Embedding Free Rings in Division Rings

Q: Why can the free algebra not be localised?

Because xA yA = 0 for distinct generators x,y: every word occurring in a nonzero element of xA starts with x, and every word in a nonzero element of yA starts with y. The Ore condition demands a common nonzero multiple, so it fails, and Ore's theorem then says there is no division ring of right fractions.

Q: What does (1-x-y)^-1 actually look like?

It is the sum of all words in x and y, each with coefficient 1 — the generating function of the free monoid. Its support is well-ordered because x,y lies in the positive cone, and each word appears in exactly one power of x+y, so every coefficient is a finite sum.

Q: Why is the centre so small in (14.26)?

Because injectivity of means that no nonidentity group element acts trivially on R. A central series must therefore have support inside 1, reducing it to a scalar, and that scalar must be fixed by every _g. The result is the fixed field of the action, which is typically far smaller than R itself.

Executive Summary

A commutative domain always embeds in its field of fractions. A noncommutative domain need not embed in any division ring at all, and even when it does, the classical construction — Ore localisation — may be unavailable. The free algebra $k ⟨ x, y ⟩$ is the standard case: it is a domain, it fails the Ore condition badly, and yet it does embed.

The proof is a two-step assembly. The free ring $R ⟨ x_{i} : i \in I ⟩$ sits inside the group ring $R [G]$ of the free group $G$ on the same generators, because distinct words in the $x_{i}$ are distinct group elements. Free groups are orderable, so the Mal'cev–Neumann construction turns $R [G]$ into a subring of the division ring $R ((G))$ .

$R ⟨ x_{i} ⟩ ↪ R ((G))$ The embedding

$G$ freeIndex group

failsOre condition, rank

\geq 2

1937–49Moufang, Mal'cev, Neumann

Overview

Let $R$ be a division ring and let ${x_{i} : i \in I}$ be independent indeterminates, each commuting with $R$ but not with one another. The free ring $R ⟨ x_{i} : i \in I ⟩$ is free as a left $R$ -module on the set of words in the $x_{i}$ , with multiplication given by concatenation.

For $| I | \geq 2$ this ring is a domain that cannot be localised. If $x \neq y$ are two of the generators then $x R ⟨ x_{i} ⟩ \cap y R ⟨ x_{i} ⟩ = 0$ , since every nonzero element of the first has all its words beginning with $x$ and every nonzero element of the second has all its words beginning with $y$ . So there is no common nonzero multiple and no Ore ring of fractions.

R ⟨ x_{i} : i \in I ⟩ \subseteq R [G] \subseteq R ((G)), G = free group on {x_{i}} .

(14.25)

The first inclusion is a statement about words; the second is the Mal'cev–Neumann construction with trivial twist.

The two prerequisites are on neighbouring pages: the orderability of free groups is a group-theoretic input, and the passage from $R [G]$ to the division ring $R ((G))$ is The Mal'cev–Neumann Construction of Laurent Series Rings. The presentation and universal property of $R ⟨ x_{i} ⟩$ are on Free Rings and Presentations.

Learning Objectives

Show that $x A \cap y A = 0$ in the free algebra, so the Ore condition fails for rank $\geq 2$ .
Identify $R ⟨ x_{i} ⟩$ with a subring of the group ring $R [G]$ of the free group $G$ .
State why free groups admit bi-invariant total orders.
Assemble the embedding $(14.25)$ from these ingredients.
Prove $(14.26)$ : the centre of $R ((G, ω))$ for injective $ω$ , and central infiniteness.
Invert $1 - x - y$ explicitly inside $k ((G))$ and identify the resulting support.

Definitions

DefinitionThe free ring over

R

Let $R$ be a ring and ${x_{i} : i \in I}$ a set of symbols, each assumed to commute with every element of $R$ . The free ring $R ⟨ x_{i} : i \in I ⟩$ has as left $R$ -basis the set $W$ of all finite words $x_{i_{1}} x_{i_{2}} \dots x_{i_{n}}$ (including the empty word $1$ ), with multiplication extending concatenation $R$ -bilinearly.

It is characterised by a universal property: any ring map $R \to T$ together with any choice of elements $t_{i} \in T$ commuting with the image of $R$ extends uniquely to $R ⟨ x_{i} ⟩ \to T$ with $x_{i} \mapsto t_{i}$ .

