Math 122, Solution Set No. 1: 1 Chapter 1.1 Problem 16

Math 122, Solution Set No.
1
1 Chapter 1.1 Problem 16
Let A be nilpotent, i.e. Ak = 0 for some k > 0. Then I claim
I − A + A2 − A3 + · · · Ak−1 = (I + A)−1 .
This is because right-multiplying by this matrix gives:
(I + A)(I − A + A2 − A3 + · · · Ak−1 ) = I + A − A + · · · + Ak−1 − Ak−1 + Ak
= I + Ak
= I
because Ak = 0. similarly, left-multiplication yields:
(I − A + A2 − A3 + · · · Ak−1 )(I + A) = I
therefore I − A + A2 − A3 + · · · Ak−1 = (I + A)−1 , i.e. (I + A) is invertible ♦
Note: I deducted points if you did not show the inverse was two-sided.
2 Chapter 2.1 Problem 1(b)

The multiplication table for S3
× e (12) (23) (13) (123) (132)
e e (12) (23) (13) (123) (132)
(12) (12) e (123) (132) (23) (13)
(23) (23) (132) e (123) (13) (12)
(13) (13) (123) (132) e (12) (23)
(123) (123) (13) (12) (23) (132) e
(132) (132) (23) (13) (12) e (123)
All multiplication tables are read (as in Artin) with the entry in row u and column v being
the product u ◦ v.♦

0 1
This problem asks for an example of 2 × 2 matrices such that A−1 B 6= B −1 A Let A =
1 0

−1 0
and B = . Then A2 = B 2 = I, i.e. A = A−1 and B = B −1 . So A−1 B = AB =
0 1

0 1 0 −1
but B −1 A = BA = 6= A−1 B.
−1 0 1 0

Let a, b be elements of a group G, such that a has order 5 and a3 b = ba3 . Left-multiplying
by a3 gives a3 (a3 b) = a3 ba3 . But a3 a3 = a6 = a, because a5 = 1. So ab = a3 ba3 , and
recalling the original condition gives ab = ba3 a3 = ba6 = ba♦
1
(a) Yes. The identity is real, and since GLn (R) is a group, it is closed and contains inverses.
(b) Yes. This was covered in class on 9/18.
(c) No. 1 is a positive integer, but −1, its inverse, is not.
(d) Yes. The identity 1 is positive, and the product of positive reals is positive implies
that this set is closed. The multiplicative inverse
of
a positive number is likewise positive.
a 0
(e) No. The set of all matrices of the form: does not include the identity element
0 0
of GLn (R) ! ♦

Assume that ∀x, y ∈ H, xy −1 ∈ H. If H 6= ∅, then ∃x ∈ H. So by assumption, e = xx−1 ∈
H. Then e, x ∈ H ⇒ x−1 = ex−1 ∈ H. Finally, if x, y ∈ H, we now know that y −1 ∈ H
which means xy = x(y −1 )−1 ∈ H, i.e. H is closed. This proves H is a subgroup. ♦

Lemma: Every subgroup of a cyclic group is cylic.
Proof: Let G = hxi and H ≤ G. Let n be the smallest postive integer such that xn ∈ H.
Since H is closed, hxn i ≤ H. Assume ∃xm ∈ H such that xm ∈ / hxn i. Then m = an + b
with 0 ≤ b < n, and x ∈ H. Therefore b = 0, i.e. x ∈ hx i. So H = hxn i, which means
b m n
H is cyclic.
Now let G have order n and r be an integer that divides n. Then H = hxn/r i is a
subgroup of G having order r. Assume H is not unique, i.e. H 0 ≤ G and |H 0 | = r. Then
by the Lemma, H 0 = hxm i for some m, and let m be the smallest positive integer such
that H 0 = hxm i. If m - n then we can write n = qm + s with 0 < s < m. Then we have
x(q+1)m = xm−s ∈ H 0 . But m − s < m is a contradiction, so m|n and thus |H 0 | = n/m.
But we already had |H 0 | = r, which tells us that m = n/r and H 0 = hxn/r i = H. Therefore
H is unique. ♦ (Note:) In a cyclic group of order n, xar = xbr ⇒ ar ≡ br (mod n) but it
does not imply a = b, even mod n.

(a) Let a, b ∈ G, an abelian group, with orders m, n respectively.
Claim: The order of ab divides the least common multiple of m, n.
Proof: Let r be the least common multiple of m, n. Then r = zm = yn for some
integers y, z. So we have
(ab)r = ar br = (am )z (bn )y = ez ey = e
Since (ab)r = e, The order of ab must divide r..

Note: The order of ab is not in general equal to the least common multiple of m, n: For
example, let G = C4 = hxi with a = x and b = x3 , in which case ab has order 1. ♦
2

0 2 0 3
(b) Let G = GL2 (R) and consider the matrices A = 1 and B = 1 . Then
2 0 3 0
we 2 2
have A = B = I, that is, A and B have order 2, which is finite. However,
n
AB =
2/3 0 (2/3) 0
which has the startling property that (AB)n = which is never
0 3/2 0 (3/2)n
equal to the identity for n 6= 0, i.e. AB has infinite order. ♦

Construct the map ϕ : R+ → P defined by ϕ(x) = 2x .
ϕ(x + y) = 2x+y = 2x 2y = ϕ(x)ϕ(y)
So ϕ is a homomorphism. Furthermore, 2x = 1 iff x = 0. So Ker(ϕ) = 0, i.e. ϕ is injective.

Finally, if x ∈ P , ϕ(log2 (x)) = x. So ϕ is surjective. Thus ϕ is an isomorphism, i.e.
R+ ∼
= P. ♦

1 1 1 0
Define A = and B = Furthermore define the permutation matrix D ∈
0 1 1 1

0 1
GL2 (R) by P = Then D = D−1 and DAD−1 = B, i.e. A and B are conjuagte in
1 0
GL2 (R).
Suppose that ∃C ∈ SL2 (R) such that CAC −1 = B. Then
CAC −1 = DAD−1 ⇒ (D−1 C)A = A(D−1 C)
So D−1 C commutes with A. Write the general form of a 2 × 2 matrix, and we have:

1 1 a b a+c b+d a a+b a b 1 1
= = =
0 1 c d c d c c+d c d 0 1
So c = 0 and a = d. This says det(D−1 C) = a2 > 0. But det(D−1 C) = (det D−1 )(det
C) = −1 · 1 < 0, a contradition. Therefore A and B must not be conjugate in SL2 (R). ♦

(a) ϕ is surjective because ϕ(x−1 ) = x, ∀x ∈ G. The identity is the inverse only of itself, so
ker(ϕ) = e, i.e. ϕ is injective.
(b) Let ϕ be an automorphism, which means ϕ is a homomorphism. Let x, y ∈ G. Then
xy = (y −1 x−1 )−1 = ϕ(y −1 x−1 ) = ϕ(y −1 )ϕ(x−1 ) = yx
and so G is abelian. If G is abelian, then
ϕ(xy) = (xy)−1 = y −1 x−1 = x−1 y −1 = ϕ(x)ϕ(y)
and so ϕ is an bijective homomorphism, i.e. an automorphism. Thus ϕ is an automorphism

if and only if G is abelian. ♦
3
Math 122, Solution Set No. 2
As a general note, all elements of Sn will be written in disjoint cycle notation
unless otherwise specified. Also, as a notational convention, H ≤ G means H is a
subgroup of G.
1 2.3.14
(a) Note that, if ϕ : Z+ → Z+ is an automorphism, ϕ is completely determined
by ϕ(1). This is because, if n ≥ 0 ∈ Z, ϕ(n) = ϕ(1) + ... + ϕ(1) (n times) and
ϕ(−n) = −ϕ(n). Thus, for arbitrary n ∈ Z, ϕ(n) = nϕ(1).
In particular, this implies Im ϕ = ϕ(1)Z. Note mZ = Z iff m = 1 or −1. Thus
ϕ(1) = −1 or 1, and the group of automorphisms is C2 , the only group of order 2.
(b) An automorphism on a cyclic group G =< a > is completely determined by
its action on a generator a (this is by the same reasoning as in a). If ϕ(a) = b, where
b is not a generator of the cyclic group, then Im ϕ =6= G. If ϕ(a) = c, where c
is a generator, then Im ϕ =< c >= G. The fact that this map is a homomorphism is
problem 2.4.5.
In this particular situation, we note that all cyclic groups of finite order 10 are
isomorphic, so it suffices to examine the case G = Z/10. It is an easy check to see
that 1, 3, 7, 9 ∈ Z/10 are the only generators. In general, the generators for a Z/n are
those integers viewed modulo n that are relatively prime to n. Let 1 be a generator
for Z/10. We can define ϕ(1) = 1, 3, 7 or 9. Thus |Aut(G)| = 4. One can check that
Aut(G) is cyclic, so we have that G is isomorphic to C4 .
(c) Note that if H is a subgroup of G (denoted H ≤ G), and ϕ ∈ Aut(G), then
|Im(ϕ|H )| = |H|. That is, since ϕ restricted to H is an isomorphism with domain H,
its image must have the same order as H.
I claim < (12), (123) >=< H, K >= S3 . To see this, note that |H| = 2, |K| = 3,
so 2 and 3 divide the order of | < (12), (123) > |, that is, 6 ≤ | < (12), (123) > | ≤ S3 .
But |S3 | = 6, so we have our result.
By an easy extension of the argument in (a), the image of these two generators
under ϕ determines the automorphism. From order considerations, we have ϕ((12)) =
(12), (13) or (23), and ϕ((123)) = (123), (132). Using the argument above, we note
that any combination of a two-cycle and a three-cycle generates the group. Thus,
since any element of σ ∈ S3 can be written as some product σ = (12)...(123)...(12)...
we simply define ϕ(σ) = ϕ(12)...ϕ(123)...ϕ(12).. It is easy to see that such a map is
an isomorphism. There are 6 such maps, and thus |Aut(S3 )| = 6.
Thus, by problem 2.6.11 below, Aut(S3 ) is isomorphic to S3 or C6 . I claim Aut(S3 )
is isomorphic to S3 . Clearly, if we show Aut(S3 ) is nonabelian, then we’re done.
Conjugation by an arbitrary element σ ∈ S3 is always an automorphism (this is easy
to check; it actually works for any group). To prove S3 is nonabelian, therefore, it is
sufficient to demonstrate σ1 , σ2 such that switching the order of conjugation by these
elements does not yield the same automorphism. That is, we must find σ1 , σ2 , σ3 such
1
that σ1 σ2 σ3 σ2−1 σ3−1 6= σ2 σ1 σ3 σ1−1 σ2−1 . One example (which you can check) is
(123)(12)(23)(12)−1 (123)−1 = (12)
(12)(123)(23)(123)−1 (12)−1 = (23).
2 2.4.5
Let x, y ∈ G. We have ϕ(xy) = (xy)n = xyxyxy..xy (n times)= xn y n (because G is
abelian) = ϕ(x)ϕ(y).
3 2.4.8
(a) A subgroup of S3 must have 1, 2, 3, or 6 elements by Lagrange’s Theorem. The
subgroup with 1 element is clearly the trivial subgroup, and the subgroup with 6
elements is clearly S3 . If any subgroup contains a 2-cycle and a 3-cycle, it must
contain the disjoint subgroups of orders 2 and 3 contained generated by these cycles,
so it must be of order at least 6, that is, it must be all of S3 . Thus the only other
subgroups are the cyclic subgroups generated by the three 2-cycles, (3 of order 2), and
the cyclic subgroup generated by the 3-cycles, (one of order 3). By problem 2.6.10(a),
the subgroup generated by the three cycles is normal. None of the 2-cycle subgroups
are normal (as you can check by looking at the group table from the last assignment).
(b) Since the quaternion group, call it G, has 8 elements, it can have subgroups of
order 1, 2, 4, 8. Again, the subgroups of order 1 and 8 are the trivial subgroup and G
itself. Denoting the elements as on p48, we have the subgroups generated by i, j, k,
each of order 4 (as one can check by simply multiplying out the matrices). Since each
of these subgroups have index 2, they’re normal by problem 2.6.10(a). We note that
< −i >=, < −j >=< j >, < −k >=< k > and −1 is an element of each of
these subgroups.
We also have the subgroup generated by −1, which is of order 2. Let g ∈ G −
{1, −1}. Then, as stated above, g has order 4, which implies g −1 (−1)g = g 3 (−1)g =
(−1)(g 4 ) = −1, so this subgroup is normal as well. Any combination of two of i, j, k
generates the group, as you can check from the relations given on p48. Thus, this is
it.
4 2.4.16
Consider restricting the homomorphism ϕ to H =< x >. Im(ϕ|H ) =< ϕ(x) > in G0 ,
and thus |Im(ϕ|H )| = |ϕ(x)|. Since |H| = |Kerϕ||Imϕ|, we have that |ϕ(x)| divides
r.
2
5 2.4.23
Suppose g 0 ∈ G0 . Because ϕ is surjective, we can write g 0 = ϕ(g) for some g ∈ G. We
have g 0 ϕ(N )g 0−1 = ϕ(g)ϕ(N )ϕ(g)−1 = ϕ(g)ϕ(N )ϕ(g −1 ) = ϕ(gN g −1 ) = ϕ(N ).
6 2.6.4
· ¸
i 0
Let γ = . We have
0 1
· ¸
ai bi
γGL2 (R) = { |ab − cd 6= 0, a, b, c, d ∈ R}
c d
and · ¸
ai b
GL2 (R)γ = { ab − cd 6= 0, a, b, c, d ∈ R}.
ci d
Clearly these two cosets are unequal.
7 2.6.10
(a) Let H ≤ G such that [G : H] = 2. Clearly hHh−1 = H for h ∈ H. Suppose
a ∈ G − H and aha−1 ∈ / H for some h ∈ H. Since [G : H] = 2, we have aha−1 ∈ aH,
which implies (a−1 a)ha−1 = ha−1 ∈ (a−1 a)H = H. Thus ha− 1 = h0 for some
h0 ∈ H, and we have a− 1 = h−1 h0 ∈ H, which implies a ∈ H, a contradiction. Thus
aha−1 ∈ H for all h ∈ H; that is, H is normal.
(b) Consider the subgroup of S3 generated by (12). This subgroup is of order 2
in a group of order 6. Thus it has index 3. By looking at the multiplication table for
S3 drawn up on the last homework, it is clear that this subgroup is not normal.
8 2.6.11
(a) If G is a group with an element of order 6, say x, and |G| = 6, then clearly
G =< x >, which is isomorphic to C6 .
(b) Before beginning this proof, I would like to note that very few of you actually
proved this result. This is understandable; it’s tricky to provide a classification of
groups of a particular order. I’ll try to point out the places where most had trouble.
Claim: If |G| = 6, G has an element of order 3, and none of order 3, then G is
isomorphic to S3 .
Proof: Let x be an element of order 3 in G. We thus have the subgroup {e, x, x2 } ≤
G. Let y ∈ G, y ∈< / x >. Then y, xy, x2 y ∈ G. This is the right coset < x > y.
Noting that cosets partition the group, and that these two cosets provide 6 elements,
we have G = {e, x, x2 , y, xy, x2 y}.
We CANNOT automatically assume |y| = 2; we must prove it. By Lagrange’s
theorem, |y| = 1, 2, 3, or 6. Clearly y 6= 1, 6. Suppose |y| = 3. Then y 2 ∈
3
{x, x2 , y, xy, x2 y, }. Using the cancellation law, we can construct contradictions for
any of these values. For example, suppose y 2 = x2 y. Then we would have y 2 y −1 =
x2 yy −1 which implies y = x2 , a contradiction. Thus, assuming we’ve done all of these
checks, |y| = 2.
There is one more relation we must prove:
yx = x2 y.
Assuming this result, we may simply reference p44 of Artin and its characterization
of S3 . The proof of this relation was another problem area for most people.
The cancellation law makes it clear that yx ∈ / {e, x, x2 , y}. Therefore, suppose for
a contradiction that yx = xy. Then x commutes with y, and we have (xy)n = xn y n ,
which implies |xy| = lcm(|x|, |y|) = 6. But we assumed that G contains no elements
of order 6, so this is a contradiction, and we have yx = x2 y. Note that this is the only
portion of the proof in which we use the fact that G contains no elements of order 6.
(Note: Thanks to William Meyerson; I stole a lot of this proof from him).
(c). Claim: If all elements of G are of order 1 or 2, then G is abelian.
Proof: Suppose a, b ∈ G. We have 1 = (ab)2 = (ab)(ab) = aba−1 b−1 which implies
ba = ab.
By order considerations, we must have e, x, y ∈ G, and xy ∈ G and is distinct from
the first three elements, by closure and the cancellation law. By commutativity and
the fact that each element has order 2, this we must have at least one more element,
call it z. Now we have e, x, y, z, xy, yz, xz, xyz ∈ G by closure. By the cancellation
law, all of these elements are distinct. Thus |G| ≥ 8, a contradiction. There are no
groups of this form.
9 2.8.2
No. Suppose it was nontrivially decomposable, say into a direct product of G1 and
G2 . This assumption is made without loss of generality, for we can always collapse
direct products involving more groups into the product of two groups. Then, by
nontriviality and Lagrange’s theorem, without loss of generality |G1 | = 2, |G2 | = 3.
Thus G1 is isomorphic to C2 and G3 is isomorphic to C3 , because these are the only
possible isomorphism classes of groups of order 2, 3. Thus we have S3 is isomorphic to
the direct product of C2 and C3 , which in turn is isomorphic to C6 (via Proposition
8.4). But this is a contradiction, see 2.6.11 above.
10 2.8.10
Let l = lcm(m, n). Then (x, y)l = (xl , y l ) = (e, e), for m, n|l. Further, by definition,
this is the least integer that m, n both divide, so it is the order of (x, y).
4
11 2.10.1
(a) Let A = (aij ) ∈ G, and let ϕ : G → R∗ be the map defined by ϕ(A) = a11 . We
have · ¸· ¸ · ¸ · ¸ · ¸
a b e f ab ∗ a b e f
ϕ( ) = ϕ( ) = ab = ϕ( )ϕ( ).
0 d 0 h 0 ∗ 0 d 0 h
Thus ϕ is a homomorphism. Because it’s kernel is exactly the set H, H ≤ G and
H is normal. This map is clearly surjective, so we have G/H is isomorphic to R∗ .
(b). H is not normal. For example, note
· ¸· ¸· ¸−1 · ¸
2 3 3 0 2 3 3 3/2
= .
0 4 0 5 0 4 0 5
(c). Let ϕ : G → R∗ be the map defined by ϕ(A) = a11 /a22 . We have

· ¸· ¸ · ¸ · ¸ · ¸
a b e f ae ∗ a b e f
ϕ( ) = ϕ( ) = ae/dh = ϕ( )ϕ( ).
0 d 0 h 0 dh 0 d 0 h
So ϕ is a homomorphism. Because its kernel is exactly the set H, H ≤ G and H is

normal. This map is clearly surjective, so we have G/H is isomorphic to R∗
(d). Let ϕ : G → (R∗ )2 be the map defined by ϕ(A) = (a11 , a22 ). We have
· ¸· ¸ · ¸ · ¸ · ¸
a b e f ae ∗ a b e f
ϕ( ) = ϕ( ) = (ab, dh) = ϕ( )ϕ( ).
0 d 0 h 0 dh 0 d 0 h
So ϕ is a homomorphism. Because it’s kernel is exactly the set H, H ≤ G and H is

normal. This map is clearly surjective, and we have G/H is isomorphic to (R∗ )2 .
12 2.10.10
z z w zw
Define ϕ : Cx → U by ϕ(z) = |z| . We have ϕ(zw) = |z| |w|
= |zw| = ϕ(z)ϕ(w), so ϕ is
a homomorphism, and from elementary complex analysis, it is a surjective map onto
z
U . Further, |z| = 1 if and only if z ∈ P . Thus, by the first isomorphism theorem,
x
C /P is isomorphic to U .
Define ϕ : Cx → P by ϕ(z) = |z|. We have ϕ(zw) = |zw| = |z||w| = ϕ(z)ϕ(w),
so ϕ is a homomorphism, and from elementary complex analysis, it is a surjective
map onto P . Further, |z| = 1 if and only if z ∈ U . Thus, by the first isomorphism
theorem, Cx /U is isomorphic to P .
5
1 3.1 Problem 1
(a) This is a subspace. The zero matrix is symmetric, and the set is closed under addition and scalar
multiplication because for symmetric matrices A, B and c ∈ F , (A + B)t = At + B t = A + B and
(cA)t = cAt = cA.
(b) The zero matrix is not invertible, hence this is not a subspace.
(c) This is a subspace. The zero matrix is upper triangular, and if we add any two matrices with
zeroes below the diagonal, their sum will have zeroes below the diagonal. Likewise the set is closed
under scalar multiplication.
Note: I required you to show that the subspaces were nonempty for (a) and (c). However, also
note that the identity matrix for these as vector spaces is the zero matrix, not In ∈ GLn (R).
2 3.2 Problem 1
√ √ √ √ √ √ √
Let a + b 2,√c + d 2 ∈ Q[ √ 2] (a)√(a + b 2) + (c + d 2) = (a + b) + (c + d) 2 ∈ Q[ 2]
(b) −(a +√b 2) = −a √ − b 2 ∈ Q[ 2] √ √
(c) (a + b √2)(c + d 2 = (ac√ + 2bd) + (ad + bc) √2 ∈ Q[ 2]
(d) (a + b 2)−1 = (a − b 2)/(a2 − 2b2 ) ∈ Q[ 2] (note there are no rational numbers a, b s.t.
(a2 − 2b2 ) = 0 √
except for the pair 0, 0.
(e) 1 ∈ Q ⊂ Q[ 2] √
Therefore by Artin (2.1), Q[ 2] is a subfield of C.
3 3.2 Problem 4
If W = V , then W is the set of solutions to AX = 0 where X is the zero matrix. If not, choose an
orthonormal basis (w1 , . . . , wk ) for W using the Gran-Schmidt process and extend this to a basis
Pk
w1 , . . . , wk , v1 , . . . , vn−k for V . Then let A = [v1 , . . . , vn−k ]t . If x ∈ W , then we have x = i=1 ai wi
Pk Pk Pk t
and so Ax = A i=1 ai wi = i=1 ai Awi = i=1 ai [v1 · wi , . . . , vn−k · wi ] = 0 since the vi are
orthogonal to the wi by construction. Pk Now let
Pn−k x ∈ V s.t. Ax
Pk = 0. Then x is perpendicular to each
vector vi , so if we write x = i=1 ai wi + i=1 ci vi = i=1 ai wi since each of the coefficients ci
must be 0. So x ∈ W , i.e. W is exactly the space of solutions for the linear system of equations
AX = 0.
4 3.2 Problem 11
The determinant of this matrix in Fp is 10 (mod p). Since we have that A is invertible when det
A 6= 0, we must have 10 6= 0(mod p) ⇔ p - 10 ⇔ p 6= 2, 5. Therefore A is invertible unless in Fp
unless p = 2, 5.
5 3.2 Problem 15
(a) If x ∈ Fp is its own inverse, then x2 = 1(mod p)⇔ (x − 1)(x + 1) = 0(mod p) ⇔ x = 1 or
x = −1 = p − 1(mod p). So since every other element has a distinct inverse, the product of all
nonzero elements must be 1 · 1 · · · 1 · (−1) = −1. (b) (p − 1)!(mod p) is the product we solved for in
(a). Therefore it is equal to -1.
1
6 3.3 Problem 5
Let Bij be the n × n matrix with a 1 in the (i, j)th and(j, i)th entries and zeroes in all other entries.
Then thePset B = {Bij |i ≥ j} is a basis for the symmetric matrices. B is linearly independent
because i ci Bij = 0 says that the matrix with entries ci is identically zero, i.e. ci = 0∀i. B spans
the set
P of symmetric matrices because if A = (aij ) then if A is symmetric (aij ) = aji and we have
A = n≥j≥i≥0 aij Bij . Therefore B is a basis.
7 3.3 Problem 7
Suppose there are scalars such that ax3 + b cos x + c sin x = 0 Then this equation must
√ be true for
all x, so if we set x = 0 we obtain b = 0. Then if we set x = π/4 we obtain x3 = (c 2)/2a unless
a = 0. But in the first case, we have x3 is a constant function, which is clearly false, so we must
have a = c = 0. Therefore these three functions are linearly independent.
8 3.3 Problem 16
(a) Let W be a subspace of V . Choose a basis w1 , . . . , wk for W and extend to a basis w1 , . . . , wk , v1 , . . . , vn−k
for V . Then let U = span(v1 , . . . , vn−k ). Then clearly W + U = V since any vector in V can
be written
Pk as the sumPn−kof the basis
Pk vectors for Pn−k U and W . Also, if w ∈ W ∩ U then we have
w = i=1 ai wi = i=1 bi vi ⇒ i=1 ai wi − i=1 bi vi = 0 ⇒ w = 0 because these vectors are
linearly independent.
(b) Let W, V be as in part (a) with the same basis for V . W ∩ U = 0 ⇒ U ⊂ Span(v1 , . . . , vn−k ⇒
dim U ≤ n − k. Therefore it is impossible to have W ∩ U = 0 and dim W + dim U ≥ dim V .
9 3.4 Problem 10
Let F be a field of 81 elements. Then it is clear that if V has dimension 3 over F, |V | = 813 .
From class, we know any one-dimensional subspace over a finite field has |F| elements. If we have
two distinct one-dimensional subspaces, since their intersection is a subspace it must be the zero
subspace and so we can partition V into sets of 80 nonzero elements in each subspace. So there are
(813 − 1)/80 distinct one-dimensional subspaces.
10 3.5 Problem 2
Let A be the space of doubly infinite sequences. Define ϕ : A → R∞ by ϕ(. . . , a−1 , a0 , a1 , . . . ) =
(a1 , a−1 , a1 , . . . ). Since ϕ is defined componentwise, it is clearly bijective and linear. Therefore ϕ is
an isomorphism.
11 4.3 Problem 4
Assume T is not multiplication by a scalar. Then T does not have a two-dimensional eigenspace, and
so we can choose some v which is not an eigenvector of T , i.e. T (v) 6= cv for any c ∈ F. So {v, T (v)}
is linearly independent ⇒ this set is a basis, since V has dimension2. The columns of T w.r.t. this
2 0 a
basis are just the coordinates of T (v), T (v) in this basis, so [T ] = where T 2 (v) = (a, b)t .
1 b
Note: The fact that T is not multiplication by a scalar does not mean T has no eigenvectors.
2
122 Solution Set 4
1 4.3.9
(a) Let {1, x, x2 , ..., xn } be a basis for P . With respect to this basis,
 
0 1 0 ··· 0
0 0 2 0 · · · 0
 
0 0 0 3 0 · · · 0
 
D :=  .. .. ..  .
. . .
 
