← LibraryVanishing Polynomials | KEVOS® MathematicsProject Delivery · Project ManagementLesson 340/352← PrevNext →
ArticlePublished 9 Aug 202619 min readBy Kevin Jogin
Skip to content

Engineering Mathematics Advanced Polynomial equations

Vanishing Polynomials

A single conjugation-and-subtract argument determines every polynomial that vanishes on a conjugacy class of a division ring: the vanishing polynomials form the two-sided ideal generated by the class's minimal polynomial, and on an infinite division ring only the zero polynomial vanishes everywhere.

Page ID
KVS-ENG-MATH-0247
Taxonomy
ENG / ENG-MATH
Collection
noncommutative-rings-core
Source
(16.5)–(16.7), §16 (pp. 265–267)
Reviewed
2026-08-08
Version
1.0.0

Executive Summary

A conjugacy class A of a division ring D is stable under conjugation by definition. That single symmetry is enough to pin down every polynomial that vanishes on A. If hD[t] vanishes on A and has minimal degree, conjugating h coefficientwise produces a second vanishing polynomial of the same degree; subtracting the two produces a shorter one unless every coefficient of h was already central.

The consequences are immediate and complete. If A is algebraic over F=Z(D) with minimal polynomial f, then (16.5) every nonzero vanishing polynomial has degree at least degf, and (16.6) the vanishing polynomials are exactly the left multiples D[t]f — which, f being central, is a two-sided ideal. If A is not algebraic, nothing nonzero vanishes on it. And if A is replaced by the whole of an infinite D, the answer is again zero (16.7).

deghdegfDegree bound (16.5)
D[t]fVanishing ideal (16.6)
0Vanishing on infinite D
1 argumentProves all three

Overview

Fix a division ring D with centre F, and let D[t] be the polynomial ring in a central variable t, with evaluation by right substitution. Over a field, the polynomials vanishing on a finite set S form the principal ideal generated by sS(ts); that is the whole story. Over a division ring the natural test sets are not finite sets but conjugacy classes, because they are the sets on which the theory of roots is actually controlled — see The Gordon–Motzkin Theorem.

The results here answer three questions about a class A: how short a nonzero vanishing polynomial can be, exactly which polynomials vanish, and what happens when A is enlarged to all of D.

{hD[t]:h(A)=0}=D[t]f(t),f=minimal polynomial of A over F,
(16.6)

Valid whenever A is a conjugacy class algebraic over F. Because f has central coefficients, this left ideal is in fact two-sided.

The reason the mechanism works is that a class is stable under the conjugation action while a polynomial's coefficient list is not. Forcing the two to agree squeezes the coefficients into the fixed ring of the action, which for an inner action on a division ring is precisely the centre.

Learning Objectives

  • Show that all elements of an algebraic conjugacy class share one minimal polynomial over F.
  • Prove that a minimal-degree vanishing polynomial on a conjugation-stable set has central coefficients.
  • Derive the degree bound (16.5) from that statement.
  • Prove (16.6) that the vanishing set is D[t]f, and observe that it is a two-sided ideal.
  • Prove (16.7): an infinite division ring admits no nonzero polynomial vanishing on all of it.
  • Decide, for a given class, whether any nonzero polynomial vanishes on it at all.

Definitions

Definition§16Algebraic class and its minimal polynomial

Let D be a division ring with centre F and AD a conjugacy class. Call A **algebraic over F** if some element of A is algebraic over F. Since conjugation by dD fixes F elementwise, g(dad1)=dg(a)d1 for any gF[t]; hence every element of A is then algebraic and all elements share one minimal polynomial fF[t], called the **minimal polynomial of A**.

h(A)=0
Shorthand for: h(a)=0 for every aA, where h(a) means substitution of a on the right of the coefficients.
Conjugation-stable set
A subset SD with eSe1=S for all eD. Conjugacy classes, unions of classes, F itself and D itself are all conjugation-stable.
he
The coefficientwise conjugate of h=diti, namely (edie1)ti, for eD.
degf
For an algebraic class, degf=dimFF(a) for any aA; it is 1 exactly when A is a central singleton.
Two-sided ideal of D[t]
An additive subgroup closed under multiplication by D[t] on both sides. Any left ideal generated by a polynomial with coefficients in F is automatically two-sided.

