Zachary Scherr Math 503 HW 2 Due Friday, Feb 5

1 Reading
1. Read sections 7.3, 7.4, 7.5, 7.6 of Dummit and Foote

2 Problems
1. 7.1.26

Solution:
(a) First note that
ν(1) = ν(1 · 1) = ν(1) + ν(1)
showing that ν(1) = 0 and hence 1 ∈ R. From this we have,

0 = ν(1) = ν(−1 · −1) = ν(−1) + ν(−1)

implying that ν(−1) = 0 as well. Now let r, s ∈ R be non-zero with r − s 6= 0. Then by

definition we have ν(r), ν(s) ≥ 0. We know that

ν(−s) = ν(−1) + ν(s) = ν(s),

and so
ν(r − s) ≥ min(ν(r), ν(−s)) = min(ν(r), ν(s)) ≥ 0
showing that r − s ∈ R. Therefore R is a subgroup of K, and since

ν(rs) = ν(r) + ν(s) ≥ 0

we also have that R is closed under multiplication. Therefore R is a subring of K and we have
already seen that 1 ∈ R.
(b) Let x ∈ K be non-zero. If ν(x) ≥ 0 then x ∈ R by definition. Otherwise, ν(x) < 0, and since ν
is a homomorphism of groups we get
 
1
ν(x−1 ) = ν = ν(1) − ν(x) = −ν(x) > 0
x

and so x−1 ∈ R.

(c) Let x ∈ R be non-zero. Since R ⊆ K, and K is a field, we know that there exists y ∈ K with
xy = 1. Then
0 = ν(1) = ν(xy) = ν(x) + ν(y).
If x ∈ R× then y ∈ R, Since both ν(x), ν(y) ≥ 0 and ν(x) + ν(y) = 0 we see that ν(x) = 0.
Conversely, if ν(x) = 0 then 0 = ν(x) + ν(y) implies that ν(y) = 0 and hence y ∈ R.

2. 7.1.27

Solution: The question doesn’t ask this, but let’s verify that νp is a valuation. Consider x, y ∈ Q×
and write them as
a c
x = pα and y = pβ
b d
Zachary Scherr Math 503 HW 2 Due Feb 5

where a, b, c, d are all relatively prime to p. Then

ac
xy = pα+β
bd
and since ac and bd are still relatively prime to p we see

νp (xy) = α + β = νp (x) + νp (y).

Since νp (pk ) = k we easily see that ν is surjective. Lastly, assume WLOG that α ≤ β. Then

ad + pβ−α bc
 
αa
c a c
β
x+y =p +p = pα + pβ−α = pα .
b d b d bd

Since bd is relatively prime to p and ad + pβ−α bc ∈ Z, we see that

νp (x + y) ≥ α = min(νp (x), νp (y))

proving that indeed νp is a valuation.

For such a prime p, let Rp ⊆ Q denote the discrete valuation ring associated to νp . If x = pα ab ∈ Q× ,
with a, b relatively prime to p, then x ∈ Rp if and only if α ≥ 0. Thus Rp is the set of all fractions
of Q who in their lowest terms have denominators relatively prime to p. By the previous exercise,
the units in Rp are precisely those elements for which α = 0, which consists of all fractions who, in
its lowest terms, have both numerator and denominator relatively prime to p.

3. 7.4.12

Solution: Let I = (a1 , a2 , . . . , an ) and J = (b1 , b2 , . . . , bm ). Then by definition of IJ we have

ai bj ∈ IJ implying that
({ai bj | 1 ≤ i ≤ n, 1 ≤ j ≤ m}) ⊆ IJ.
Conversely, we have that IJ = ({ij | i ∈ I, j ∈ J}). Let i ∈ I and j ∈ J. Then there are elements
r1 , . . . , rn ∈ R and s1 , . . . , sm ∈ R with

i = r1 a1 + . . . + rn an
j = s1 b1 + . . . + sm bm .