$W$: The free monoid on ${x_{i}}$ : all finite words, with concatenation.
$G$: The free group on ${x_{i}}$ : reduced words in the $x_{i}$ and their inverses. $W \subseteq G$ as a submonoid.
Right Ore domain: A domain $A$ with $a A \cap b A \neq 0$ for all nonzero $a, b$ ; equivalently, one possessing a right division ring of fractions.
$R [G]$: The group ring: finite formal sums $\sum a_{g} g$ with $a_{g} \in R$ .
$R ((G))$: The Mal'cev–Neumann series ring with trivial twist: formal sums with well-ordered support.
$R^{ω}$: The fixed ring ${r \in R : ω_{g} (r) = r for all g \in G}$ .

When R is a field k, the free ring is the free associative k-algebra, usually written with angle brackets. The generators do not commute with each other, only with the coefficients.

Core Concepts

Three distinct questions

It is worth separating what is being asked.

**Does $A$ embed in some division ring?** For a general domain the answer is no — Mal'cev constructed a domain with no such embedding.
**Does $A$ have a division ring of fractions, i.e. one generated by $A$ with every element of the form $a b^{- 1}$ ?** This holds exactly when $A$ is a right Ore domain.
Is there a universal division ring of fractions? Cohn's theory of free ideal rings answers this affirmatively for free algebras, and the resulting object — the free field — is not the Mal'cev–Neumann ring.

The theorem on this page settles question (1) for free rings, by an explicit and rather large ambient division ring. It says nothing about (2), which is false, and it is not the answer to (3).

Why the free group and not the free monoid

The free ring is already a subring of the monoid ring $R [W]$ — in fact it is $R [W]$ . But $W$ is not a group, so no series construction applies. Passing to the free group $G \supseteq W$ costs nothing, because distinct words of $W$ remain distinct as reduced words of $G$ , so $R [W] \subseteq R [G]$ is an inclusion of rings.

free monoid

W

free group

G

ordered group

(G, \leq)

division ring

R ((G))

Where the ordering comes from

A free group is residually torsion-free nilpotent: the Magnus embedding $G ↪ k ⟨ ⟨ X_{i} ⟩ ⟩$ , $x_{i} \mapsto 1 + X_{i}$ , separates the terms of the lower central series and all quotients are torsion-free. A residually torsion-free nilpotent group carries a bi-invariant total order, obtained by ordering each quotient and combining lexicographically. This is the input Lam quotes as $(6.19)$ .

Key Results

PropositionThe free algebra is not Ore

Let $R$ be a nonzero ring and $| I | \geq 2$ , and let $A = R ⟨ x_{i} : i \in I ⟩$ with $x \neq y$ two of the generators. Then $x A \cap y A = 0$ and $A x \cap A y = 0$ . In particular, if $R$ is a division ring then $A$ is a domain that is neither a right nor a left Ore domain, so $A$ has no division ring of fractions.

Proof

Write elements of $A$ in the left $R$ -basis $W$ of words. Every word occurring in a nonzero element of $x A$ begins with the letter $x$ , and every word occurring in a nonzero element of $y A$ begins with $y$ . Since $W$ is a basis, a nonzero element cannot lie in both, so $x A \cap y A = 0$ . The left-handed statement is symmetric, reading words from the right. A right Ore domain requires $a A \cap b A \neq 0$ for all nonzero $a, b$ , which fails here.

LemmaThe free ring inside a group ring

Let $G$ be the free group on ${x_{i} : i \in I}$ and $W \subseteq G$ the submonoid of words in the $x_{i}$ with no inverses. Then $W$ is a free monoid and the $R$ -span of $W$ inside $R [G]$ is a subring isomorphic to $R ⟨ x_{i} : i \in I ⟩$ .

Proof

Distinct words in the $x_{i}$ are already reduced as words in $G$ , hence are distinct group elements; so $W$ injects into $G$ and is free as a monoid. The group elements of $G$ form an $R$ -basis of $R [G]$ , so the span of $W$ is free as a left $R$ -module on $W$ , and multiplication in $R [G]$ restricted to $W$ is concatenation. That is exactly the defining data of $R ⟨ x_{i} ⟩$ .

Corollary(14.25)Mal'cev–Neumann–Moufang

Let $R$ be a division ring and ${x_{i} : i \in I}$ any set of independent indeterminates commuting with $R$ . Then the free ring $R ⟨ x_{i} : i \in I ⟩$ can be embedded in a division ring.