0 ··· ··· n
0 ··· ··· 0
The characteristic polynomial of D is easily calculated to be xn . By the Cayley-

Hamilton theorem (that is, the last problem on this set), we have Dn = 0, which
implies that D is nilpotent (quick proof by Jason Adaska).
(b). Define Pi := span{1, x, ..., xj } for 1 ≤ j ≤ n. These Pi are clearly invariant
subspaces, as almost everyone noted. However, most people didn’t show (or didn’t
show correctly) that these were the only invariant subspaces. Here we go:
Suppose W is a nontrivial invariant subspace of P . Pick an element of maximal
degree in W , say p = ak xk + ... + a0 , where necessarily ak 6= 0. I claim W = Pk . We
proceed by induction. Dk−1 p ∈ R, so < 1 > is contained in W . Suppose we’ve shown
Pn−1 ≤ W . Then Dk−n−2 p − ck−1 xk−1 + ... + c0 = dn xn for some ci , d0 ∈ R, which
implies xn ∈ W . Hence Pn ≤ W . Continuing inductively, we have that W = Pk .
2 4.3.10
If a 6= d, B clearly has the distinct eigenvalues a, d (note that the eigenvectors of an
upper triangular matrix are the entries along the diagonal). Thus, by Proposition
(6.3) and Theorem (6.4), we may write B, via a change of basis, as the diagonal
matrix A, where the new basis is the basis of eigenvectors for B guaranteed to exist
by Proposition (6.3). Saying that two matrices are the same up to a change of basis
is the same as saying that they are similar.
Conversely, suppose a = d. Then, for all matrices P ∈ GLn (R), P AP −1 = B im-
plies that AP P − 1 = B (because A is diagonal) which implies A = B, a contradiction,
for b is nonzero.
3 4.4.2
· ¸
a b
(a) Let A = be a symmetric matrix in GL2 (R). The characteristic polynomial
b c
of A has discriminant (a−c)2 +4b2 which is always nonnegative, thus the characteristic
polynomial has real roots. The roots of the characteristic polynomial are the
1
· ¸
a b
(b) Let A = ∈ GL2 (R) such that b, c > 0. The characteristic polynomial
c d
of A has discriminant (a − d)2 + 4bc ≥ 0 because b, c > 0, thus the characteristic
polynomial has real roots.
4 4.4.4
A real 3 × 3 matrix has a characteristic polynomial of degree three. A degree three
polynomial has at least one real root (from elementary analysis), hence the 3 × 3
matrix has at least one real eigenvalue.
5 4.5.4
(a) Suppose A, B ∈ On .
Closure: AB(AB)T = ABB T AT = AIAT = AAT = I.
Inverses: AAT = AT A = I implies A−1 = AT .
Identity: II T = I.
Note SOn = SLn ∩ On . The intersection of two subgroups is also a subgroup, so
SOn is a subgroup.
Note that det : On → {−1, 1} is a homomorphism. Therefore [On : SOn ] =
|On /SOn | = 2 by Lagrange.
(b) O2 is NOT isomorphic to SO2 × {I, −I}. The only elements of order two in
SO2 × {I, −I} are ρπ × I, ρπ × −I, I × I, I × −I. That is, there are four of them;
any other rotation ρ in SO2 has order greater than 2, which forces the order of ρ × I,
ρ × −I to be greater than two. In contrast, O2 has infinitely many elements of order
two; any reflection through the origin is such an element. Because an isomorphism
preserves order, the two groups cannot be isomorphic.
O3 is isomorphic to SO3 × {I, −I}. −I ∈ / SO3 , so SO3 ∩ {I, −I} = I Considering
{I, −I} as a subgroup of O3 , it is easy to see that it is normal because both of
these elements commute with any matrix. We’ve already commented that SOn is
normal in On above. Because −I ∈ / SO3 and [On : SOn ] = | = 2, we have that
SO3 {I, −I} = O3 . Thus, by Proposition 2.8.6.c, O3 is isomorphic to SO3 × {I, −I}.
6 4.5.7
The product of the two eigenvalues of A is equal to the determinant of A. Further,
we proved in class that if λ is an eigenvalue of A, then so is λ−1 . The only possible
eigenvalues satisfying these two stipulations are 1, −1. Let the associated eigenvec-
tors be v1 , v2 , respectively. Because they are eigenvectors corresponding to different
eigenvalues of A, they are linearly independent. I claim that they are orthogonal. We
have
< v1 , v2 >=< Av1 , Av2 >=< v1 , −v2 >= − < v1 , v2 >
2
which implies < v1 , v2 >= 0. Noting that Av1 = v1 and Av2 = −v2 , we have that A
is a reflection of the plane fixing the span of the vector v1 .
7 4.5.8
We proved in class that any element A of SO3 fixes a vector v1 (where we may
assume v1 is a unit vector) and rotates the plane normal to that vector. Extend v1
to an orthonormal basis for R3 , say {v1 , v2 , v3 }. With respect to this basis A can be
represented in matrix form as  
1 0 0
0 ∗ ∗
0 ∗ ∗
The lower right hand corner of this matrix is just a rotation of θ degrees the plane
normal to the vector v1 . We have already chosen an orthonormal basis; and hence
the bottom right corner takes on the form below:
 
1 0 0
0 cos(θ) − sin(θ)
0 sin(θ) cos(θ)
(see 4.5.2). The trace of this matrix clearly satisfies the given relation. Because trace
is invariant under similarity transformations, we cos(θ)) = 21 (trA − 1)
8 4.6.1
(a) The characteristic polynomial of this matrix is (2 − x)2 − −1 = (x − 3)(x − 1), so
the eigenvalues are 3, 1. The corresponding eigenvectors are (1, 1)T , (1, −1)T , which
you can readily check (this is just high school algebra).
(b) Because the basis for the diagonal matrix will be (1, 1)T , (1, −1)T , we know
that these two vectors form the columns of the matrix P , the change of basis matrix
from the new basis to the standard basis. Thus we have:
· ¸· ¸· ¸−1
−1 1 1 2 1 1 1
P AP =
1 −1 1 2 1 −1
· ¸ · ¸
3 3 1 −1 −1
= −
1 −1 2 −1 1
· ¸
3 0
= .
0 1
(c) · ¸
−1 3 0
A=P P
0 1
3
which implies · ¸
30 −13 0
A = (P P )30
0 1
· ¸
−1 3 0 30
=P ( ) P
0 1
· 30 ¸
−1 3 0
=P P
0 1
· ¸
1 330 + 1 330 − 1
= .
2 330 − 1 330 + 1
9 4.8.1
(a) · ¸ · ¸ · ¸ · ¸ · ¸
1 1 1 −1 1 −1 1 1
+ + + + + ... =
1 −1 2 −1 3! 1 4! 1
" P∞ (−1)n P∞ (−1)n #
Pi=0 (2n)!
(−1)n Pi=0 (2n+1)!
∞ (−1)n =
− ∞ i=0 (2n+1)! i=0 (2n)!
· ¸
cos(1) sin(1)
.
− sin(1) cos(1)
(b) · ¸ · ¸ · ¸ · ¸ · ¸
1 a b 1 a2 ab 1 a3 a2 b 1 a4 a3 b
+ + + + + ... =
1 2 3! 4!
· P ∞ ai b P ∞ ai ¸
i=0 i! a i=1 i! =
1
· a b a ¸
e a (e − 1)
.
0 1
10 4.MISC.15
µ ¶
a b
(a) Let A = . Then charpoly(A) = (a−x)(d−x)−bd = x2 −(a+d)x+ad−bc.
c d
We have
A2 − (a + d)A + (ad − bc)I =
µ 2 ¶ µ 2 ¶ µ ¶
a + bc ab + bd −a − ad −ab − bd ad − bc 0
+ +
ac + cd bc + d2 −ac − cd −ad − d2 0 ad − bc
µ ¶
0 0
.
0 0
4
 
λ1 0 0 0
 0 λ2 0 · · · 0 
 
 0
(b) Let A =  0 0 λ3  where the λi are not necessarily unique. Further,
 .. .. 
 . . 
0 0 0 · · · λn
let p(x) = x + an−1 xn−1 + ... + a0 be the characteristic polynomial of A. Then we
n
have
An + an−1 xn−1 + ... + a0 I =
   
λn1 0 0 0 an−1 λn−1
1 0 0 0
 0 λn 0 · · · 0   0 an−1 λ2n−1 0 ··· 0 
 2   
0 0 λ n
0  0 0 n−1
an−1 λ3 0 
 3 +  + ...
 .. ..   .. .. 
 . .   . . 
n n−1
0 0 0 · · · λn 0 0 0 · · · an−1 λn
 
a0 0 0 0
 0 a0 0 · · · 0 
 
 0 0 a0 0 
... +  =
 .. .. 
 . . 
0 0 0 · · · a0
 
p(λ1 ) 0 0 0
 0 p(λ2 ) 0 ··· 0 
 
 0 0 p(λ ) 0 
 3 
 .. .. 
 . . 
0 0 0 · · · p(λn )
But the λi , being the diagonal entries of a diagonal matrix, are the eigenvalues of A.
Hence p(λi ) = 0 for all i, and we have our result.
(c) By hypothesis, we can find some n × n matrix P such that P AP −1 is diagonal.
Again, let p(x) = xn + an−1 xn−1 + ... + a0 be the characteristic polynomial of A. We
have, from (b),
p(P AP −1 ) = (P AP −1 )n + an−1 (P AP −1 )n−1 + ... + a0 I =
P An P −1 + an−1 P An−1 P −1 + ... + a0 I =

P (An + an−1 An−1 + ... + a0 I)P −1 = 0.
This directly implies, by multiplying on the left by P −1 and on the right by P , that
p(A) = 0.
(d) For this problem, we will choose the operator norm || · || on n × n matrices.
Let A be an arbitrary n × n matrix. By Proposition 6.1, there is an n × n matrix
P such that C := P AP −1 is upper triangular. The diagonal entries of C are the
eigenvalues of C, and hence also A. Call these eigenvalues λ1 , ..., λn . Pick some
² > 0, and then choose ²1 , ..., ²n such that λ1 + ²1 , ..., λn + ²n are all distinct and
5
 
²1 0 0 0
 0 ²2 0 · · · 0 
 
 0
²i < ( 2n ) for all i. Let E =  0 0 ²3
² 1/2
. Then C + E is a diagonalizable
 .. .. 
 . . 
0 0 0 · · · ²n
matrix, for it has distinct eigenvalues. This implies that D := P −1 (C + E)P is also
a diagonalizable matrix. Now, we have
||D − A|| = ||P −1 (C + E)P − A|| = ||P −1 CP + P −1 EP − A||
≤ ||P −1 CP − A|| + ||P −1 EP || = 0 + ||P −1 ||||E||||P ||

²
= ||E|| ≤ n(( )1/2 )2 < ².
2n
Thus, we have a diagonalizable matrix D arbitrarily close to our original matrix A.
Let p(x) be the characteristic polynomial of A. From continuity of polynomials, we
have that p(D) = 0 (from c above) implies lim²→0 p(D) = p(A) = 0.
6
1 5.2.13
(a) (⇒) if x is on l, then glide reflection acts on points on l as a translation; therefore, x, m(x) and
m2 (x) lie on l and are colinear. (⇐) If x is not on l, then the line joining x to m(x) crosses l (because
m is a glide reflection) and so does the line from m(x) to m2 (x). If these points are distinct, then
this implies x, m(x), m2 (x) are not colinear, since any line intersects l at most once. (b) From Artin,
if m is orientation-reversing, it is either a reflection or glide reflection.If x is a reflection, m2 (x) = x
and so m is a glide reflection; by (a) it is along the line l.
2 5.2.14
By Artin (2.4) etc., we know that any m ∈ SM has the form m = ta ρθ with a ∈ R2 andθ ∈ R. Write
a as (a1 , a2 ) and define the map ϕ : SM → GL2 (C) by
iθ
e a1 + ia2
ϕ(ta ρθ ) =
0 1
This is clearly a bijection between SM and the desired subgroup of GLc (C), and it is easily checked
that it is a homomorphism and therefore an isomorphism. Note: we may not assume that if ta is a
translation, a ∈ R.
3 5.3.3
Let Dn = hx, yi|xn = y 2 = 1andxy = yx−1
For D13 : by Lagrange, proper normal subgroups have order 2 or 13; the order 2 hyi subgroup
is not normal, since xyx−1 = 2x ∈ / hyi. The order 13 subgroup hxi has index 2, and is therefore
normal; the quotient group has order 2 and is therefore C2 .
For D15 ; we have the normal subgroups hx1 i, hx3 i, hx5 i which are normal because they are the
kernels of the homomorphisms ϕn : D15 → Dn , n = 1, 3, 5 defined by ϕn (x) = x, ϕ(y) = y. The
first Isomorphism Theorem tells us that the quotient groups are Dn , n = 1, 3, 5. All other subgroups
contain the element y, which (as above) implies that they are not normal.
Note: the dihedral group is not only the symmetries of an n-gon, but also exists as an abstract
group; computations can be made in the dihedral group and its structure analyzed without reference
to rotations and reflections.
4 5.4.2
Let x be a rotation around point a and y a rotation around point b. Consider the motion xyx−1 y −1 .
This is clearly nonzero, since rotations around different points do not commute. However, if we take
the image of this motion under the derivative map ϕ : M → O2 (R), we see that it is sent to the
identity, since the motions x and y are orientation-preserving and hence ϕ(xyx−1 y −1 ) ∈ SO2 (R),
which is an abelian group. Therefore xyx−1 y −1 ∈ ker ϕ, i.e. it is a translation.
5 5.4.4
Numbering the patterns 1-17, we have the following point groups: 1) C1 ; 2) C2 ; 3) C3 ; 4) C4 ; 5) C6 ;
6) D1 ; 7) D1 ; 8) D1 ; 9) D2 ; 10) D2 ; 11) D2 ; 12) D2 ; 13) D4 ; 14) D4 ; 15) D3 ; 16) D3 ; 17) D6 .
1
6 5.4.6
√ √ √ 2
Note that 2 − 1 < 1, so for some n, ( 2 − 1)n < for any √ > 0. Furthermore, since ( 2) = 2, by
the binomial expansion this product is of the form a + b 2 with a, b ∈ Z. So we can find elements
of this subgroup arbitrarily close to zero, and therefore can add arbitrarily small elements to find
+
an element√ of the subgroup arbitrarily close to any real number, i.e., the subgroup of R generated
by 1 and 2 is dense in R.
7 5.4.10
(a) this point group contains a π/2 rotation and a reflection, and is therefore isomorphic to D2 . (b)
There are four cosets in the point group: they are (1) translations of the frieze pattern, (2) trans-
lations of the frieze pattern with a rotation, (3) translations with a reflection, and (4) translations
with both a reflection and rotation. (c) [G : G ∩ T ] = |G/G ∩ T | = |point group| = 4.
8 5.4.12
As in 5.4.10, Note that [G : G ∩ T ] = |point group|. The point group of symmetries of an equilateral
triangular lattice is not in fact D3 , but rather D6 (as in class). Therefore [G : G ∩ T ] = |D6 | = 12.
9 5.4.14
Many people had incorrect answers for parts of this problem. In addition to reading the solution set,
it would be helpful to look at the Japanese wallpaper patterns on the course website. Numbering
the patterns on page 190 as (a) through (f), and those on p. 173 as (1) through (17), we note the
following; (a) has 90-degree rotational symmetry about the center of a square, but no reflectional
symmetry; it is therefore has the same symmetry type as pattern (4). (b) has 90-degree rotational
symmetry and reflections, but NO center of 90-degree rotation lies on an axis of reflection; therefore
this is the same symmetry type as pattern (14). (c) This pattern has bilateral but not rotational
symmetry, and its lattice is that of an equilateral triangle, as is pattern (8). (d) this pattern has a
glide reflection and no other point symmetries, as does pattern (7). (e) has the same symmetries
as pattern (b), and (14) on page 173. (f) has 180-degree rotational symmetry and translational
symmetry but no others; likewise is pattern (2) on page 173.
10 5.5.5
Note that the group of symmetries of the square acts transitively on the vertices, edges, and diagonals
of the square. The symmetry group has order 8, so using the counting formula we can immediately
classify the stabilizer of an edge or vertex as C2 - in the case of an edge, this corresponds to the
identity and reflection about the perpendicular line through the center of that edge. In the case of
a vertex, the identity and reflection about the diagonal line through that vertex. There are only
two diagonals, so the stabilizer of a diagonal has 4 elements; the identity, rotation of 180 degrees,
reflection about the diagonal, and the composition of the latter two elements.
2
122 Solution Set 6
1 5.5.8
(a) g(0) = 0 for all g ∈ G. Thus 00 = {0}. Let e1 be the first vector of the standard
basis for Rn and v ∈ Rn be an arbitrary vector, not equal to zero. We may construct
an element h ∈ G such that he1 = v by simply letting the first column of the matrix
of h be v (with respect to the standard basis. Thus there are two orbits; we have
Rn = O0 ∪ Oe1 . As a check, note that gv = 0 implies v = 0 for g ∈ G.
(b) Write g ∈ G with respect to the standard basis, in terms of its column vectors
v1 , ..., vn as
g = (v1 , ..., vn ).
We have ge1 = v1 . Thus, the stabilizer of e1 is all matrices of the form g =
(e1 , v2 , ..., vn ) ∈ G.
2 5.6.1
gaH = aH ⇐⇒ gah = ah0 (for some h, h0 ∈ H ) ⇐⇒ ga = ah00 (for some h ∈ H )
⇐⇒ g = ah00 a−1 ⇐⇒ g ∈ aHa−1 . Thus the stabilizer is aHa−1 .
3 5.6.4
(a) The stabilizer, by definition, is all permutations of {1, ..., n} that leave 1 fixed,
that is, it is all permutations of {2, ..., n}, which is clearly isomorphic to Sn−1 .
(b) Sn is generated by the all of the transpositions of the n elements. The only
transpositions that are not in H = Stab({1}) are (12), (13), ...(1n). Thus, S =
H ∪ {∪1<j≤n (1j)H}.
I claim that these cosets of H are distinct. By the counting formula, we have
|Sn | = n! = |O{1} ||H| = |0{1} |(n − 1)! which implies |O{1} | = n. By Lagrange,
|Sn | = |H|[G : H], so we have |O{1} | = [G : H] = n. If these cosets were not
distinct, we would have less than n cosets of H in G, which contradicts the fact that
[G : H] = n.
(c) We have ϕ((1j)H) = (1j) · 1 = j.
4 5.7.2
As stated in the problem (and as you can check by drawing it), there are exactly two
ways to inscribe a regular tetrahedron in a cube. If we denote one such tetrahedron
by ∆, we thus have |O∆ | = 2. From 5.9.1, the order of rotational symmetries of a
cube is 24. By the counting formula, we have |Stab(∆)| = 24/2 = 12.
1
5 5.7.5 (thanks to Jason Adaska)
First we assume that all indices [G : K], [H : K], [G : H] are finite, otherwise this
equation doesn’t make sense. Consider the action of G on G/K. The stabilizer of
K for this action is K, so we have |OK | = [G : K] by the counting formula. Now
0
consider the orbit of K under the action of H on G/K, denote this as |OK |. Again,
0
the stabilizer is K, so we have |OK | = [H : K]. Let a1 , ..., an be elements of G such
that G = ∪1≤i≤n ai H is a disjoint union. Thus we have
OK = {ai HK : 1 ≤ i ≤ n}
= {a{hK : h ∈ H} : 1 ≤ i ≤ n}
0
= {ai OK : 1 ≤ i ≤ n}.
0
Thus |Os | = |G/H||OH |, which implies [G : K] = [G : H][H : K].
6 5.8.5
Let G be the group acting on S. There are 2! possible permutations of the elements
of the orbit with order two, and 3! possible permutations of the elements of the orbit
order three. These correspond to all of the possible actions of a group on this set;
any action just permutes the elements in some way. Because the action is faithful, no
two elements of G correspond to the same action. Thus |G| ≤ 12.
Suppose |G| = 12. G must have elements corresponding to all of the permutations
of S, namely e, (12), (13), (23), (123), (132), (45) or any combination of these elements.
Thus G is completely defined by its action on S, and we see that G is isomorphic
to S3 × S2 . Without loss of generality, let the two orbits be O3 = {1, 2, 3} and
O2 = {4, 5}. We let the nontrivial element of S2 transpose 4, 5 and the elements of
S3 act as usual on {1, 2, 3}.
Suppose |G| 6= 12. Then, since |O3 |||G| and |O2 |||G|, we have |G| = 6. From a
previous problem set, the only groups of order 6 are C6 and S3 .
C6 : We have that C6 is isomorphic to C3 × C2 ; if < x >= C3 and < y >= C2 ,
we define an action by sending x to the permutation (123) and y to the permutation
(12).
S3 : We define a homomorphism S3 → S3 × C2 by σ → σ × sgn(σ). Then let S3
act as usual on {1, 2, 3} and C2 as before on {4, 5}.
7 5.8.7
The subspaces are: µ ¶ µ ¶ µ ¶
0 1 2
1 := { , , }
0 0 0
µ ¶ µ ¶ µ ¶
0 0 0
2 := { , , }
0 1 2
2
µ ¶ µ ¶ µ ¶
0 1 2
3 := { , , }
0 2 1
µ ¶ µ ¶ µ ¶
0 1 2
4 := { , , }
0 1 2
.
By Prop. 5.8.2, to show that we have a homomorphism from GL2 (F ) → S4 , it is
enough to show that we have an action of GL2 (F ) on this subspace. This is trivial
(and we’re given this fact in the statement of the problem). The kernel of this map
is elements of g ∈ GL2 (F ) such that g · 1 = 1, g · 2 = 2, g · 3 = µ3, g · 4¶= 4. µ
But the
¶
1 0 2 0
only such g are scalar matrices. There are 2 in GL2 (F ), namely and .
0 1 0 2
Now, |GL2 (F )| = (32 − 1)(32 − 3) = 48, so |im(ϕ)| = |GL2 (F )|/|ker(ϕ)| = 24. But
|S4 | = 24, so im(ϕ) = S4 .
8 5.9.1
I’m just going to list some answers; most didn’t have any trouble with the explana-
tions.
Octahedron: Edge pole orbit order-6. Face pole orbit order-4. Vertex pole orbit
order-3.
Icosahedron: Edge pole orbit order-15. Face pole orbit order-10. Vertex pole orbit
order 6.
9 5.9.3
There are 4 diagonals, all of which are in the same orbit under rotations. Thus
H := |Stab(diagonal)| = 24/4 = 6 by the counting formula, which implies that
Stab(diagonal) is isomorphic to S3 or C3 × C2 . Thinking geometrically, we see that
H is generated by two rotations, each of order three, about the vertices defining
the diagonal, and a symmetry of order two which flips the two vertices defining the
diagonal. Further, if we conjugate the flip by a rotation, we get another rotation, not
a flip. Thus the subgroup of H of order two, generated by this flip, is not normal.
This rules out the possibility that H is isomorphic to C3 × C2 , for every subgroup of
order two in C3 × C2 is normal. Thus H is isomorphic to S3 .
10 5.9.5
Let S be the set of pairs of opposite vertices of the icosahedron. Then S has order 6.
Further, any pair can be taken to any other pair via some element of the icosahedral
group. Pick some pair p. The order of the icosahedral group is 60 (from Artin), so we
have 60 = |Op ||Stab(p)| = 6|Stab(p)|. Thus |Stab(p)| = 6, and Stab(p) is a subgroup
of the icosahedral group.
3
1 6.1.1
This rule does define a (left) operation of G on itself. 1 · x = x(1−1 ) = x, and (gh) · x = x(gh)−1 =
xh−1 g −1 = g · (h · x). (Here · denotes the group action).
2 6.1.4
Let G be a p-group acting on S such that p - |S|. Assume there is no fixed point. Then for each
P s ∈ S,
Stab(s) 6= G ⇒ |Os | = pe , e ≥ 1 by the Counting Formula. Then since we can write S = s∈S Os ,
and p|Os ∀s ∈ S, we have p|S, a contradiction. Therefore there is a fixed point for the action of G
on S.
3 6.1.5
Note: since any rigid motion can be written as the composition of translations, rotations, and
reflections, it suffices to consider conjugating by these elements.
Consider translations: Conjugating ta by r gives tr(a) ; conjugating by ρθ gives tρθ (a) ; and con-
jugating by another tb gives ta . So if we conjugate by any rigid motion of the plane, ta will be
conjugate to a translation by the same distance. Also, ta is conjugate to any translation of the
same magnitude, so all translations of the same magnitude form a conjugacy class (the identity is
translation by the zero vector).
Consider rotations: Conjugating ρθ by another rotation gives ρθ ; by a translation ta gives rotation
of angle θ around the point a, and conjugating by r gives ρ−θ . So the conjugacy class of a rotation
is rotations by ±θ around any point in the plane.
Consider reflections: if we conjugate r by the motion ta ρθ , we obtain ta+ρ2θ (r(−a)) ρ2θ r. Let
ϕ = 2θ. Also note that since rigid motions are bijections of the plane, by suitably choosing a we
can obtain b = a + ρ2θ (r(−a)) for any b ∈ R2 . Therefore r is conjugate to any orientation-reversing
motion, including glide reflections of all magnitudes. Note that since conjugacy is an equivalence
relation, all of the orientation-reversing elements of M are in the same conjugacy class.
4 6.1.6
(a) 1 + 1 + 1 + 2 + 5. This is not a valid class equation, since it implies that |Z(G)| = 3 - 10, a
contradiction since Z(G) ≤ G. (b) 1 + 2 + 2 + 5 is the class equation for D5 . (c) 1 + 2 + 3 + 4 is not
valid, since it implies ∃s such that |Os | = 3 - 10, a contradiction. (d) 1 + 1 + 2 + 2 + 2 + 2. This is not
a valid class equation. |Os | = 2 ⇒ |Gs | = 5. But Z(G) ≤ Gs and |Z(G)| = 2 - 5, a contradiction.
5 6.1.8
(a) Clearly 1, −1 are in the center of Q8 since they are the only elements of order 2. Furthermore,
it can be checked by computation that Ci = {±i} and likewise for j, k. Therefore the class equation
is 10 = 1 + 1 + 2 + 2 + 2.
(c) 10 = 1 + 2 + 2 + 5, a special case of part (e)
(e) Consider Dn = hx, y|xn = y 2 = 1 and xy = yx−1 i. Clearly the identity is in its own conjugacy
class, and if we conjugate an element xj by xm , we get xj , while if we conjugate by xm r we get
x−j . Now consider xj xm rx−j = xj rxm rrx−j = x2m x−j r. If n is odd, then x2 m cycles through all
elements xj as m does, and so all elements xj r form a single conjugacy class. If n is even, x2 m
cycles through all elements xj , j even as m does, and it is easily seen that all elements xj r with j
1
odd also form a conjugacy class (also note that in this case, xn/2 is its own inverse). So the class
equations are:
n even: 2n = 1 + 1 + 2 + · · · + 2 + n/2 + n/2
n odd: 2n = 1 + 2 + · · · + 2 + n
(g) While it is possible to divide the matrices into categories based on their trace, the quickest way
to do this problem is to compute the conjugacy classes by direct computation (using, for example,
mathmatica). The result is that all trace 0 matrices are conjugate, while there are three conjugacy
classes of order 4, 4, and 1 of the matrices of both trace 1 and trace 2. The class equation for
SL2 (F3 ) is then 24 = 1 + 1 + 4 + 4 + 4 + 4 + 6. Note furthermore that SL2 (F3 ) is NOT isomorphic
to A4 × Z/2Z.
6 6.1.12
Let N be normal in G and |G| odd. Since N is normal, it is a union of conjugacy classes contained in
N ; the orders of these classes must be odd, since they necessarily divide G. So since the identity is
in its own conjugacy class, the only possible class equations are 5 = 1+1+1+1+1 and 5 = 1+1+3.
The second implies that |Z(G) ∩ N | = 2, a contradiction since Z(G) ∩ N is a subgroup of G and
so by Lagrange’s Theorem, its order is necessarily odd. Therefore every element of N is in its own
conjugacy class, i.e. N ≤ Z(G).
2
122 Solution Set 8
We take the convention that sp is the number of Sylow p- subgroups of a particular
group G.
1 6.2.4
Suppose A5 had a subgroup of order 30, say H. Then [A5 : H] = 2 which implies H
is normal. But A5 is simple, so this is a contradiction.
2 6.2.5
I claim A5 is the only proper normal subgroup of S5 . Suppose for a contradiction
that S5 had another normal proper subgroup H. Then H ∩ A5 is normal in A5 , which
is simple. Thus H ∩ A5 = A5 or {e}. If H ∩ A5 = A5 , then |H| ≥ 60 which implies,
by Lagrange’s theorem, that |H| = 60, so we have H = A5 , a contradiction. Thus
H ∩ A5 = {e}, which implies, since HA5 ≤ S5 , that |HA5 | = |H||A5 | = 60|H| ≤ 120,
the order of S5 . Because H is nontrivial we have |H| = 2. Thus H is the cyclic
subgroup generated by a single element of S5 . The only such subgroups that are
not contained in A5 are those generated by a single transposition (ij). Because all
transpositions are in the same conjugacy class, as proven in class, H is not normal,
which provides us with our needed contradiction.
3 6.2.7
We prove the assertion. Suppose the abelian group G has prime order p. Then,
because the order of a subgroup H of G must divide the order of G, we have |H| = 1
or p. Thus H = {e} or H = G.
Now we prove the contrapositive of the other direction. Suppose an abelian group
G has composite order greater than 1 (the trivial group is automatically non-simple).
Then there is a prime p such that p divides |G|, which implies by Corollary 4.3 that
there is a cyclic subgroup H of order p in G. Because G does not have prime order,
H is a proper subgroup of G. In fact, H is normal in G, because all subgroups of an
abelian group are normal.
4 6.2.8
(a) From a previous problem set, |T | = 12. There are three subgroups of order 2; the
stabilizers of edges. These all have trivial intersection, so these subgroups give us 3
distinct order 2 elements. Action by conjugation is transitive on this set of elements,
hence we have one conjugacy class of order 3.
1
Similarly, there are 4 subgroups of order 3, the stabilizers of vertices. Since the
intersection of these subgroups is the identity, we have 8 distinct order 3 elements.
The action of conjugation by T breaks this into two separate orbits of order 4 (you just
have to think about this geometrically; draw a few pictures and convince yourself).
Throwing in the trivial conjugacy class, we have
|T | = 1 + 3 + 4 + 4.
(b) The center Z(T ) is a subgroup of T whose order is the sum of all the 1’s in
the class equation. Thus |Z(T )| = 1, which implies Z(T ) = {e}.
(c) |T | = 22 3. By the Sylow Theorems, there is a subgroup H of T of order 4.
Any element of H must have order 1, 2 or 4, and there are exactly 4 such elements in
T , as discussed above. Hence H is the sole Sylow 2-subgroup.
(d) Suppose there was a subgroup H ≤ T such that |T | = 6. Then H would be
normal, and hence would be some union of conjugacy classes of T . But no combination
of 1, 3, 4, 4, the orders of the 4 distinct conjugacy classes of T , sum to 6, so such a
subgroup cannot exist.
5 6.4.1
We have 20 = 22 5, so s5 |4 and is congruent to 1 (mod 5), so we have s5 =1. This
implies that there is exactly one subgroup of order 5 in a group G of order 20. This
subgroup must be isomorphic to C5 , and hence has exactly 4 elements of order 5. If
there were any other elements of order 5 in G, they would generate their own cyclic
subgroup of order 5, which can’t happen; there’s only one such subgroup. Thus there
are exactly 4 elements of order 5 in G.
6 6.4.3
(Thanks to Zhe Lu) It suffices to show sp or sq is equal to one, so assume for the sake
of contradiction that both are not one. The possibilities, by the third Sylow theorem,
are then sq = p, p2 and sp = q. Since q = 1 (mod p), p|q − 1, which implies q ≥ p + 1.
If sq = p, p = 1 (mod q) which implies p2 = 1 (mod q), so the third Sylow theorem
implies that p2 = 1 (mod q) whether sq = p or sq = p2 . Thus q|p2 −1 = (p−1)(p+1).
But,from above, q ≥ p + 1, which implies q does not divide p − 1 and q = p + 1. The
only way this equation can hold for p, q prime is if p = 2, q = 3, which corresponds
to the case 22 3 = 12. Artin classifies the groups of order 12 in theorem 5.1, and in
particular he proves that all such groups have a proper normal subgroup; that is, they
are not simple.
7 6.4.6
From class, |GL2 (Fp )| = (p2 − 1)(p2 − p) = p(p − 1)2 (p + 1). The largest power of p
dividing this order is p, so we wish to find a subgroup of order p. Consider the cyclic
2
µ ¶
1 1
subgroup generated by . Because, as you can check,
0 1
µ ¶k µ ¶
1 1 1 k
=
0 1 0 1
for all k ∈ Fp , this group is exactly
µ ¶
1 k
{ : k ∈ Fp }
0 1
which has order p.
8 6.4.14
It’s obvious that a group G of prime order is simple; by Lagrange’s theorem, it can
only have the two trivial subgroups, so it can’t possibly have a nontrivial normal
subgroup. To prove that the rest of the groups G aren’t simple, we start with a few
lemmas (note p, q denote primes in these lemmas):
Lemma 1: If |G| = pk , k > 1, G is not simple.
Proof: By Prop. 6.1.11, Z(G) 6= {e}. If Z(G) = G, then G is abelian, and hence,
by problem 6.2.7, is not simple. If Z(G) 6= G, then Z(G) provides a proper normal
subgroup.
Lemma 2: If |G| = p2 q, then G is not simple.
Proof: Given above.
Lemma 3: If |G| = pq, then G is not simple.
Proof: If p = q, we’re done by Lemma 1, so assume without loss of generality that
p > q. Then sp = 1 (mod p) and sp |q, which implies sp = 1. Thus the sole Sylow
p-subgroup of G provides a nontrivial normal subgroup.
With these lemmas in place, we’ve dealt with every order less than 60 other than
24, 30, 36, 40, 42, 48, 56. We deal with these using case-by-case analysis.
24: We proved in class that no group of order 24 is simple.
30: By the Sylow theorems, s3 = 1 (mod 3),s3 |10. Thus s3 = 1 or 10. Suppose
s3 = 10. The Sylow 3-subgroups are all of order 3, and have trivial intersection. Thus
we have accounted for (3 − 1)(10) + 1 = 21 elements of the group. Similarly, s5 = 1
(mod 5), s5 |6. Suppose s5 = 6. Then, again, we have 6 subgroups of order 5, all with
trivial intersection, and, by Lagrange, all of these subgroups have trivial intersection
with the Sylow 3-subgroups. We’ve thus found 21 + (5 − 1)(6) = 45 distinct elements
of the group, which is a contradiction.
36: We have s3 = 1 or 4. Suppose s3 = 4. Then the action of G by conjugation
on the set S of Sylow 3-subgroups induces a homomorphism
ϕ : G → S4
Assuming G is simple, ker ϕ = G or {e}. If it’s G, then conjugation by G fixes

