Math 562 Spring 2012

Homework 1 Drew Armstrong

1. Let R be a ring. We say that a ∈ R is nilpotent if an = 0 for some n. If a is nilpotent, prove that
1 + a and 1 − a are units (i.e. invertible).
Proof. Recall that in any ring we have (−a)(−b) = −(ab) (see HW 3.7 from MTH 561). Thus in
any ring with 1 (commutative or not) we have the following identities:
1 − an = (1 − a)(1 + a + a2 + · · · + an−1 ),
1 − (−1)n an = (1 + a)(1 − a + a2 − · · · + (−1)−1 an−1 ).
If an = 0 then we obtain inverses for 1 + a and 1 − a. 

2. Let I ⊆ R be an ideal. Prove that I = R if and only if I contains a unit.

Proof. First suppose that I = R then 1 ∈ I so I contains a unit. Conversely, suppose that I contains
a unit u, say uv = 1 for u, v ∈ R. But since I is an ideal we have uv = 1 ∈ I. Then for any a ∈ R
we have a = 1a ∈ I. Hence I = R. 

3. Let ϕ : R → S be a ring homomorphism.

(a) Prove that ϕ(0R ) = 0S .
(b) Prove that ϕ(−a) = −ϕ(a) for all a ∈ R.
(c) Let a ∈ R. If a−1 ∈ R exists, prove that ϕ(a) is invertible with ϕ(a)−1 = ϕ(a−1 ).
Proof. To prove (a) note that ϕ(0R ) = ϕ(0R +0R ) = ϕ(0R )+ϕ(0R ). Then subtract ϕ(0R ) from both
sides to get 0S = ϕ(0R ). To prove (b) consider a ∈ R. Then use part (a) to write 0S = ϕ(0R ) = ϕ(a−
a) = ϕ(a)+ϕ(−a). Now subtract ϕ(a) from both sides to get ϕ(−a) = −ϕ(a). To prove (c) consider
a ∈ R and suppose that there exists a−1 with aa−1 = a−1 a = 1R . Applying ϕ to the three parts of
this equation and using the fact that ϕ is a homomorphism gives ϕ(a)ϕ(a−1 ) = ϕ(a−1 )ϕ(a) = 1S .
We conclude that ϕ(a−1 ) = ϕ(a)−1 . 

[Note that the property ϕ(ab) = ϕ(a)ϕ(b) does not imply ϕ(1R ) = 1S for rings, so we just assume
ϕ(1R ) = 1S (because we want it).]

4. Let I ⊆ R be an ideal and consider a, b, c, d ∈ R with a + I = c + I and b + I = d + I. Prove

that (a + b) + I = (c + d) + I and ab + I = cd + I. This shows that addition and multiplication of
cosets is well-defined.
Proof. Since a + I = c + I and b + I = d + I there exist x, y ∈ I with a − c = x and b − d = y. To
prove that (a + b) + I = (b + d) + I, first consider an arbitrary element a + b + u ∈ (a + b) + I with
u ∈ I. Then we have a + b + u = (c + x) + (d + y) + u = (c + d) + (x + y + u) ∈ (c + d) + I. Hence
(a + b) + I ⊆ (c + d) + I. Similarly we find (c + d) + I ⊆ (a + b) + I and hence (a + b) + I = (c + d) + I.
To prove that ab + I = cd + I, first consider an arbitrary element ab + u ∈ ab + I with u ∈ I. Then
we have ab + u = (c + x)(d + y) + u = cd + (cy + xd + xy + u). Since cy, xd, xy, u are all in I we
conclude that ab + u = cd + (cy + xd + xy + u) ∈ cd + I, hence ab + I ⊆ cd + I. The proof of
cd + I ⊆ ab + I is similar. We conclude that ab + I = cd + I. 

