Math 504, Fall 2013

HW 1

1. Let R be the ring of continuous functions [0, 1] → R with point-wise addition and
multiplication. Prove that the set of functions vanishing at a point x ∈ [0, 1] is a maximal
ideal in R, we denote it by m x . If m is a maximal ideal of R that is not equal to m x for
any x ∈ [0, 1], show that there are a finite set of elements f 1 , . . . f n in m that have no
common zero on [0, 1]; by considering f 02 + · · · + f n2 , show that there in no such m; i.e,
the maximal ideal in R are the ideal m x , x ∈ [0, 1].

Let ex be the ring homomorphism R → R, f 7→ f ( x ). It’s surjective since R contains all

the constant functions, and by definition ker ex = m x . Since R is a field, it follows that m x
is a maximal ideal of R.
Suppose now M is a maximal ideal different from all the m x ’s. In particular, M * m x
for any x, i.e. for any x there is f x ∈ M such that f x ( x ) 6= 0. By continuity, we can
assume f x 6= 0 in an open neighborhood Ux of x. Since the Ux ’s form an open cover of
the compact set [0, 1], there is a finite subcover U1 , . . . , Un , corresponding to the elements
f 1 , . . . , f n . Let g := f 12 + . . . + f n2 ∈ M. Since the f i ’s have no common zeroes, g is never
zero, hence invertible in R. This implies that M = R, hence every maximal ideal is of the
form m x for some x. 

2. Factor x8 − 1 and x12 − 1 in Q[ x ].

x8 − 1 = ( x − 1)( x + 1)( x2 + 1)( x4 + 1)

x12 − 1 = ( x − 1)( x + 1)( x2 + x + 1)( x2 + 1)( x2 − x + 1)( x4 − x2 + 1)
That equality holds above over Q is just computation, and that x − 1 and x + 1 are irre-
ducible over Q[ x ] is immediate. What remains is to show the irreducibility of the quadrat-
ics and quartics. We note that we are implicitly using the fact quadratics in Q[ x ] are re-
ducible if and only if they have linear terms, a fact which follows immediately by the
division algorithm. For quartics, they could have a linear factor or be the product of ir-
reducible quadratics. By the rational roots test, the only possible rational roots for all
polynomials are ±1. A computation yields that neither 1 nor −1 are zeros of any of the
quadratics, so they are all irreducible in Q[ x ]. If the quartic x4 + 1 is the product of irre-
ducible quadratics, then x4 + 1 = ( x2 + bx + c)( x2 + dx + e) (we can take both leading
coefficients to be 1), which furnishes the system of equations
d+b =0
e + c + db =0
dc + be =0
ec = 1,

1
and d = −b implies the third equation becomes −b(c + e) = 0. If c = −e, then the fourth
equation cannot be solved since −c2 = 1 has no solutions in Q. If b = 0, then d = 0,
then the second equation gives e = −c again, which is impossible. Thus x4 + 1 is not the
product of quadratics and is hence irreducible over Q[ x ]. We now seek a representation
of x4 − x2 + 1 = ( x2 + bx + c)( x2 + dx + e), which furnishes a similar system of equations
d+b =0
e + c + db = −1
dc + be =0
ec = 1.
If d = −b, the third equation again becomes −b(c + e) = 0. If c = −e, then the fourth
equation cannot be solved since −c2 = 1 has no solutions in Q. If b = 0, then d = 0 and
the second equation reads e + c = −1. If e = c + 1, then c(c + 1) = 1 has no solutions
over Q (quadratic formula), so we conclude this quartic is irreducible. 

3. If d and e are greatest common divisors of { a1 , . . . an } in a domain R, show that d and

e are associates, i.e. unit multiples of one another.

Since they are both greatest common divisors, d | e and e | d. Therefore, e = xd and d = ye
for some x, y ∈ R. Therefore, e = xd = x (ye) = ( xy)e and it follows that 1 = xy since R is
a domain, hence e and d are associates. 

4. Let k [ x, y] be the polynomial ring on two variables with coefficients in the field k.
Show that the ideal J = k [ x, y]≥n = span{ xi y j | i + j ≥ n} can be generated by n + 1
elements, but not by n elements. (Hint: Think of degree).

