Chapter 14
Hard

Polynomial Rings

00 · Symbol Glossary

$R[x]$Polynomial ring over R

The ring of all formal polynomial expressions a0+a1x+⋯+anxna_0+a_1x+\cdots+a_nx^n with coefficients ai∈Ra_i\in R, under the usual addition and multiplication of polynomials.

$\deg f$Degree of f

The highest power of xx with a nonzero coefficient in ff. By convention deg⁡0:=−∞\deg 0 := -\infty, so that degree-arithmetic rules (deg⁡(fg)=deg⁡f+deg⁡g\deg(fg)=\deg f+\deg g, etc.) hold without exception.

$a_n$Leading coefficient

The coefficient of the highest-degree term of ff (assuming f≠0f\neq0). A polynomial is monic if its leading coefficient is 11.

$f(a)$Evaluation of f at a

Substitute x=ax=a into ff and compute the resulting element of RR. The map f↦f(a)f\mapsto f(a) is exactly the evaluation homomorphism from Chapter 13, specialized to polynomial rings.


01 · Constructing R[x]

Definition — Polynomial Ring

For a ring RR, the polynomial ring R[x]R[x] consists of all formal expressions f=a0+a1x+⋯+anxnf=a_0+a_1x+\cdots+a_nx^n (ai∈Ra_i\in R, n≥0n\geq0), with:

(∑iaixi)+(∑ibixi)=∑i(ai+bi)xi,(∑iaixi)(∑jbjxj)=∑k(∑i+j=kaibj)xk\Big(\sum_i a_ix^i\Big) + \Big(\sum_i b_ix^i\Big) = \sum_i(a_i+b_i)x^i, \qquad \Big(\sum_i a_ix^i\Big)\Big(\sum_j b_jx^j\Big) = \sum_k\Big(\sum_{i+j=k}a_ib_j\Big)x^k

Verification is routine but real

Addition is componentwise, so (R[x],+)(R[x],+) is an abelian group directly from (R,+)(R,+) being one. Multiplication's associativity and distributivity over addition follow by expanding both sides of the relevant identities and matching coefficients — genuinely mechanical, but a real check, exactly parallel to verifying (Rn,+)(R^n,+) or Mn(R)M_n(R) satisfy ring axioms. R[x]R[x] is commutative exactly when RR is, and has unity 1R1_R (the constant polynomial) exactly when RR does.


02 · Degree

Definition — Degree

For nonzero f=a0+⋯+anxnf=a_0+\cdots+a_nx^n with an≠0a_n\neq0, deg⁡f:=n\deg f := n. By convention, deg⁡0:=−∞\deg 0 := -\infty.

Theorem 14.1 — Degree Inequalities

For f,g∈R[x]f,g\in R[x]: (a) deg⁡(f+g)≤max⁡(deg⁡f,deg⁡g)\deg(f+g) \le \max(\deg f,\deg g); (b) deg⁡(fg)≤deg⁡f+deg⁡g\deg(fg) \le \deg f + \deg g.

Proof

(a) Coefficients of f+gf+g above max⁡(deg⁡f,deg⁡g)\max(\deg f,\deg g) are sums of zero coefficients from both ff and gg, hence zero. So f+gf+g has no nonzero coefficient beyond that maximum.

(b) Writing f=∑aixif=\sum a_ix^i (deg⁡f=m\deg f=m) and g=∑bjxjg=\sum b_jx^j (deg⁡g=n\deg g=n), every coefficient of fgfg at a power xkx^k with k>m+nk>m+n is a sum of terms aibja_ib_j with i+j=k>m+ni+j=k>m+n; since i≤mi\le m forces j=k−i>nj=k-i > n, giving bj=0b_j=0 (or symmetric), every such term vanishes.

Theorem 14.2 — Degree Is Additive Over Domains

If RR is an integral domain and f,g∈R[x]f,g\in R[x] are both nonzero, then deg⁡(fg)=deg⁡f+deg⁡g\deg(fg) = \deg f+\deg g.

Proof

Let am,bna_m,b_n be the leading coefficients of f,gf,g (both nonzero). The coefficient of xm+nx^{m+n} in fgfg is exactly ambna_mb_n (every other pairing i+j=m+ni+j=m+n with i≠mi\neq m would need j>nj>n, giving bj=0b_j=0; similarly for j≠nj\neq n). Since RR has no zero divisors and am,bn≠0a_m,b_n\neq0, ambn≠0a_mb_n\neq0. So fgfg has a nonzero coefficient at exactly degree m+nm+n, and by Theorem 14.1(b) no nonzero coefficient beyond it: deg⁡(fg)=m+n=deg⁡f+deg⁡g\deg(fg)=m+n=\deg f+\deg g.

