Homework 3

9.1.4. Claim: The ideal (x) of Q[x, y] is prime but not maximal. The ideal (x, y) in Q[x, y]
is maximal.

Proof: Consider first the map

φ : Q[x, y] → Q[y]
where
φ(f (x, y)) := f (0, y).
First, I claim φ is well-defined. If
j
n X
X
f (x, y) = ai,j xi y j−i
j=0 i=0

is an element of Q[x, y], then

X j
n X X
φ(f (x, y)) = ai,j 0i y j−i = a0,j y j ∈ Q[y].
j=0 i=0 j=0,n

Now I claim φ is a ring homomorphism. Suppose f (x, y) = nj=0 ji=0 ai,j xi y j−i and
P P
Pm Pj i j−i are in Q[x, y], and say n ≥ m. For convenience, define
g(x, y) = j=0 i=0 bi,j x y
bi,j = 0 for i + j > m. Then
n
X
φ(f (x, y) + g(x, y)) = (a0,j + b0,j )y j = φ(f (x, y)) + φ(g(x, y)).
j=0

Also,
j
n X
X
φ(f (x, y)g(x, y)) = (a0,k b0,j−k )y j = φ(f (x, y))φ(g(x, y)).
j=0 k=0

This shows φ is a ring homomorphism.

The kernel of φ is the set of polynomials which φ sends to 0. These are polynomials all
of whose terms contain an x. That is, the kernel of φ is exactly (x). Hence by the First
Isomorphism Theorem for rings, Q[x, y]/(x) ∼ = Q[y]. We know that Q[y] is an integral
domain. Hence (x) is prime. Since Q[y] is not a field, (x) is not maximal.
Now consider the map
ρ : Q[x, y] → Q

1
defined by
ρ(f (x, y)) := f (0, 0).
First, I claim this map is well-defined. If
j
n X
X
f (x, y) = ai,j xi y j−i ,
j=0 i=0

then
ρ(f (x, y)) = a0,0
is an element of Q.
Now I claim this map is a ring homomorphism. Suppose f (x, y) = nj=0 ji=0 ai,j xi y j−i
P P
Pj
and g(x, y) = m i j−i are in Q[x, y], and say n ≥ m. Again for convenience,
P
j=0 i=0 bi,j x y
define bi,j = 0 for i + j > m. Then

ρ(f (x, y) + g(x, y)) = a0,0 + b0,0 = φ(f (x, y)) + φ(g(x, y)).

Also,
ρ(f (x, y)g(x, y)) = a0,0 b0,0 = φ(f (x, y))φ(g(x, y)).
This shows φ is a ring homomorphism.
The kernel of this ring homomorphism is the set of polynomials with no constant term.
This is exactly the ideal (x, y). Hence, by the First Isomorphism Theorem, Q[x, y]/(x, y) ∼
=
Q. Since Q is a field, (x, y) is a maximal ideal.

9.2.1. Claim: If F is a field, and f (x) ∈ F [x] is a polynomial of degree n, then for
every g(x) ∈ F [x]/(f (x)), there is a unique polynomial g0 (x) of degree ≤ n − 1 such that
g(x) = g0 (x).

Proof: Suppose g(x) ∈ F [x] is nonzero. Then by Theorem 9.3, there exist unique polyno-
mials q(x) and r(x) in F [x] so that

g(x) = q(x)f (x) + r(x),

and either r(x) = 0 or the degree of r(x) is less than n. Set g0 (x) = r(x). Then since
g(x) − g0 (x) = q(x)f (x) ∈ (f (x)), g(x) = g0 (x) ∈ F [x]/(f (x)).
Now suppose there were another g00 (x) whose degree were smaller than n, and such that
g(x) = g00 (x) ∈ F [x]/(f (x)). Then there would exist q 0 (x) such that g(x) = q 0 (x)f (x) +
g00 (x). This would contradict the uniqueness of this decomposition in Theorem 9.3, hence
g00 (x) = g0 (x) is unique.

9.2.5. Claim: Suppose F is a field, and p(x) ∈ F [x]. Then all ideals of F [x]/(p(x)) are of
the form

2
Proof: First, by the Fourth Isomorphism Theorem for rings, I/(p(x)) is an ideal of
F [x]/(p(x)) if and only if I is an ideal of F [x] containing p(x). By Theorem 9.3 F [x]
is a Euclidean Domain, and hence a Principal Ideal Domain. Hence all ideals of F [x] are
of the form I = (f (x)) for some f (x) ∈ F [x]. Then (f (x)) ⊇ (p(x)) if and only if f (x)
divides p(x).
Since F [x] is a Unique Factorization Domain, we can write p(x) = p1 (x)p2 (x) · · · pn (x)
for some irreducible polynomials pi (x) which are unique up to associates. Since associate
elements of a ring generate the same ideal (Proposition 8.3), we then know that the ideals
of F [x] containing p(x) are exactly the ideals of the form (pi1 (x) · · · pis (x)) for some subset
{i1 , . . . , is } of {1, . . . , n}. We can then conclude that the ideals of F [x]/(p(x)) are all of
the form (pi1 (x) · · · pis (x))/(p(x)).