all Sylow 3-subgroups. This is a contradiction, for the Sylow 3-subgroups are all
3
conjugate, and we’ve assumed that there’s 4 of them. If ker ϕ = {e}, then ϕ is
injective, which is a contradiction, for |G| = 36 > 24 = |S4 |.
40: We have s5 |8, s5 = 1 (mod 5), which implies s5 = 1.
48: We have s3 = 1 or 4. Suppose s3 = 4. Then the action of G by conjugation
on the set S of all Sylow 3-subgroups induces a homomorphism
ϕ : G → S4
Assuming G is simple, ker ϕ = G or {e}. If it’s G, then conjugation by G fixes

all Sylow 3-subgroups. This is a contradiction, for the Sylow 3-subgroups are all
conjugate, and we’ve assumed that there’s 4 of them. If ker ϕ = {e}, then ϕ is
injective, which is a contradiction, for |G| = 348 > 24 = |S4 |.
56: We have s7 = 1 or s7 = 8, and s2 |7. Suppose s7 = 8 and s2 = 7. In this case a
Sylow 7-subgroup has order 7, so any two Sylow 7-subgroups have trivial intersection.
Thus, the set of all Sylow 7-subgroups account for (7 − 1)(8) + 1 = 49 elements of G.
Consider any two Sylow 2-subgroups, say K, H. These subgroups are of order 8, and
they can’t share all of their elements. Thus |K ∪ H| > 8. Further, K, H must have
trivial intersection with any Sylow 7-subgroup by Lagrange’s theorem. Thus K, H
give us an additional k > 7 distinct elements of G, which means we’ve 49 + k > 56
elements of G, a contradiction.
9 6.4.15
Let G be a group such that |G| = 33. By the Sylow theorems, s3 = 1 (mod 3) and
s3 |11, which implies s3 = 1. Similarly, s1 1 = 1 (mod 1)1 and s1 1|3, which implies
that s1 = 1. Thus we have two subgroups H, K ≤ G such that |H| = 3, |K| = 11, and
H, K are both normal in G. From Lagrange’s theorem, H ∩ K = {e}. We also have
HK ≤ G. Noting that |HG| = |H||G| = 33, this implies HK = G. Putting these
facts together and applying Prop. 2.8.6, we have that G is isomorphic to H × K.
Since H is isomorphic to C3 and K is isomorphic to C11 , we have G is isomorphic to
C3 × C11 . Thus there is only one group (up to isomorphism) of order 33.
4
1 6.5.2
Note if |G| = 4, then G ∼ = C4 or G ∼= D2 ∼= V4 So assume p 6= 2. Then by the Sylow Theorems, ∃
a subgroup C2 of order 2 and a unique (therefore normal) subgroup Cp of order p. So C2 ∩ Cp = e
and so by Artin 2.8.6 G = Cp C2 . So as in class, G ∼= Cp nϕ C2 , where ϕ : C2 → Aut(Cp ) ∼= Cp−1 . If
ϕ is trivial, then we have G ∼= pC × C ∼ C
2 = 2p . The only other automorphism of Cp of order dividing
2 is ϕ : x 7→ x−1 . In this case we have G is generated by x, y with xp = y 2 = e and xy = yx−1 and
in this case G ∼= Dp .
2 6.5.4
(a) Let |G| = 55. By the Sylow Theorems, we have that there is a subgroup of G isomorphic to C5 ,
and a unique (therefore normal) subgroup isomorphic to C11 . As in 6.5.2 this implies G ∼ = C11 n C5 ,
i.e. G is generated by x, y with x11 = y 5 = e and yxy ∈ C11 e ∴ yxy = xr , r ∈ 1, . . . , 10.
5
(b) Note x = y 5 xy 5 = xr ⇒ r5 = 1 (mod 11). It is easily checked that 25 = 65 = 75 = 85 =
105 = 10 (mod 11) and so r cannot be any of these integers.
(c) Note also 15 = 35 = 45 = 55 = 95 = 1 (mod 11) and so r can take these values. If r = 1
then we have G ∼ = C11 × C5 ∼= C55 . If r = 3 then we have G = hx, y|x11 = y 5 = e and yxy = x3 i.
If r ∈ {4, 5, 9}, then G is isomorphic to this case by the map that sends x to x and y to y n , where
n ∈ {2, 3, 4}.
3 6.6.2
(34)(123)(45)(34) = (241)(35)
4 6.6.3
Note qp = p(qp)p−1 and so they have cycles of equal length.
5 6.6.4
Note first that the order of a permutation is the least common multiple of the lengths of its component
cycles (an easy lemma). So, (a) (12345) has order 5, and (12345)(67) has order 10. Any element of
S7 is the least common multiple of the length of its cycles, and so (b) the largest possible order is
12. So (a) there is no element of order 15.
6 6.6.5
Note that the cycle (a1 · · · ak ) = (ak ) · · · (a1 ) and therefore has parity k − 1. So if we write a
permutation as a product of cycles, each cycle of even (respectively odd) length will have sign -1
(resp. 1). Since the sign is multiplicative, this implies sign(σ) = (−1)n , where n is the number of
cycles of even length in the cycle decomposition of σ.
7 6.6.11
Let σ = (12). Note Cσ has order 6, and so |Z(σ)| = 24/6 = 4. Since e, (12), (34), and (12)(34)
stabilize σ, this is the entire centralizer.
1
8 6.6.14
(a) Two permutations are conjugate iff their cycle decompositions have the same orders. It is easy
to check that in S5 there are 24 5-cycles, 30 4-cycles, 20 3-cycles, 10 2-cycles, 20 products of disjoint
2- and 3-cycles, 15 products of 2 disjoint 2-cycles, and the identity. So the class equation is: 120
= 1+ 10 + 15 + 20 + 20 + 24 + 30. (b) Note that as in 6.6.5, we can easily check that the
elements in A5 are the 5-cycles, 3-cycles, identity, and products of disjoint transpositions. Note
that for a 5-cycle σ, Cσ ⊂ S5 has order 24 and so Z(σ) ≤ S5 has order 5. In fact, it is clear that
Z(σ) = e, σ, σ 2 , σ 3 , σ 4 ⊂ A5 and so in A5 , |Cσ | = 60/5 = 12. Similar computations show that the
conjugacy class of a 3-cycle or product of disjoint transpositions is the same in S5 and A5 . Therefore
the class equation for A5 is the same as the icosahedral group, i.e. 60 = 1 + 12 + 12 + 15 + 20.
9 6.6.15
Note for a < b , (ab) = (b-1, b) ... (a+1, a+2) (a, a+1) (a+1, a+2) ... (b-1, b). As in 6.6.5, any
cycle is the product of transpositions and so since any permutation is the product of cycles, the
symmetric group Sn is generated by the elements (12), (23) ... (n-1, n).
2
122 Solution Set 10
1 6.7.1
The free group on two generators is nonabelian, whereas the product of two infinite
cyclic groups is. Therefore, the two are not isomorphic.
2 6.7.3
Let S be the underlying set of generators for the free group. Further, let Cl(S) be
the set of closed words in S, and Con(S) be the set of conjugacy classes of words in
S. We define a map f : Con(S) → Cl(S) by joining the ends of any element in a
given conjugacy class. To check that this map is well defined, suppose that x, y are
in the same conjugacy class. Then wxw−1 = y for some word w ∈ S. Clearly closing
y is equivalent to closing x; the two extra words w just cancel out.
Surjective: Given any closed word, break it at an arbitrary point, and consider
the conjugacy class of the word that arises. This conjugacy class will clearly map to
the given closed word.
Injective: Suppose x, y ∈ S are two reduced words such that f (x) = f (y), where
we think of x, y as representative elements for two conjugacy classes. The only sort
of cancellation that can occur when we close x or y is of the form x = w−1 zw being
sent to the closed word formed from z. It follows that x, y are conjugate.
3 6.8.2
If a normal subgroup contains S, then it will contain all gsg −1 for all g ∈ G, s ∈ S.
The subgroup generated by the given set, by definition, is the smallest subgroup
containing these elements.
4 6.8.10
Let ta ρθ ri , tb ρϕ rj , where i, j = 0 or 1, be arbitrary elements of M (we know any
element of M can be written in this form from §5.2). We have, using the formulae in
§5.2:
ta ρθ ri tb ρϕ rj (ta ρθ ri )−1 (tb ρϕ rj )−1 =
ta ρθ tr(b) ρ−ϕ ri rj ri rj (ta ρθ )−1 (tb ρϕ )−1 =
ta ρθ tr(b) ρ−ϕ (ta ρθ )−1 (tb ρϕ )−1 =
ta tρθ (r(b)) ρθ ρ−ϕ ρ−θ t−a ρ−ϕ t−b =
ta+ρθ (r(b))+ρ−ϕ (−a)+ρ−2ϕ (−b) ρ−2ϕ
If we vary a, b, ϕ, θ, we get all of SM .
1
5 6.8.16
Prove. Let |G| = n. We may choose a finite number of generators; just take all the
elements of the group as generators (there’s only n).
Note that the multiplication table for G completely defines the law of composition
in G. Any entry in the table, say xy = z, can equivalently be written as the relation
xyz −1 . The set of relations formed in this manner from the group table is clearly a
defining set of relations for the group, and there are only n2 entries. This provides a
finite set of generators and relations that present G.
6 7.1.1
Let A = [aij ], B = [bij ], and denote by e1 , ..., en the standard basis for Rn . Then we
have, for all 1 ≤ i, j ≤ n, that aij = eTi Aej = eTi Bej = bij , so A = B.
7 7.1.4
√ −1
Normalizing (1, 0), for one basis vector we can use ( 2 , 0). One of the conditions
the other vector must satisfy is:
µ ¶
√ −1 T 2 1
( 2 , 0) (x, y) = 0.
1 2
√ −1 √ −1
One such vector is (x, y) = ( 5 , −2 5 ), which is actually orthonormal to the
√ −1 √ −1 √ −1
first, as you can check. These two vectors, ( 2 , 0), ( 5 , −2 5 ) are clearly
linearly independent, and so we have an orthonormal basis.
8 7.1.6
If the field F does not have characteristic 2, then 2−1 ∈ F . We have
2−1 (q(v + w) − q(v) − q(w)) = 2−1 (< v + w, v + w > − < v, v > − < w, w >) =
2−1 (< v, v > +2 < v, w > + < w, w > − < v, v > − < w, w >) =
< v, w > .
9 7.2.4
The given vector along with the two vectors (1, 0, 0), (0, 1, 0) clearly form a basis for
R3 . We apply the Gram-Schmidt process to this basis. Let u1 = √13 (1, 1, 1).
1 1
w2 = (1, 0, 0)− < √ (1, 1, 1), (1, 0, 0) √ (1, 1, 1) =
3 3
2
2 1 1
( , − , − ).
3 3 3
q
1
Normalizing, we have the second vector of our orthonormal basis as 6
(2, −1, −1) =
u2 . For the third vector, we compute:
1 1
w3 = (0, 1, 0)− < √ (1, 1, 1), (0, 1, 0) > √ (1, 1, 1)
3 3
r r
1 1
−< (2, −1, −1), (0, 1, 0) > (2, −1, −1) =
6 6
1 1
(0, , − ).
2 2
√ √
2 2
Normalizing, we have u3 = (0, 2
, − 2
). Thus u1 , u2 , u3 form the required orthonor-
mal basis.
10 7.2.6
Because P is invertible, we have the equation (P −1 )T A = A0 P . This equation implies
that if x ∈ kerA then P x ∈ kerA0 . Note P is invertible, so this sets up a bijective
correspondence between the kernel of A and the kernel of A0 , which implies their
ranks are equal.
11 7.2.7
Disprove. Let A = I be the matrix of the standard dot product in Rn . It clearly
has n ones for eigenvalues. The matrix of A with respect to an arbitrary basis is
P T AP for some invertible P . One such change of basis matrix is P = 2I. We have
P T AP = 4I, which clearly has n fours for eigenvalues. Therefore, the eigenvalues of
A are not independent of basis.
3
1 7.2.10
2
Note that Mn (R) ≡ Rn under the identification [A1 , . . . , An ] ↔ ((A1 )t , . . . , (An )t ). Using this
2
identification, we see that the given form is just the standard dot product on Rn and therefore
positive-definite, with an orthonormal basis given by the matrices eij , 1 ≤ i, j, ≤ n where eij is the
n × n matrix having a 1 in the (i, j)th position and zeroes elsewhere.
2 7.2.15
(a) Let w ∈ (W1 + W2 )⊥ ⇒ hw, w1 + 0i = 0∀w1 ∈ W1 ⇒ w ∈ W1⊥ . Likewise w ∈ W2⊥ and so
(W1 + W2 )⊥ ⊆ W1⊥ ∪ W2⊥ . If w ∈ W1⊥ ∪ W2⊥ , then it follows from bilinearity that w ∈ (W1 + W2 )⊥ .
Therefore (W1 + W2 )⊥ = W1⊥ ∪ W2⊥ .
(b) Let w ∈ W . Then for any v ∈ W ⊥ , hv, wi = 0 and so w ∈ W ⊥⊥ ⇒ W ⊆ W ⊥⊥ .
(c) Let W1 ⊆ W2 and w ∈ W2⊥ . Then hw, w2 i = 0∀w2 ∈ W2 ⇒ hw, w1 i = 0∀w1 ∈ W1 since
W1 ⊆ W2 . Therefore W2⊥ ⊆ W1⊥ .
3 7.2.17
(a) It is easy to compute that the matrix of the form with respect to the standard basis is:
 
1 0 0 0
0 0 1 0
Meij =   
0 1 0 0
0 0 0 1
(b, c) It is easy to check that the basis mij is orthonormal:

√ √
1 0 0 2/2 0 2/2 0 0
(m11 , m12 , m21 , m22 ) = ( , √ , √ , )
0 0 2/2 0 − 2/2 0 0 1
and that the signature with respect to this basis is (3, 1).
(d) Likewise check that {Ci } is a basis for the trace-zero subspace, where:
√ √ √
2/2 √0 √ 0 2/2 √0 2/2
(c1 , c2 , c3 ) = ( , , )
0 − 2/2 2/2 0 − 2/2 0
Then the matrix of the form with respect to this basis is computed to be:
 
1 0 0
Meij = 0 1 0
0 0 −1
Thus the signature is (2, 1).
4 7.2.23

0 1
Let A =
1 0
1

a b 2ab ad + bc
(a) Let P ∈ GL2 (F2 ) and consider P AP t . If we let P = , then P AP t = =
c d ad + bc 2ab
0 1
b/c det (P ) = 1 and 2 ≡ 0 in F2 . Therefore A is not diagonalizable as the matrix of a bilinear
1 0
form.
(b) By explicit computation, we can check that the orbits are:

0 0
{ },
0 0

0 1
{ },
1 0

1 1 1 0
{ , },
0 1 1 1

1 0 1 1 0 1
{ , , },
0 1 1 0 1 1

1 1 1 0 0 0
{ , , },
1 1 0 0 0 1

1 1 1 0 0 1 0 0 0 1 0 0
{ , , , , , },
0 0 1 0 0 1 1 1 0 0 1 0
2
122 Solution Set 12
1 7.4.5
Let A = [aij ], B = [bij ] be two hermitian matrices, and c, d ∈ R. Clearly the zero
matrix, which will function as the additive identity, is hermitian. We must show that
cA + dB is hermitian. This is equivalent to showing caij + dbij = caji + dbji , which is
trivial because aij = aji , bij = bji and c, d ∈ R.
Let Eij denote the n × n matrix with a one in the ijth entry and zeros elsewhere.
Then an obvious basis for this space is {Eii } ∪ {Eij + Eji } ∪ {i(Eij − Eji )}.
2 7.4.16
Let f, g, h ∈ P and a, b ∈ C.
Linearity in the second variable:
< f, ag + bh >=
Z 2π ¡ ¢
f (eiθ ) ag(eiθ ) + bh(eiθ ) dθ =
0
Z 2π Z 2π
f (eiθ )ag(eiθ )dθ + f (eiθ )bh(eiθ )dθ =
0 0
Z 2π Z 2π
a f (eiθ )g(eiθ )dθ +b f (eiθ )h(eiθ )dθ =
0 0
a < f, g > +b < f, g > .
Conjugate linearity in the first variable:
< af + bh >=
Z 2π ³ ´
af (eiθ ) + bh(eiθ ) g(eiθ )dθ =
0
Z 2π Z 2π
af (eiθ )g(eiθ )dθ + bh(eiθ )g(eiθ )dθ =
0 0
a < f, g > +b < h, g >

Positive definite: Z 2π
< f, f >= f (eiθ )f (eiθ )dθ =
0
Z 2π
|f (eiθ )|dθ ≥ 0
0
1
where the last Rinequality holds because |f (eiθ )| ≥ 0. Now, suppose < f, f >= 0.
2π
Then we have 0 |f (eiθ )|dθ = 0, which implies by elementary analysis (because f is
a polynomial and thus is continuous) that |f (eiθ )| = 0 ⇒ f (eiθ ) = 0 for 0 ≤ θ ≤ 2π.
Thus f has infinitely many zeros, and because it is a polynomial, this implies in turn
that f is identically zero.
k
(b) One can easily check that the set { √x2π }k=0,..,n is an orthonormal basis for this
space.
3 7.5.4
One can easily check that µ ¶
1 1 i
P =√
2 1 −i
has the property that P AP ∗ is diagonal and P P ∗ = I. In general for this sort of
problem, find a basis of eigenvectors for A and normalize them so that they give a
unitary change of basis matrix.
4 7.5.9
By the Spectral theorem, we can find some orthogonal matrix P such that P AP T =
[dii ] is diagonal, with dii ∈ R. P is orthogonal, so P T = P −1 . Thus P eA P T =
−
P eA P −1 = eP AP 1 = [edii ](*). This final matrix is diagonal, and clearly positive
definite because the dii are real and thus the edi i are all greater than zero. Conjugating
by the orthogonal matrix P doesn’t change the signature of eA , so we’re done.
(*) This calculation also offers a proof eA = P T [edii ]P is orthogonal; it’s the
product of three orthogonal matrices and orthogonal matrices form a group.
5 7.5.10
We take ⊥ to be with respect to the standard hermitian product <, >. Suppose
x ∈ ker A. Then Ax = 0 ⇒< Ax, y >= 0 for all y which implies < x, A∗ y >= 0 for
all y, which implies x ∈ (ImA∗ )⊥ .
Conversely, suppose x ∈ (ImA∗ )⊥ . Then, for all y, < x, A∗ y >= 0 ⇒< Ax, y >=
0, which implies, since <, > is nondegenerate, that Ax = 0.
6 7.6.1
We may write this quadric as X T AX + BX + c, where
 
1 2 1
A = 2 0 0 .
1 0 1
2
and B = (3, 0, 1). By the Spectral theorem, there exists an orthogonal matrix P such
that P AP T is diagonal. Take
 
1 −1 −1
P = 1 −1 1 
0 5 1
We will use this matrix to transform the quadric.

 
5 0 0
P T AP = 0 20 0 
0 0 −4
Note (3, 0, 1)P = (3, −6, 1).