Common mistake — Degree can drop unexpectedly outside a domain

In Z4[x]\mathbb{Z}_4[x]: (2x+1)(2x+1)=4x2+4x+1=0⋅x2+0⋅x+1=1(2x+1)(2x+1) = 4x^2+4x+1 = 0\cdot x^2+0\cdot x+1 = 1 (since 4≡04\equiv0 in Z4\mathbb{Z}_4). Here deg⁡(2x+1)=1\deg(2x+1)=1 on both sides, but deg⁡((2x+1)2)=deg⁡(1)=0≠1+1\deg\big((2x+1)^2\big) = \deg(1) = 0 \neq 1+1. Theorem 14.2 genuinely requires RR to be a domain — Z4\mathbb{Z}_4 is not (Chapter 11), and the degree formula fails exactly because the leading coefficients 2×2=4=02\times2=4=0 collide to zero.

Corollary 14.3 — R[x] Is a Domain iff R Is a Domain

R[x]R[x] is an integral domain if and only if RR is an integral domain.

Proof

(⇐\Leftarrow) Suppose RR is a domain. If f,g∈R[x]f,g\in R[x] are both nonzero, Theorem 14.2 gives deg⁡(fg)=deg⁡f+deg⁡g\deg(fg)=\deg f+\deg g, a finite nonnegative number — so fg≠0fg\neq0 (which would need deg⁡(fg)=−∞\deg(fg)=-\infty). No zero divisors: R[x]R[x] is a domain (commutative and unital, inherited directly from RR).

(⇒\Rightarrow) Suppose R[x]R[x] is a domain. The constant polynomials form a subring of R[x]R[x] isomorphic to RR (an easy check: constants add and multiply exactly as their coefficients do). If RR had a zero divisor ab=0ab=0 with a,b≠0a,b\neq0, the corresponding constant polynomials would give ab=0ab=0 in R[x]R[x] too, with both factors nonzero — contradicting R[x]R[x] being a domain. So RR has no zero divisors: a domain.

Theorem 14.4 — Units of R[x] Over a Domain

If RR is an integral domain, R[x]×=R×R[x]^\times = R^\times (the units of R[x]R[x] are exactly the constant polynomials that are units in RR).

Proof

If u∈R×u\in R^\times, the constant polynomial uu has inverse the constant polynomial u−1u^{-1} in R[x]R[x], so u∈R[x]×u\in R[x]^\times.

Conversely, suppose fg=1fg=1 in R[x]R[x]. By Theorem 14.2 (domain, both f,gf,g nonzero since their product is nonzero), deg⁡f+deg⁡g=deg⁡(1)=0\deg f+\deg g = \deg(1) = 0. Since degrees are nonnegative integers, deg⁡f=deg⁡g=0\deg f=\deg g=0: both f,gf,g are constants, say f=a,g=b∈Rf=a,g=b\in R with ab=1ab=1. So a∈R×a\in R^\times, and f=af=a is (the constant polynomial corresponding to) a unit of RR.


03 · The Division Algorithm

The single most useful structural fact about polynomials over a field — the direct polynomial analog of ordinary long division of integers.

Theorem 14.5 — Division Algorithm for F[x]

Let FF be a field and f,g∈F[x]f,g\in F[x] with g≠0g\neq0. There exist unique q,r∈F[x]q,r\in F[x] with:

f=qg+r,r=0 or deg⁡r<deg⁡gf = qg+r, \qquad r=0 \text{ or } \deg r < \deg g

Proof

Existence, by strong induction on deg⁡f\deg f (treating f=0f=0 or deg⁡f<deg⁡g\deg f<\deg g as base cases with q=0,r=fq=0,r=f). If deg⁡f≥deg⁡g\deg f\geq\deg g, write f=amxm+⋯f=a_mx^m+\cdots, g=bnxn+⋯g=b_nx^n+\cdots with m≥nm\geq n, am,bn≠0a_m,b_n\neq0. Since FF is a field, bn−1b_n^{-1} exists; form:

f1:=f−ambnxm−ngf_1 := f - \frac{a_m}{b_n}x^{m-n}g

The xmx^m terms cancel by construction, so deg⁡f1<deg⁡f\deg f_1 < \deg f. By the induction hypothesis, f1=q1g+rf_1 = q_1g+r with r=0r=0 or deg⁡r<deg⁡g\deg r<\deg g. Then f=f1+ambnxm−ng=(q1+ambnxm−n)g+rf = f_1 + \frac{a_m}{b_n}x^{m-n}g = \big(q_1+\frac{a_m}{b_n}x^{m-n}\big)g + r, giving the required qq.