No chain condition and no finiteness over F is assumed. A may be — and usually is — infinite.

Core Concepts

The mechanism: conjugate, subtract, contradict

Suppose SD is conjugation-stable and some nonzero hD[t] vanishes on S. Multiplying on the left by the inverse of the leading coefficient, we may take h monic; choose it of least degree m among monic vanishing polynomials.

Fix eD. For aS, conjugating the relation h(a)=0 by e gives he(eae1)=0, and eae1 again runs over all of S. So he is monic of degree m and also vanishes on S. Then hhe vanishes on S and has degree <m, so it must be zero. Since e was arbitrary, every coefficient of h commutes with every element of D — that is, lies in F.

NormaliseScale on the left so that the shortest vanishing polynomial is monic of degree m.
Conjugatehe vanishes on S too, because S is conjugation-stable and conjugation is a ring automorphism.
Subtracthhe is a vanishing polynomial of degree <m, hence zero by minimality.
ConcludeAll coefficients are fixed by every inner automorphism, so they lie in F=Z(D).

From centrality to the three theorems

Once the shortest vanishing polynomial is known to lie in F[t], each of the three results follows by asking what a nonzero element of F[t] vanishing on S can be. For an algebraic class it must be a multiple of the class's minimal polynomial; for a non-algebraic class it cannot exist; for S=D it forces F to be finite and D to be algebraic over F, which Jacobson's theorem then forbids.

Key Results

Lemmacf. (16.5)Minimal vanishing polynomials are central

Let D be a division ring with centre F and let SD be stable under conjugation by D. If some nonzero element of D[t] vanishes on S, and h is monic of least degree among such, then hF[t].

Proof

Write h(t)=tm+d1tm1++dm and suppose djF for some j. Pick eD with edje1dj and set b=ebe1. For aS, conjugating am+d1am1++dm=0 by e gives

(a)m+d1(a)m1++dm=0.

As a runs over S so does a=eae1, so the monic polynomial h(t)=tm+d1tm1++dm also vanishes on S. Hence H(t):=h(t)h(t)=i=1m(didi)tmi vanishes on S. Its coefficient in degree mj is djdj0, so H0, while degH<m. Scaling H on the left by the inverse of its leading coefficient contradicts the minimality of m.

Lemma(16.5)Degree bound on a class

Let D be a division ring with centre F and let A be a conjugacy class of D that is algebraic over F, with minimal polynomial fF[t]. If hD[t]{0} vanishes on A, then deghdegf.

Proof

Some nonzero polynomial vanishes on A — namely f — so a monic vanishing polynomial h0 of least degree m exists, and mdegh for every nonzero vanishing h. By the previous lemma h0F[t]. Pick aA; then h0(a)=0 with h0F[t] nonzero, so the minimal polynomial f of a over F divides h0 in F[t]. Hence deghm=degh0degf.

Theorem(16.6)The vanishing ideal of an algebraic class

With D, F, A and f as in (16.5): a polynomial hD[t] vanishes on A if and only if hD[t]f(t).

Proof

() Let aA. Since f(a)=0, the remainder theorem (16.2) gives fD[t](ta), hence h=qfD[t](ta) and therefore h(a)=0, again by (16.2). As aA was arbitrary, h(A)=0.

() Assume h(A)=0. Because f is monic, division on the right by f is available: h=qf+h1 with h1=0 or degh1<degf. Evaluation is additive, and qf vanishes on A by the first half, so h1 vanishes on A. By (16.5) a nonzero such h1 would satisfy degh1degf, which is false; hence h1=0 and h=qf.

CorollaryThe vanishing set is a two-sided ideal

In the situation of (16.6), D[t]f is a two-sided ideal of D[t]. Indeed fF[t] commutes with every element of D[t], so D[t]fD[t]=D[t]D[t]f=D[t]f. Contrast the vanishing set of a single root r, which is only the left ideal D[t](tr) and is two-sided precisely when rF.

PropositionEx. 16.6Non-algebraic classes admit no vanishing polynomial

Let A be a conjugacy class of D that is not algebraic over F=Z(D). Then no nonzero hD[t] vanishes on A.

Proof

If some nonzero polynomial vanished on A, the lemma would supply a monic vanishing h0F[t]. Evaluating at any aA would exhibit a as algebraic over F, contradicting the hypothesis.

