counterexample of Amitsur, recorded as Exercise 9 of §13: with $F = ℚ (x)$ and $A = F [t; δ]$ the differential polynomial ring for $δ = d / d x$ , the ring $A$ is a simple domain whose only units are $F^{*}$ , and $F$ is invariant under every automorphism of $A$ without being central. Generalisations of $(13.17)$ to simple rings exist, but they require additional hypotheses.

Relationship Map

$D^{*}$ — the unit groupSubgroups arbitrary

Normal subgroups of $D^{*}$ Include

Z (D)^{*}

, the commutator subgroup, and every

K^{*}

for

K

normal

Those of the form $K^{*}$ , $K$ a division subringConstrained by

(13.17)

With $K \neq D$ Forced inside

Z (D)

Subfields of $Z (D)$ The only possibilities

Read downwards this is a rigidity chain: the more structure a normal subgroup carries, the closer it is forced to the centre. Read upwards it explains why $(13.18)$ and $(13.19)$ are corollaries — both construct a normal division subring and then invoke the bottom of the chain.

(13.13)

–

(13.14)

(13.15)

(13.16)

(13.17)

(13.18)

(13.19)

$(13.17)$ additionally feeds The Multiplicative Group of a Division Ring, where the same identity $(13.14)$ is used again — this time to show that the upper central series of $D^{*}$ stops at the first term.

Applications and Industry Use

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

Group theory

Linear groups over division rings

$(13.17)$ is the base case for normal-subgroup structure in $G L_{n} (D)$ . The classification of normal subgroups of the general linear group over a division ring — they are essentially central, or contain the special linear group — rests on knowing which division subrings can be normal.

Algebraic groups

Reduced Whitehead groups

$(13.19)$ says the commutator subgroup of $D^{*}$ generates $D$ . The quotient $D^{*} / [D^{*}, D^{*}]$ and the reduced Whitehead group $S K_{1} (D)$ are central objects in the arithmetic of division algebras, and $(13.19)$ is the statement that the commutators are not confined to a proper subring.

Skew field constructions

Testing candidate subrings

When building a division ring by Ore localisation or by free constructions, $(13.17)$ is the cheapest obstruction to check: a noncentral division subring that is visibly invariant under conjugation means the construction has gone wrong.

Internal to algebra

Simplicity arguments

The honest summary: these results are infrastructure. They are the reason one may assume, in a noncommutative division ring, that commutators are plentiful and that no proper division subring is conjugation-stable.

Failure Modes and Common Mistakes

Do not assume the set of multiplicative commutators is a subgroup of $D^{*}$ ; it generates one, which is a different statement, and $(13.19)$ is about the division subring generated, not the subgroup.
Do not extend $(13.17)$ to simple rings. Amitsur's differential polynomial ring $ℚ (x) [t; d / d x]$ is the standing counterexample.
Do not overlook that $c$ must be nonzero in $(13.13)$ : the identity is stated for noncommuting $a$ and $c$ , and $c = 0$ commutes with everything.
Do not confuse $b = a - 1$ with $a^{- 1}$ . The proofs use both, and the identity is false if they are interchanged.

Best Practices

When you need to force an element into a subring, look for an identity expressing it as a quotient of two elements you already control; $(13.13)$ is the template.
State Cartan–Brauer–Hua with both hypotheses every time, and say which normality convention you are using — $x K x^{- 1} \subseteq K$ for all $x \in D^{*}$ , not merely for $x \in K^{*}$ .
Prefer the multiplicative results when the characteristic is unknown or equal to $2$ ; the additive analogues carry a restriction that the multiplicative ones do not.
When verifying an example, compute at least two conjugations before concluding a subring is normal.

Quick Reference

Commutator

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

; also

a^{- 1} c a c^{- 1}

Substitution

b = a - 1 \in D^{*}

when

a \neq 1

Identity (13.13)

a (a^{- 1} c a - b^{- 1} c b) = c - b^{- 1} c b \neq 0

Identity (13.14)

a (a^{- 1} c a c^{- 1} - b^{- 1} c b c^{- 1}) = 1 - b^{- 1} c b c^{- 1} \neq 0

(13.15)Centralizer of all commutators

= Z (D)

(13.16)All commutators central

\Rightarrow D

a field

(13.17)

K

normal,

K \neq D

\Rightarrow K \subseteq Z (D)

(13.18) / (13.19)Conjugates of a noncentral

d

generate

D

; commutators generate

D

Which hypothesis fails when a conclusion fails
Failed conclusion	Which hypothesis to check first
$K$ normal but not central	$K = D$ ? Or is the ring not a division ring?
Commutators do not generate $D$	Is $D$ commutative? Then every commutator is $1$
Conjugates of $d$ do not generate	Is $d$ central? Then its class is ${d}$
Identity $(13.13)$ gives $0 = 0$	Do $a$ and $c$ actually fail to commute?

Frequently Asked Questions

Why does the proof need to form

a - 1

at all?