Uniqueness. Suppose f=q1g+r1=q2g+r2f=q_1g+r_1=q_2g+r_2 with each ri=0r_i=0 or deg⁡ri<deg⁡g\deg r_i<\deg g. Then (q1−q2)g=r2−r1(q_1-q_2)g = r_2-r_1. If q1≠q2q_1\neq q_2, the left side has degree deg⁡(q1−q2)+deg⁡g≥deg⁡g\deg(q_1-q_2)+\deg g \geq \deg g (Theorem 14.2, valid since FF is a domain), while the right side has degree <deg⁡g<\deg g (Theorem 14.1(a), since both rir_i do) — a contradiction. So q1=q2q_1=q_2, forcing r1=r2r_1=r_2 too.

Step-by-step — Dividing $f=x^3+2x-3$ by $g=x-1$ in $\mathbb{Q}[x]$
1
Match leading terms: x3/x=x2x^3 / x = x^2. Multiply: x2(x−1)=x3−x2x^2(x-1) = x^3-x^2.
2
Subtract: (x3+2x−3)−(x3−x2)=x2+2x−3(x^3+2x-3) - (x^3-x^2) = x^2+2x-3.
3
Repeat — match x2/x=xx^2/x=x: x(x−1)=x2−xx(x-1)=x^2-x. Subtract: (x2+2x−3)−(x2−x)=3x−3(x^2+2x-3)-(x^2-x) = 3x-3.
4
Repeat — match 3x/x=33x/x=3: 3(x−1)=3x−33(x-1)=3x-3. Subtract: (3x−3)−(3x−3)=0(3x-3)-(3x-3)=0.
5
Conclusion: q=x2+x+3q=x^2+x+3, r=0r=0. So x3+2x−3=(x−1)(x2+x+3)x^3+2x-3 = (x-1)(x^2+x+3) exactly.

04 · Roots and the Factor Theorem

Theorem 14.6 — Factor Theorem

For a field FF, f∈F[x]f\in F[x], and a∈Fa\in F: f(a)=0f(a)=0 if and only if (x−a)(x-a) divides ff.

Proof

By Theorem 14.5 (division algorithm with g=x−ag=x-a, degree 1): f=q(x−a)+rf = q(x-a)+r with r=0r=0 or deg⁡r<1\deg r<1, i.e. rr is a constant c∈Fc\in F. Evaluating at x=ax=a: f(a)=q(a)(a−a)+c=cf(a) = q(a)(a-a)+c = c. So r=f(a)r=f(a) always. Then (x−a)∣f  ⟺  r=0  ⟺  f(a)=0(x-a)\mid f \iff r=0 \iff f(a)=0.

Corollary 14.7 — A Degree-n Polynomial Has at Most n Roots

A nonzero polynomial f∈F[x]f\in F[x] (FF a field) with deg⁡f=n\deg f=n has at most nn distinct roots in FF.

Proof

By induction on nn. If n=0n=0, ff is a nonzero constant, with no roots — consistent with "at most 0."

Suppose deg⁡f=n≥1\deg f=n\geq1 and ff has a root a∈Fa\in F. By Theorem 14.6, f=(x−a)hf=(x-a)h for some h∈F[x]h\in F[x], and by Theorem 14.2 (domain), deg⁡h=n−1\deg h = n-1. If b≠ab\neq a is another root of ff: 0=f(b)=(b−a)h(b)0=f(b)=(b-a)h(b). Since FF is a field (no zero divisors) and b−a≠0b-a\neq0, h(b)=0h(b)=0: every root of ff other than aa is a root of hh. By the induction hypothesis, hh has at most n−1n-1 distinct roots, so ff has at most 1+(n−1)=n1+(n-1)=n distinct roots (counting aa itself). If ff has no roots at all, the bound "at most nn" holds trivially.

Common mistake — The root bound requires a field (or at least a domain)

In Z8[x]\mathbb{Z}_8[x] (not a domain, since Z8\mathbb{Z}_8 isn't prime), consider f(x)=x2−1f(x)=x^2-1, degree 2. Check every element of Z8\mathbb{Z}_8: 12−1=01^2-1=0; 32−1=8≡03^2-1=8\equiv0; 52−1=24≡05^2-1=24\equiv0; 72−1=48≡07^2-1=48\equiv0. Four roots — 1,3,5,71,3,5,7 — for a degree-2 polynomial, violating Corollary 14.7's bound. The corollary's proof used FF being a field (specifically, no zero divisors) at the exact step "b−a≠0b-a\neq0 and h(b)=0  ⟹  h(b)=0 \implies contradiction unless..." — that step fails entirely once zero divisors are available.