Our original quadric was X T AX +BX −6 = 0. With the change of basis X = P X 0
we have the new, equivalent quadric
X 0T (P T AP )X + (BP )X + c = 0 = 5x02 + 20y 02 − 4z 02 + 3x0 + 2y 0 − 2z 0 − 6 = 0
We complete the square:
3 2 9 12 1 1 1
5(x0 + ) − + 20(y 0 + − − 4(z 0 + )2 + − 6 =
20 20 20 20 4 4
3 2 1 1 25
5(x0 + ) + 20(y 0 + )2 − 4(z 0 + )2 − =0
10 20 4 4
Thus we have a one-sheeted hyperboloid. Notice that in completing the squares,
we affected the constant coefficient. Therefore, it is not sufficient to compute the
eigenvalues of the original form to distinguish between the one-sheeted and two-
sheeted hyperboloids.
7 7.8.4
We prove it. If A is a real skew symmetric matrix, then iA is hermitian, and thus
there is some unitary matrix P such that P iAP ∗ = P iAP −1 is diagonal with all real
diagonal entries. That is, the eigenvalues of iA are all real. If λ ∈ R is an eigenvalue
of A, then Ax = λx ⇐⇒ iAx = iλx. But iλ, being an eigenvalue of iA, is real, so λ
must be purely imaginary.
8 10.1.2
Let R be the smallest subring of C that contains the real cube root of 2. 1 ∈
R ⇒ Z ∈ R because C has characteristic zero. Clearly all elements of the form
a + b21/3 + c41/3 , a, b, c ∈ Z are in R. If we show that the set of such elements,
call it S, form a ring, we’re done. Showing that S is an additive group is trivial, the
3
multiplicative identity is 1, and distributivity follows from distributivity of C provided
that we show S is closed under multiplication. We have, for a, b, c, d, e, f ∈ Z:
(a + b21/3 + c41/3 )(d + e21/3 + f 41/3 ) =
(ad + 2bf + 2ed) + (bd + ae + 2cf )21/3 + (cd + af + be)41/3 ,

which demonstrates closure.
9 10.1.9
(a) Not a ring. Addition is clearly commutative, associative, and the identity is ∅.
However, if A 6= ∅, A−1 doesn’t exist. This is because A is contained in A ∪ B for all
B. The multiplicative identity is U , and it is clearly associative.
Finally, distributivity:
A(B + C) = A ∩ (B ∪ C) = (A ∩ B) ∪ (A ∩ C) =
(B ∩ A) ∪ (C ∩ A) = (B ∪ C) ∩ A = (B + C)A
by DeMorgan’s laws.
(b) A ring. Addition is clearly commutative. It has identity element ∅. A series
of Venn diagrams will check that it is associative. The inverse of an element A is
A. Multiplication follows as before, and distributivity can be checked using Venn
diagrams.
(c) Not a ring. Addition clearly forms an additive group with the zero function as
the additive identity. Multiplication is associative because composition of functions
is associative, with the function which sends x to x serving as the identity.
However, distributivity fails. Let f (x) = x, g(x) = x, h(x) = x2 . Then we have
h(f (x) + g(x)) = (2x)2 = 4x2 whereas h(f (x)) + h(g(x)) = 2x2 .
10 10.2.6
(a) The fact that this is a ring follows from the fact that R[t], the set of formal
polynomials, is a ring, if we just note that in any sum or product of formal power
series, the coefficient of tn is just a linear combination of coefficients of powers tk ,
k ≤ n. For example, if a0 + a1 t + ... and b0 + b1 t + ... are two formal polynomials in
R[[t]], then the coefficient of tn in their product can be computed by determining the
coefficient of tn in
(a0 + a1 t + ... + an tn )(b0 + b1 t + ... + bn tn )
if we consider the ring axioms in terms of coefficients, we just apply the fact that the
formal polynomials form a ring.
(b) If p(t)(b0 + b1 t + ...) = 1 for some power series b0 + b1 t + ..., then a0 b0 = 1
which implies a0 is invertible.
4
Now, suppose a0 is invertible. Then we want to find b0 , b1 , b2 , b3 , ... ∈ R such that
p(t)(b0 + b1 t + ...) = 1. Set b0 = a−1
0 . We have that the nth coefficient in the product
is given by
n
X
(ai bn−i )
i=0
Pn
We may define bn inductively as − i=1 a−1
0 (ai bn−i ). Thus we have that the nth
coefficient in the product is
n
X n
X
(ai bn−i ) + a0 (− a−1
0 (ai bn−i )) = 0
i=1 i=1
5
1 10.3.4
Let f ∈ (2) ∪ (x) ⊂ Z[x]. Since x - 2 and 2 - x, we have f = 2xg(x) and so f ∈ (2x). Conversely, if
f ∈ (2x) ⊂ Z[x] we have f = 2xg(x) ⇒ f ∈ (2) ∪ (x). ∴ (2) ∪ (x) = (2x) ⊂ Z[x].
2 10.3.6
Pn
Let I := { 0 ak xk ∈ Z[x]|2k+1 |ak }. Note 2 ∈ I, but x ∈ Z and x · 2 ∈
/ I. So I is not an ideal.
3 10.3.8
(a) Let F : R[x, y] → R sends f (x, y) to f (0, 0). Then ker F = (x, y). (b) If F : R[x] → C is given by
F (f (x)) = f (2 + i), then ker F is all real polynomials having (2 + i) as a root. Any such polynomial
necessarily has (2 − i) as a root, and so ker F = ((2 + i)(2 − i)) = (x2 − 4x + 5).
4 10.3.24
(a) Let N = N (R) be the nilradical of a commutative ring R. Let x, y ∈ N ⇒ xn = y n = 0 for
some sufficiently
P large n.p Then if z ∈ R, (zx)n = z n xn = 0 and furthermore −x ∈ N . Finally,
2n q
(x + y) = p+q=2n ci x y so we always have p > n or q > n and so this sum is zero. Therefore N
is an additive subgroup closed under multiplication by arbitrary elements
Q of R, i.e. N is an ideal.
(b) Let R = Z/nZ and N = N (R). Write the prime factorization n = pi ei and let I be the image
of the ideal ( pi ) in R (by correspondence). Then I ⊂ N because if x ∈ I, xmax ei = 0. Also
Q
Q N ⊂I
because any nilpotent element must clearly contain all the prime factors of n. So N = ( pi ) ⊂ R.
As a corollary, we have N (Z/12Z) = (6) and N (Z) = N (Z/0Z) = (0).
5 10.6.2
Let R be a finite integral domain. Then if x 6= 0 is an element of R, the set {x, . . . , xn , . . .} must
contain some nonzero duplicate xn = xn+k since R is a domain. Therefore 1 = xk = (x)(xk−1 ) and
so all nonzero elements of R have inverses, i.e. R is a field.
6 10.6.5
If R is a domain of order 10, note that R must have the additive structure of Z/10Z which is the only
abelian group of order 10. It follows from the Sylow Theorems that there are nonzero elements x, y
such that x+x+x+x+x = y +y = 0. By the distribution property, x(1+1+1+1+1) = y(1+1) = 0
and since R is a domain, 1 + 1 + 1 + 1 + 1 = 1 + 1 = 0 ⇒ 0 = 1 ⇒ R = {0} which is nonsense.
Therefore such a ring does not exist.
NOTE: Finite fields necessarily have order that is the power of a prime (I required you to prove
this if you used it), and not just prime order.
7 10.7.1
We have that all ideals in Z are principal, and that Z/nZ is a field exactly when n is prime. Therefore
corollary (7.3) implies the desired result.
1
8 10.7.7
By Cor. (7.3) this amounts to showing whether the appropriate ideals are maximal. Consider
I = (x3 + x + 1) = (f ) ⊂ F2 [x]. From class, F2 [x] is a PID, and so I is maximal iff (f ) is an
irreducible polynomial. However, since it has degree 3, it would have to have a linear factor and
therefore a root in F2 ir it were reducible. However f (0) = f (1) = 1 and so I is maximal. Now
consider I(x3 + x + 1) = (x − 1)(x2 + x − 1) in (F3 [x] and so I ⊂ ((x + 2)) and is therefore not
maximal.
9 10.7.9
Define the bijective correspondence by identifying the plane with C and identifying the ideal (x − p)
for any point p on the real line and the ideal (x − z)(x − z) with any point z = a + bi with b > 0.
Because R[x] is a principal ideal domain, these ideals are clearly maximal since they are generated
by irreducible polynomials; it is easy to check that any maximal ideal in R[x] has this form.
10 10.7.10
Let R, M be as given and suppose N is an ideal properly containing M . Then by hypothesis, N
contains a unit and so N = R. Therefore M is maximal. Suppose N is a maximal ideal of R. Then
N must contain no units and so N ⊆ M ⇒ N = M . Therefore M is the unique maximal ideal of
R, and R is called a local ring with maximal ideal M .
11 10.8.11
Define f1 = x2 + y 2 − 1, f2 = x2 − y + 1, andf3 = xy − 1.
(a) Set all three polynomials equal to 0. Then subtracting f2 from f1 gives y 2 + y = 2. This
gives y = 1 or y = −2.√In the first case, we have by f2 that x = 0 and then f3 6= 0. The second
implies by f2 that x = 3i and then again f3 6= 0. So the polynomials have no common zeroes.
(b) The following polynomials, supplied by Edo Gallo, are such that p1 f1 + p2 f2 + p3 f3 = 0:
p1 = x − xy + 1, p2 = y − x + xy + 1 , p3 = y − x − 1.
2
Math 123: Abstract Algebra II
Solution Set # 1
Gregg Musiker
February 11, 2002
1 Chapter 11, Section 1

1.1 Problem 1
a + b = p where p is prime, a, b, p > 0. Assume d|a and d|b. Then d|a + b so
d|p. Since p is prime d = 1 or p. but p|a and p|b implies a, b ≥ p. But then
a + b > p. ⇒⇐
1.2 Problem 8
a) In F2 [X], if x = 0 or x = 1, x3 + x + 1 = 1 thus x3 + x + 1 = 1 has no
linear factors. A reducible cubic must factor into three linear irreducibles or a
quadratic and linear irreducible. Since it has no linear factors, x3 + x + 1 = 1
is irreducible. √
b) By the quadratic formula, x2 − 3x − 3 has the roots x = 3±2 21 so in F5 [X],
√
it has the roots x = 3±2 1 = 1 or 2 thus x2 − 3x − 3 = (x − 1)(x − 2).
√
c) By the √quadratic formula, x2 + 1 has the roots x = ± −1 since F7 does not
contain ± −1, (Try squaring 0, . . . 6 modulo 7 or apply quadratic reciprocity,
etc.) x2 + 1 is irreducible in F7 [x].
1.3 Problem 9
a) Suppose f1 , f2 , . . . , fn are all of the monic irreducible polynomials of F [x].
Let g = f1 f2 · · · fn + 1. fi 6 |g∀i. Since F [x] is a U.F.D. (in fact it is a Euclidean
domain by Proposition 2.18, which implies it is a P.I.D. and thus a U.F.D.), one
can factor g into irreducibles, hence ∃fa s.t. fa is irreducible and fa |g. fa must
be different than any of the initial fi since fi cannot divide both N = f1 f2 · · · fn
and N + 1. Also, N + 1 is not a unit because deg(fi ) ≥ 1 so deg(N + 1) ≥ 1
but units in F [x] have degree zero. In fact x is the only prime in F [[x]] up to
associates.
b) This fails for F[[x]] because the last line fails (i.e. there exist irreducibles
fi that DO divide both N = f1 f2 · · · fn and N + 1) This works because x is
1
irreducible and x + 1 is a unit in F[[x]] thus N = x implies N + 1 = x + 1 but
x + 1 is a unit and thus there are no extraneous irreducibles dividing it.
1.4 Problem 14
Let f (x) ∈ F [x] have a multiple root (f (x) = (x − α)2 g(x). Then the derivative
of f ,
f 0 = 2(x − α)g(x) + (x − α)2 g 0 (x).
Assuming f is irreducible in F [x], gcd(f , f 0 ) = 1 or f . (F is a field thus F [x]
is a Euclidean domain, and thus gcd is well-defined.)
f 6 |f 0 because deg(f) ≥ deg(f’). Thus gcd(f , f 0 ) = 1 hence ∃h1 , h2 s.t.
f h1 + f 0 h2 = 1 for all x (*).
But since F is a subfield of the complex numbers, it is valid to let x = α ∈ C
but f (α) = f 0 (α) = 0 and thus (*) cannot hold. Thus f must be reducible if it
is has a multiple root and if f is irreducible, it cannot have a multiple root.

2.1 Problem 2
2πi
a) Prove that R = Z[ζ], ζ = e 3 is a Euclidean domain
Picturing R as a lattice (as Artin does for Z[i])√ in the complex plane where
formed by additions of the vectors (1,0) and (− 3/2, 1/2). Thus the grid will
be composed of rhombuses of base length 1, and perpendicular height 1/2. The
2
distance
√
from the center (worst case) of a rhombus to any of the corners is
3
1 − 4 < 1. Thus if one looks at the multiples of x ∈ R one looks at a sub-
lattice and find that the remainder will have norm less than the norm of x.
√
b) Prove that R = Z[ −2 is a Euclidean domain
Again, this can be proved algebraically or geometrically. Proceeding √
geometri-
cally, one find the grid made√up of rectangles of length 1 and width 2. Since
the center is a distance of 3/2 from any of the corners, by logic similar to
example (a), N (r) ≤ 3/4N (x).
2.2 Problem 4
if d = gcd(a, b) ∈ R (R is a PID), then ∃c1 , c2 ∈ R s.t. d = ac1 + bc2 . Assume
a, b ∈ Z, and d is the gcd in Z, d0 is the gcd in Z[i].d0 |a, d0 |b ⇒ d0 |d in Z. In Z[i],
there also exist c3 , c4 ∈ Z[i] s.t. ac3 + bc4 = d0 . d|a, d|b ⇒ d|d0 inZ[i]. Thus d
and d’ are associates in Z[i] (Note: gcd is only defined up to associate).
2.3 Problem 5
Assume R is an integral domain. thus if xy = 0, x = 0 or y = 0. Let p be a
prime in R. p|ab implies p|a or p|b. assuming p is reducible, p = vw where v, w
2
are not units in R.
p|vw so p|v or p|w. W.L.O.G. assume p|v. Then there exists a non-unit x s.t.
px = v. Thus p = vw = pxw and p(1 − xw) = 0.
Since p 6= 0, and R is an integral domain, 1 − xw = 0. thus 1 = xw which means
w is a unit, thus contradicting our assumptions. Thus all primes in an integral
domain are irreducible.
(NOTE: important to show that R must be an integral domain for primes to be
irreducibles.) The cancellation law is equivalent to the statement p = pxw ⇒
1 = xw and only holds in an integral domain. Also, the Artin defines irreducible
and associate only in an integral domain. To see what goes wrong outside an
integral domain, consider 3, 9 ∈ Z/(12). 3 ∗ 9 = 3 thus 3|9 and 9|3 but 3 and 9
do not differ by a unit. Also 3 is clearly not irreducible nor even reducible into
finitely many irreducibles.
2.4 Problem 8
Find the gcd of (11 + 7i, 18 - i) in Z[i].
N (11 + 7i) = 160, and N (18 − i) = 325. So dividing 11 + 7i into 18 - i, we
get (1 - i) with a remainder of 3i.
Since N (3i) = 9 ≤ 1/2N (11+7i), we have found a valid version of b = aq +r.
Dividing 3i into (11 + 7i) (since gcd(a,b) = gcd(a,r) if N (b) > N (a)) we get (-4i
+ 2) with a remainder of (-1 + i).
N (−1 + i) = 4 ≤ 1/2N (9i). Dividing (-1 + i) into (3i), I get (-i + 1) with a
remainder of i. Since N (i) = 1 ≤ 1/2N (−1 + i), this is a valid remainder.
In fact, N (i) = 1, means that i is a unit and thus gcd of (11 + 7i, 18 - i) is
1 up to a unit.

3.1 Problem 3
• Proof 1. Let f, g ∈ C[x, y] s.t. f is irreducible and suppose V(f) ⊂ V(g).
By the classical Nullstellensatz (Artin Thm 10.8.7) since g is identically zero
on the set V(f), g n ∈ (f ) for some positive integer n. Thus f |g n . Since f is
irreducible, f |g.
(NOTE: f |ab ⇒ f |a or f |b requires f to be prime, and irreducible only implies
prime in a PID which C[x, y] is not. However, in the special case of f |g · · · g,
irreducibility is enough:
Assume not, that there exists positive integer j s.t. f |g j+1 but f 6 |g j . Then
writing out the RHS as irreducibles, g · g j = f h1 · · · hn Since C[x, y] is a UFD,
g must factor into associates of a subset of f, h1 , . . . , hm . Cancelling those from
both sides, g j = hi1 · · · him where f does not equal any hij otherwise f |g j .
However, this implies f |g.)
3
• Proof 2. If f is a constant, f |g trivially. Any non-constant bivariate complex
polynomial has infinitely many zeros. To see this, imagine letting fα (x) =
f (x, α) and vary α to get an infinite number of zeros. Thus g has infinitely
many zeros, hence f and g share an infinite number of zeros. By (Artin Prop.
10.8.8), f and g have a nonconstant polynomial factor in common. Since f is
irreducible, f |g.
• Proof 3 (Thanks to Ed Dean). Let g(x, y) = p1 (x, y) · · · pk (x, y) be a
factorization of g into irreducibles. Then V (g) ⊂ ∪ki=1 V (pk ). Let I(V (f )) and
Qk
I(V (g)) be the ideal of these varieties. I(V (g)) = ( i=1 pk ) ⊂ I(V (f )). Since
f is irreducible, I(V (f )) = (f ) Hence f is an associate of pk thus f |g.
3.2 Problem 4
Let f (x), g(x) ∈ Z[x] ⊂ Q[x].
• Assume (f(x), g(x)) contains an integer c ∈ Z. Then ∃p(x), q(x) ∈ Z[x]
s.t.
f (x)p(x) + g(x)q(x) = c.
Thus ∃po (x), qo (x) ∈ Q[x] s.t.
f (x)po (x) + g(x)qo (x) = 1
and thus the gcd of f(x) and g(x) is 1.

• Assume f(x) and g(x) are relatively prime in Q[x]. Thus
f (x)po (x) + g(x)qo (x) = 1
and clearing the denominators of po (x) and qo (x), f (x)p(x) + g(x)q(x) = c, and
(f(x), g(x)) contains an integer c.
4
Solution Set 2 Math 123
February 20, 2002
1. Artin §11.3 #8
Let R = C[y, z, w], so we want to show that f (x) = xw − yz is irreducible in R[x]. Since
gcd(w, yz) = 1 (the prime factorization of yz is y · z), f is irreducible, so by Artin Prop.
11.3.6, since f is irreducible in C(y, z, w)(x) (it is linear), f is irreducible in R.
2. Artin §11.4 #11
a) We have f (x) = x2 + 27x + 213 ≡ x2 + x + 1 (mod 2); since the latter is irreducible in
F2 [x] (f (0) = f (1) = 1), f is irreducible in Q[x].
b) Since 3 does not divide 1, 3|6 and 12, and 9 does not divide 12, by Eisenstein’s Criterion,
x3 + 6x + 12 is irreducible in Q[x].
c) We have f (x) = 8x3 −6x+1 ≡ x3 +x+1 (mod 7). Since f is cubic, if it is not irreducible
it has a linear factor and thus a root in F7 . We see by inspection that f (x) 6= 0 for all
x ∈ F7 , so f is irreducible.
d) We have f (x) = x3 + 6x2 + 7 ≡ x3 + x2 + 2 (mod 5). As above, f has no root in F5 , so

f is irreducible.
e) This f (x) = x5 − 3x4 + 3 is irreducible by Eisenstien’s criterion with p = 3.
3. Artin §11.4 #11
First we prove the following useful fact.
L EMMA 3.1. If M is any n × n complex matrix with M k = In for some k, then M is

diagonalizable.
P ROOF.
We will also consider M to be an operator on an n-dimensional complex vector space
V = Cn . For vectors v = (v1 , . . . , vn ) and w = (w1 , . . . , wn ), define hv, wiP
0 = v̄1 w1 +
n
· · ·+v̄n wn to be the standard complex dot product of v and w, so hv, vi0 = i=1 |vi |2 =
1 Taken from Noam Zeilberger’s solutions
1
P ROBLEM 4 2
|v|2 > 0. Now let

k−1
X
hv, wi = hM i v, M i wi0 .
i=0
We claim that h·, ·i is a positive-definite Hermetian form (see Artin 7.4 and 7.5). Clearly
h·, ·i is a Hermetian form, and for nonzero v ∈ V ,
k−1
X
hv, vi = |M i v|2 > 0.
i=0
Now, for any v, w ∈ V , hM v, M wi = hv, wi. Thus M is unitary, so M M ∗ = In =

M ∗ M , and M is normal. By Artin Theorem 5.4, then, M is diagonalizable.
❒
Let A be as in the problem, and let c(x) ∈ Z[x] be the characteristic polynomial of A.
Considering A as a complex matrix, find a basis for V = Cn with respect to which A is
diagonal. Since A 6= In , at least one diagonal entry (eigenvalue) λ is not 1; suppose that
v is a corresponding eigenvector (so Av = λv). Well, v = Ap v = λp v, so λ is a nontrivial
root of xp − 1—that is, λ is a root of the cyclotomic polynomial fp (x) = xp−1 + · · · + x + 1.
But since λ is an eigenvalue of A, we have c(λ) = 0 as well, so since fp (x) is a primitive
irreducible polynomial in Z[x], fp divides c, and n = deg c ≥ p − 1.
4. Artin §11.4 #12
As before, we know that any reducible cubic polynomial in F3 [x] has a root in F3 . Since the
arbitrary monic cubic is of the form f (x) = x3 + ax2 + bx + c, there are 27 such polynomials,
but nine of them (the ones with c = 0) have a root at 0, so we only have to test eighteen of
them. By listing all 18 polynomials and checking each at x = 0, ±1, we find that the only
irreducible ones are x3 − x + 1, x3 − x − 1, x3 + x2 − 1, x3 + x2 + x − 1, x3 + x2 − x + 1, x3 −
x2 + 1, x3 − x2 + x + 1, x3 − x2 − x − 1.
5. Artin §11.5 #3
Using the criteria for whether a prime in Z is prime in Z[i], we can find the prime factor-
ization for an element of Z[i] by looking at the integer prime factorization of its norm, and
trying various combinations of the factors. Thus we can calculate:2
a) (1 − 3i)(1 + 3i) = 10 = 2 · 5 = (1 − i)(1 + i)(1 − 2i)(1 + 2i). After some investigation,

we obtain 1 − 3i = −(1 + i)(1 + 2i).
b) From the previous part, 10 = (1 − i)(1 + i)(1 − 2i)(1 + 2i).
2 From Seth Kleinerman’s calculations.

P ROBLEM 6 3
c) We see immediately that 6 + 9i = 3(2 + 3i). Well, (2 + 3i)(2 − 3i) = 13 which is prime
and 3 6≡ 1 (mod 4), so 3 and 2 + 3i are prime in Z[i] and our factorization is complete.
6. Artin §11.5 #6
√
We would like to say that p is prime in R = Z[ 3] if and only if R/(p) is a field. This does not
follow immediately from by Artin, Proposition 11.2.14 (almost everybody did this wrong)—
that proposition only applies to principal ideal domains R, and it is not at all obvious that R
is a PID. One can prove without too much trouble that R is indeed a PID, but the following
is even easier:3 more or less by definition, p is a prime element if and only if (p) is a prime
ideal, which is true if and only if R/(p) is an integral domain, and since R/(p) is clearly
finite, this is true if and only if R/(p) is a field (recall that a field is an integral domain and
that a finite integral domain is a field).
Now we can proceed as in lecture. We have canonical isomorphisms
√
Z[ 3]/(p) ∼ = Z[x]/(p, x2 − 3) ∼
= Fp [x]/(x2 − 3).
Since Fp [x] is a principal ideal domain, we can now use Artin Prop. 11.2.14 to conclude that
R/(p) is a field if and only if x2 − 3 is irreducible in Fp [x].
7. Artin §11.6 #3
√ √ √
We know that if n ∈√ Q is not square in Q then n ∈ / Q. Suppose
√ that d0 ∈ Q( d). Since
every element of Q( d) = Q[x]/(x2 − d) is of the form a + b d for a, b ∈ Q, we can write
√ √
d0 = a + b d (1)
p
for a, b ∈ Q. Now, if a = 0 then we have d0 /d = b ∈ Q, which is a contradiction since d0 /d
0
√ not a square in Q since d and d are distinct squarefree integers. So a 6= 0, and b 6= 0 since
is
0
d is irrational; thus squaring both sides of (1) gives
√ √ d0 − a2 − b2 d
d0 = a2 + b2 d + 2ab d =⇒ d =
2ab
√ √ √ √
0 / Q( d) and Q( d) 6=
√ is contradiction since d is irrational as well. Thus d ∈
which
Q( d0 ).
8. Artin §11.6 #7
√
Let R = Z[η] be the ring of algebraic integers in Q( d) for some squarefree integer d < 0.
We know that R = {a + bη | a, b ∈ Z} is a lattice in C since η is not real; we wish to show
that R is maximal, i.e. that there is no strictly larger lattice that is also a subring.
3 Thanks to Richard Louis Rivero for this one.
P ROBLEM 9 4
Let S ≤ C be a ring containing R that is also a lattice. By Artin Corollary 5.4.15, we

can assume that S is generated (as a lattice) by 1 and another element α—there can be no
element of S with norm less than 1, since otherwise there would be an infinite number of
elements in the unit circle and S would not be a lattice. Since S is a ring, we can write
α2 = bα + √ c for b, c ∈ Z, so α is√an algebraic integer satisfying x2 − bx − c = 0. Thus
α = 2 (b ± b2 − 4c), so α ∈ Q( d0 ), where d0 is the squarefree part of b2 − 4c. Now, if
1
√ √ √
d 6= d0 , then since d ∈ / Q( d0 ) ≥ S by the previous problem, we would have d ∈ / S
0
which
√ would contradict the fact that R ≤ S. Thus d = d and α is an algebraic integer in
Q( d), which means S = R, and R is maximal.
9. Artin §11.7 #4
First we note that for any nonzero α ∈ A, αᾱ ∈ A is a positive integer, so A ∩ Z 6= 0. Select
the smallest positive integer n in A and apply Artin Corollary 5.4.15 to conclude that there
must be a lattice basis for A one of whose elements is n.
10. Artin §11.7 #5

√
Let δ = −5 and let L be the lattice generated by (3, 1 + δ). By definition, L is an abelian
subgroup of R; we must show that RL = L. It suffices to show that L is closed under
multiplication by δ, and by linearity, all we have to show is that 3δ and δ(1 + δ) are elements
of L. Indeed, 3δ = 3(1 + δ) − 3 and δ(1 + δ) = δ − 5 = (1 + δ) − 2 · 3.
By drawing a picture, one can quickly convince oneself that α = 1+δ has minimal absolute
value in L. Since
1 1
(α + αδ) = (2δ − 4) = δ − 2 = (δ + 1) − 3
2 2
we have L = (α, 12 (α + αδ)) (note that (δ + 1) − (δ − 2) = 3).
Joe Rabinoff
MATH 123: ABSTRACT ALGEBRA II
SOLUTION SET # 3
GREGG MUSIKER
1. Chapter 11, Section 6

Problem 8
a) Let S = Z[α], where α is a complex root of a monic polynomial of degree 2.
Prove that S is a lattice in the complex plane.
Let f (x) = x2 + bx + c ∈ C[X] have α as a root. By Prop. 10.5.7 {1, α} or {1}
is a basis for S. This is because α2 = −b · α − c hence further powers of α can be
rewritten in terms of α. (Many people got points off for missing this step.) Thus
S = {1 · Z + α · Z}, and is a lattice by definition.
b) Prove the converse, a subring S of C which is a lattice has the form given in (a).
Since S is a subring of C, 1 ∈ S. Since S is a lattice, suppose S = α · Z + β · Z
where α, β ∈ C. Let mα + nβ = 1. Then zmα + znβ = z ∈ S for any integer z,
thus Z ⊂ S. Furthermore, β can be rewritten as n1 − m k
n α. Assuming n > 1, β ∈ S
1+γ
for any k, and consequently, nk ∈ S where γ ∈ Z[α]. However, this means that S
is not discrete even though S is a lattice. ⇒⇐
Another way of saying this is that 1 ∈ S is a basis element of the lattice for the
following reason: if 1 were not of minimal norm, there would exist γ s.t. |γ| < 1 but
then taking powers of γ, one gets arbitrarily small elements, which again contradicts
the fact that S is discrete.
Many people got points off because they asserted that 1 is a basis element without
proving it.
Thus mα + β = 1, and β can be rewritten in terms of the basis {1, α}. S is a lattice
so it can only have two basis elements. Lastly, α2 ∈ S which implies there is a way
to write α2 = (−b)α + (−c) for integers b and c. Thus f (α) = 0.