[Note that (a + b) + I = (c + d) + I only requires that I is closed under addition. The proof that
ab + I = cd + I really requires that I is an ideal. In other words, if S ⊆ R is an additive subgroup we
can always define R/S as an additive group, but we can only define multiplication on R/S when S is an
ideal.]
5. When does ab = 1 imply ba = 1? Consider a, b ∈ R where R is a finite ring, and suppose that
ab = 1. Show that b + (1 − ba)ai is a right inverse of a for all i ≥ 0. Use this and the finiteness of R
to show that ba = 1. [Recall: We have also seen that AB = I implies BA = I for square matrices
over a field. Now we have two results of this sort...]
Proof. Suppose that ab = 1 and note that for all i ≥ 0 we have
a[b + (1 − ba)ai ] = ab + (a − aba)ai = 1 + ai+1 − abai+1 = 1 + ai+1 − ai+1 = 1.
Hence b + (1 − ba)ai is a right inverse of a for all i ≥ 0. Since our ring is finite there must exist i < j
such that b + (1 − ba)ai = b + (1 − ba)aj . Multiply both sides on the right by bj and use the fact
that ab = 1 to get b + (1 − ba)bj−i = b + (1 − ba). Now subtract b from both sides and use the fact
that (1 − ba)b = b − bab = b − b = 0 to find 0 = 1 − ba. We conclude that ba = 1 as desired. 

6. Recall that a group G is simple if for any group homomorphism ϕ : G → H we have ker ϕ = G
(the whole group) or ker ϕ = 1 (the trivial group). We can define a simple ring similarly in terms
of ring homomorphisms. Prove that a ring is simple if and only if it is a field. (Hence the term
“simple ring” is unnecessary.) [Hint: Look in the book.]
Proof. Recall that I ⊆ R is an ideal if an only if I is the kernel of a ring homomorphism. Thus we
can say that a ring R is simple if it has only two ideals: (1) = R and (0) = {0}.
First suppose that R is a field and let I ⊆ R be an ideal. If I 6= (0) then I contains a nonzero
element a. But since R is a field, a is a unit, and we conclude by Problem 2 that I = (1) = R.
Hence R is a simple ring.
Conversely, suppose that R is a simple ring and let a ∈ R be a nonzero element (if R = (0) then R
is not really a field, but I forgot to worry about this silly case when I wrote the question). Since (a)
is an ideal and (a) 6= (0) we must have (a) = (1). That is, a is a multiple of 1, which means that a is
a unit. Since this is true for all nonzero a ∈ R, R is a field (or, I guess, a division ring — I also forgot
to say that R is commutative (oh well); in any case, the term “simple ring” is unnecessary). 

7. Prove Descartes’ Factor Theorem. Let F be a field and consider the ring F[x] of polynomials.
Given f (x) ∈ F[x] and α ∈ F such that f (α) = 0, prove that f (x) = (x − α)h(x) where h(x) ∈ R[x]
with deg(h) = deg(f )−1. [Hint: Observe that xn −αn = (x−α)(xn−1 +αxn−2 +· · ·+αn−2 x+αn−1 )
for all n ≥ 0. Consider the polynomial f (x) − f (α).]
Proof. To save space, we define the polynomial [n]x,α := (xn−1 + xn−2 α + · · · + xαn−2 + αn−1 ) for
each positive integer n and real number α. Suppose that f (x) ∈ R[x] has degree d and write
f (x) = ad xd + ad−1 xd−1 + · · · a1 x + a0
for a0 , . . . , ad ∈ R with ad 6= 0. Then applying the identity xn − αn = (x − α)[n]x,α we can write
f (x) − f (α) = ad (xd − αd ) + ad−1 (xd−1 − αd−1 ) + · · · + a1 (x − α)
= ad (x − α)[d]x,α + ad−1 (x − α)[d − 1]x,α + · · · + a1 (x − α)[1]x,α
= (x − α)(ad [d]x,α + ad−1 [d − 1]x,α + · · · + a1 [1]x,α )
= (x − α)(ad xd−1 + lower order terms ).
If f (α) = 0 then we obtain f (x) = (x − α)h(x) where h(x) ∈ R[x] has degree d − 1. 