05 · Exercises

EXERCISE 14.1

Multiply out (x+1)(x2−x+1)(x+1)(x^2-x+1) directly and compare degrees to Theorem 14.2's prediction.

(x+1)(x2−x+1)=x3−x2+x+x2−x+1=x3+1(x+1)(x^2-x+1) = x^3-x^2+x+x^2-x+1 = x^3+1. deg⁡(x+1)=1\deg(x+1)=1, deg⁡(x2−x+1)=2\deg(x^2-x+1)=2, sum =3=deg⁡(x3+1)=3=\deg(x^3+1). Matches Theorem 14.2, as expected since Q\mathbb{Q} is a field (hence a domain).

Verify Theorem 14.2 directly for f=x+1f=x+1, g=x2−x+1g=x^2-x+1 in Q[x]\mathbb{Q}[x], by computing fgfg and comparing degrees.

EXERCISE 14.2

Perform the same style of polynomial long division as the Section 03 example.

Divide x3−1x^3-1 by x−1x-1: match x3/x=x2x^3/x=x^2, subtract x2(x−1)=x3−x2x^2(x-1)=x^3-x^2, leaving x2−1x^2-1. Match x2/x=xx^2/x=x, subtract x(x−1)=x2−xx(x-1)=x^2-x, leaving x−1x-1. Match x/x=1x/x=1, subtract 1(x−1)=x−11(x-1)=x-1, leaving 00.

q=x2+x+1q=x^2+x+1, r=0r=0: x3−1=(x−1)(x2+x+1)x^3-1=(x-1)(x^2+x+1).

Divide f=x3−1f=x^3-1 by g=x−1g=x-1 in Q[x]\mathbb{Q}[x], finding qq and rr.

EXERCISE 14.3

Apply the Factor Theorem: evaluate ff at the proposed root first.

Evaluate f(x)=x3−4x+3f(x)=x^3-4x+3 at x=1x=1: 1−4+3=01-4+3=0. By Theorem 14.6, (x−1)∣f(x-1)\mid f. Dividing: x3−4x+3=(x−1)(x2+x−3)x^3-4x+3 = (x-1)(x^2+x-3) (check: matching leading terms and subtracting, as in prior examples).

Show that (x−1)(x-1) divides f=x3−4x+3f=x^3-4x+3 in Q[x]\mathbb{Q}[x], using the Factor Theorem, and find the quotient.

EXERCISE 14.4

Apply Corollary 14.7 directly — note that Q\mathbb{Q} is a field.

deg⁡f=5\deg f = 5 over the field Q\mathbb{Q}. By Corollary 14.7, ff has at most 5 distinct rational roots.

A degree-5 polynomial f∈Q[x]f\in\mathbb{Q}[x] is given. Using Corollary 14.7, what is the maximum number of distinct rational roots it can have?


06 · Chapter Summary

ConceptStatement
R[x]R[x]Formal polynomials over RR; commutative/unital iff RR is
Degree inequalitiesdeg⁡(f+g)≤max⁡(deg⁡f,deg⁡g)\deg(f+g)\le\max(\deg f,\deg g); deg⁡(fg)≤deg⁡f+deg⁡g\deg(fg)\le\deg f+\deg g (Thm 14.1)
Degree over a domaindeg⁡(fg)=deg⁡f+deg⁡g\deg(fg)=\deg f+\deg g exactly (Thm 14.2)
R[x]R[x] is a domainIff RR is a domain (Cor 14.3)
Units of R[x]R[x]Exactly R×R^\times, when RR is a domain (Thm 14.4)
Division algorithmUnique q,rq,r with f=qg+rf=qg+r, deg⁡r<deg⁡g\deg r<\deg g, over a field (Thm 14.5)
Factor Theoremf(a)=0  ⟺  (x−a)∣ff(a)=0 \iff (x-a)\mid f (Thm 14.6)
Root boundDegree-nn polynomial over a field has ≤n\le n roots (Cor 14.7)

Next: Chapter 15 — Euclidean Domains & Principal Ideal Domains abstracts the division algorithm just proved for F[x]F[x] into a general framework covering both Z\mathbb{Z} and polynomial rings at once, and shows every ideal in such a ring is principal.