√
Problem 7.d Use the method of Prop (7.9) to describe the ideals in Z[ −7].
Like the proof of Theorem 11.7.9, let A be a nonzero ideal of R and let α ∈ A be
a nonzero element of A of minimal absolute value
√ d. The principal ideal (α) = Rα
contains complex numbers
√ of the form (a + b −7) · α where a, b ∈ Z. Thus it has
the lattice basis (α, α −7). A contains (α) thus A = (α).
Let A be an ideal that contains more than (α), e.g. β ∈ A but √ β 6∈ (α). Using√ the
same trick as Artin, we create a rectangle with vertices 0, α, α −7,
√
and α√
+ α −7.
Letting n = 2, we draw disks of radius 2r around the points α 2−7 , α+α2 −7 , and
√
α + α 2−7 and also draw disks of radius r around the four vertices of the rectangle.
1
2 GREGG MUSIKER
These disks will cover the entire rectangle. β must be in the rectangle by Lemma
5.4.14. By Lemma 11.7.11, β must be one of the points at the center of these disks
and β is not in (α), hence β is one of the three points in the center.
√ √
Assume α 2−7 ∈ A. Multiplying by −7, −7α α
2 ∈ A hence 2 ∈ A. But α was chosen
to be the element of minimum absolute value. ⇒⇐
√ √
α + α 2−7 ∈ A also implies α2 ∈ A which leaves us with the fact β = α+α −7
2 ∈ A.
√
Consequently, A = (α, β) = (α, α 2−7 ).
√
Problem 7.f Use the method of Prop (7.9) to describe the ideals in Z[ −10].
This proof follows the same logic as 7.d except one use the three original circles
and
√ needs to√add eight circles of radius r/3 around points with coordinates in
( −10α/3, 2 −10α/3) × (0, α/4, 3α/4, α). Otherwise, the circles will not cover
the rectangle. √
1
√ Then analyzing the cases , one finds that if β = −10α/3, then multiplying by
−10, −10α/3 ∈ A ⇒ α/3 ∈ A. This contradicts the fact that α was chosen to be
the element of minimal norm. Using a similar √ technique one finds that most of the
centers of the circles cannot be β ∈ A. But, 12 −10 · α will work.
√
Thus the possible ideals are multiples of (α) and (α, 21 −10 · α).
√
Problem 9 Let d ≤ −3. Prove that 2 is not a prime element in the ring Z[ d],
but that 2 is irreducible in this ring.
If 2 was prime, then 2|ab ⇒ 2|a OR 2|b.
√ √
If d is odd, let 2α = (1 + d)(1 − d) = 1 − d. Thus α = 1−d
2 ∈ Z and ≥ 2 for
d ≤ −3. √ √
and 2 does not divide (1 + d) or (1 − d).
√ √ √ √
If d is even, let 2α = (2 + d)(2 − d) = 2 − d OR d · (− d).
However, 2 is irreducible
√ because N (2) = 4 = N (α)N (β) if αβ = 2.
N (α) = N (a + b d) = a2 + b2 |d| > |d| ≥ 4 whenever b 6= 0 and a 6= 0. If b 6= 0, α is
an integer, but 2 is irreducible over the integers. If a 6= 0, α2 is a multiple of d ∈ Z
which will not divide 2. We would need N (α) = 2 so that N (α)N (β) = 2 · 2.

√
Problem 1 Let R = Z[ −6]. Factor the ideal (6) into prime ideals explicitly.
√ √ √
√ = ( −6)(− −6). For a common refinement,
(6) = (2)(3) √ √consider A = (2, −6),
B = (3, −6). In fact, A = A and B = B since − −6 ∈ ( −6).
Now we need to show A and B are prime ideals. The usual way to prove P is a
prime ideal is to prove R/P is a field. In this case R/A ∼ = F2 , R/B ∼= F3 . Another
way to do this is to show the norms of elements in this ideal are prime, thus the
elements will be irreducible. √
Also, we can show by linear identities that A2 = (2), B 2 = (3), and AB = ( −6).
So (6) = A2 B 2 .
1Thanks to Elizabeth Schemm

MATH 123: ABSTRACT ALGEBRA II SOLUTION SET # 3 3
√
Problem 2 Let δ = −3 and R = Z[δ]. (This is not the ring of integers in the
imaginary quadratic number field Q[δ].) Let A be the ideal (2, 1 + δ). Show that
AA is not a principal ideal, hence that the Main Lemma is not true for this ring.
AA = (4, 2 + 2δ, 2 − 2δ) since the product of two ideals is the ideal which is the
product of its generators. Clearly AA ⊃ (4) and 2 divides all the generators so
(2) ⊇ AA. But since 2 6∈ AA by virtue of the fact no linear combination (using
integers and integer multiples of δ) of the generators equals 2.
√
Problem 3 Let R = Z[ −5]. Determine whether or not 11 is an irreducible
element of R and whether or not (11) is a prime ideal in R.
To prove 11 is irreducible, use the same method as problem 11.7.9 . Namely a
2
√ = 11 ⇒ N (α)N (β)2 = 112 ⇒ N (α), N (β) = 11.
nontrivial factorization αβ
However, if α = a + b −5, then N (α) = a + 5b 6= 11 for any choice of a, b.
To prove (11) is a prime ideal, here are two good strategies:
• Any ideal is a sublattice. One can show using norms, and geometry, (11) is a
maximal sublattice which implies it is a prime ideal.
√
• We could show R/(11) is a field. R/(11) ∼= Z[ −5]/(11) ∼ = Z[x]/(x2 + 5, 11) ∼
=
2 2
F11 [x]/(x + 5). This last ring is a field ⇔ x + 5 is irreducible modulo 11. By
inspection, one can find 6 ≡ −5 is not a square modulo 11.
SOLUTION SET # 4
GREGG MUSIKER

Problem 5 Prove that A ⊃ A0 implies that AB ⊃ A0 B.
P
AB = Pi αi βi where αi ∈ A and βi ∈ B.
A0 B = i γi βi where γi ∈ A0 and βi ∈ B.
Since γi ∈ A0 ⊂ A, A0 B ⊂ AB.
Problem 6√ Factor the principal ideal (14) into prime ideals explicitly in R = Z[δ],
where δ = −5.
(14) = (2)(7) in the integers. Is there another way to factor 14 in Z[δ]? In fact
there is: by √looking at√the equation 14 = a2 + 5b2 , a = 3, b = 1 is a solution. So
(14) = (3 + −5)(3 − −5).
Ideal factorization
√ will allow us√to have a common refinement, letting
A = (2, 3 + −5), B = (7, 3 + −5), (14) = AABB = ABAB.
√ √
Proof that this is the right factorization. AA = (4, 14, 6+2 −5, 6−2 −5).
By the main lemma, AA = (n), an integral principal ideal. 2 divides every generator
so AA ⊂ (2). 14 − 3(4) = 2 so AA ⊃ (2) implies AA = (2). Similar logic proves
BB = (7).
Proof that this A, A, B, B are prime. Since AA = (2) and BB = (7), the
norm of these two ideals are prime, thus they are prime ideals.
Problem 7 Let P be a prime ideal of an integral domain R, and assume that
existence of factorizations is true in R. Prove that if a ∈ P then some irreducible
factor of a is in P .
By the definition of prime ideal given in section 9, if α, β ∈ R s.t. αβ ∈ P then
α ∈ P or β ∈ P .
If a is irreducible, we’re done (a ∈ P ) so assume a is reducible. Then a = α1 α2 · · · αn
as an irreducible factorization. (Note: This is not the factorization, it is a possible
factorization, R need not be a UFD.)
By repeated application of the definition, α1 or α2 or . . . αn ∈ P .

Problem 2 Let d = −14. For each of the following primes p, determine whether
or not p splits or ramifies in R, and if so, determine a lattice basis for a prime
ideal factor of (p) : 2, 3, 5, 7, 11, 13.
By proposition (11.9.3), p stays prime in R if and only if x2 + 14 is irreducible
mod p.
1
2 GREGG MUSIKER
Modding by p one gets the following equations: √ √

If p = 2, x2 ≡ 0 ≡ x · x mod 2. In fact (2) ramifies as (2) =√(2, −14)(2, √−14).
If p = 3, x2 + 2 ≡ 0 ≡ (x + 1)(x + 2) mod 3. (3) = (3, 1 + √−14)(3, 1 − √−14).
If p = 5, x2 + 4 ≡ 0 ≡ (x + 1)(x + 4) mod 5. (5) = (5, 1 + −14)(5,√ 1 − √−14).
If p = 7, x2 ≡ 0 ≡ x · x mod 7. In fact (7) ramifies as (7) = (7, −14)(7, −14).
If p = 11, x2 + 3 ≡ 0 mod 11 is irreducible by inspecting 12 through 62 . Thus
(11) stays prime. √
2
√ If p = 13, x + 1 ≡ 0 ≡ (x + 5)(x + 8) mod 13. (13) = (13, 5 + −14)(13, 5 −
−14).
One must check these ideal equivalences on the right. One can do this by the
method problem 11.8.6 for example. Many people noticed the relation between the
factors of x2 + 14 and the lattice basis of the ideal factorization.
Proof√ of this correspondence.
√ √ Let p ≡ √ 2 or 3 mod 4. Suppose (p) =
(p, a + b d)(p, a − b d) = (p, p(a + b d), p(a − b d), a2 − b2 d) = AA. If a2 − b2 d
and p are relatively prime, 1 ∈ AA which is a contradiction, thus a2 − b2 d ≡ 0
mod p. Assuming
√ that√b = 1 like in all the above examples, a2 − d ≡ 0. Thus
(p) = (p, a+ d)(p, a0 + d) only if there are a, a0 that satisfy the equation a2 −d ≡ 0
mod p. It turns out that a and a0 are the two roots of √ this equation, √
which will be
additive inverses of each other. But the ideal (p, −a + d) = (p, a − d).
Problem 3 (a) Suppose that a prime integer p remains prime in R. Prove that
R/(p) is then a field with p2 elements.
Two good ways to this:
1) Since (p) remains prime ideal, and maximal since we are dealing with the ring
of integers. However, in the ring of integers, prime ideals are maximal so R/(p) is
a field. The number of elements is p2 by counting lattice points. Also, one could
use the fact that R/(p) is a finite integral domain which is a field.
2) Since p stays prime, by Prop 11.9.3, g(x) = x2 −d (x2 −x+ 14 (1−d) respectively
if d ≡ 1 mod 4) is irreducible over Fp . But letting F = Fp [x]/(g(x)) = Fp [α] such
that g(α) ≡ 0. F will be a field made up of linear combinations of a + bα for
a, b ∈ Fp . p2 such elements.
Problem 3 (b) Prove that if p splits in R, then R/(p) is isomorphic to the product
ring Fp × Fp .
By proposition 11.9.3, g(x) ≡ (x − a)(x − b) for a, b ∈ Fp . One can create an
isomorphism between R/(p) and Fp × Fp by the following map:
Let φ : R/(p) = Fp [x]/(g(x)) → Fp [x]/(x − a) × Fp [x]/(x − b) by φ(w) = (y, z)
such that y is the remainder when we set x ≡ a and z is the remainder when we
set x ≡ b. Notice that either of these equivalences implies g(x) ≡ 0.
And Fp [x]/(x − a) ∼
= Fp since setting x ≡ a an element which is already in the
field doesn’t change Fp .
Problem 4 Let p be a prime which splits in R, say (p) = P P , and let α ∈ P be
any element which is not divisible by p. Prove that P is generated as an ideal by
(p, α).
√ √ √
Let α = a + b d. Then P P − (p2 , p(a + b d), p(a − b d), a2 − b2 d). If we want
(p) = P P , we need p divides all the generators (or alternatively no element outside
of (p) in P P ) and p ∈ P P .
α ∈ P implies that a2 − b2 d = αα ∈ P P = (p). Since p clearly divides all the

other three generators, we know have that p divides all the generators. Thus no
element which p doesn’t divide could be in P P .
Now we can run into a problem if p does not equal a linear combination of the
generators. Then p 6∈ P . Notice we already used the condition α ∈ P . Here we use √
p does not divide α. Assume p 6= 2 and p does not divide d. Then p 6 |α = a + b d
implies that p 6 |a or p 6 |b.
If p 6 |a, then C = p(α + α) = 2ap is not divisible by p2 since p 6= 2. Thus
gcd(p2 , C) = p over the integers which means there is a Z-linear combination such
that p ∈ R.
√
If p 6 |b, then C 0 = p(α − α) d = 2bdp is not divisible by p2 , thus gcd(p2 , C 0 ) = p
over the integers.
Here we need additional reasoning for the cases of p = 2 and p|d.
However there is another way1 to do this without the need to break it into cases.
Let P = (p, α). Then (p2 , pα, pα, N (α)) = P P = (p2 ) or (p) by Prop (9.1). We
require P P = (p) by the given hypotheses. Assume P P = (p2 ). Then since P is a
prime ideal (factorization into prime ideals is unique even when R is not a UFD),
P = (p). But then (p, α) = (p). It is clear that p ∈ P but P ⊂ (p) only if p divides
all the generators. But by hypothesis, p 6 |α. ⇒⇐
Problem 7 Assume that d ≡ 2 or 3 mod 4. Prove that a prime integer p ramifies
in R if and only if p = 2 or p divides d.
√ √ √
if p = 2 and d ≡ 3 mod 4, then (2, 1 − d)2 = (4, 2 − √ 2 d, 1 + d − 2 √d) = (2) since
d ≡ 1 mod 2 implies 2 divides all generators and (2−2 d)−(1+d−2 d) = 1−d ≡ 2
mod 4.
Some students tried to use the fact that (p) ramifies if and only if x2 −d = (x−α)2 .
However, one can only use this fact if one shows it.
√
if p|d
√ (notice√ this covers√ the case d ≡ 2 mod 4 and p = 2) then (p, d)2 =
(p, d)(p, − d) = (p2 , p d, d) = (p) since p|d implies that p divides all the gener-
ators, and d square-free implies that gcd(p2 , d) over the √ integers is p which means
that there is a Z-linear combination such that p ∈ (p, d)2 .
√ there exists a p√6= 2 such that
Converse: Assume √
(p) = (p, a + b d)2 = (p2 , p(a + b d), a2 − 2ab d + b2 d) = A.
Then A ⊂ (p) only if p divides all the generators, so p|a2 + b2 d and p|2ab. p 6 −2
so p|a or p|b. If p|a, p|a2 + b2 d implies p|b2 d which implies p|b or p|d. And p|b
implies p|a. But p|a and p|b implies that p2 divides all the generators and thus
p 6∈ A. ⇒⇐ Thus it must be the case p|a and p|d.
Problem 11 Prove Proposition (9.1), i.e. If P is a nonzero prime ideal of R, then
there is an integer prime p so that either P = (p)orP P = (p) and that conversely
there is a prime ideal P of R so that either P = (p) or P P = (p).
This proof followed reasoning from class on February 22nd. If N (P ) = p, then
P P = (p). If not assume N (P ) = p1 · p2 · · · pn . However, if n > 2 then there exists
a factorization of P P into further prime ideals. ⇒⇐ Hence, N (P ) = p1 · p2 and
N (P ) = N (P ) implies N (P ) = p2 and P P = (p)(p) and P = (p).
1Thanks to Elizabeth Schemm

4 GREGG MUSIKER
For the converse, (p) − P1 P2 . . . Pn . p = p implies N (p) = p2 = N (P1 ) . . . N (Pn ).

Thus there are at most two prime ideals dividing (p). If there is one, P = (p). If
there are two, N (P1 )N (P2 ) = p2 implies N (P1 ) = N (P2 ) = p and thus (p) = P P .

Problem 3 Let R = Z[δ] where δ 2 = −6.
Problem 3 (a)Prove that the lattices P = (2, δ) and Q = (3, δ) are prime ideals
of R.
See Problem 11.8.1 on PS #3.
Problem 3 (b)Factor the principal ideal (6) into prime ideals explicitly in R.
See Problem 11.8.1 on PS #3.
Problem 3 (c)Prove that the ideal classes of P and Q are equal.
P ∼ Q if there exists τ and σ such that τ P = σQ. In this case let τ = δ and
σ = −2. τ P = (2δ, −6) = (−6, 2δ) = σQ.
Problem 3 (d)The Minkowski bound for R is [µ] = 3. Using this fact, determine
the ideal class group of R.
First some review: The Minkowski bound√can be found by the computation,
µ < π4 ∆(R). Since 6 ≡ 2 mod 4, ∆(R) = 6 the area of the fundamental rectangle.
Now using this bound, we need to look at the prime ideals dividing primes ≤ 3,
namely 2 and 3. Since these primes were already analyzed, there are a priori five
equivalence classes of ideas: < (1) >, , , < Q >, and < Q >. However,
(2) and (3) both ramify so P = P and Q = Q. Furthermore, by part (c), P ∼ Q.
Thus there are two classes: < (1) > and . The only group structure a
two-element group can have is C2 .

!#"%$'&)(*(
+,(
-/.0(
13254687789#:6;9=<>6;?@/<A
6;B2DCFEG'E7%6;BHEI*68?J@LK?7MNEI9:J25E%BHE%@J25687;4768COCF25CP68BF2 9:J2DCF6;G'25Q;E,R 2JRHST<VUXWYS<ZA[U\R
]^259V_a`2b6B28c*B25CO25?J96;9Ed52'KIeSHf< U#QXCOKgST<VUS#_hUbWiSjkmlFU\R'n=:J25?oST<U'WpST<ZAqUsr/t S<US_hUuW
ST<A[US_hUvWYSFjkmlFUwr/t <V_x6;?@<A_y68BF2cBE?4Ec*6;7zR
!#"%$'&)(*(
+,(
-/.Z{}|#~z\
WYhk
n=:JE%CE@2868747%6;CFC^BKc/ECX468747689#25@uE?uPB9#E%?bK?'c*68^25CX}
8**39#:25BF2E9HECXCO:K?'9#:6;93E%9
ECD9:J2L4M547%E4;BFKcuKI}KB@J25BXIKBQJ^25?28B689#25@`Me68?E%@J25687CFJ4:9:J689v WjT;lRP*259#9E?
Wj*kv ¢¡5l3:28B2£¡W¤ hkQ*2:6;d52
¥WYj¦*km§Ojk¨ ©¡5lOjk ª¡5llvWjT^l
COE%?J42«^¬3j¦*km§8lhW®RVP7CFKQÏBKG¯}BF9E?°2'±^?K²9:J689Zh³ECCOE%G'E768BZ9#K>´µW¶j#·O¡5lQ68?J@
COE%?J42¸h¹0E%C0c*BFE%?J4Ec*687Qhhº/E%C£CFEG'E%768B09#K /Rwn=:25CO2»ITKJBLE%@J256874768COC0B25c*BF25CO25?9#689#E%d825CL68BF2
@JB68?'E?'¼E%^BF2h}R*R
½) WYhk
¾
*E?42¸hk¿¾ªÀÁxjÃÂ/ÄlQvKB`K?J@aÅÃÆ)ÇeEC/§RbÈ'2b±^?KÉITBKG¶c*BK`725GÊ*R ËwÌ0ÍbK?ª6
c*BF25d5EKC:JKG'25KB±h9#:6;9j·8l3BF68G'EÎJ28CE%?J9Kb´h³OQ68?@b9:J689jT^l368?J@ªj§;l3@K0?K9=B6;G'E%IMR*E?42
Ï jTÐÑlDWÒÐJ³D Ók¿¾NECvBF25@J4E`7%2hE?»Ô º Qj;lDW¥ ÓCOc*7%E9#CFQ)68?J@bCOE%?J42 Ï jTÐÑlPE%CE%BFB25@J4E%`72LE?»Ô ÕOQ
j#§8l3B28G'68E?CE?25BF9QCFK02V468?'E;?JKB2hE9\R
n=:2 BF25CF9JKI*9#:2 68?J687M5COE%CEC}COE%G'E768BP9K9#:2 WYhkv468CO25Q;Ed525?£E%?£}BF9E?PR^È»E9:u¡LWo¤ hk
¾Q
2:6;d52jkv ©¡5ljkZw¡5lWYkÖLWÓ·V×m×mQV:JE%4:';Ed525Cv68?'25ØJ689#E%K?'E?uE@256;7%C
jkv ©¡5ljk w¡5lvWÓ´ ³ ³ ³8Ù
ÚDKQ9:J2Z?JKBGYKI3Û¢jk} ¡5lÑEC¨9#:2ZCO68G'268CÛ¢jk¨¡5l#QCFK£9:J2Nc*BEG'2ZI6849KBFE%Ü
689#EK?£KIjk} ¡5l
468?ª4K?J96;E%?Ý´ÞK?J7%MK?J42JRb*E?42u@K25C?JK9 @Ed5E%@J2¸k o¡°ß#EzR 2JRoºà jkV ¡5lNE%Ch?JK9 E%?ªKJB
BFE%?J;áQJV2h468?J?K93:J68d52h`K9#:Ó68?@ Ó6;CXc*BFE%G'2I6849#KBCXKIjkv Ý¡5lQâXCOK£4:KKCOE%?JuÓ68?J@
6;c*c*BKcBE689#257%M,Q246;?'4K?J47@29#:6;9
jkv ¢¡5lvWy´L ³ W3t SLU ³ WS´LU\ã à WS´LU Ù
n=:JC9#:2VE%@J25687J4768COCX^BKJc'äuE%Cv;28?25BF689#25@/`M£/RåN?2468?uØE4±^7Mh4K?47%J@29#:689¨´gEC?K9
c*BFE%?J4Ec*687`25468JCF2E%IH´¥WajTæX °ç¡5lÑ9#:28?'Û¢jæX »ç¡5lWoæ^³= wk¿¾5ç³NW·Q:E4:0:68C?JKLCFK79#EK?
è héêìë
í^îOézíï[î;ðîFñ#ïÃîOë#îOòJëó^ô ê*êõë#îTí^ö

÷PøJù}úû*ü
ý ýþPÿ
ý }ÿ ýFþPÿ

;ô ñmî= ¿^îíéêTêô ïÃîÑêõ^í^îOêé¿ëóî3ô òîOïÃê;ô ð !#"%$'&(õðézë¿ë#éê,TïÃî*)+
E?u9#:2E%?J928;25BFC
R ,/n=:JC yECX6;7%COK/?K93c*BFE%?J4Ec*687QÑCFK0CFE?420´h³vECXc*BFE%?J4Ec*687Q9#:2KB@J25BKIÓE%C
6;?@'268^68E?uÎJ?J@'9:J689Hä.-W0/21
Q;25?J25B6;928@0`MS#LU\R
}779#:6;9=B25G'6;E%?JCXEC9#K£468747689#2h9#:2VE%@J25687C
R*È'2CF2529#:6;93E%9#:´¥Waj·*kv ©¡5l3V2:J68d52
´ ´WYjTkmÖ*\·*jk ©¡5lO·jk ª¡5llXWaj·;l
COE%?J42£«^¬=jÃJkmÖ;l¨Wo·m)CFEG'E%768BF7%M,QJE%9#:WjTO·N ©¡5lH2:J68d52
¥WYj¦*\·k8Oj#· ©¡5lOj·Vª¡5llvWjT^l
COE%?J42«^¬3j¦*O·*kmlhW *RNP^68E?QD ³ ECCOEG'E%768BN9#K>´¥6;?@¸ º E%CCFEG'E7%6;BZ9#KZ9#:2/ITKJBNE%@J25687
CO:6;c*25C6;B2;Ed525?'E?u¼E;JB2}R }R
3 héêìë
í^îOézíï[î)ïÃêTéJðîFñ#ïÃîOë#îOòë%^ô êîOOê54XêìëzîÃí
÷PøJù}úû*ü
ý ýþPÿ6
8ý }ÿ ýFþPÿ67