Thus using distributivity and commutativity we have

X
ij = (ri sj )(ai bj ) ∈ ({ai bj | 1 ≤ i ≤ n, 1 ≤ j ≤ m}).
1≤i≤n
1≤j≤m

Therefore
IJ = ({ij | i ∈ I, j ∈ J}) ⊆ ({ai bj | 1 ≤ i ≤ n, 1 ≤ j ≤ m})
and hence we get equality.

4. 7.4.13

Solution:

Page 2
Zachary Scherr Math 503 HW 2 Due Feb 5

(a) Let ϕ : R → S be a ring homomorphism, let P be a prime ideal of S and let I = ϕ−1 (P ).
We’ve already proven in the previous homework that I is an ideal of R. Assume a, b ∈ R satisfy
ab ∈ I. By definition of I we have

ϕ(a)ϕ(b) = ϕ(ab) ∈ P.

Since P is prime, this means that ϕ(a) ∈ P or ϕ(b) ∈ P and hence a ∈ I or b ∈ I. Thus either
I = R or I is a prime ideal.
If R ⊆ S is a subring, then the inclusion map ι : R → S is a homomorphism. If P is a prime
ideal of S, then ι−1 (P ) = P ∩ R. Thus either P ∩ R = R or P ∩ R is a prime ideal of R.
(b) Now we suppose that ϕ is surjective and M is a maximal ideal of S. The “sexy” way of
proving this is to use the first isomorphism theorem. Consider the quotient map π : S → S/M .
Composing, we have a map
ψ : R → S/M
given by ψ = π ◦ ϕ. This map is surjective since both ϕ and π are surjective. Moreover, we
know that ker(ψ) = ϕ−1 (M ). Thus by the first isomorphism theorem,

R/ϕ−1 (M ) ∼
= S/M.

The latter ring is a field since M is maximal in S, and so R/ϕ−1 (M ) is a field and hence
ϕ−1 (M ) is maximal in R.
A more bare bones approach is as follows. First we need to prove that ϕ−1 (M ) is a proper
ideal of R. Because ϕ is surjective, there exists x ∈ R with ϕ(x) = 1 ∈ S. Since M 6= S we
have 1 6∈ M and so x 6∈ ϕ−1 (M ). Next, let I be any ideal of R containing ϕ−1 (M ). Then ϕ is
surjective so we know that
M = ϕ(ϕ−1 (M )) ⊆ ϕ(I).
Because ϕ is surjective, the set ϕ(I) is an ideal of S containing M and hence ϕ(I) = M or
ϕ(I) = S. I claim that ϕ−1 (ϕ(I)) = I. We know that I ⊆ ϕ−1 (ϕ(I)). For the reverse inclusion,
let r ∈ ϕ−1 (ϕ(I)). Then ϕ(r) ∈ ϕ(I) and so there exists s ∈ I with ϕ(r) = ϕ(s). Thus
ϕ(r − s) = 0 showing that r − s ∈ ϕ−1 (0). Since

I ⊇ ϕ−1 (M ) ⊇ ϕ−1 (0)

we get that r − s ∈ I showing that r ∈ I as well. Applying this to the above, we see that
I = ϕ−1 (M ) or I = ϕ−1 (S) = R, proving indeed that ϕ−1 (M ) is maximal.
For an example, consider the inclusion map ι : Z → Q. Since Q is a field, the ideal (0) ⊆ Q is
maximal. It’s preimage in Z is the zero ideal which is prime but not maximal.

5. 7.4.14

Solution:
(a) Let f (x) = xn + bn−1 xn−1 + . . . + b1 x + b0 . The goal is to prove that for every g(x) ∈ R[x]
there exists p(x) ∈ R[x] with deg(p(x)) < n so that
f (x) ≡ p(x) (mod f (x)).
We prove this by induction on deg(g(x)). There is nothing to prove if deg(g(x)) < n. If
deg(g(x)) = n, then write g(x) = cxn + ĝ(x) where deg(ĝ(x)) < n. Modulo f (x), we have
f (x) = xn + bn−1 xn−1 + . . . + b1 x + b0 ≡ 0 (mod f (x))