First of all, it’s clear that J can be generated as an ideal by the n + 1 monomials x n , x n−1 y, . . . , yn .
We’ll show that it can’t be generated by less than n + 1 elements.
Let G be any finite generating set for J, and let G0 be the set consisting of the degree
n part of the polynomials in G. We claim that J = ( G0 ). Since one containment is clear, it
will suffice to show that xi yn−i ∈ ( G0 ) for all i ≤ n.
Indeed, we know that
x i y n −i = ∑ p j (x, y) gj (x, y) with g j ∈ G
If we write p j ( x, y) = p j (0, 0) + p0j and g j = g̃ j + g0j where g̃ j is the degree n part of g j .
Then the only degree n term in the product p j ( x, y) g j ( x, y) is p j (0, 0) g̃ j , and since in the
above sum the terms of degree > n cancel, we have

x i y n −i = ∑ p j (0, 0) g̃j
This shows that the monomials xi yn−i are in the k span of G0 .
Assume now has at most n elements: then we would have that the k span of G0 ,
contains an n + 1-dimensional subspace, impossible.


2
 √
5. Show that the ring of Gaussian Integers, Z[i ] = Z[ −1], is a Euclidean domain with
respect to the functions δ : Z [i ] → Z defined by δ( x ) := x x̄, where x̄ denote the complex
conjugate of x.
Let f , g ∈ Z[i ], with g 6= 0: we have to define a way to divide f by √ g. We know that
f /g ∈ C. Since in C a point can never be further than a distance of √ 2/2 from a lattice
point, then there must be √ q ∈ Z[i ] at a distance less than or equal to 2/2 from f /g. Thus
f /g = q + r0 with |r0 | ≤ 2/2. This implies that f = qg + r0 g with r0 g = f − qg ∈ Z[i ].
Call r0 g = r. Then δ(r ) = δ( gr0 ) = gr0 gr0 = ggr0 r0 = δ( g)(|r0 |)2 ≤ δ( g)/2 < δ( g). So
f = qg + r with δ(r ) < δ( g) that is what we wanted to prove.

6. Factor 2, 3 and 5 in Z[i ] as products of primes.

We claim that 2 = (1 + i )(1 − i ), 3 = 3, and 5 = (2 + i )(2 − i ) are prime factorizations in

Z[ i ].
First, suppose 1 + i |( a + bi )(c + di ). But

( a + bi )(c + di ) = ( a − b)(c − d) + (1 + i )(bc + da + (i − 1)bd)

so (1 + i )|( a − b)(c − d). This means (1 + i )(e + i f ) = ( a − b)(c − d). Comparing i coeffi-
cients, we see e + f = 0, so in fact e(1 + i )(1 − i ) = 2e = ( a − b)(c − d). Suppose without
loss of generality that 2| a − b. Then (1 + i )|2| a − b. But a + bi = a − b + b(1 + i ) so in fact
1 + i | a + bi. Hence 1 + i is prime. By symmetry, 1 − i is prime as well.
To show 3 is prime in Z[i ], suppose 3|( a + bi )(c + di ). Then 3|( a2 + b2 )(c2 + d2 ) after
multiplying by conjugates. So without loss of generality, 3| a2 + b2 . In F3 , the only solution
to a2 + b2 = 0 is a = b = 0. So 3| a, b hence 3| a + bi. So 3 is prime.
Lastly, 2 + i is prime for the same reason 1 + i is prime, but we repeat the proof for
completeness. Suppose 2 + i |( a + bi )(c + di ). But

( a + bi )(c + di ) = ( a − 2b)(c − 2d) + (2 + i )(bc + da + (i − 2)bd))

so (2 + i )|( a − 2b)(c − 2d). This means (2 + i )(e + i f ) = ( a − 2b)(c − 2d). Comparing i

coefficients, we see e + 2 f = 0, so in fact e(2 + i )(2 − i ) = 5e = ( a − 2b)(c − 2d). Suppose
without loss of generality that 5| a − 2b. Then (2 + i )|5| a − 2b. But a + bi = a − 2b + b(2 + i )
so in fact 2 + i | a + bi. Hence 2 + i is prime. By symmetry, 2 − i is prime as well.
In conclusion, 2 = (1 + i )(1 − i ), 3 = 3, and 5 = (2 + i )(2 − i ) are the corresponding
prime factorizations.
One can also also argue that Euclidean domains are UFDs, so prime is equivalent to
irreducible, and use the norm of problem 9 to show that 1 + i, 3 and 2 + i are irreducible.