Proof

Let $G$ be the free group on ${x_{i}}$ . By the lemma, $R ⟨ x_{i} ⟩$ is a subring of $R [G]$ . Free groups admit bi-invariant total orders, so fix one, making $(G, \leq)$ an ordered group. Taking $ω$ to be the trivial homomorphism $G \to Aut (R)$ , the Mal'cev–Neumann theorem says $R ((G))$ is a division ring, and the twisted group ring $R [G, ω] = R [G]$ is a subring of it. Composing the two inclusions gives the embedding.

Corollary(14.26)The centre, and central infiniteness

Let $R$ be a field, $(G, \leq)$ a nontrivial ordered group, and $ω : G \to Aut (R)$ an injective homomorphism. Set $A = R ((G, ω))$ . Then

Z (A) = R^{ω} = {r \in R : ω_{g} (r) = r for all g \in G},

and $A$ is a centrally infinite division ring.

Proof

Scalars constrain the support. Let $α = \sum_{g} a_{g} g \in Z (A)$ and $r \in R$ . Then $α r = \sum_{g} a_{g} ω_{g} (r) g$ while $r α = \sum_{g} r a_{g} g$ . Comparing the coefficient of $g$ and using that $R$ is commutative, $a_{g} (ω_{g} (r) - r) = 0$ for every $r \in R$ . So if $a_{g} \neq 0$ then $ω_{g} = {id}_{R}$ , and injectivity of $ω$ forces $g = 1$ . Hence $α = a_{1} \in R$ .

Group elements constrain the coefficient. For $h \in G$ , $α h = a_{1} h$ and $h α = ω_{h} (a_{1}) h$ , so $ω_{h} (a_{1}) = a_{1}$ for all $h$ , i.e. $a_{1} \in R^{ω}$ .

Conversely $R^{ω} \subseteq Z (A)$ : such an $r$ commutes with all of $R$ because $R$ is commutative, and $g r = ω_{g} (r) g = r g$ for every $g$ . Hence $Z (A) = R^{ω}$ .

Central infiniteness. An ordered group is torsion-free, so a nontrivial $G$ is infinite. The group elements are $R$ -linearly independent in $A$ , so $\dim_{R} A$ is infinite; since $R^{ω} \subseteq R$ , $\dim_{Z (A)} A$ is infinite as well.

RemarkWhat the embedding is not

The division ring $R ((G))$ is enormously larger than any division ring generated by $R ⟨ x_{i} ⟩$ , and the embedding depends on the choice of order on $G$ . Cohn's theory produces a canonical alternative — the universal field of fractions of the free algebra, the free field — which is generated by the free algebra and has a universal property. The Mal'cev–Neumann embedding is the elementary existence proof, not the canonical one.

Proof Techniques and Method

How these proofs work, and which move to reuse.

Recognise the obstructionLocalisation is unavailable: the Ore condition fails for the free algebra on two or more generators.

Enlarge the index structureReplace the free monoid of words by the free group, at no cost since words stay distinct.

Import an orderFree groups are residually torsion-free nilpotent, hence bi-orderable.

Apply the series construction

R ((G))

is a division ring by Mal'cev–Neumann, and contains

R [G] \supseteq R ⟨ x_{i} ⟩

Move 1

Embed rather than localise

When fractions are unavailable, look for a larger ring with a known division property that already contains the object. This is the standard escape when the Ore condition fails.

Move 2

Argue by leading word

In a ring free on a basis of words, statements about intersections of ideals reduce to statements about first letters. The failure of the Ore condition is a one-line observation once the basis is in view.

Move 3

Test centrality twice

Commute against scalars to constrain the support; commute against group elements to constrain the surviving coefficient. Exactly the pattern used to compute the centre of a twisted Laurent series ring.

Move 1 has a limit worth recording: it cannot be pushed to arbitrary domains, because Mal'cev exhibited a domain that embeds in no division ring at all. The free ring is embeddable because it happens to sit inside a group ring of an orderable group; that is a special feature, not a general principle.

Worked Example

Inverting $1 - x - y$

Let $k$ be a field, $G$ the free group on $x, y$ , and fix a bi-invariant order in which $x > 1$ and $y > 1$ . Consider $α = x + y \in k [G]$ , whose support ${x, y}$ lies in the positive cone $P$ . By the convergence corollary of the Mal'cev–Neumann theory,

(1 - x - y)^{- 1} = \sum_{n \geq 0} (x + y)^{n} = \sum_{w \in W} w,

(E.1)

$W$ is the free monoid on $x, y$ : the sum runs over every word, each with coefficient $1$ .