Page 3
Zachary Scherr Math 503 HW 2 Due Feb 5

implying that
xn ≡ −(bn−1 xn−1 + . . . + b1 x + b0 ) (mod f (x)).
Thus
g(x) = cn xn + ĝ(x) ≡ −c(bn−1 xn−1 + . . . + b1 x + b0 ) + ĝ(x) (mod f (x))
n−1
and so if we let p(x) = −c(bn−1 x + . . . + b1 x + b0 ) + ĝ(x) then either p(x) = 0 or deg(p(x)) is
less than n. For the inductive step, assume we know the statement to be true for all polynomials
of degree at most k ≥ n, and deg(g(x)) = k+1. Write g(x) = cxk+1 +ĝ(x) where deg(ĝ(x)) ≤ k.
By the induction hypothesis, there exist polynomials p(x), q(x) of degree at most n − 1 with

ĝ(x) ≡ p(x) (mod f (x))

k
x ≡ q(x) (mod f (x)).

Thus
f (x) ≡ xq(x) + p(x) (mod f (x))
and since deg(xq(x)) ≤ n, we can again appeal to induction to say that xq(x) ≡ p̂(x) (mod f (x))
for some p̂(x) of degree at most n − 1. Thus

g(x) ≡ p̂(x) + p(x) (mod f (x))

and the induction step is complete.

(b) Let p(x), q(x) be distinct polynomials in R[x] of degree at most n − 1. Assume for the sake
of contradiction that p(x) ≡ q(x) (mod f (x)). Then p(x) − q(x) ∈ (f (x)) implying that f (x)
divides p(x) − q(x) in R[x]. Let g(x) ∈ R[x] be such that f (x)g(x) = p(x) − q(x). Write
g(x) = bm xm + . . . + b1 x + b0 . Since f (x) is monic, we see that the leading term of f (x)g(x) is
bm xn+m , proving that deg(f (x)g(x)) = deg(f (x)) + deg(g(x)).
Therefore

deg(p(x) − q(x)) = deg(f (x)g(x)) = deg(f (x)) + deg(g(x)) ≥ deg(f (x)) = n

which is a contradiction since deg(p(x) − q(x)) < n.

(c) If both a(x) and b(x) are non-zero have degree less than n then a(x), b(x) 6≡ 0 (mod f (x)) by
part b). However,
a(x)b(x) = f (x) ≡ 0 (mod f (x))
showing that a(x) and b(x) are zero-divisors in R[x]/(f (x)).
(d) Let f (x) = xn −a where a ∈ R is nilpotent. Then there exists a positive integer m with am = 0.
As a consequence of xn ≡ a (mod xn − a) we have

xnm = (xn )m ≡ am ≡ 0 (mod xn − a)

provingthat x is a nilpotent in R[x]/(xn − a).

(e) Let p be a prime and let R = Fp . Since the characteristic of R[x] equals p, we know that
(α + β)p = αp + β p for any α, β ∈ R[x]. In particular,

(x − a)p = xp − ap

for any a ∈ Fp . We know that ap = a for any a ∈ Fp so in fact

(x − a)p = xp − a ≡ 0 (mod xp − a)

proving that x − a is nilpotent in R[x]/(xp − a).

Page 4
Zachary Scherr Math 503 HW 2 Due Feb 5

6. 7.4.15

Solution:
(a) By problem 14, every element in E = F2 [x] is equivalent modulo (x2 + x + 1) to a polynomial of
degree at most 1. The only such polynomials are 0, 1, x, x+1 showing that E = F2 [x]/(x2 +x+1)
consists of the four elements {0, 1, x, x + 1}.
(b) We have
+ 0 1 x x+1
0 0 1 x x+1
1 1 0 x+1 x
x x x+1 0 1
x+1 x+1 x 1 0
Since E is an abelian group of order 4 under addition, in which every element is its own inverse,
we have the E is isomorphic to the Klein-4 group.
(c) We have
× 0 1 x x+1
0 0 0 0 0
1 0 1 x x+1
x 0 x x+1 1
x+1 0 x+1 1 x
×
Since all non-zero elements of E are invertible we see that E is a field. Moreover, E is cyclic
and is generated by x.