;ô ñmî= ¿^îíéêTêô ïÃîÑêõ^í^îOêé¿ëóî3ô òîOïÃê;ô ð !#"89&(õðézë¿ë#éê,TïÃî*)+
!#"%$'&)(*(
+,(*(.0(
È»E9:¢¡»W ¤ ·e6;?@;:ÁW /=< ¡?>õQ2u468?@J25ÎÑ?2'6u?JKBGA@¨jTæ' Óç¡5l£WAB æ ³ x·;ç ³ BK?DC < ¡?>=:JE%4:
`254KG'25C¨6hCFEÜ
2hITJ?49#EK?'K?7: )?JK92V9:J6893IKBFEX8GIHJC < ¡?>2:6;d52K@¨j*ELG}lvWM@¨j*El%@¨j*G}l\R
];259}Ð}O NPH7:ÝE9Q: NJ0 W R SUR Tv25ÎÑ?2WV¿AJWXY;)CFJc*c*KCO2WV¿AÑWMV¿àA ZV¿³A ¡E9:PV¿àA 8V¿³A H.CRT25ÎJ?J2[V\9K
`29:J247%KCO25CO93E?9#25;28B¨9QK V¿\A ß#IK^B ]¨W¥k;O·áQ6;?@'725_9 VhM W Và D V ³ R*];289 `¿AÑa
W V¿AZ
VhM
W `
Aà b
`¿³A ¡Q*COK
@¨*j ` A lW cc ` Aà ³ ¸· ` ³A ³ cc d ` Aà ³ Ý· ` ³A ³ d
COE%?J42 `¿àA O `¿³A d ³à RÚDKQ*2468?'BFE%9#2
VW
Ð Z ` A W3t ÐM W VeNhI `
N
E%9#Z : `»f W N`¿ATRZÚDK^Q @¨*j `¿l0g W @¨*j NOl @¨hj `¿AÃjl ik@¨*j N*lvCOE%?J472 @¨hj `¿AÃKl iÉkQ}6;?Z @ `lHm:COE%?J42ÐH VeNbHn:eR
n=:JoC :xEC[6 pÑ47%E@E68?@KG'68E?PR
÷PøJù}úû*ü
!#"%$'&)(*(
+,(*(.rq
¼EBCO9v2bcBKd52'6;28?25BF687IT6;49hß#V:JE%4:ECc*BKd5E@28@:28B2uG'6;E%?J7%MITKBZBF25IT28B25?J42'68?J@>`28468JCF2Qs
68CDE?9#25B25CO9#25@}E%9EC?JK99254:?JE%46877%M£c*68BF9}KI}9:J2Lc*BK`725G'á8RH];289X¡/W ¤ ITKB J t SCFØ6;B25IB2528Q
72592:`209:J2£BFE%?JuKI}6;7%^25`BF68E4hE%?J925^25BFCNE%; ? C < ?¡ >óQH6;?@ITKuB EyWÞæe yçv¡ Hb:@J25ÎJ?J2 EyWÞæ/Ýç¡^R
ÚDK929#:6;9 :xwzy°R
{ ?û |} XRT= R ~# " F8$'"$ Ñ$"!%~ J~ "%#~ )8 U* =5$J5O ;!6 ?u! $J"%! E 8J~ '! ;! EIiÒk 8$
G;H.~ GIizE G t k H"+5+ E " X! 8 =6 O!H6 6 '5$J!=! 8!3" ^6 X! 8$ k +
÷PøJù}ù R
È'246;7%7J9:JE%C E9#:2 j ' ¡£¢ ¤¥¡FR
];28§9 ¦¨ :ª© «¬yb³ `2e9#:2e28G'`25@@JE%?J0K§I :oE?9#Ku9:J2hj PO ®l ¯Tc*7%6;?2JRÈ'2£?JK92e9:J68F9 ¦
*j ELG}lZW
j EG EG¨luWs*j ELG E GPl#QXCOKG'J7%9#E%c7%E46;9EK?¢ED ? :ÞE%Ch4KKB@E?6;92 ¯ÃE%CO2G'J7%9#Ec*7%E46;9EK?¢Ez ? yb³¿R
n=:JCE%Î EX8 G68BF2LE?9#:2hÎÑBCO9}ØJ68@BF68?J9Q)E,R 2J#R Ev EO G G t SmQÑ9:J25J ? ELGECD9KKQÑCOJ K ~_DECD47KCF28@
J?@28B¨G'79Ec*7E%4689#E%K?u6;?@'9#68±Ê%?JeE?d528BCO25C
R
ÚDKa72529 E°Hm~ `2/9:J2eI?@J68G'25?J96;73?JE%9#QH6;?@»725F9 G0Hn~ `2068?M/25725G'25?J9\Rv*Jc*c*KCO2
9#:6;=9 G H² 1 *S EUVCFJc*c*KCO2bIB9#:28B9#:6;= 9 Gfi kbß#K9#:25BFE%CO72 G ã à i¶kL68?D @ G ã à HÞ1 S EPU CO9E77%á8R
]^25F9 ± t SL`209:J2£E?9#25^25B CFJ4:»9#:6;29 EL² t G t E²6³ à Q}68?J@»7%25§9 ´Ýk W G 1 EL²ÑR *7K ńikZ68?J@
´ t EL²6³ à 1 EL²k W E=QH:E4:4K?J9BF68@JE%49C 9#:20IT6;499:J6829 E68CZ9:J2eITJ?@J68G'25?J96;73?JE%9Rvn=:C
~#WYS EPU#=?K9#2V9#:6;# 9 ~_vECXE?ÎÑ?E92COE%?J4K2 E²6³ à inEL²hITKB256;4P: ±R
❒
3
µ Wo
¶=Mh7KK±;E?0689H¼E%^BF2[·}R%¿QJK?J2V468?2568COE%7Mh4K?Jd5E?42hK?J25CO257%IH9:J689¨·ª¡NE%C9#:2ITJ?@6;G'25?9#687
J?E9\R
¸ W§
¶=MN@JB6;VE?6c*E%49JB2 COE%G'E768BP9K¼E;JB=
2 ·}R
QK?2Z468?£CF282N9:J689jT/¡5l 1 ·EC9:J2 ITJ?@6;G'25?9#687
J?E9\R
!#"%$'&)(*(
+,(*(.r¹
n=:2Vc*E49#BF2VECÊ%d825?uE?'¼E%^BF2[·}R%R
!#"%$'&)(*(
+,(eº/.Z{
n=:BFK^:KJ9=9:JE%CcBK`725G'Q9:J26;G'È%28?9=BE?£E%C / R*È'2V±^?Ko9:J689jTæÑ\çl3E%CXc*BE?4E%c*6;7%QJ;28?25BF689#25@
`MZ9#:2Z;BF25689#25CO9J4KG'G'K?e@Ed5ECFKB WÓ«;¬=jTæÑOçl#Q68?@09#:689Hæ^ÐD ç+N'WM»P:68C6VCFK79#EK?£EIJ6;?@0K?7M
EI»uHªjTæÑ\çlWYj lRn=:JC9#:25BF2ECX6hCOK79#E%K?uEI=68?J@'K?7ME%I}«^¬3jæJ\çlB »
R
ÚKQ=EIvæ;Ðu g+ç N»x W »D9:J25?IKBv6;?QM ¼8 ½gE9:æ ¼Wa'ç ½y2e:J68d52uæ)jÃÐun ¼^lP gç¿hj N/b ½VlZx W » 68C
2577%P9:JE%CvE%CD9#:2h^25?28B687ÑITKBFGIKB9#:2hCOK7%J9EK?CDCOE%?J42LÊ%d825?b6;?MLK9#:25BXCFK79#E%K?°æ^ÐJA Ý'ç NA3°
W »FQ
÷PøJù}úû*ü Í ·

¾5¿
¾
¾ ¾5¿ ¿
;ô ñ mî ¿ ¿^î*ô ðñL À &óî*Á^;îOòmòîOòô ðmë#éJë%^î#ÂÃUÄ*ÅÆ¿íïÃðîOÇvô ëó§ðô ë#êô ðïÃTñîHòézëzê*
2°46;?ÝCO`9#B6;49V9:J2»9K¸COK7%J9EK?C/9#K¸K`96;E%?Óæ)jÃÐ¸ÓÐÑAql ç5jhN»MNAqlWÈSmQN:E4:ÝE%C/9#K¸CF6;M

9#:689Næ)jÃÐ»©ÐÑA[leWÉç5hj NA)D N*l\RVÚDKQ9:J2'CFG'6;7%725CO9v6;`CFK79#2'd56;7%J2uKIDæ¼WÉç'½ÒE%CWÉq¬Â»jTæÑOçlQ68?J@
COE%?J42 Éq¬Â»jæJ\çl}@JE%d8E%@J25CD68?M£?JG'`25B CFJ4:9:J689XÊæ B 68?J@6ç B QÑ2LG'JCF9P:6;d52L9#:6;9}6;7%7)COK7%J9EK?C
68BF2uKIX9#:2bITKBG æ)jTÐa ±Ë¼l ç¿hj Nz ±Ê½lD:25BF72 ±xH / ECh68?MbE%?J925^25BoQ ¼¸Ì W É¬ÂjæÑOçl 1 æ5Q¨68?J@
½n W Éq¬Â°jæÑOçl 1 çmR
!#"%$'&)(*(
+,(eº/.r¹
È'2VB25BE9#29#:2V25Ø6;9EK?bÐ ³ ¸·6N ³ WIÍ£68CjTÐe ª¡5ljTÐ'¡5lXWjÎÍÑl#Q*:25BF2L¡hW ¤ Z·^R*n=:JC¨V2VCO252
9#:689#Q)CFE?4Ï2 :ÞW /r< ?¡ >JEC 6£÷s*TQÑ9#:28B2LECD68?E%?J928BXCOK7%J9EK?E%IP68?J@K?J7%M£E%IËÍ'COc*7E%9#C E?.:NQ=:E4:E%C
9#B2£E%IP68?J@K?7M£EI¨ÐÑ³N g·VEC B25@J4E%`72£E?ÔFÐQ=:JE%4:EC 9#:2e468CO2eEI68?@K?7M0E%IXÐJ³0WZ·V:6;C 6
COK7%J9EK?uG'K@ Í3R
n=:689P68CD9#:2L6;7%^25`B60c*6;B9HKI}9:J2Lc*BFK`7%25G'P9#:2LBF25CF9PEC6£?JG'`25Bv9#:28KBMhØ25CO9#E%K?RP}COCFJG'2
9#:68Ê9 ÍbW¥ R ·£ß#E I ÍW¥·9:J25?u47256;B7M0ÐJ³D Ý·NECB25@J4E`725á8RÑ*E?42e9:J2h25Ø689#EK?ÐJ³hWk:6;CX62 Ñ;68497M
9#VKCOK7%J9EK?CE%?ªÔ Ð QP9:J2£±^25BF?257}KI¨9:J2£:JKG'KG'KBc*:JE%COGÁD Ð Ò« Ð ³ ¨3rÔ ÐÓ « rÔ ÐÓ EjC Ô6Õ£6k ÖQHCFK9#:2
EG'68^2LE%C:6;7%I}KI39#:2h25725G'25?J9CKIÔ ÐÓ QÑE,R 2JR)EI jTÔ ÐÓ l³× W ÔmÐJK³ B8Z
Ð HÔ ÐÓ ÖQÑ9:J25I ? B%jTÔ ÐÓ l£³ BWÞÎj Íbgkml 1 ·R
*E?42¢Ô ÐÓ EC'68`257E68?QjÔ ÐÓ l³bE%C'?JKBG'687QVCFKgÔ Ð 1 jTÔ ÐÓ l³uE%C'^BKJcÝECOKG'KBc*:JE%49° K ÔÕ£6k ÖQNCOK©9#:2
c*BFK@49HKI39#K'Ø68@JB6;9E4Z?JK?B28CFE@J25Cß#287%25G'28?9#CXKI¨Ô ÐÓ yjTÔ ÐÓ l³\á E%C6eØJ68@BF689#E%4VB28CFE@J2hG'K@
Í3RPn=:JCLZ·ECD6£COØ68BF2eG'KÏ @ Í/EI}6;?@K?7MeEI}28E%9#:25BZ·6;?@¢hkD68BF2LCOØ^J68BF25CDG'K@ Í'KBv`K9#:u68BF2
÷PøJù}úû*ü Í Ø
? K9PCOØ^J68BF25CvG'K@Í3R3È'257%7Q)2L±;?JK¥9:J689DhkDE%CD6eCOØ6;B2LG'K@Í0E%I}68?J@bK?7MeEIÙÍ°ÀÞkªjÃÂ/Ä^lQ

68?J@9:J2eCOc*254E687=468CO2eKIPØ68@JB6;9E4B254E%c*BFK4E9Me92577C C 9#:689·E%C 6/COØ68BF2£G'K@ÏÍuE%IP68?J@K?J7%M0EI
ÍÀ¥kvKB ¾©jÃÂ/ÄÖ;lR
n8KhCOG'G'68BFE%Ü
25Q9#:28?QÐ ³ · N ³ I W ÍL:6;CP6VCOK7%J9EK?/V:J25?025E%9#:28ÙB ÍW·vKËB ÍÀ¥kKBxjÃÂ/ÄÖ;l\R
Ú£Û ËÜ#Ý¤¥ Û¥

! "$#&%('*)
+-,.)0/1)
2436587:9<;>=?;:@3BA0C(DEGFH30IDC(DC(DJLK-IMNOCPRQM&SGT305$I30ART5$MDSUDV*WYX;ZDS\[]X];^M&N_Wa`[cbdW
[fe
@g5IG3EG36hiSM5MDS(DV*AjC(DEFGH&3IDC(DC(DJLK-IMNOClk3IGARm63VDJnA<SgoWYXp;Be
7rqsW6tubv7rqsW`<wxtybzW
7{qwxt}|
~IFGN17zM&NjEG365$36JC(MS36E@g7{qw
tY$M&SZ436SG36JA<H&e{ARSGg]C(DEFGH&3IDC(DC(DJLK-IMNOC?79r;=VDJ
ARSGg(;BC(DEGFH3(M&N{EG3653RJC(MS36E\@g(7rqwxtaVsDJy5$I3UNLA<C(3jJ36A<NLDSGRP85MN{36ARNOgj5$D(NL3R3j5IGAR5nk3jTA<S
TIGDDNL3A<Sgm6A<H&FG3VsDJu7{qwxt$esuk3IGARm63THARNONOM&hi36EA<H&HGC(DEFGH&3IGDC(DC(DJK-IGM&NOC(Nr;=?;A<Ny<Mm636S
@gC(FH5MK-HM&TAR5$M&DS@gNODC(3JMSj3RH&36C(3RS5}P
! "$#&%('*)
+-,.)0/
Q-FGK-K-DNO35$IA<5aaqsat_bz0kI3RJ3jaqs[it_bd4y ^-[U```Zff8[ [ MNyAC(DSMT! M&JLJ36EGFTM&@H3
MS5$36<3RJuK-DHg6SGDC(M&A<H fPy8~ I36SZwR¡Lu¡|||¡O <3RS36JLAR5$3(;vfAR8N{ AYVsJL363\¢nC(DEFGH&3£$M.P 3GP_ b¥¤jq
-}]>`` `
¦ff t$}eGNODp;l§b ¢©¨¢y£¨v```ª¨¢_ A<NuA<SA<@36H&MARSp4JDFKeGA<SE];j«ª¬p;§b
q$¢r«
¬£¢ut IA<Nr¬ 36H36C(36SG5NªP
!"$#&%('*)
+-,.)0/®
¯8° 24365±?@3A]NOM&C(K-H3;!sC(DEFH36erARSGE©H365²³X´±µ@3cARSgSGDS¶
36JD36H&3RC(36S5}Pc~I3C(DEGFH3
436S3RJAR5$36Ec@gY²$M P 3GP·ARHH*hGSGM&5$3jNOFC(N¸z¹RW ¹ ²cVsDJrW ¹ X;B!M&NrAYNLFG@C(DEFH3jDVy±ceGNOD\M5C(FGNL5
@3ARHHDV±cº*5$IFGN!±¥M&Nu<36SG36JA<53RE]$A<NyARS];!sC(DEFH36r@g²aP
»u3RhGSG3rAC(DEFGH&3!IGDC(DC(DJK-IGM&NOC7:9R;¼=?±l<Mm636So@g7rqwxt_bz²p$ARSE\7{qW6t_bv7rqsW` wxtyb
Wo`6²GRPB½(3(¾4SDk³5$IAR5B¿b¥ÀRÁÂG7MNANLFG@C(DEFH3(ARSGE£5$IFNARS£ME3RARHnDVr;!enARSGE£NOM&SGT37
MNoNOFJÃ$36T5Mm63¦$NOMST3²¦<36SG36JA<53RN(±c}eB@g]5$I3ÄM&JLNL5sNODC(DJK-IGM&NOCµ~I36DJL36C(e;j« §b ±
$A<N;BC(DEGFH36NORPÅF5o;j«ÆMN(ARHNLD¦AJLM&SG<eARSGE¦M5$N;BC(DEGFH3NO5$JFT5FGJ3£MN(ARHNOD^<Mm636S
@gcJMSC(FH5MK-HM&TAR5$M&DSaP(Q-MST3]±µARSE©5$I36JL36VsDJL3p;«ÇARJL3NOM&C(K-H36e{;j«ÇIARNSGD£K-JLDK-36J
NOF@C(DEGFH36NLeGM P 3GPfSGDoKJDK-36JaME36A<H&NOeA<SE(MNy5$I36JL36VsDJL3jÈM&NyC(A<ÉRMC(ARH.P
Ê ° Ë-ÌÍÎ xÏ Ð_Ñ<Ò6Ó(Ó(ÔÕ
Q-FGK-K-DNO35IGAR5y7]MNrSDS¶
36JLD·P*24365yÖ!436S36JLAR5$31×£ARSGE£ÖxØf436SG36JA<531×ÙØÚº·SGD535$IA<5·k3jTARS\5$AR¾43
ÖxØGbv7{qÖ
tiNOMST3jÖa436SG36JA<536Nu×A<SE(7MNnSDS¶
36JLD\$NLDY7MNNOFJÃ$36T5Mm636RP4ÛDSNOM&EG36JuÜbdÀRÁÂG7_P
Ý
ò ó ÞßÚëà
á4ô ßxâOõRÞ áâsìyãäâ<ìåÞ.8æ}é ßÞ à6é ìyà$Þiá4Þ.çäÞ à$à â4çààè4é çÙé ß.êâsáëxàªâOàèßsfâsâçà áà$Þ.é êÞ çaãâàèìZfâLíÞ ÞäöâLð ßfßÚà à$âOîsíxâsà$âsäìé çfâ*ã ãìÞíëxãäâ<fÞ ìÞ ìÞ.äá4é ßïìßsð ñ
÷
ønùGúû õ-ü} ¼ý þ
½(3¾4SDk5$IAR5_ÿM&NrAjNOF@C(DEFGH&3UDVÙ×·eiNLD136M5IG36J!Çb¼DJBÇb×rP*ÅF5_ b´×£NOM&SGT317

MN_SGDS¶
36JDeNOD7pMN_MSªÃ$3RT5$M&m63A<SE(MN_5$IFN_@M Ã 36T5$M&mR3GP

! "$#&%('*)
+-,+Y/1)

Åg5IG30SG36ÉR5K-JLD@H&3RC(e*Üblq[·¡ -t8M&NyVJ363M&V*A<SE(DSHgM&VM5M&N_AKJMSTMK-A<H-M&EG36ARHn$SD5$305$IAR5 [·¡
MN_ARS(MS5$364JA<H-EDC(A<M&S\A<SE(5$IFGNyIA<NySDY¶
36JDoEMm6M&NODJNORPÅFG58k3¾<SGDkv5$IAR55$IMNyMN_SGD585IG3TARNO3GP
! "$#&%('*)
+-,+Y/

Q-FGK-K-DNO35$IAR5 \M&NApVsJL363;BC(DEFGH&3Re_NOD] k3cIA<m63ApC(DEFGH&3cM&NODC(DJK-IGM&NOCµ7 9y; = PY2<365

¡|||¡ c@3\5$I3cNL5$ARSGEA<JE©@ARNOM&NVDJ; P(Q-MST3 \MNApNOF@NO365{DVr5IG3\JMS©;Bey5$I3c36ÉRK-JL36NONLMDS

7{q ¹ ta` C(A<¾<3RNyNL3RSNO3GP nNLNOFC(35$IA<5n² >w6Pf½(3IGARm63

7{q7rq ta` ¤©7{q tn` tybv7{q tÙ`7{q tÙ¤©7{q ta`7{q t_bz

@F5G5IG3 ¹ VDJC¼AB@ARNOMNVDJn; A<SEjA<J3!5$IFNH&MS3RARJHgBMSEG36K-36SGE36SG5e<NLDo7{q tª` ¤j7{q t6` bd
NOM&SGT3Y7]M&NuM&S6Ã$36T5Mm63GPÅF55IGM&NuTDS5$JA<EMT5$Ny5$I3ARNONLFGC(K-5$M&DS\5$IA<5a7]M&NuM&S6Ã$36T5Mm636eGNLDYk3C(FNO58IA<m63

² bwRP!~IGFN bweARSE5DcNLA<g(5IGAR5r7´98; = YMNBA(C(DEGFH3(M&NODC(DJK-IGM&NOC MN03 4FGM&mRARH36S5u5D

NOARg6MS\5$IA<5 MN!5$I3UKJMSTMK-A<HM&EG36ARH436S3RJAR5$36Ep@g(bl7{qwxtP {DkneMVuM&N!AY¶
36JDEMm6M&NODJrDVy;
5$I36S\5$I3RJ336ÉRMNO5NuNLDC(3SGDS¶
36JDW1X;zkM&5$IpWª^b¼xeNODc7{qW6t!b>W
7{qwxt{b>W
^b¼
ekIMTI\A<<A<M&S
TDSG5JLAREGM&T5Ny5$I3ARNONOFC(K-5$M&DS(5$IA<5a7pMN_MSªÃ 36T5$Mm63GP~IGFNrM&NySGD58A¶
36JDYEMm6MNLDJ.P
!
"
Û·DSmR36JNO36Hg.eiNOFKK-DNL3U5IGAR5 (b qsatnMN{5IG3K-JMSTM&KARH*M&EG36ARHi436S36JLAR5$36E@gUARSp36H&3RC(36S5y]kIGM&TI

MN_SD5Aj¶
36JDoEMm6MNLDJ.PG»u36hiS3o7¦94;´= @go7rqwxtrbv_PGÛ·H36ARJLH&g(7MN_NOFJÃ$36T5Mm636º8MVn7{qW6t{bv5IG36S
Wª©bvxe-kIGM&TI(C(36ARSGNrWbdPf~IGFN{7]MNyAC(DEFGH&3MNLDC(DJLK-IMNOCP
! "$#&%('*)
+-,+Y/®
243658;^@3BAJMSNOFTIo5IGAR5i36m636JLgBhiSM536HgB436S3RJAR5$36E\;!sC(DEFH3BMNaVsJ3R36e<A<SEoARNONLFGC(305$IAR5·;^M&NnSD5
$
#
5$I3B¶
36JLDJMSPQ-FGK-K-DNO305$IAR5·;:IA<NaNLDC(3K-JDK-3RJ8M&EG36ARH ªekIMTIYTDS5$ARMSNaASDSG¶x3RJDj36 H36C(36SG5W4P
½(36HH&e·;j« M&NrARS;BC(DEGFH36eiNLD1M5·MNuVJ3636º5$IFGNu5$I3RJ3M&NrAjC(DEGFH3M&NODC(DJK-IGM&NOC 794; = ;j« P $
ÅF5_7{qWª¡O-¡|||¡O<t{b>W`x7{qw<¡L¡|||¡L4tubNLMST3YWob¼MS£;j« ªekIGM&TI\TDSG5JLAREMT5$Nr5IG3VsART5·5$IA<5

7pMN_MSªÃ 36T5$Mm63GP~IGFNr;TARSIARmR3SGDoK-JLDK-36JaME36A<H&NuARSE(MNyAjhi36HEnP
&% '%
! "$#&%('*)
+-, Y/
,+ # - .- ( / *0 *) 1 2 3+
243650aqs[ ¹ t(Xv¢ [6 ¡|||¡[i 8@3(AcK-DHg6SDC(MARHaME36SG5M5g YkIDNO3\TDJLJ36NOK-DSEGM&SGJL36ARHnK-DHg6SDC(MARH
VsFGST5MDScMN :9 = £PaÅg nJ5$MSpÛIA<K-53RJ ÷ e·ø-JDK-DNOM5MDS nP -e·M&V MN!M&EG36S5$M&TARHH&gY¶
36JD5IG36S
465 87 #9;=: <?>A@ B
Þ.à$âfà&4à {é ßçfÞ àªçâO.âLßßsé ã öy é íâsçxà$é àèöð
ønùGúû õ-ü} 2 1
bl$5$IMNrMN!5$I3jDSGH&gYM&C(KDJ5$ARS5aK-JLDK-36JL5gjDVn5$I3jJL36ARHSGFC(@36JLNr5$IA<5ak3oFNO36ºÙ5$I3UJL36NO5nDVn5$I3
C
K-JLDDV·4M&m63RSpMS JL5MSpVsDJ_5$I3oTDC(K-H36ÉYSFGC(@36JN{MN!M&EG36S5$M&TARHRPa~IFGN{5IG3UMC(AR43oDVuFSGE36Jr5$I3
D
TA<SDSGM&TARHJLM&SGUIGDC(DC(DJK-IGM&NOC ¢ [ ¹ =µ; [ ¹ fMN_¶
36JLDeNLDY5$I3M&EG36S5$M&5$gIDHENyMS;oP
!"$#&%('*)
+-, Y/1) &E
¯8° FG w =
F K F K F wÆ
¤ w IH*J ¤wIH*J = ¤w *J = KLJ
Ê ° F wIH G F w G G F w G F w G F w G
M NPO J = M ¤ O J = M ¤ ¤ O J = M ¤ $J = *J
fQ ° RG R R
HS ¤ M R ¤ Ow G ¤¤w MHLTU = S ¤w< Hw RH ¤ Ow G ¤¤w MHLTU = S wR R w ¤ w G ¤ Vw M TU
wG wG w
= S wRw ¤ ¤wRw = S ¤ ¤w<w = S wRw
TU TU TU
W ° 24NODc365Y@X¥g1EGb[M&A<<Z]D8\ SG AR H^M&¶
ARMSN!cMS]XvK-kARJL3150T$ARARS]RPuhi½(SE:31T$HA<ARS]5$5JMT36364ARJL@5!ARvXNO36N1A<NBqed<A(
¡C(d DtyEFGVDH&3YJ'I_DC(A<SDE C(qgDfsªJLK-¡hf IMtyNOC`VsDJ'_ a:=bNOFTanI e
5$IA<5]f bid A<SEjf bkK$d P½(3(TDFHEEGDp5$IMNNLg6NO5$36C(AR5$M&TARHH&g@g\¾<3R36K-MS5$JA<T¾cDVykIAR5

36A<TIoJDkARSGEoTDH&FGC(SoDK-3RJAR5$MDSUED3RN5$D36ARTIU@A<NLMNLe@FG5GA<NaM&Na5$I3BTARNO30kM&5$IUC(DNO5*DV5$I36NO3
K-JLD@H&3RC(NLefM5gl NnNOMC(K-H&g36A<NLM36Jn5DUEJA<kzAKM&T5FGJ30A<SE1hiSEYNLFGTIY @ARNO36NÙ@g0IGARSEaPf½(3hGSGEo5$IA<5
dA<S Erb³f qw<bs¡h<H K*td AR SGMECN_Ad @ARb NOMNyqw<Vs¡Dw
JYt·aBM&NrP AU@A<NLMN{VsDJ]_ $SD5$35IGAR5!m ÁonpZ y ^ b ¤w}eiARSGE5$IAR5 f bqd
tvu wyxz({|~}u~
SOLUTION SET # 7
GREGG MUSIKER

Problem 4 Let d1 , d2 , . . . be the integers referred to in Theorem (4.3).
a) Prove that d1 is the greatest common divisor of the entries of aij of A.
b) Prove that d1 d2 is the greatest common divisor of the determinants of the 2 × 2
minors of A.
c) State and prove an extension of (a) and (b) to di for arbitrary i.
I only do part (c) since (a) and (b) are the special cases i = 1 and i = 2 respectively.
 
d1 0 0 ... 0 ... 0
 0 d 2 0 . . . 0 . . . 0
 
 0 0 d 3 . . . 0 . . . 0
Claim Let M =   . Then d1 d2 · · · di is the gcd of
0 0 0 . . . d n . . . 0
 ... 
0 0 0 ... 0 ... 0
any i × i minor of M .
Since d1 |d2 | . . . |dn , an i × i minor which is zero or the product of i distinct dj
will be divisible by the product of the first i dj (d1 · · · di ) and this product is the
greatest since d1 · · · di is a possible i × i minor.
By Theorem 4.3, any integer matrix can be diagonalized into a form like M by
invertible row and column operations. Thus it suffices to show the following:
If N is a matrix where d1 · · · di is the gcd of all of the i × i minors, then P N Q also
has this property where P is an elementary row operation and Q is an elementary
column operation.
1) Assume P (Q) is the row (column) operation which adds an integer multiple of
row (column) I to row (column) J.
For any i × i minor not involving I or J, the minor is unchanged. For any minor
containing both I and J, by the properties of determinants (pages 20-23 Artin),
the determinant is unchanged. Since gcd(a, b) = gcd(a, a + b), even if only one out
of I and J is in minor, the gcd of all the minors will be unchanged.
2) Assume P (Q) is the row (column) operation which interchanges row (column)
I with row (column) J.
For any i × i minor not involving I or J, the minor is unchanged. For any minor
containing both I and J, the determinant is changed by −1 which is a unit, hence
it does not affect gcd. If only one out of I or J is in the minor, then we have just
switched the labeling of two minors in our list of minors and gcd is unaffected.
1
2 GREGG MUSIKER
3) Assume P (Q) is the row (column) operation which multiplies row (column) I
by a unit, ±1.
For any i × i minor not involving I, the minor is unchanged. For any minor
containing both I, the determinant is changed by ±1 which is a unit, hence it does
not affect gcd.
Problem 5 Determine
all integer solutions to the system of equations AX = 0,
4 7 2
when A = .
2 4 6
We first diagonalize this matrix:

1 −2 0 −1 −10 1 0 0 −1 −10 0 −1 −10
A= , = ,
0 1 2 4 6 4 1 2 4 6 2 0 −34
   
1 0 17 1 0 0
0 −1 −10  0 −1 −10 0 −1 −10 
0 1 0 = , 0 1 −10 =
2 0 −34 2 0 0 2 0 0
0 0 1 0 0 1

0 −1 0 0 −1 0 −1 0 1 0 0
, = .
2 0 0 1 0 2 0 0 0 2 0

1 0 0 0 −1 1 0 1 −2 −4 8
Thus = P AQ for matrices P = =
0 2 0 1 0 4 1 0 1 1 −2
    
1 0 17 1 0 0 1 0 17
and Q = 0 1 0  0 1 −10 = 0 1 −10
0 0 1 0 0 1 0 0 1

1 0 0
X = 0 only if x1 = x2 = 0 but x3 can be any number.
0 2 0
    
1 0 17 0 17x3
0 1 −10  0  = −10x3  so the integer solutions to AX = 0 are of the
0 0 1 x3 x3
 
17x3
form −10x3 .
x3

1 1 0 −1
Problem 7 Prove that the two matrices A = and B = generate
0 1 1 0
the group SL2 (Z).
Any matrix M in SL2 (Z) has rank 2 (is invertible) and a determinant of 1
hence M can be diagonalized as the 2 × 2 identity. In other words there exists row
operations P and column operations Q such that M = P IQ = P Q.
The row (column) operation of adding an integer multiple of the second row
(first
column)
to the first row (second column) is represented by the matrix
1 n
= An for all integers n and the row (column) operation of adding an
0 1
integer multiple of the first row (second column) to the second row (first column)
is represented by the matrix

1 0
= (ABA)n for all integers n. While multiplying one row or column by
n 1
−1 would result in a matrix of determinant −1, which is not allowed. We thus
combine this operation with the operation permuting the rows or columns to get
an operation with determinant equal to (−1)(−1) = 1, B. A, B, I ∈ SL2 (Z) and
all matrices in SL2 (Z) are equal to a composition of As and Bs, consequently, A
and B generate SL2 (Z).