The support of the right-hand side is $W = {1} \cup S^{\infty}$ with $S = {x, y}$ , which is well-ordered by the key lemma, and each word occurs in exactly one $(x + y)^{n}$ — namely $n$ equal to its length — so every coefficient is a finite sum. Thus $(E . 1)$ is a legitimate element of $k ((G))$ .

Why no fraction representation exists

Inside any division ring containing $k ⟨ x, y ⟩$ , consider $x^{- 1} y$ . If the free algebra were right Ore, every element of the generated division ring would have the form $a b^{- 1}$ with $a, b \in k ⟨ x, y ⟩$ ; but $x^{- 1} y = a b^{- 1}$ would give $y b = x a$ , a nonzero element of $x A \cap y A$ , contradicting the proposition above. So the elements of $k ((G))$ generated by the free algebra are genuinely more complicated than single fractions.

A centrally infinite instance

For $(14.26)$ take $R = ℚ (t_{1}, t_{2}, \dots)$ , $G = ℤ$ generated by $g$ , and $ω_{g}$ the automorphism shifting $t_{j} \mapsto t_{j + 1}$ . This $ω$ is injective — no nonzero power of the shift is the identity — so $Z (A) = R^{ω} = ℚ$ , and $A = R ((G, ω))$ is a centrally infinite division ring containing $R$ , itself of infinite dimension over $ℚ$ .

Comparison and Classification

Routes from a domain to a division ring
Method	Applies when	Produces	Canonical?
Field of fractions	commutative domain	smallest field containing it	yes
Ore localisation	right (or left) Ore domain	division ring of fractions $a b^{- 1}$	yes, up to isomorphism
Mal'cev–Neumann	subring of $R [G, ω]$ with $G$ orderable	a very large series division ring	no — depends on the chosen order
Cohn's universal field of fractions	free ideal rings, including free algebras	the free field, generated by the ring	yes, by a universal property
No method exists	Mal'cev's 1937 domain	nothing — no embedding exists	—

Properties of the free algebra $k ⟨ x, y ⟩$
	$k [x]$	$k [x, y]$	$k ⟨ x, y ⟩$
Domain	yes	yes	yes
Noetherian	yes	yes	no
Ore domain	yes	yes	no
Has a division ring of fractions	yes	yes	no
Embeds in a division ring	yes	yes	yes
Every left ideal free of unique rank	yes	no	yes

Properties of the free algebra $k ⟨ x, y ⟩$

Relationship Map

All domainssome embed in no division ring at all (Mal'cev, 1937)

Domains embeddable in a division ringincludes every subring of

R [G, ω]

with

G

orderable

Free rings $R ⟨ x_{i} ⟩$ embeddable via

R ((G))

, but not Ore for rank

\geq 2

Ore domainspossess an honest division ring of fractions; free rings of rank

\geq 2

are excluded

The picture to keep is that embeddability and localisability are independent properties. Ore domains are the localisable ones; free rings are embeddable without being localisable; and Mal'cev's example is neither.

free groups orderable

R ((G))

a division ring

R [G]

embeds

R ⟨ x_{i} ⟩

embeds

Applications and Industry Use

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

Ring theory

Existence proofs for generic constructions

Generic matrices, generic division algebras and Amitsur's non-crossed-product examples all begin by embedding a free-like ring in a division ring; this corollary is the licence to do so.

Combinatorics on words

Rational and algebraic power series

The identity $(1 - x - y)^{- 1} = \sum_{w} w$ is the generating function of the free monoid. Formal language theory works in $k ⟨ ⟨ x, y ⟩ ⟩$ , where rational series correspond to regular languages.

Free probability

Noncommutative rational functions

Evaluating noncommutative rational expressions on matrices and operators requires a well-defined free field; existence of some ambient division ring is the first step.

Control theory

Noncommutative transfer functions

Multidimensional and noncommutative systems theory uses formal series in noncommuting variables, and realisability results are statements about rational elements of such a division ring.

Group theory

Orderable groups and the zero-divisor conjecture

Kaplansky's conjecture that $k [G]$ is a domain for torsion-free $G$ is known for orderable $G$ precisely by this route, and remains open in general.

Symbolic computation

Noncommutative Gröbner bases

Computation in $k ⟨ x_{i} ⟩$ uses word orders and non-terminating Gröbner procedures; the failure of the Ore condition is the structural reason no fraction-based normal form exists.

The honest summary is that the theorem is used as a licence. It is rarely computed with; it is cited to guarantee that an expression involving inverses of elements of a free ring has a meaning at all.