Theorem(16.7)No polynomial vanishes on an infinite division ring

Let D be an infinite division ring. Then no nonzero hD[t] vanishes identically on D.

Proof

Suppose one did. The set S=D is conjugation-stable, so a monic vanishing polynomial h of least degree m1 exists and lies in F[t] by the lemma. In particular h is a nonzero polynomial of degree m over the field F vanishing at every element of FD, so |F|m and F is finite.

But h(D)=0 with hF[t] says that every element of D is algebraic over F, i.e. D is an algebraic division algebra over the finite field F. By Jacobson's theorem (13.11), D is commutative, so D=F is finite — contradicting the hypothesis that D is infinite.

RemarkFiniteness is essential

For a finite division ring — a finite field 𝔽q by Wedderburn's little theorem — the polynomial tqt vanishes identically. So (16.7) is exactly a statement about infinite D, and its proof necessarily routes through a commutativity theorem.

Proof Techniques and Method

How these proofs work, and which move to reuse.

Two ideas do all the work, and both generalise well beyond §16.

  • Symmetrise and subtract. When a set is invariant under a group of ring automorphisms, apply an automorphism to a minimal-degree relation and subtract. The difference is shorter, so it must vanish, and the original coefficients are forced into the fixed ring. Here the group is the inner automorphisms and the fixed ring is Z(D); the same move works for Galois actions and for σ-twisted settings.
  • Divide by a monic central polynomial. Right division by a monic f is available over any ring, and when f is central the quotient ideal D[t]f is two-sided. This turns a vanishing condition into a divisibility condition, which is what makes (16.6) usable in Wedderburn's factorisation argument.

The pattern to carry away: to compute a vanishing ideal over a noncommutative ring, first show the shortest generator is central, then compute in the commutative subring. Attacking the noncommutative problem directly is far harder and usually unnecessary.

Worked Example

The class of i in the real quaternions

Take D=, F=, and A the conjugacy class of i, that is, the unit sphere of purely imaginary quaternions. Every aA satisfies a2=1 and no linear equation over , so f(t)=t2+1 and degf=2.

  • (16.5) predicts no nonzero polynomial of degree 1 vanishes on A. Directly: at+b with a0 has the unique root a1b, and |A|>1.
  • (16.6) predicts the vanishing polynomials are [t](t2+1). So the only monic quadratic vanishing on A is t2+1 itself.
  • The polynomial h(t)=it2+i=i(t2+1) vanishes on A and has noncentral coefficients — permitted, because it is not of minimal degree.

A concrete check of the two-sidedness corollary: (t2+1)j=j(t2+1) in [t] since the coefficients 1 and 1 are real, so multiplying a vanishing polynomial on the right by any quaternion again yields a vanishing polynomial.

A class on which nothing vanishes

Let D be the division ring of fractions of the first Weyl algebra A1()=x,y/(xyyx1). In characteristic zero Z(D)=, and x is transcendental over because A1() is a domain containing the polynomial ring [x].

So the conjugacy class A of x is not algebraic over Z(D), and by the proposition above no nonzero hD[t] vanishes on A. Concretely: however many conjugates dxd1 one imposes as roots, no finite-degree polynomial identity results. This is the generic situation in a centrally infinite division ring.

Frameworks and Models

Conjugacy classes of a division ring sort into exactly three kinds by their vanishing ideal.

All conjugacy classes of Dvanishing set is a two-sided ideal of D[t], possibly zero
Algebraic over Fminimal polynomial fF[t] exists; vanishing set =D[t]f0
Noncentral, degf2the class is infinite (Herstein); f is irreducible over F of degree 2
Quadratic classesdegf=2; the case governing quaternion algebras and result (16.17)
Central singleton {a}, aFf=ta; vanishing set =D[t](ta), which here is two-sided
Not algebraic over Fvanishing set =0; no polynomial identity of any degree

Given a class A, what vanishes on it?

A={a} with a centralExactly the multiples of ta. The class is a point and the theory degenerates to the commutative case.
A noncentral and algebraicExactly D[t]f with f the shared minimal polynomial, irreducible over F of degree n2. Nothing of degree <n vanishes.
A not algebraicNothing but 0. Any attempt to interpolate through A by a polynomial fails at every degree.