Because a hypothesis about commutators constrains conjugation, and a single element $a$ gives only one conjugation to work with. The pair $(a, a - 1)$ gives two conjugations of the same $c$ whose difference can be compared with $c$ itself, and the comparison is possible only because $a$ and $a - 1$ differ by a central element. This is precisely the step that has no analogue for abstract groups.

Is the set of multiplicative commutators closed under multiplication?

Not in general — a product of commutators need not be a commutator, in $D^{*}$ as in any group. $(13.19)$ therefore speaks of the division subring generated by them. The quaternion computation illustrates the point: $- 1$ and $j$ are both commutators, and one must argue via generation to reach all of $ℍ$ .

How does

(13.17)

differ from the additive result

(13.7)

in strength?

They have the same shape but incomparable hypotheses. $(13.7)$ assumes $K$ is a Lie ideal — closed under $x \mapsto a x - x a$ — and needs $char K \neq 2$ . $(13.17)$ assumes $K^{*}$ is normal in $D^{*}$ and needs nothing about the characteristic. Neither hypothesis implies the other, but in practice normality is the easier one to verify, which is why $(13.18)$ and $(13.19)$ go through the multiplicative version.

Can Cartan–Brauer–Hua be strengthened?

Yes, in a group-theoretic direction. Faith showed that for a division subring $K$ with $K \neq D$ and $K not \subseteq Z (D)$ , the normalizer $N_{D^{*}} (K^{*})$ has infinite index in $D^{*}$ — so $K^{*}$ is not merely non-normal, it is very far from normal. Versions for simple rings, semisimple rings and twisted settings also exist, each with its own extra hypotheses.

What is the status of Herstein's conjecture?

Herstein proved it when $D$ is countable. Lam's 1991 text records the general case as open, and this page follows that record rather than asserting a resolution. The hypothesis is a substantial weakening of $(13.16)$ : instead of demanding that commutators be central, it demands only that some power of each be.

Why can't the results be transplanted to matrix rings over

D

The proofs divide by elements that are only known to be nonzero, and in $M_{n} (D)$ nonzero does not mean invertible. Identity $(13.13)$ still holds formally wherever $a$ , $a - 1$ and $c$ are units, but the step from "nonzero" to "invertible" is exactly what a division ring supplies and a matrix ring does not.

References

T. Y. Lam, A First Course in Noncommutative Rings, Graduate Texts in Mathematics 131, Springer-Verlag, 1991, §13, (13.13)–(13.19), pp. 221–223, with Exercises 7 and 9.
I. N. Herstein, Noncommutative Rings, Carus Mathematical Monographs 15, Mathematical Association of America, 1968, Chapter 3.
L. K. Hua, “Some properties of a sfield”, Proceedings of the National Academy of Sciences of the USA 35 (1949), 533–537.
C. Faith, “On conjugates in division rings”, Canadian Journal of Mathematics 10 (1958), 374–380.
P. Draxl, Skew Fields, London Mathematical Society Lecture Note Series 81, Cambridge University Press, 1983.
L. H. Rowen, Ring Theory, Volume II, Academic Press, 1988 (division algebras and their multiplicative structure).

AI Suggested Questions

Verify identity $(13.13)$ symbolically and determine for which central shifts $b = a - ζ$ the argument still works.
Give the details of Amitsur's example showing Cartan–Brauer–Hua fails for simple rings.
State and prove Faith's theorem on the index of $N_{D^{*}} (K^{*})$ .
Describe the normal subgroups of $G L_{n} (D)$ for $D$ a division ring and explain the role of $(13.17)$ .
What is known about $S K_{1} (D)$ for a division algebra $D$ over a number field?
Prove Exercise 7 of §13: for $h \notin {0, 1}$ , both $h$ and $1 + h$ normalise $K^{*}$ if and only if $h$ lies in $K$ or centralises $K^{*}$ ; then deduce $(13.17)$ .
Determine the multiplicative commutator subgroup of $ℍ^{*}$ explicitly.

01Executive Summary

02Overview

03Learning Objectives

04Definitions

05Core Concepts

Deriving the identity

Two ways to read the identity

Why b=a−1 and not some other shift

06Key Results

07Proof Techniques and Method

08Worked Example

A noncentral commutator in the quaternions

Testing Cartan–Brauer–Hua

Testing (13.18)

09Comparison and Classification

10Relationship Map

11Applications and Industry Use

Linear groups over division rings

Reduced Whitehead groups

Testing candidate subrings

Simplicity arguments

12Failure Modes and Common Mistakes

13Best Practices

14Quick Reference

15Frequently Asked Questions

16Related KEVOS Topics

17References

18AI Suggested Questions

Continue learning

Executive Summary

Overview

Learning Objectives

Definitions

Core Concepts

Why $b = a - 1$ and not some other shift

Key Results

Proof Techniques and Method

Worked Example

Comparison and Classification

Relationship Map

Applications and Industry Use

Failure Modes and Common Mistakes

Best Practices

Quick Reference

Frequently Asked Questions

Related KEVOS Topics

References

AI Suggested Questions