Standards and Notation

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

Free ring

R ⟨ x_{i} : i \in I ⟩

; also

R {x_{i}}

and

R * ℤ^{* I}

in some sources

Free power series

R ⟨ ⟨ x_{i} ⟩ ⟩

— completion at the augmentation ideal, distinct from

R ((G))

Free fieldCohn's universal field of fractions, written

R ⟨ ⟨ x_{i} ⟩ ⟩

𝒟 (R; I)

depending on the source

Ore conditionStated on the right here:

a A \cap b A \neq 0

. Left and right versions are independent in general

GAP / MagmaFree associative algebras and noncommutative Gröbner bases; no support for the ambient division ring

SageFreeAlgebra, FreeGroup; series rings over general ordered groups are not implemented

MarkupPresentation MathML per ISO/IEC 40314; angle brackets rendered as fenced operators, not as the less-than and greater-than signs

Computational Notes

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

The embedding is not effective. It requires a bi-invariant order on the free group; the Magnus construction gives one, but comparing two group elements means comparing Magnus expansions, which is expensive and unbounded in the degree needed.
**Arithmetic in $k ⟨ x_{i} ⟩$ is fine**: words are strings, multiplication is concatenation, and noncommutative Gröbner basis machinery exists — though it need not terminate, since the free algebra is not noetherian.
**Arithmetic in $k ((G))$ is not fine**: elements have infinite, order-type-complicated supports. Practical computation is done in $k ⟨ ⟨ x_{i} ⟩ ⟩$ with truncation by total degree, which suffices for rational and algebraic series.
Rational expressions in noncommuting variables are handled by linear representations: a noncommutative rational function is encoded by matrices $(A_{i})$ and vectors, and identity testing reduces to linear algebra over $k$ — this sidesteps the free field entirely.
Deciding whether a given domain embeds in a division ring is not decidable in general; Mal'cev's counterexample shows the property is not automatic even for finitely presented domains.

Failure Modes and Common Mistakes

Do not confuse $R ⟨ x_{i} ⟩$ with $R [x_{i}]$ : the generators commute with $R$ but not with each other.
Do not assume the result extends to all torsion-free groups; whether $k [G]$ is always a domain for torsion-free $G$ is Kaplansky's open zero-divisor problem.
Do not expect $(14.26)$ without injectivity of $ω$ : if $ω$ has a nontrivial kernel then central elements can have support outside ${1}$ .
Do not treat $k ⟨ ⟨ x, y ⟩ ⟩$ as a division ring — it is a local ring whose units are the series with nonzero constant term.

Historical Notes and Lessons Learned

1931Ore's conditionOre characterises the domains possessing a division ring of fractions, and observes that the condition is restrictive.
1937Mal'cev's counterexampleA cancellative semigroup whose semigroup algebra is a domain not embeddable in any division ring — settling that embeddability is a real hypothesis.
1937MoufangConstructs a division ring containing the group algebra of a free group of rank two, motivated by the coordinatisation of projective planes.
1948–49Mal'cev and NeumannThe general series construction over ordered groups, giving the embedding for free rings over arbitrary division rings and arbitrary index sets.
1963–71Cohn's free ideal ringsCohn develops firs and universal fields of fractions, producing the canonical free field and a structure theory that the series construction does not supply.

The lesson is that the two 1937 papers frame everything that follows: Mal'cev's counterexample shows that no general theorem is available, and Moufang's construction shows that the free case is nonetheless tractable. The 1948–49 work is the systematic version of Moufang's idea.

Quick Reference

Statement

R ⟨ x_{i} : i \in I ⟩

embeds in a division ring, for

R

a division ring

Route

R ⟨ x_{i} ⟩ \subseteq R [G] \subseteq R ((G))

G

free on

{x_{i}}

Group inputfree groups are bi-orderable (residually torsion-free nilpotent)

Ring inputMal'cev–Neumann:

R ((G))

is a division ring

Ore fails

x A \cap y A = 0

for distinct generators, rank

\geq 2

Centre

Z (R ((G, ω))) = R^{ω}

for

R

a field,

G

nontrivial,

ω

injective

Consequence

A

is centrally infinite

Not canonicaldepends on the chosen order; contrast Cohn's free field

Statements and their hypotheses
Statement	Hypotheses	Reference
$x A \cap y A = 0$	$A = R ⟨ x_{i} ⟩$ , $\| I \| \geq 2$ , $R \neq 0$	elementary
$R ⟨ x_{i} ⟩ \subseteq R [G]$	$G$ free on ${x_{i}}$	before (14.25)
free groups are orderable	none	(6.19)
$R ⟨ x_{i} ⟩$ embeds in a division ring	$R$ a division ring	(14.25)
$Z (A) = R^{ω}$	$R$ a field, $G$ nontrivial ordered, $ω$ injective	(14.26)
$A$ centrally infinite	same	(14.26)

Frequently Asked Questions

Why can the free algebra not be localised?