Comparison and Classification

Which conclusions hold for which test set S
Vanishing set is nonzeroMinimal vanisher is centralVanishing set is two-sidedDegree bound available
Single noncentral root {r}yesnonono
Algebraic conjugacy class Ayesyesyesyes
Non-algebraic conjugacy classnon/ayesn/a
Infinite division ring Dnon/ayesn/a
Finite field 𝔽qyesyesyesyes

Which conclusions hold for which test set S

Commutative analogue versus division ring statement
Commutative factDivision ring analogueReference
Polynomials vanishing on a finite set S form (s(ts))Polynomials vanishing on a class A form D[t]f(16.6)
A nonzero polynomial of degree n has at most n rootsA nonzero polynomial vanishing on A has degree at least degf(16.5)
No nonzero polynomial vanishes on an infinite fieldNo nonzero polynomial vanishes on an infinite division ring(16.7)
The vanishing ideal is always two-sidedTrue for classes; false for a single noncentral rootCorollary above
tqt vanishes on 𝔽qSame, and Wedderburn's little theorem says there is no other finite caseRemark above

Relationship Map

This page supplies the divisibility statement that the rest of §16 consumes.

conjugate-and-subtract(16.5) degree bound(16.6) vanishing ideal(16.9) Wedderburn factorisation
  • Wedderburn's Factorisation Theorem uses (16.5) to force a maximal chain of linear right factors to exhaust the whole minimal polynomial.
  • The Gordon–Motzkin Theorem is logically independent of this page but supplies the complementary bound, on classes rather than degrees.
  • The Niven–Jacobson Theorem uses the quadratic case (16.17), which is (16.6) specialised to degf=2 plus a remainder argument.
  • (16.7) is the source of the standard fact that D[t] is neither left nor right primitive for a centrally finite D, developed in the exercises of §16.

Applications and Industry Use

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

  • **Structure of D[t].** Knowing that a vanishing set is D[t]f with f central identifies the two-sided ideals of D[t] that arise from evaluation conditions, and is the mechanism behind the failure of D[t] to be primitive when D is centrally finite.
  • Polynomial identities. (16.7) is a one-variable non-identity statement: an infinite division ring satisfies no nontrivial one-variable polynomial condition with left coefficients. It marks the boundary of what PI theory can detect at the level of a single variable.
  • Symbolic computation. Deciding membership in a vanishing ideal reduces, by (16.6), to right division by a central polynomial — an operation that behaves exactly as in the commutative case and can be delegated to standard library routines.
  • Skew polynomial codes. In D[t;σ] the analogous vanishing ideals are generated by the minimal σ-polynomial of a σ-conjugacy class; encoders and decoders for rank-metric codes are built from exactly this divisibility statement.

As with most of §16, the honest description is that these results are internal machinery. Their value is that they make an apparently infinite condition — vanishing on an infinite set — equivalent to a single divisibility, which is finitely checkable.

Standards and Notation

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

Class notationA or Δ(a) for the conjugacy class of a
Minimal polynomialf or minpolyF(A); Lam writes simply f(t)F[t]
Vanishingh(A)=0; some authors write AV(h)
Ideal generatedD[t]f=(f) when fF[t]; write D[t](tr) for the one-sided case
CentreF=Z(D); the letter k is common in the algebra literature
ImplementationsMagma and Sage provide quaternion algebras with minimal polynomials of elements; skew polynomial division is available in Sage's OrePolynomialRing

Failure Modes and Common Mistakes

  • Do not assume every class has a minimal polynomial; non-algebraic classes have none, and then no nonzero polynomial vanishes on them at all.
  • Do not divide on the left by f and expect the same remainder; only right division is used, and only because f is monic.
  • Do not read (16.6) as saying f is irreducible over D — it factors into linear factors over D, which is exactly Wedderburn's theorem.

Quick Reference

SettingD division ring, F=Z(D), t central, right evaluation
Class minimal polynomialfF[t], shared by all aA
Degree bound0h, h(A)=0deghdegf
Vanishing idealh(A)=0hD[t]f
Two-sidednessD[t]f is two-sided because fF[t]
Non-algebraic classvanishing set is 0
Whole ringD infinite only h=0 vanishes on D
Finite case𝔽q: tqt vanishes identically
Statements and their exact hypotheses
ReferenceHypothesesConclusion
Lemma (mechanism)S conjugation-stable, some nonzero vanisher existsThe monic vanisher of least degree lies in F[t]
(16.5)A a class algebraic over F, minimal polynomial f, h0, h(A)=0deghdegf
(16.6)Same, hD[t] arbitraryh(A)=0hD[t]f
Ex. 16.6A a class not algebraic over FOnly h=0 vanishes on A
(16.7)D an infinite division ringOnly h=0 vanishes on D

Frequently Asked Questions

Why must all elements of a conjugacy class share one minimal polynomial?