7. 7.4.16

Solution:
(a) Modulo f (x) = x4 − 16 we have x4 ≡ 16 (mod f (x)). Thus if k ≥ 0 and 0 ≤ r < 4 we get

x4k+r = (x4 )k xr ≡ (16)k xr (mod f (x)).

Therefore

7x13 − 11x9 + 5x5 − 2x3 + 3 ≡ 7(16)3 x − 11(16)2 x + 5(16)x − 2x3 + 3 (mod f (x))

which simplifies to
−2x3 + 25936x + 3 (mod x4 − 16).

(b) We have
(x4 − 16) = (x2 − 4)(x2 + 4) = (x − 2)(x + 2)(x2 + 4).
Therefore
0 ≡ (x − 2)(x + 2)(x2 + 4) (mod x4 − 16).
Since the terms in the product are all non-zero, we see that x − 2 and x + 2 are zero-divisors.

8. 7.4.37

Page 5
Zachary Scherr Math 503 HW 2 Due Feb 5

Solution: Let R be a local ring. We would like to prove that if x ∈ R − M then x ∈ R× . Assume
for the sake of contradiction that x 6∈ R× . Then (x) ⊆ R is a proper ideal of R. But then (x) must
be contained in some maximal ideal of R, and since M is the only maximal ideal of R, we see that
(x) ⊆ M or x ∈ M which is a contradiction.
Conversely, let R be a commutative ring with 1 such that R − R× is an ideal of R. If I is any proper
ideal of R, then I ∩ R× = ∅. Thus I ⊆ R − R× , proving that R − R× is a maximal ideal of R.
Moreover, since all proper ideals are contained in R − R× it must be the unique maximal ideal and
hence R is a local ring.

9. 7.4.39

Solution:
(a) For k ≥ 0 define Ak = {r ∈ R | ν(r) ≥ k} ∪ {0}. Since ν : K × → Z is surjective, there exists
π ∈ K with ν(π) = 1. Then
ν(π k ) = k
for all k ≥ 0, showing in fact that π k ∈ Ak for k ≥ 0. I claim in fact that Ak = (π k ). To see
this, let α ∈ Ak . Then ν(α) ≥ k and so

ν(απ −k ) = ν(α) + ν(π −k ) = ν(α) − k ≥ 0.

This shows that απ −k = r ∈ R and so α = rπ k ∈ (π k ). Conversely, we’ve seen that π k ∈ Ak

and if s ∈ S then ν(s) ≥ 0 showing that

ν(sπ k ) = ν(s) + ν(π k ) = ν(s) + k ≥ k.

Thus sπ k ∈ Ak for any s ∈ R and hence (π k ) ⊆ Ak .

Therefore each Ak = (π k ) is a principal ideal. Moreover, π k+1 = π · π k ∈ (π k ) proves that

Ak+1 = (π k+1 ) ⊆ (π k ) = Ak

as desired.

(b) Let I be a non-zero ideal of R. Choose α ∈ I so that ν(α) = k is minimal, we will prove that
I = Ak . To do so, first note that since every element of I has valuation at least k, we get
I ⊆ Ak by definition. Now, since ν(π k ) = k, we also have

ν(α−1 π k ) = ν(α−1 ) + ν(π k ) = −k + k = 0

implying that α−1 π k = r ∈ R. Thus π k = rα showing that Ak = (π k ) ⊆ (α) ⊆ I.

Thus we’ve just shown that every ideal in R is of the form Ak , and since A1 is the unique
proper ideal of R not contained in any other proper ideal, we see that R is a local ring with
maximal ideal A1 .

10. Find all ideals of Z[x] containing the ideal (4, x2 ).

Solution: By the fourth isomorphism theorem, it suffices to find all ideals in the ring Z[x]/(4, x2 ).
By the second isomorphism theorem, this ring is isomorphic to R = (Z/4Z)[x]/(x2 ), and by problem
14 above we see that this ring contains the 16 elements of the form ax + b where a, b ∈ Z/4Z.