Because $x A \cap y A = 0$ for distinct generators $x, y$ : every word occurring in a nonzero element of $x A$ starts with $x$ , and every word in a nonzero element of $y A$ starts with $y$ . The Ore condition demands a common nonzero multiple, so it fails, and Ore's theorem then says there is no division ring of right fractions.

Is the resulting division ring canonical?

No. It depends on the choice of bi-invariant order on the free group, and it is far larger than anything generated by the free algebra. Cohn's universal field of fractions — the free field — is the canonical object, characterised by a universal property, and it is a proper subobject of the picture here in the sense of being generated by the free algebra.

Does every torsion-free group work in place of the free group?

Not known. The construction needs a bi-invariant total order, and orderability is strictly stronger than torsion-freeness. Whether $k [G]$ is a domain, let alone embeddable in a division ring, for every torsion-free $G$ is Kaplansky's zero-divisor conjecture, still open.

What does

(1 - x - y)^{- 1}

actually look like?

It is the sum of all words in $x$ and $y$ , each with coefficient $1$ — the generating function of the free monoid. Its support is well-ordered because ${x, y}$ lies in the positive cone, and each word appears in exactly one power of $x + y$ , so every coefficient is a finite sum.

Why is the centre so small in

(14.26)

Because injectivity of $ω$ means that no nonidentity group element acts trivially on $R$ . A central series must therefore have support inside ${1}$ , reducing it to a scalar, and that scalar must be fixed by every $ω_{g}$ . The result is the fixed field of the action, which is typically far smaller than $R$ itself.

Does this say anything about whether a given finitely presented domain embeds?

Only if you can exhibit it inside a twisted group ring over an orderable group. There is no general criterion, and Mal'cev's 1937 example shows none can simply say "every domain". Embeddability is a genuine hypothesis, not a formality.

References

T. Y. Lam, A First Course in Noncommutative Rings, Graduate Texts in Mathematics 131, Springer-Verlag, 1991, §14, (14.25)–(14.26), pp. 247–249; orderability of free groups is (6.19).
B. H. Neumann, “On ordered division rings”, Transactions of the American Mathematical Society 66 (1949), 202–252.
A. I. Mal'cev, “On the immersion of an algebraic ring into a field”, Mathematische Annalen 113 (1937), 686–691.
P. M. Cohn, Free Rings and Their Relations, 2nd edition, London Mathematical Society Monographs 19, Academic Press, 1985.
P. M. Cohn, Skew Fields: Theory of General Division Rings, Encyclopedia of Mathematics and its Applications 57, Cambridge University Press, 1995, Chapters 2 and 6.
D. S. Passman, The Algebraic Structure of Group Rings, Wiley-Interscience, 1977, Chapter 13.

AI Suggested Questions

Construct an explicit bi-invariant order on the free group of rank two using the Magnus embedding.
Describe Mal'cev's 1937 domain that embeds in no division ring, and identify which axiom of embeddability it violates.
Compare the Mal'cev–Neumann division ring containing $k ⟨ x, y ⟩$ with Cohn's free field.
State Kaplansky's zero-divisor conjecture and summarise the classes of groups for which it is known.
Show that the free algebra of rank at least two is not left or right noetherian.
How are noncommutative rational functions represented by linear systems, and what identity-testing algorithms result?
Which subrings of $k ((G))$ for $G$ free are themselves free ideal rings?

01Executive Summary

02Overview

03Learning Objectives

04Definitions

05Core Concepts

Three distinct questions

Why the free group and not the free monoid

Where the ordering comes from

06Key Results

07Proof Techniques and Method

Embed rather than localise

Argue by leading word

Test centrality twice

08Worked Example

Inverting 1−x−y

Why no fraction representation exists

A centrally infinite instance

09Comparison and Classification

10Relationship Map

11Applications and Industry Use

Existence proofs for generic constructions

Rational and algebraic power series

Noncommutative rational functions

Noncommutative transfer functions

Orderable groups and the zero-divisor conjecture

Noncommutative Gröbner bases

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

Worked Example

Inverting $1 - x - y$

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