Because conjugation fixes the centre F pointwise. For gF[t] and dD one has g(dad1)=dg(a)d1, so g kills a exactly when it kills every conjugate of a. The sets of annihilating polynomials over F therefore coincide across the class, and so do their monic generators.

Does (16.6) say the minimal polynomial of a class is irreducible over D?

No. It is irreducible over the centre F, but over D it splits completely into linear factors — that is Wedderburn's factorisation theorem (16.9). The statement h(A)=0hD[t]f is about divisibility in D[t], not about irreducibility there.

Is the vanishing set of a class really a two-sided ideal?

Yes, and the reason is cheap: it equals D[t]f with f having coefficients in Z(D), so f is a central element of D[t] and the left ideal it generates is automatically two-sided. This is the exception rather than the rule — the set of polynomials vanishing at a single noncentral element is only a left ideal.

How does (16.7) relate to polynomial identities?

It is a one-variable statement, and a strong one: an infinite division ring satisfies no nontrivial condition of the form aixi=0 for all x. It does not say anything about multilinear identities in several variables, where centrally finite division rings do satisfy identities — the standard identity of degree 2n for algebras of degree n.

What breaks if the class is replaced by an arbitrary infinite subset?

The conjugate-and-subtract mechanism, which is the only tool available. It needs eSe1=S for every eD. For a general infinite subset the conjugated polynomial need not vanish on the same set, and nothing forces the coefficients into the centre.

Can a nonzero polynomial vanish on two different conjugacy classes?

Certainly — take the product of their minimal polynomials, or any common left multiple. What (16.5) constrains is the degree: vanishing on A alone already costs degf, and vanishing on several algebraic classes costs at least the degree of the least common left multiple of their minimal polynomials.

References

  1. T. Y. Lam, A First Course in Noncommutative Rings, Graduate Texts in Mathematics 131, Springer-Verlag, 1991, §16, (16.5)–(16.7) (pp. 264–267).
  2. N. Jacobson, Structure of Rings, American Mathematical Society Colloquium Publications 37, revised edition, 1964.
  3. B. Gordon and T. S. Motzkin, “On the zeros of polynomials over division rings”, Transactions of the American Mathematical Society 116 (1965), 218–226.
  4. I. N. Herstein, Noncommutative Rings, Carus Mathematical Monographs 15, Mathematical Association of America, 1968.
  5. P. K. Draxl, Skew Fields, London Mathematical Society Lecture Note Series 81, Cambridge University Press, 1983.
  6. L. H. Rowen, Ring Theory, Volume I, Academic Press, 1988.

AI Suggested Questions

  • Work out the vanishing ideal of a union of two distinct algebraic conjugacy classes and compare it with the product of their minimal polynomials.
  • Does the conjugate-and-subtract argument survive when inner automorphisms are replaced by a group of outer automorphisms of D?
  • For the division ring of fractions of the Weyl algebra, which conjugacy classes are algebraic over the centre?
  • Formulate and prove the analogue of (16.6) for skew polynomial rings D[t;σ] with σ-conjugacy classes.
  • Give a proof that D[t] is neither left nor right primitive when D is centrally finite, using (16.7) and the exercises of §16.
  • How large can the degree of the minimal polynomial of a class be, relative to dimFD, for a centrally finite division ring?
Page
KVS-ENG-MATH-0247
Path
Engineering / Mathematics
Template
kevos-knowledge-article-v2
KEVOS® Knowledge Library — reviewed 2026-08-08

Continue learning

Absolutely Irreducible Modules | KEVOS® MathematicsArticle · Project ManagementAdditive Commutators | KEVOS® MathematicsArticle · Project ManagementAlgebraic and Geometric Multiplicities of Eigenvalues | KEVOS® MathematicsArticle · Project ManagementAlgebraically Closed Division Rings | KEVOS® MathematicsArticle · Project Management