Problem 2 Find a ring R and an ideal I of R which is not finitely generated.
A good example is the following: R = F [x1 , x2 , . . . ] a polynomial ring in a
countably infinite number of variables. (Note: we need a ring like this since any
polynomial ring (over a field) with a finite number of variables) is Noetherian by
the Hilbert Basis Theorem, and thus all ideals would be finitely generated.)
Let I be the ideal generated by all (or an infinite subset of) the variables. One
can check that I is an ideal by showing that for a, b ∈ I, r ∈ R. then a + b and
ra ∈ I. xi ∈ I for all i ∈ N, and 1 6∈ I thus the smallest basis for I is (x1 , x2 , . . . , )
an infinite collection of all the variables. The span of any smaller basis would not
include the element xi for some xi since there are no relations that allow xi to be
written as a R-linear combination of the other variables.
Another good example is R = { Continuous functions from R→ R}. This is a
ring using (f + g)(x) = f (x) + g(x) and (f · g)(x) = f (x) · g(x). Remember that
addition and multiplication preserve continuity. I = {f : f (0) = 0}. To see that I
is an ideal, note that if f (0) = 0 and g(0) = 0 then (f + g)(0) = f (0) + g(0) = 0
and (f · g)(0) = f (0)g(0) = 0.
To see that I is not finitely generated, one can show that sin(x), sin(2x), sin(3x), . . .
are all linearly independent and have no common divisor but are all in I.
Interesting note: if we had looked at the subring of polynomial functions from
R→ R (R[x]), then I would still be an ideal, but in this case it would have been
finitely generated, in fact I = (x) the ideal of polynomials with no constant terms.
Problem 3 Prove that existence of factorization holds in a noetherian integral

domain.
By Proposition (12.5.17), every submodule of V , a finitely generated module
over a noetherian ring R, is finitely generated.
By Proposition (12.5.13), the ascending chain condition holds, there is no infinite
strictly increasing chains of submodule of V .
In light of (12.1.3), we can let V be R, itself, which changes these statements
to the fact that there is no infinite strictly increasing chain of ideals of R. Thus
there is no infinite increasing chain of principal ideals, and by Proposition (10.2.3),
factorization exists.
Problem 4 Let V ⊂ Cn be the locus of zeros of an infinite set of polynomials

f1 , f2 , . . . . Prove that there is a finite subset of these polynomials whose zeros
define the same locus.
4 GREGG MUSIKER
V is a variety and as we saw earlier in section 3 of chapter 11, we can associate

an ideal of polynomials to this variety. In particular, let I = (f1 , f2 , . . . ). Then the
variety associated with I, V (I) is precisely V (I) = V,
(f1 ) ⊆ (f1 , f2 ) ⊆ (f1 , f2 , f3 ) ⊆ . . .
V (f1 ) ⊇ V (f1 , f2 ) ⊇ V (f1 , f2 , f3 ) ⊇ . . .

Since Cn is Noetherian, there can be no infinitely ascending chains of ideals (the
chains must stabilize) and there exists n such that (f1 , . . . , fn ) = I and V (f1 , . . . , fn ) =
V.
Problem
√ 6 Determine a presentation matrix for the ideal (2, 1 + δ) of Z[δ] where
δ = −5 (over R).
As explained in class, we consider the surjective map
R2 →φ (2, 1 + δ)
(x, y) → 2x + (1 + δ)y.
ker φ has two generators. There is the relation that 2(−1 − δ) + (1 + δ)2 = 0 and
also the relation 2(−3) + (1 + δ)(1 − δ) = 0
These two relations cannot be derived from each other.
We find ker φ by finding (x, y) ∈ R such that 2x + (1 + δ)y = 0 which means
(1 + δ)
ker φ = {y : y ∈ R and y is also in R}.
2
−3 −1 − δ
Thus the presentation matrix is
1−δ 2

Problem 1 Find a direct sumof cyclicgroups which is isomorphic to the abelian
2 2 2
group presented by the matrix 2 2 0.
2 0 2
We use the operations more than just the elementary row and column operations
for manipulating presentation matrices as described in proposition 12.5.12
         
2 2 2 2 0 2 2 0 2 2 0 0 2 0 0
2 2 0 → 2 0 0 → 0 0 −2 → 0 0 −2 → 0 2 0.
2 0 2 2 −2 2 0 −2 0 0 −2 0 0 0 2
So the presented abelian group is Z/2Z × Z/2Z × Z/2Z.
Problem 4. Determine the number of isomorphism classes of abelian groups of

order 400 = 24 · 52 .
By Theorem (12.6.13), the elementary divisor form of the Structure Theorem, any
e e
abelian group G such that |G| = N = q1e1 · · · qnen looks like Z/q11,1 × · · · × Z/q11,m ×
en,1 en,m
· · · × Z/qn × · · · × Z/qn where the qi s are distinct primes. Consequently, the
number of isomorphism classes of abelian groups of order N is P (e1 )P (e2 ) · · · P (en )
where P (k) is the number of partitions of k. For example P (5) = 7 since 5 = 4+1 =
3 + 2 = 3 + 1 + 1 = 2 + 2 + 1 = 2 + 1 + 1 + 1 = 1 + 1 + 1 + 1 + 1.
For this specific problem, the number of isomorphism classes is P (2)P (4) =
2 × 5 = 10.
More concretely, the possibilities are
Z/16Z × Z/25Z,
Z/8Z × Z/2Z × Z/25Z,
Z/4Z × Z/4Z × Z/25Z,
Z/4Z × Z/2Z × Z/2Z × Z/25Z,
Z/2Z × Z/2Z × Z/2Z × Z/2Z × Z/25Z,
Z/16Z × Z/5Z × Z/5Z,
Z/8Z × Z/2Z × Z/5Z × Z/5Z,
Z/4Z × Z/4Z × Z/5Z × Z/5Z,
Z/4Z × Z/2Z × Z/2Z × Z/5Z × Z/5Z, and
Z/2Z × Z/2Z × Z/2Z × Z/2Z × Z/5Z × Z/5Z.
Note that terms such that are relatively prime such as Z/2Z × Z/5Z ∼
= Z/10Z, but
Z/2Z × Z/2Z ∼6 Z/4Z.
=

"!$#&%
')(+*-,/.
0)1324576 855913:;592<9=?>"@A5CBD8=EGF$@>36H13I)>31"41"6 JCKL1"2M6 85748A=NO=46 59NO1"EP6H134$I)Q>3J92QF$13=>3RS:?13<9592M6 8=?6T6 85
BUV591":;592AEOI)=?45W13ETXUVY13F$592EP1"Q2A=>AZ[5W4=2$4Q2A4>3@YA5[6 8=?6\6 85[2QNF$=?>)]VQNOF^]HQN_6H8A1"E`F$=?6HNO1"aW13E
bcc Xfegehege jlkk
cc i Xhehege kk
ehe Bfege
d eh e i Bhe
ehegehe B mn
"!$#&%
')(+*-,/o)p
qr:?592A59N=?>s=?>":;QN136H8AFt]VQNu6 813EGv;132YwQ]-I)NQx>359Fy1"Ez:?13<9592{132|q}NO6H132~6H8A1"Ez4=?EO513Ez59=EPJ592AQ@:;8
6 QEPQ>"<95759a?I)>"134136 >"J70)56H6 1"2A:CwVDV}VV-ZW578=<5-wLVeVA
ZW81348F$59=?2ET6 8=?6}W^W1"ET4Q2EP6H=?26 R\=2AY/^+A/9/xJz1"2A6H59:;N=?6H132:S8A@E/wV/
O+ | \]VQNEPQF$5 O
"!$#&%
')(+*-,%

85INQx>359F=?E}EP6 =6 59Y1"E=F$x13:;@Q@E=E136YQ5E 2Q6AEOI59413]VJTZW8A=6NO1"2A:-QN\¡¢59>3Yu6 85592A6HNO1"59E}Q])6 85
F$=6 N13az=N54Q2A6H=?1"2A59Y1"2A~1"613EN59=?EOQ2A=x>356 QD=?EOEP@F$5z6H8A=6_6 85G4Q59]V¡¢413592A6HES=NO5G132=2=NOx1"6 N=?NJ
¡¢59>"Y£z¥¤G¦;596s§¨r£©Gx5=?2ªAUYA1"F$592AEO13Q2=?>&<95946 QNÈPI)=45u=2AYL>3596_«¬§¯®°§±x5u6 85QI)5N=6 QN
=EPEPQ413=6 59YZW136H86H8A5DF$=6 N13az813E[Y59¡¢2136H13Q2:;13<959E§^6 85EO6 N@A46 @NO5DQ]s=£$² ³lUF$QYA@>359~TZW5
ZWQ@>3Y$>"13v;56HQ¡A2AY$=I)N59EP592A6H=?6H13Q2Q]6H8A5/F$QYA@>35A
´ >359=NO>"J§^13EW¡A2A1"6 59>3JL:;59259NO=6 59Y=?EW=2w£$² ³}F$QY@A>"5LEO13245L136µ13EW¡A2A1"6 59>3JL:;5925N=6 59Y=EW=2w£-U
F$QY@A>"5·¶ 1 5A<5946 QN-EPI)=459¸¹H2Y559YR`>3596-º9»`x5L6 857¼VU6 8C4QQNYA1"2A=6 5L<9546 QN-Q]T§t½£ © R=2AY
>3596/¾ » x5D6 857¼VU6 8C4QQNYA1"2A=6 5D<95946HQN/Q]£$² ³¿©M¶ 1 5AL¾9Ày i Pe) Pe?OH~TZW5L6 8@AEW8=<5D=F$=?I
Á ¬?£$² ³Â©D®t§ÃEO592AY132:¾
» Ä®yº9»¶ =?2Y6 8@ESÅ_V\ÆÇ¾
» ®t Ä Å_«[+º9»3¸ÈµQZ±Z[5[ZWQ@>3Y$>"13v;5/6 QG¡¢2Y6 85
N5>"=?6H13Q2EPR)1+ 5A59>"5F$5926 EQ]&6H8A5/v;59N2A59>Q] Á )É$59>3>"RA1"] «Tº » ÃÊ » Z[5[n8nn =?<95 Á V¾ »AË Ê » sÃe¶ ZW85N5
ZW5=NO5[6 8132v;132:zQ]Ê»¢=E=259>"5F$5926\Q] xQ6H87£©u=2AY£$² ³¿©ÌÃ£©?¸AÈSQZ_RZW5v?2AQZ6 8=?6\Q245
ZW5D8=?<95$Y56H59NOF$1"2A59Y6 85$QI)5N=6 QN[«ZW5$8=?<95DYA596H5NF$13259YM§¨=E[=2|£$² ³lUF$QY@A>"5·¶ =2AY8=?<95
6 8@E]HQ@2AYz=?>">AN59>3=6 13Q2EP¸~AEP1"2A45W=>31"2A59=N_QI)59NO=6 QN}13E`4QF$I)>"56H59>3J[Y596 59NF$1325YxJ[8QZÍ136\=?46 EQ2
=zx=EP1"EPRZW5uv?2AQZÎ6H8A=6ZW5G8=?<95u=z4QF$I)>3596 5uEP596}Q]}N59>3=6 13Q2EÏ8@ES13]_Ê»ÐÊ» À Ê» © Q@AN
ÑÕlÜ "ÒÓÔ¥Õ Ù ;Ö Ó ÂÖç × ÔHÓ Ô¥ØÚÙ ÖOÛ Õ Õ"Ü Õ × Õ9Ô}ÕlÜ ç ÖÝÖÛ¢Þ× Ô)ß Ó×HÓ àÛ ØÓ Ô ÖPÛ¢á Ó ä Û×HÕ àà Ó Ù ×Vâ Óä ØÚÙÕsÕ ê?ÓÕlÜ ÔHÓ ã
Ö Õ"Õ"Ü Ü Ö)Õ\Ù ÓäÔVÔ Ö Ô)Ù/ÕlåPÜ næÏnÜ n ÔÓ ç+ÓÕ`ÙsÓ ä àÕlÜ ÖPÛ
Þ Ó+ä Ù/Ö å ×VàÔVÓ ÖVà ÓÖ ØµìâíHî Ûï ï ÖVð à Ö Õ ÓA×Häàè
ä ÖVÂÛWÕ ×`Ó¢Ø â¢âÔ¥Õ T×HàLÕlÜ ×H Ô Ó×HÓ Âèå àé àÚë × Ö Ö_ à â Ö
Ö ? è àÖPÖPÛ Ö\× Ö àÛÖV ×VàÛ â
ñ
òóAô}õö)÷ Òrø ù
I)NO59EO526 =6 1"Q2F$=6 N13a13E

bcc Ë Ê ÀOÀ Ë Êû À ÆÆÆ Ë Ê © À jlkk
c Ë Ê-À û Ë Ê ûPû ÆÆÆ Ë Ê © û k
ú Íý © Ë
d Ë Ê À Ë Êû ÆÆÆü Ë Ê m
© © ©9©
=2AYCþÿÏ ú Ws13ET6 8548A=N=?46 59N13EP6H134WI)Q>"J2QF$13=> Q]µuÈSQZ_R xJ´ NO=F$59N ETNO@>35Z[5v;2À QZ
ZW5M4=?2±¡¢À2 YÍ=F$=6 N13a¶ 6H8A5M=Y Q1326GF$=?6HNO1"a?¸ ú µZW1"6 8 úÎú þÿ
ú ý © R¶ 1 5A
ú
þÿÚ ú þ ú H¸EOQ ú ú l¾
» ¾
»ZW8A1"48|F$59=2AE ¾
» ú Ç
@A6-EP1"2A45 ú 1"E
=I)NO59EO526 =6 1"Q2$F$=?6HNO1"aZW5W8=?<95z § £$² ³Â© ú £$² ³¿©R)EPQ1"]} V&À9 O £©1"E=2J<5946 QN
ZW5W8=<5 } À }V+ © ú £$² ³Â©ÏRAEO!Q }V«/+Íe-1"2§Ínn]Hn QN_© =>3>&RA1 5A" }«[_±e
nnn
"!$#&%
')(+*-,/'}%
¹V6}EO@A]V¡¢459E6HQEP8QZr6H8A=6}=$#QNOY=?2$x>"Q4v13EEP13F$1">3=N6HQ136HE6HNO=2EPI)QEP59Rx54=@AEO5G6H8A592D13]\Q@N`<5946 QN
EPI)=45§^13&E % » '» ZW859NO5 '» 13EEO6 =x>35D@2AY59NGÎÀ =2YC6 85DF$=?6HNO1"a » ]VQNu Q2 '» 13E= #QNY=?2
x>3Q4v;RA6 8592LZ[5G4=2¡A2A) Y (»&EP@48L6 8=*6 (}» -+» ( » r ,» ~}Y59¡¢2132:z=2QI)59NO=6 QN (±Q2§6 8=?6 =46HE
=-E (»)Q2 ' »ZW1">3>:?13<9!5 (. ( À Ã ,Ç +¶ /TN"R;F$QN5-4Q24N596 59>3J+R>"5906 (x5/6H8A5Tx>3Q4vWY13=:;Q2=?>F$=6 N13a
ZW1"6 8$59=?418 ( » Q2$6H8A5/YA1"=?:?Q2A=>+ ¸
0)Q=EPEO@AF$5D6 8=6Tr13E/= #QNOY=2x>"Q4vLZW1"6 851":;592<=>3@)5 2R1 5Au1"]µº?À9 º 13E/6 85$x=?EO13E/]HQN
6H8A592wTº À 32 º À Íº9û PTº © À 42}º © À ±º © O-º © 32 º © LÉ$59n>3n>"Rn 48A© QQEO132:6 85x=EP1"E
Ê » Íº © » ZW5W8=?<95GTÊ-ÀT5 nn2 n Ê-À9OTÊ û 5
2}Ê û MÊ-À9 O-Ê © 5 2}Ê © |Ê © ÀÏ¹H2$Q6H8A59N_Z[QNOYEPR
bcc 2 i ÆÆÆ e e jlkk bcc ege ÆÆÆ e i jlkk bcc 2 nnn e ÆÆÆ e e jlkk bcc ehe ÆÆÆüe i jlkk
cc e 2 ÆÆÆ e e kk cc ege ÆÆÆ i e kk cc i 2 ÆÆÆ e e kk cc ehe ÆÆÆ i e kk
c k c k c k c k
d ee ee ÆÆÆÆ6 Æ 2 i m d e i ÆÆÆ ege m d e e ÆÆ7
Æ e 2
Æ 2 e e i ÆÆÆüege
i e ÆÆÆ ege e e ÆÆÆ i 2 m d i e ÆÆÆüege m
ZW81348zEO8AQZWE6 8=?6} . ,OqF$QN5T=?xEO6 N=?46¢Z[=?J/Q]AEO5591"2A:6H8A1"E_13E6H8A=6 ,¢=?>"EPQG8A=E_48A=NO=46 59NO1"EPU
6 1"4µI)Q>3J92QF$13=>V Ë 2_©W=2Y8A=E_=/Q2A59UYA1"F$592AEO13Q2=?> 2U5913:?592AEOI=45R)ZW81348zF$59=2AE136F$@EP6&8=?<95
6 85[EO=?F$85 #QNYA=24=?2Q2A1"4=>]HQNF
"!$#&%9)(+%-,:9
813E[I)NQQ]1"EW25=N>3JL13Y526 1"4=>_6HQ6 85LI)NQQ]`6 8=?6=7¡¢2136H5D1326 59:;N=?>YQF$=13213EW=¡¢59>3Y ¶ 132YA5959YAR
=¡A2A1"6 5z1"2A6H59:;N=?>YQF$=?1"213ET=>3ZW=J9E=$¡¢2136H5Y13F$592EP13Q2=> <95946 QNµEOI)=?45zQ<959NµEOQF$5zEP@xN132:$13EPQU
F$QNI81346 < Q ;-=Ú¸L¦?59?6 >5@¨x52AQ2 A
59NQ~/ZW5ZW13EO8M6 QEP8QZÐ6 8=??6 >8=?E=2·1"2A<959NOEO5ABµ59¡¢25
CED ¬ F ^
@ ® G @ÍxJ CED &TH >;\}´ >359=?N>3J CED 1"E/£-U>313259=?N~_=>3EOQR&13] CED &T e[6 85E 2 >; e[ZW8A1"48
F$59=?2E6 8=6-r eDEO1324I 5 @^13E=2·1326 59:;N=>sYQF$=132·=2AJ Y >H K e)D8A@E C D 1"EG1"2 5946H13<959~[EO13245
13613E/=>"1325=N-QI59N=?6HQNQ2=L¡¢2136H59UVY13F$592EP1"Q2A=><95946HQNTEPI)=?459R136s13EW=>3EOQEP@LN 5946H13<959~S6 859NO59]VQNO59R
i 1M C D \EPQ6 85N5[1"EÈPQF$O5 NZ[136 8 i C D N`5 >PNÚ
òóAô}õö)÷ ÒQ R
"!$#&%9)(3'z,%
¦;596TS| i U2 û WVµEP1"2A:z=z25=6_>"1325=Nµ=>3:?59xNO=6 N134v6 8=?û 6}J9Q@7v;2QZÎ13]}J9Q@74=?F$5G6 Q$F$JGEP5946 13Q2
QNxJEO13û F$I)>3J2Q6 132:6 8=??6 X °Y S ËÎi [ ZM\ S]Z Ë ^K S r^K S ËÎi RZW5¡¢2YM6 8=6-Z[136 8 Å_¢
_Z Ë K wK9 Ë BR)ZW5W8=?<95Å_L S}s±e¹V2$QNYA59N_6HQEO8AQZ±6 8=?6Å13Es13NNO59Y@A413x>"5WQ<959N `ÎZW5WQ2>3J
2559Y6HQEO8AQZÍ6H8A=6Å8A=E`2AQGNO=6 1"Q2A=>NQQ6 EP~AEP1"2A45zÅ13E`F$Q2A1"4R1"6EP@]H¡A459E6HQEO8AQZÍ6 8=6_Åz8=?E2Q
NQQ6 EsF$QYAû @>3) Q aR¢Z[8A1"48D13E6 N@A5xJu=uEP1"F$I)>35u4=>34@>3=6 13Q2 8@E-Å$1"Eµ6H8A51"NON59YA@41"x>35I)Q>3J92QF$13=>
]VQN i b 2 W+¶ /TN"R @EP1"2A:D6H8A5]H=466 8=?66H8A5GYA59:?NO595zQ]}=$¡¢59>3Y75a6 592EP1"Q2713ETF$@>36H13I)>31"4=6 1"<E5 c ZW8A1"48
J9Q@ZW59N52 6¢=>3>3QZ[5Yz6 Qu@AEO5[859NO5 dREO13245uK13EÌ)N13F$5-=?2eY SfE `RZW5-F$@EP68=?<95[6H8A=6&6H8A5-YA59:?NO595
Q]] SQ<959N `Î13EKÚREPQÅz13E`=@A6HQF$=?6H134=?>">3J13NN5Y@41"x>35A ¸
"!$#&%9)(3'z,:9
VµEO132:6 857EP=F$57>"13259=?N=>3:?5xN=6 N134v=EGx59]VQNO59R`QNWxJ2AQ6H132:6H8A=6?gewY2
û Ëbh û i20j Ë
û
i gk2 < g^X
RZW5W¡A2AYz6 8=?W6 213E`=WNQQ6Q]Å_&`Ã_ j Ëi g l û X)m VµEP1"2A:u6H8A:5 n@A=YNO=6 1"4]HQNF$@A>"=WZW5
¡¢2Y6 8=66 85zNQQ6 EQ]SÅ_&`=NO15 oOp h okX q i B~2Q6 5z6 8=?6xJE n@=?N132:xQ6H8EP1"YA59ETZW5EO5956 8=?6
q B q Kr p h X q i B=2AY q B Ë q Kr p hWË X q i B;T¦?56 SÃ q B Ë q KTÈSQZR13]SÅ713ET2Q6
13NN5Y@41"x>35L¶ Q<959N_EPQF$5¡A5>"YA¸6 8592Å$8A=Es=G>"13259=?N`QN!= n@=?YNO=6 1"4T]H=46HQN+)¹H]Å8=E2QNQQ6 6H8A592
Q25WQ]&6 85W]VQ>3>"QZW132:uF$@AEO6}x5/$= n;@A=YAN=6 134]V=46HQN s
Å À &sÎV Ë 2J2_`Ã û Ë 2 û Ã û Ë h X q i B9
Å û V¢s Ë 2_ Ë S`Ã û Ë 27t S}+$t 20S û Ë XÏ q BTX
Å Z V¢s Ë 2_$< S`Ã û Ë 2 Ë S}+ Ë 20S û Ë XÏ q K Ë X
uwv ´}>"59=?N>3JÅ8=E/2QN=?6H13Q2=?> NOQQ6HET=?2Y2AQ25$Q]SÅ À PÅÏû9PÅ Z 13E-=LN=6 13Q2=>}I)Q>3J92n QF$13=>3R_EOQCÅ1"E
13NNO59Y@A413x>"5WQ<959N `
x v 0)13245 q y K y ` q Bu¶ =EJ9Q@CEP8QZW59YCQ2=7I)N5<91"Q@AEW8QF$59ZWQNv7=EPEP1":;2F$592A6H¸R-Å8=?EW2Q
NOQQ6HES13< 2 ` q B?Ç- zTQZW59<959N3R_ÅÏeû {` q B² )³lR EPQÅÏû-13E6H8A5F$13213F$=> I)Q>"J2QF$13=>\]HQN 2Q<59N
`+ q B
|mv 0)13F$13>"=?N>3J+-R o?2} o:S~U ` q i eHR}EOQÅ8A=ET2AQ>313259=?N]V=46HQNOEO~8AQZW59<959N3R\2Q2A5Q]µÅ À PÅÏûPÅ

` q i e² ³ RAEOQÅ1"E6H8A5/F$13213F$=>AI)Q>"J2QF$13=>]HQ- N 27Q<959N ` q i e\=?EsZW59>3> Z
v 0)13245 q K) q B ` q i B9 R Å$8=?Es2AQ$>"13259=?Ns]H=46 QNEP~\8AQZ[5<959NSÅ À l` q i B9² )³lR&EOQ7Å À 1"Eµ6 85
F$13213F$=>AI)Q>"J2QF$13=>]HQ-N 2`
"!$#&%9)(3'z,:
QYQz6H8A1"EsI)NQx>359F$RQ25W4Q@>3YDEO13F$I)>3JWEOQ>3<956 85[5n@=6 13Q2
i J 27J 2 û >W{ N2J 2 û ` i
]VQWN >¢} N
O~A6H8A1"E}ZWQ@>3YuN5 n@A1"NO5EOQ>3<9132:[EOQF$5>31"2A59=N}5 n@=?6H13Q2EÏmzTQZW59<95N"R?6H8A59N5S13E =-EP>31"4v?59N3R;F$QN5
=>3:;QN136H8AF$1"4WZW=J6HQD:;596_=6_6 859EP5G^5 n;@A=6 1"Q2AE
¦?59-6 ~ ` 2 ÏQNs=?2J Sb ZW5G4=27Y59¡¢25=
òóAô}õö)÷ Ò ø
F$=I C ¬ ® 6H=?v?132:MÄ®S\~6 813ETZW13>">}x5=EÙ>"1325=NF$=IZW136H8132<95NEP5 C À C F"À
13] S K eWÈSQZ_R C ZW13>">8=?<95D=DF$=?6HNO1"aÀ Z[136 8N59EPI)5946s6 Q6H8A5$x=EP1"EL i 22 û H~6 85N59]HQN59R C
ZW1">3>8A=<95W6 85[1"2A<959NOEO5[F$=6 N13a¢)0)1"2A45 C i `fS À ZW5W4=?2$N59=?YQ]V]]S À 132D6 59NF$EQ] i 22 û
]VNOQF6 85W592A6HNO1"59EQ]6 85[¡¢NEP64Q>3@F$2
¹V2Q@Nµ4=?EO59R S i U 2b 2 û ÈSQ6 1"2A:$6H8A=6 C) i -2s C) Y2_/2 û R=2AY C) Y2 û -
2WZ
K 2 Ë XÏR)1"6 1"E=2$59=?EOJW4=?>"4@>3=6 1"Q2L6HQz¡A2AY$6H8A=6
^
b i ehe j b ege Ë X j b e Ë X e j b igË X Ë X j