8. Let R and C be the real and complex fields. Let ϕ : R[x] → C be the map that sends a polynomial
f (x) to its evaluation f (i) ∈ C at x = i.
(a) Prove that ϕ is a surjective ring homomorphism.
(b) Recall the definition of complex conjugation: a + ib := a − ib for a, b ∈ R. Prove that
f (−i) = f (i) ∈ C for all f (x) ∈ R[x].
 (c) Use Descartes’ Factor Theorem to prove that the kernel of ϕ is the principal ideal generated
by x2 + 1:
ker ϕ = (x2 + 1) := {(x2 + 1)g(x) : g(x) ∈ R[x]}.
Proof. The multiplicative identity of R[x] is the constant polynomial 1(x) = 1, so clearly ϕ(1) =
1(i) = 1 ∈ C, which is the multiplicative identity of C. To prove (a) we must show that Pϕ(f +k g) =
ϕ(f ) + ϕ(g) and ϕ(f g) = ϕ(f )ϕ(g) for all f, g ∈ R[x]. To this end, let f (x) = k ak x and
k
P
g(x) = k bk x . Then we have
X X X
ϕ(f ) + ϕ(g) = f (i) + g(i) = ak ik + bk ik = (ak + bk )ik = (f + g)(i) = ϕ(f + g)
k k k
and also
! !
X X X X
ϕ(f )ϕ(g) = f (i)g(i) = (au iu )(bv iv ) = au bv ik = (f g)(i) = ϕ(f g).
k u+v=k k u+v=k
Notice that the proof of ϕ(f )ϕ(g) = ϕ(f g) uses the fact that C is commutative. (For this
reason we will only consider polynomials over commutative rings.) Finally, note that the map is
surjective since for any a + ib ∈ C we have a + ib = ϕ(f ) with f (x) = a + xb ∈ R[x].
Given complex numbers a + ib and c + id note that
a + ib + c + id = (a − ib) + (c − id) = (a + c) − i(b + d)
= (a + c) + i(b + d) = (a + ib) + (c + id)
and
(a + ib)(c + id) = (a − ib)(c − id) = (ac − bd) − i(ad + bc)
= (ac − bd) + i(ad + bc) = (a + ib)(c + id).
Combined with the fact that 1 = 1 we conclude that complex conjugation z → z is a ring isomor-
phism C → C (we call it a field Pautomorphism). Furthermore, we have z = z for all z ∈ R ⊆ C. Now
we will prove (b). Let f (x) = k ak xk and consider any complex number z ∈ C. Then using the
homomorphism properties of conjugation we have
X X X
f (z) = ak z k = ak (z)k = ak (z)k = f (z).
k k k

In particular, taking z = i gives f (−i) = f (i).

Finally consider the surjective homomorphism ϕ : R[x] → C given by ϕ(f ) = f (i). To prove (c)
we will show that ker ϕ = (x2 + 1). Indeed, if f (x) ∈ (x2 + 1) then we can write f (x) = (x2 + 1)g(x)
and then ϕ(f ) = (i2 + 1)g(i) = 0 · g(x) = 0, hence f ∈ ker ϕ and (x2 + 1) ⊆ ker ϕ. Conversely,
suppose that f ∈ ker ϕ; i.e. f (i) = 0. By Descartes’ Factor Theorem applied to f (x) ∈ C[x] (a
slightly tricky point) we have f (x) = (x − i)g(x) for some g(x) ∈ C[x]. But by part (b) we know
that f (i) = 0 implies f (−i) = 0 hence f (−i) = −2i · g(−i) = 0, which implies that g(−i) = 0.
Then Descartes’ Factor Theorem implies that g(x) = (x + i)h(x) for some h(x) ∈ C[x]. Putting this
together we get
f (x) = (x − i)(x + i)h(x) = (x2 + 1)h(x)
for some h(x) ∈ C[x]. The only problem left is to show that h(x) ∈ R[x]. But since f (x) and
(x2 + 1) are in R[x] we must also have h(x) ∈ R[x] (for example, we could do long division to
compute f (x)/(x2 + 1) = h(x)). We conclude that h(x) ∈ R[x] and hence f (x) is in the ideal
(x2 + 1) as desired.