7. Prove that a Euclidean domain is a PID.

3
Let R be a Euclidean domain with respect to the function δ : R → Z. Let I ⊂ R be an
ideal that is not 0. Choose s ∈ I such that s 6= 0 and δ(s) = min{δ(r ) : r ∈ I }; such an
element is guaranteed to exist because δ(r ) ≥ 0 for all r ∈ R.
Choose any other r ∈ I. By definition, there exists q1 , q2 ∈ R such that r = q1 s + q2
with q2 = 0 or δ(q2 ) < δ(s). Since I is an ideal, q1 s ∈ I and r − q1 s = q2 ∈ I. We choose
s to be of minimal norm among elements in I, so it must be that q2 = 0. Then for all
r ∈ I, there exists q ∈ R such that r = sq. That is, I ⊂ (s). It’s already true that (s) ⊂ I,
so (s) = I and I is a principal ideal. R and I were arbitrary, so this shows that every
Euclidean domain is a PID. 

8. Let A = k [ x, x −1 ] be the subring of k ( x ) generated by x, x −1 and k. Is k [ x, x −1 ] a PID?

Why?

Let I be an ideal of A, and let J := I ∩ k [ x ]. Then J is an ideal of k[ x ], hence it’s principal,

say J = ( p). We claim that I = ( p) in A, thus showing that every ideal in A is principal.
Clearly, ( p) ⊆ I. Conversely, let f ∈ I. Then we can write f = x n f 0 ( x ) where f 0 ( x ) ∈
k [ x ] and n ∈ Z. Since f 0 = x −n f is also in I, then f 0 ∈ J so f 0 is a multiple of p in k[ x ]. It
follows that f is a multiple of p, so I = ( p) as claimed. 

√
9. Let d be a square-free positive integer. Define the norm function N : Z[ −d] → Z
given √
N ( a + b − d ) = a2 + b2 d2

1. Establish some important properties of N.

√
2. Show that u is a unit in Z[ −d] if and only if N (u) = 1.

3. Show that the only units in Z[i ] are ±1 and ±i.

√
4. If d > 1, show that the only units in Z[ −d] are ±1.

1. The fundamental property of the norm is that N ( a) N (b) = N ( ab), as a simple cal-
culation shows. Also, it’s clear from the definition that N has values in N.
√
2. Suppose that u is a unit in Z[ −d] and let u be its inverse. By part (1) be know that
N is multiplicative, so N (u) N (u−1 ) = N (1) = 1. As each of N (u) and N (u−1 ) are
in N, both must be 1.
√
Conversely, if N (u) = uu = 1, then since u ∈ Z[ −d] we have that u is a unit.

3. It’s easy to see that the only elements whose norm is 1 are ±1, ±i, and by part (2)
they are the only units.

4. As above, if d > 1 the only elements with norm one are ±1.


4
10. Find an element in R = C[ x, y, z]/( xy − z2 ) that is irreducible but not prime.

Since in R we have xy = z2 , then x |z2 . We’ll show that x does not divide z, thus implying
that x is not prime.
Suppose z = xp in R for some p. Then this means that

z = xp + q( xy − z2 ) for some q

where this is an equality in C[ x, y, z].

Now write p = ∑ pi and q = ∑ qi as the sum of their homogeneous components.
Every term of q( xy − z2 ) has degree at least 2, and they have to cancel with the terms of
x ( p1 + p2 + . . .). It follows that z = xp0 , absurd.
We now claim that x is irreducible. First of all, observe that the automorphism z 7→ −z
of C[ x, y, z] descends to an automorphism φ of R. Define N : R → R as N ( p) = pφ( p),
much like the norm in problem 9. For any other element in R, note that it can be written
uniquely as p( x, y) + zq( x, y), thus N ( p( x, y) + zq( x, y)) = p2 − xyq2 . We can then regard
N as having values in C[ x, y]. As in problem 9, the units are characterized by the fact that
their norm is invertible in C[ x, y], and one can check directly that N ( ab) = N ( a) N (b).
We have that N ( x ) = x2 and if x = αβ were not irreducible then N ( x ) = N (α) N ( β).
If we can prove we can’t have N (α) = x, then this would force N (α) = x2 so that N ( β)
would be invertible, hence β would be a unit in R.
If N (α) = x, then
x = p2 − xyq2
for some polynomials p, q. Since x | p2 , then x | p hence we can divide by x to get

1 = x ( p0 )2 − yq2

Evaluating at x = y = 0 yields a contradiction.

This means that x is irreducible but not prime.