9.3.4. Let R = Z + xQ[x].

(a) Claim: R is an integral domain, whose units are ±1.

Proof: The ring R is a subring of Q[x], which is an integral domain. Hence R is an

integral domain. Further, suppose f (x)g(x) = 1. Since we have a degree norm on R
which is additive, the degree of f (x) and g(x) must be 0. Then f (x) and g(x) must be
constants which are units in Z. Hence the units in R are ±1.

(b) Claim: The irreducibles in R are ±p where p is prime in Z, and polynomials f (x) which
are irreducible in Q[x] and have constant term ±1. These irreducibles are prime in R.

Proof: Suppose first f (x) is of the form ±p for some prime p ∈ Z. If f (x) could be a
written as a product f (x) = a(x)b(x) ∈ R, the degrees of the terms would need to add
to 0. Hence this would give a factorization of the prime p into a product of integers.
Since primes in Z are irreducible, this implies a(x) or b(x) is a unit. Hence f (x) is
irreducible in R.
Now suppose f (x) is an irreducible polynomial in Q[x] of degree at least 1, and constant
term ±1. If f (x) = a(x)b(x) in R, then since f (x) is irreducible in Q[x], one of a(x) or
b(x) is a unit in Q[x]. Say a(x) is a unit. The units in Q[x] are the nonzero elements
of Q. Hence a(x) ∈ Q. However, since a(x) ∈ R, and a(x) is degree 0, a(x) ∈ Z.
The constant term of f (x) = ±1, and the constant term of b(x) is an integer, thus
a(x) = ±1. Hence f (x) is irreducible in R.
Now suppose f (x) is an irreducible polynomial in R. If the degree of f (x) is 0, then
f (x) must be irreducible in Z, and so f (x) = ±p for some prime number p ∈ Z. If the
degree of f (x) is at least 1, then its constant term c of f (x) may only be ±1. Clearly, c
cannot be 0 or it would be possible to factor f (x) = d1 xg(x) where d is the denominator

3
 of the linear term of f (x), and g(x) = dfx(x) . Otherwise, if c were nonzero, then c would
not be a unit, and f (x) would factor as f (x) = c( 1c (f (x) − c) + 1) in R.
Suppose the degree of f (x) is at least one, and suppose by way of contradiction that
f (x) is reducible in Q[x] and factors as a product of nonunits f (x) = a(x)b(x). Both
of a(x) and b(x) must have nonzero constant terms, call them a0 and b0 in Q. Further,
a0 b0 = ±1, as seen previously Then f (x) = a0 b0 ( a10 a(x))( b10 b(x)) is a factorization of
f (x) into nonunits. Hence f (x) must be reducible in f (x).

(c) Claim: The element x is not irreducible in R, and cannot be written as a product of
irreducibles.

Proof: Since 21 x and 2 are nonunits in R which multiply to x, x is not irreducible in

R. However, if we could write x = p1 (x) · · · pn (x) for irreducible elements pi (x), then
by the additivity of degrees, all but 1 would have degree 0, and the other, say p1 (x),
degree 1. So p1 (x) would be of the form ax + b, for a ∈ Q and b = ±1, and for i > 1,
pi (x) = pi would be an irreducible in Z. It is not possible for such polynomials to
multiply to a polynomial with 0 constant term, and so x is not a product of irreducible
elements of R.

(d) Claim: The element x is not prime in R. R/(x) is not an integral domain in which
all nonzero elements which are not units are zero-divisors. Its elements can all be
represented uniquely by polynomials of the form ax + b where a ∈ [0, 1) ∩ Q and b ∈ Z.

Proof: First, consider the elements 2 and 21 x of R. Neither is contained in (x), since
the elements of (x) may only have integer coefficients for their degree 1 term. But
2( 21 x) = x ∈ (x). Hence (x) is not a prime ideal, and x is not prime in R.
In particular, the ring R/(x) is not an integral domain. Two elements f (x) and g(x)
are equal in R/(x) if f (x) − g(x) is a polynomial with no constant term, and whose
degree 1 term has an integer coefficient. Therefore every element can be represented
uniquely by a polynomial ax + b, where a ∈ [0, 1) ∩ Q, and b ∈ Z. Further, all elements
other than ±1 are zero divisors. This is because if ax + b 6= ±1, 0, then if b 6= 0,
1 1 1
b x 6∈ (x) and b x(ax + b) ∈ (x). If b = 0, 2 x(ax) ∈ (x).