Page 6
Zachary Scherr Math 503 HW 2 Due Feb 5

Since this ring is finite, this already reveals to us that there should be finitely many ideals of Z[x]
containing (4, x2 ). Let’s start by finding the units in R. Given ax + b, cx + d ∈ R we have
(ax + b)(cx + d) = acx2 + (ad + bc)x + bd = (ad + bc)x + bd.
This calculation shows that ax + b ∈ R× if and only if b ∈ (Z/4Z)× because if bd = 1 in Z/4Z then
(ax + b)(−ad2 x + d) = (ad − abd2 )x + bd = (ad − ad)x + bd = 1.
Since (Z/4Z)× = {1, 3} we see that R has 8 units. Thus proper ideals of R cannot contain these 8
elements.
Among the other 8 elements of the form ax and ax + 2 for a ∈ Z/4Z, let’s try to figure out what
principal ideals they generate. We should use to our advantage the fact that if α, β ∈ R and α = uβ
for some u ∈ R× then (α) = (β). The relation α ∼ β ∈ R if and only if α = uβ for some u ∈ R× is
easily verified to be an equivalence relation with equivalence classes
[0] = {0}
[1] = {1, 3, x + 1, x + 3, 2x + 1, 2x + 3, 3x + 1, 3x + 3} = R×
[2] = {2, 2x + 2}
[x] = {x, 3x}
[2x] = {2x}
[x + 2] = {x + 2, 3x + 2}.
Therefore, the non-zero proper principal ideals of R are
(2) = {0, 2, 2x, 2x + 2}
(x) = {0, x, 2x, 3x}
(2x) = {0, 2x}
(x + 2) = {0, x + 2, 2x, 3x + 2}.
The ideal (2x) is contained in the other non-zero principal ideals. Any non-principal ideal of R must
then contain at least two of (2), (x), (x + 2), but it’s easy to verify that
(2, x) = (2) + (x) = (2) + (x + 2) = (x) + (x + 2).
Thus any non-principal ideal of R contains (2, x), but the ideal (2, x) is maximal since
R/(2, x) ∼
= Z/2Z
is a field. Therefore the only ideals of R are
(0), (2x), (2), (x), (x + 2), (2, x), (1).
The ideals of R all correspond to ideals in Z[x] containing (4, x2 ). Therefore, the ideals of Z[x]
containing (4, x2 ) are
(0, 4, x2 ) = (4, x2 )
2
(2x, 4, x ) = (2x, 4, x2 )
(2, 4, x2 ) = (2, x2 )
(x, 4, x2 ) = (4, x)
2
(x + 2, 4, x ) = (x + 2, 4, x2 )
2
(2, x, 4, x ) = (2, x)
2
(1, 4, x ) = (1).

Page 7
Zachary Scherr Math 503 HW 2 Due Feb 5

3 Challenge Problems
Challenge Problems tend to be harder than the rest of the problems (and sometimes more interesting). You
do not need to turn these in, but you should get something out of thinking about these.
1. Let F be a finite field. Show that F has characteristic p for some prime p, and deduce that |F| = pn for
some positive integer n. Prove that Fp := Z/pZ is isomorphic to a subfield of F.

Solution: Let 1F denote the multiplicative identity of F. Then we know that there exists a map
ϕ : Z → F given by sending 1 7→ 1F and extending. The map ϕ is clearly non-zero, and since F is
finite, this map cannot be injective so its kernel equals ker(ϕ) = nZ for some positive integer n ≥ 2.
By the first isomorphism theorem, this says that Z/nZ is isomorphic to a subring of F. Of course
all subrings of F are integral domains, and so necessarily n = p is prime. Therefore Fp = Z/pZ is
isomorphic to a subgroup of F.
This also tells us that the characteristic of F is p. Then for all x ∈ F, we have

x + x + . . . + x = (1F + 1F + . . . + 1F ) x = ϕ(p)x = 0.
| {z } | {z }
p times p times

As a group under addition, this shows us that the additive order of every element in F is 1 or p since
p is prime. Of course 0 is the only element of order 1, and so in fact every non-zero element of F has
additive order p. By Cauchy’s theorem, we conclude that |F| = pn for some positive integer n ≥ 1.

Page 8