C d eegei ei m d ei ei Ke m d ei Ke Ë K X m d ii X i ËX i m
n
VµEO132:=2JWF$=?6HNO1"a132<95NEP1"Q2$=?>":;QN136H8AF^ZW5W6 8@Es¡A2AY
À i b Ë i a B i h X Ë Xe K j
Xk d
Ë K Ë B h m
C
EP&Q S À Î i a Ë B 2 Ë ^K 2 û X"
"!$#&%9)(3'z,:
É$5W8=?<95
eG2 © t> © À2 © À ÍÆÆÆ
J>)À27J>
w Ë > 2µY2 © À t> © À2 © û ÃÆÆÆÚJ>)À
w 2 À Ë > i 2 © À J> © À 2 © û ÍÆÆÆÏt> À
À n
ÈSQ6H5W6 8=?6W>?±
K eTx594=?@EP5GA©ut>
© À ¢
© ÃÆÆÆÚJ> À zJ>;1"E1"NON59YA@413x>35A
"kwm^F+
SOLUTION SET # 9
GREGG MUSIKER

Problem 1 Let F be a field, and let α be an element which generates a field
extension of F of degree 5. Then, α2 generates the same extension.
Since α2 ∈ F (α), this implies F ⊂ F (α2 ) ⊂ F (α). By multiplicativity of degree,
5 = [F (α) : F ] = [F (α) : F (α2 )][F (α2 ) : F ].
Since α is a root of f (x) = x2 − α2 , where f (x) ∈ F (α2 )[x], [F (α) : F (α2 )] =
1 or 2 depending on whether or not f is irreducible in F (α2 )[x]. But 2 6 |5, so
[F (α) : F (α2 )] = 1 which implies f is reducible and α ∈ F (α2 ) which implies
F (α) ⊂ F (α2 ), and by mutual inclusion, F (α) = F (α2 ).
NOTES: if α generated an extension of degree n where n was any odd number, the
same proof would imply that the two extensions F (α) and F (α2 ) were identical.
Also since 5 is a prime, Corollary 13.3.7 of Artin implies that K = F (α) is the same
as F (β) for β = α2 .
Problem
√ 6 Let a be a positive rational number which is not a square in Q. Then
4
a has degree 4 over Q.
Three methods for doing this problem:
√
1) Clearly, b = 4 a is a root of f (x) = x4 − a. Thus, it suffices to show that f (x)
is irreducible. First, assume that f (x) has a linear factor. Thus would imply that
the factor is (x − bζ4 ) where ζ4 is√a fourth root of unity, namely, 1, i, −1, or −i. But
this implies bζ4 ∈ Q but b2 = ± a is not even in Q thus f (x) has no linear factor.
Next, assume that f (x) breaks into two quadratics. √ Since f√(x) has no coefficients
√
for x, x2 and x3 , it must be the case f (x) = (x2 + a)(x2 − a) but again a 6∈ Q
so this is impossible. Thus f (x) is irreducible, and [Q(b) : Q] = 4.
√
2) We build the √ tower Q ⊂ Q( α) ⊂ Q(b). Since √ b is a root of f (x), [Q(b) : Q] ≤ 4.
By assumption α 6∈ Q which means that [Q( α) : Q] = 2 and by multiplicativity
of degree, this implies 2|[Q(b) : Q]. We conclude
√ that
√ [Q(b) : Q]√= 4 and not
[Q(b) : Q] = 2 by considering f (x) = (x2 + a)(x2 − a) but again a 6∈ Q so this
is impossible.
3) Let a = m/n where m and n are relatively prime. Thus if prime p|m then p 6 |n.
since a is not a square in Q, this implies there exists some prime factor p of either
m or n s.t. p|m but p2 6 |m. (NOTE: This does not imply that m or n is square-free,
only that at least one of the prime factors has an exponent of one.) Assume p|m
but p2 6 |m. Then g(x) = nx4 − m is irredubible over Z[x] by Eisenstein’s criterion
and consequently, f (x) = n1 g(x) is irreducible in Q[x]. If m is a square, then n is
1
2 GREGG MUSIKER
not, and consider α = a1 . Then x4 − α is irreducible, which means that F = Q(1/b)

is a degree 4 extension over Q and since F is a field, F = Q(b) as well.
Problem 7 Decide whether i is in the field F .

√ √
i ∈ F then√α = i −2 ∈ F . But α2 − 2 = 0. Which implies
a) F =√Q( −2). If √
α = ± 2, and thus 2 = a + b −2 ∈ F for some a, b ∈ Q. But then
√ √ a2 − 2b2 − 2
a2 + 2ab −2 − 2b2 = 2 =⇒ −2 =
−2ab
√
but −2 is not rational. Thus i 6∈ F .
√
b) F = Q(4 −2). By Problem 6, [F : Q] = 4. [Q(i) : Q] = 2. But i ∈ F =⇒
F = F (i) and 4 = [F : Q] = [F (i) : Q]√= [F (i) : Q(i)][Q(i) : Q] = 2[F (i) : Q(i)]
which implies
√ (i) : Q(i)] = 2. Since 4 −8) ∈ F there exists a, b ∈ Q(i) such that
[F √
4 4
a+b −2) = −8).
√ √ √
a2 + 2ab 4 −2) − −2b2 = 2 −2).
4
√ √
Reorganizing these terms,√ we find that
√ i ∈
√ F implies −2)
√ ∈ Q(i, √−2), but
squaring both sides of 4 −2) = a + b 2 + c −2 + di we get 4 −2) ∈ Q( 2) which
we know is impossible since [F : Q] = 4. Consequently, i 6∈ F .
c) F = Q(α) where α3 + α + 1 = 0. Any factorization of this polynomial would
involve a linear term so by the rational root (for ±1) this polynomial is irreducible
and [F : Q] = 3. If i ∈ F , 3 = [F : Q] = [F (i) : Q] = [F (i) : Q(i)][Q(i) : Q] =
2[F (i) : Q(i)]. But 2 6 |3 so that is impossible, and i 6∈ F. (NOTE: Even though this
cubic has a real root which and if α ∈ R then i 6∈ F , the cubic also has complex
roots as well, and we don’t get to choose α.)
Problem 10 Let α, β be complex numbers. If α + β and αβ are both algebraic,

then show α and β are also algebraic.
It is given that Q ⊂ Q(α + β, αβ) is algebraic. We will show that Q(α + β, αβ) ⊂
Q(α, β) is algebraic hence Q ⊂ Q(α, β) is algebraic. Consider f (x) = x2 − (α +
β)x + αβ ∈ Q(α + β, αβ)[x]. This clearly has both α and β in roots, enough said.

Problem 1 Write cos 15◦ in terms of square roots.
◦
By the half-angle formula from your favorite trigonometry class, cos2 15◦ = 1+cos
2
30
√ q√
and cos 30◦ = 23 . From cos 15◦ > 0 we conclude cos 15◦ = 3+2
4 . We could
√ √
6+ 2
simply this further into 4 .
Problem 2 Show that the regular pentagon can be constructed.

a) Using Field Theory One can construct an angle θ as long as one can construct
cos θ. So it suffices to construct x = cos θ where θ = 72◦ = 360◦ /5. This can be
done via similar triangles where one arrives at the equation
x 1
=
1 x−1
and
√
thus x satisfies an irreducible quadratic equation hence [Q(x) : Q] is 2. (x =
5−1 ◦
4 ) Additionally, one can use the fact that 5θ = 360 to conclude cos θ = cos 4θ,
cos 2θ = cos 3θ and factor
x5 + x4 + x3 + x2 + x + 1
into a product of quadratic and linear factors. Lastly, 5 = 2n + 1 so according
to Gauss, the regular 5-gon is constructible and a similar argument involving the
eulerphi function holds.
b) Using an Explicit Construction I am going to spare you all my drawings since
all of you had this or something very close. If you have a specific question about
this, feel free to ask.
Problem 5 Is it possible to construct a square that has the same area as a given
triangle?
Yes. Pick a base (one of the sides) and construct the associated perpendicular
bisector. You know can construct bh where b is the length of the base,
pand h is the
length of the height. Thus you can construct a segment of length bh/2. Thus
create
p a square using parallels and perpendicular lines whose side lengths are all
bh/2. The area of the square and the triangle will both be bh/2.
Problem 10 If a prime integer p has the form 2r + 1, then it actually has the form
k
22 + 1.
k
Let r = 2k s where s is odd. Then the polynomial xr + 1 = yP s
+ 1 where y = x2 .
s−1
Since s is odd, −1 is a root of y + 1 and y + 1 = (y + 1)( i=0 (−y)i ). If s > 1
s s
s−1
and y ≥ 4 then since (−y) ≥ 4 will dominate all the lower terms of the sum,
p = y s + 1 factors into non-units, this contradicts the primality of p. Note that as
long as k ≥ 1, y ≥ 4 for x = 2. Thus s = 1 and r = 2k .
There is another way that totally bypasses field theory and uses group theory1.
Since 2r ≡ −1 (mod p), the order of 2 is 2r in Z/pZ× . Thus by Lagrange’s
Theorem, 2r|p − 1, the order of the group. However, 2r|p − 1 = 2r implies that the
only prime factors of r must be 2, hence r = 2k for some k.

p
Problem 2 Let f (x) = x − x ∈ F [x] and p is a prime. For what fields F and
primes p will f (x) have a multiple root?
By Prop 15.5.6, α is a multiple root of f (x) iff f (α) = 0 and f 0 (α) = 0 where f 0 (x)
is the formal derivative of f (x). In this case, f 0 (x) = pxp−1 − 1. Assume
αp − α = 0 and
pαp−1 − 1 = 0.
Let β = αp−1 . β = p1 in F and α(β − 1) = 0 in F . 0p−1 6= p1 thus α = 0 is not a
multiple root. Since F is a field, it is an integral domain and we fight have there
is another α that is a multiple root where p1 = β = 1 in F . This will occur if and
only if p − 1 ≡ 0 in F which occurs if and only if char(F ) divides p − 1. char(F )
1Thanks to Noam Zeilberger.
4 GREGG MUSIKER
is always a prime or zero. If char = 0, this will be impossible, and if p = 2, no

such field exists with char(F ) = 1. However, if p ≥ 3, as long as char(F ) divides
p − 1, then f (x) has a multiple root. (NOTE: F need not be a finite field, consider
Z/2Z(t) with char equal to 2.)
Problem 4 Let α1 , . . . , αn be the n roots of a degree n polynomial f (x) in F [x].

Then what is the best upper bound for [F (α1 , . . . , αn ) : F ]?
The best upper bound is n! for the following reasoning:
First, assume that f (x) is irreducible. Then [F (α1 ) : F ] = n. If f (x) were
reducible then α1 would be the root of an irreducible polynomial with degree < n
and thus [F (α1 ) : F ] = n. Let f (x) = (x − α1 )g(x)). g(x) will have degree n − 1. By
the same argument, F (α1 , α2 ) : F (α1 )] ≤ n − 1 where the bound is achieved if g(x)
is irreducible.2 By continuing this logic inductively, and use the multiplicativity of
degree extensions we get an upper bound of n!. (NOTE: This field F (α1 , . . . , αn ) is
very important in field theory, it is known as the splitting field for the polynomial
f (x).)
2We also have to insure that α is a root of g(x), i.e. that α 6= α . Over any extension of
2 2 1
Q, an irreducible polynomial is separable, i.e. has no multiple roots. But over fields of positive
characterisitc, irreducible does not imply separable. However, there is still plenty of irreducible
separable polynomials to use to achieve the upper bound. This is a subtle point which is why I
included it only in a footnote.

!#"%$'&
(*),+.-/(
02143657!8/7!9:7<;,=?>A@CB'DFEGB*HIKJF7!98MLN>'OP@QB<DFESRB*HUTWVXEZY\[]^5_>`:Lab>'Oc:cde8/Lf57'd8Mg,:c7!;#hc8Mg#ikj*de;#l%8M[
:La%7!9m]F7P9f8Md8k;#]n:pof8cg#a%8Mq7P8cL]^;#5Lr5=?st;#L%uv9;,w9>r9:]x:pl%mg,7!;#j*g,8pd557zy
{ 1}|X5uv8Mhc8Md#['>r]^j*g#;,7!]F;,Lb~ H. st;,L7!5>A@QB<DFE@CBIJ\DH[]e5.;#7! ]F7P9f8/7!d;,hM;,:cgo<8Mg,a%8Mqc7!8MLf]e;,5L?y
!#"%$'&
(*),.-/(
iU8cdl%:c7! ];,7P7!g,8W98c5d8Ml%[?5di_Lf57P;,L7!9:7n7!98%d5mfj5=xmfL;,7P]v;,LS~ V 9f:c]k5daf8Md.J\.:cLfa
7!9m]VEJz[uk8vL5u7P9f:c7AVXEy
!#"%$'&
(*),.-r
;,d]^7L57!8v7P9f:c7![:]F:c5hc8M[WkEp=Q5d8M:w9N~ yW9fm]BfXIKB.uv;,g#g<9:hM8.Y\[J:cLfaX:c]d557!]Uy
;#hM;,a;,L`5m7![*uv8vofLfa.7!9:7
BfI B
B?@CBNI¡J
D@CBIKD E@CB ¢ J\D@CB£ ¢ J
Dz¤
W98ko<d]^72=Q:w7!5d?;,]F;,dde8Mamfw;,g#8v]^;#Lfw8b@¥`J\DEJMy*¦*;,l%;#g,:cdeg#i§[<7P9f8v]e8Mw5La%9f:c]xLf5`g,;#Lf8M:cdF=Q:cw7P5d,[m7
;,72=Q:w7!5d];,L7!5
@QB £ IJ\DxE¨@CB ¢ BI¡J
D@CB ¢ B_I¡J
Dz¤
W9mf]BfXI©BbEtBª@QBf ¢ J\D@CB< ¢ BI¡J
D@CB< ¢ B ¢ J
D«y
¬ ]6uv:c]6w5l%jm7!8Ma;,LN7!98]e8Mw5Labj*d5g,8Ml®]^8M7![<uv8¯L5u:g#g'5=?7P9f8¯;,dd8camw;#g,8¯j*5g,iMLf5l%;#:g#]
5='a8Md8c8X5hM8cd~ ^° i ¬ de7P;,L%±ªy ²ª³!8M´z[uv8vw:L%7P9fm]xuvd;,7!8
B eµ I©BNE¨@CB I©B ¢ B I B%I¡J
D@CB ¢ B IJ\D@CB ¢ B ¢ B_I¡J
D@QB I B I©B_IJ\D
@CB ¢ B I B ¢ J\D@QB I B ¢ J
D@CB I B ¢ B ¢ J\D¤
!#"%$'&
(*),.-r&
¶
¦*;,Lw8k7!98kj*5g,iML5l%;,:cg'BfISJx9:c]F8Mqc:w7!g#ik7Puv5`d557!]¥`J;,L%:cLfiko<8Mg,ar5=<w9:d:cw7P8cd;,]e7!;#w6Lf57A[uv8
L5u7!9:7?fTWVE · k=Q5dA:g#g2_¸ ¹UºcE»· ¥`JMyW9fm]
¼ ½¾ rEJ¿@IpJ\D¿c@Q V T2V D@P TWV D'¿¿¿@PÀ ÀT2V DFEIpJc¤
V
Á
Â?ÃfÄªÅÆ*ÇzÈ»É Ê
Ë =?¸Ì9f:c]xw9:cde:cw7P8Mde;#]^7!;#wv7!98ML%7!98k:c5hM8vj*de5amfw7Wuv;#g,g8JEIpJMy
!#"%$'&
(*),.-r&*&
¬ :;#L iU8Mdel%:c7! ]/;,7P7!g,8_W98M5de8Ml%[Auv8_Lf5uÍ7P9f:c7MH EÎ_=Q5d/:cLfiGZ~ H yW9fm]¯BHrISE
@CBI D!H\[*;§y 8fyfB*HkIKk9:]x8Mq:cw7!g,ik5Lf8vd557zy
!#"%$'&
(*),.-r&
+
8M7A_¸Ï8k:pd557W5=A>*[*]^5_ ¢ ¢ JEZY*yW9mf]
@Q ¢ J\D n¢ @Q ¢ J\D n¢ J/E@Q n¢ n¢ ¢ J\D ¢ @Q n¢ J\D ¢ J/EZ F¢ ¢ JEZY
]^5 ¢ Jn;#]:¯de55725=ÐfyfW9fm]Ñ 8Ml%8Ma]F;#Lf7P5N¸b[f;y 8fy2ÑÒE ~
Ó @ÔÐ*DFÒE ~ @Q ¢ J
D ¸b[:cLaN]e;,Lw8
Õ Ñ Õ E Õ ¸ Õ E×Ör³!]^;#Lfw8%>%:cLfabÐ:cde8p57P9_;,dde8Mamfw;,g#8wm;,w]^´z['uv8pl%m]^7ª9f:chM8rÑ ÒE ¸Øy2Ùfqj*g,;#w;#7!g#i§[
7!9;,]F;,]e5l%5dej*9;,]elÚ]^8MLaf] ¢ Jn7!5Ûe[uv98Mde8`k;,]F:¯de55725=A>.;,L¸Ï:LaÛ;,]F:¯de55725=Ðv;,LÑ³!:cLfa
]^;#Lfw8r¯:cLfaÛx8ML8cd:c7!8.¸Î:LaÑ?[de8M]^j*8Mw7!;,hM8Mg,i§[f7!9;,]n;,]8ML5mf9N7P5ra8c7P8Mdel%;#Lf8p7P9f8p;#l%:85=:cLfi
8Mg,8Ml%8MLf7Pćy
!#"%$'&
(*)§-/
021 i ¬ de7P;,LW98c5d8MlÝÜ*y Á ²:cLfaÞ5d5g,g,:cdiÜ*y Á É [ª7P9f8Lfml%8Md65=?;#]^5l%5dj*9f;#]^lÏwg#:]e]^8M]X5=Aaf8Mß
d8M8.o<8Mg#a©8Mqc7!8MLf]e;,5L]¸à5=sÝEâáN@QB<D67P9f:c7Fde:cl%;,=Qi_:7/¥`Jp;#]k7!98_]^:cl%8_:]v7P9f8%Lml%8cd
5=W;,]e5l%5dej*9;,]^lÚwg#:]e]^8M]5=Ww5LfL8Mw7P8Ma\ßC]^98M8c7P8Ma%de:cLfw98ca%w5hM8Mde;#Lf]n5=?sZ7P9f:c7de:cLfw9%:c7
¥`Jcy¦*;#Lfw8%uv8rw:Lw5L]^7!dmw7:cLfi%]emfw9d:Lw98Maw5hc8Md;,L_irwm7!7P;,LN:cLaj:c]^7P;,L_7!9de8M8
w5j*;#8c]x5=Aák[uk8vLf8M8Ma%5Lg,iv]e5g,hM8p7!98kL8Mq72j*:d7zy
{ 1}Þªm7áã:cg,5LZ@I/äZºIpJå?:La×æ#JcºzäçDzyW9f8bw5hM8Mdv;,]7!9m]w5l%j*g,8M7!8Mg#iaf8M7P8cdl%;,L8MaØi7!98
j*8cdl%m7!:c7!;,5L]`è V :cLfa¡è TWV 5=7P9f8]^98M8c7P]v57!:c;,L8Ma i5;#Lf:cde5mLfa©7!98Nj*5;#Lf7P] ¢ J¯:cLfa
IpJz[de8M]^j*8Mw7!;,hM8Mg,i§[;#L%:vw5mfL7!8MdßQwg,5wuk;,]^8/af;#de8Mw7!;#5L i Â de5j*5]e;,7!;#5LÜ*y Á
é [*7!uv5¯]^mw9%w5hM8Mde]
:d8;,]e5l%5dej*9;,w;,=:La¡5Lg,i;,=7!98:c]^]^5w;,:c7!8Maçj*8cd° l%m7!:c7!;,5L].:cd8w5LUêm:c7!8fyç357!87!9:7
w5LUê!mf:7P;,Lp:kj*8Mdel%m7!:c7!;#5L;#]A7!98X]^:cl%8/:c]A:cw7P;,Lv7!98Xj*8cdl%m7!:c7!;,5L¯5L`:ka;,=Q=P8Md8cL7g,:c8cg#;,L
5=W7!98k]e9f8M8M7!]
y
8c7këEì@J\cD![í Eì@PcD![Xí Eì@JD![6:cLaGí Eì@J\DP[6]e5¡î EðïJºë<ºë ºí ºí ºí y
ò 8w:LK]e;,l%j*g,i7PdeiN:V g#gxw5l%;#Lf:c7!;#5Lf]¯5=/èfó V 7P5©5 7P:;#LK7!98b=Q5g,g#5 uv;,Lo<hM8b;#]^ 5l%V 5dj* 9f;#]Ûl ñ
³!w5LUê!mf:wiM´wg,:c]^]^8M]F5=@Pè V ºeè T2V Dzô
õ @QëfºeëD2de8Mj*de8M]e8cL7!]7P9f8kwg#:]e]nuv98Mde8¯è<ó V :d8k7P9f8k]^:cl%8`
ßCwiMwg#8<y
õ @Qëfºeë\Dd8cj*d8M]^8MLf7P]7!98kwg,:c]^]Fuv;,7P9bè ó V af;#=P=Q8Mde8ML7x³!;§y 8fy*5j*j*5]^;#7!8M´6
ßCwicwg,8M]
y
õ @Qëfºeí V D2d8Mj*de8M]^8ML7!]7P9f8/wg#:]e]nuv;#7!9bè V :.
ßCwicwg,8p:cLabè TWV :¯7!d:L]^j*5]^;#7!;#5L
Â?ÃfÄªÅÆ*ÇzÈ Ü ö
õ Q@ í V ºëD2d8Mj*de8M]^8ML7!]7P9f8/wg#:]e]nuv;#7!9bè V :p7!d:L]^j*5]^;#7!;#5L%:Labè TWV :.

ßCwiMwg#8<y
õ @Qí V ºí Dd8Mjd8M]^8ML7!]7!98kwg,:c]^]Fuv98Mde8¯è ó V :d8ka;,=Q=P8Md8MLf7W7Pde:cLf]ej*5]^;,7P;,5L]Uy
W9f8Md8¯:d8`L5_l%5de8¯w:]e8M]67!5_w5Lf]e;,a8cd#[W:]=P5g#g,5uv]
y;#de]e7A5=:g#g,[2Lf8M;#7!98cdF5=Fè ó V w:cLb87!98
;,a8MLf7P;,7!i§[?8cw:cmf]e8%7!98cL7P9f8%w5hM8Mduv5mfg#a©L57Fde:cLfw9:7F5L8%j5;#Lf7«y ¬ g#]^5[è ó V w:LL57n8
7!98k]e:l%8p7Pde:cLf]ej*5]^;,7P;,5Lr8Mw:m]^8k7!98cL%7P9f8/w5hM8Mduv5mg,a%L5728kw5LfL8Mw7P8ca?y
÷1 ò 8o<LaS7!98j5g#iMLf5l%;#:g#]N>A@CBªºø*D;,Lç7!98]^:cl%85dea8Mdp:]`:5hM8fy ¬ Lf57P85LSjd5hM;,L ;#deß
de8Mamfw;,;#g,;#7!ifôFm]^;#Lf ¬ d7!;#Lf ]8ML8Mde:cgWW9f8M5d8Ml®ö?y öªy ù:cLfa Â d5j*5]^;#7!;,5Lbö?y ²ªy,Ü
[2uv8.w:Lb:j*j*g,i
l%5]^7f5=7P9f8Xl%:cw9;,L8MdeiX=Q5d2jd5hM;,Lv;,dd8camw;#;,g,;#7!iX5=;,L7!8M8Mdj5g#iMLf5l%;#:g#]A7!5v7P9f8Xd;,L.áNæ Båú[
uv9;,w99f:c]Xo<8Mg,ab5=A=Pd:cw7P;,5L]páN@QB<D«y3657!8`:g#]^5b7P9f:c7A;#L5dea8Md7!5j*de5hM87!9:76>Nd:l%;#o<8M]8Mqß
:w7!g#ib:c7v¥`J«[F5L8w:cLKo<La 7P9f8d557!]v5=ûW> Ó û<ø:cLfaK]e5g,hM8=P5d/7!98NhM:cg,m8c]p5=B]^mw9K7!9:7
>A@QBºø*DxEY:c]xuv8Mg,g§y
õ >A@QBºeøDxEZøxIØ@CB ¢ J\D@CB.I J
DP[;§y 8<yuv8/uv5mg,a%g,;#8/7!5¯]^:ciøW@CB'DFEìý ü B ¢ Jý ü B_IJMyW9f;#]
9f:c]X7!d;,j*g,8rd557!]:c7BçEÌ¥`J.³!]^;#Lfw8NûW> Ó û<øEÍMø%EÍYÿþ ø©EÍYÿþ BçEÌ¥`J«´
:La ;#]p;,dde8Mamfw;,g#8iÙ<;#]^8ML]^7!8M;#Lf ]vwde;#7!8Mde;#5LG³!uk;,7!9©j*de;#l%8B ¢ Jz´ ° 7!9m]v;,7P] A;,8Ml%:cLfL
]^mde=Q:w8%;,]/;,]e5l%5dej*9;,w¯7!55L8%5=x7!98ro<d]^7F7!uk5w:c]^8M]/:5hM8fyk365uA[A;#=B;#]khM8Mdei%g#:d8
7!98cL©ø'@QB<DFuv;,g#gª8`:j*j*de5qc;,l%:c7!8Mg#iB< ^^[?]^5N5;#Lfw5mL7!8MdeßCwg,5wuv;#]^8%:cde5mLa:%w;#dewg,8
5=AhM8cdig,:cde8rde:ca;,m]uv;,g#g;#Lfwd8c:c]^8`7!98.:cdemfl%8ML75=FøìN ¿M Ó N
E \[W]^5ø.uv;,g#g
8%af;#=P=Q8Mde8ML7F:c=Q7!8Md5;#LfN:d5mfLa7!9;,]g,55jªy mf7F;,L7!98%]^8Mw5Lfaw:]e8%:5Ó hM8M[A]e7!:cde7P;,L
uv;,7P9tBç=Q:cd`=Qde5l 7!985d;,;,L¡5L7!98Lf8M:7P;,hM8©;#l%:;,L:di:qc;,]r:cLa5;#Lf :d5mfLa:
w;#dewg,8_w8cL7!8Md8Ma:7n7!98r5de;#;#L ;,L:_w5mL7!8MdeßCwg,5wuv;#]^8_a;,d8cw7!;#5Luv;,g#gAde;#LfNiM5m:cw
7!5_uv98cd8¯iM5mb]^7P:d7!8Ma ³!iM5mb5a5uvLb5L8¯]^98c8M7A:cLab7!98MLb:cw.mfj*´z[']e5>A@CBºeø*DAl%mf]e7
9f:chM8v: A;,8Ml%:cLfL%]emfd=Q:w8v;,]e5l%5dej*9;,w67P5.7P9f8/5Lf8v;#L_7!98ko<d]^7Ww:c]^8fy
õ >A@QBºeøDxEZøIN@CB ¢ J\D@QBnI_J
D«y*¦*;#l%;,g,:cdg,iX7!5v:c5hM8M[7!9;,]9:]ª7!d;,j*g,8d557!]:7'BbE¥`J?:cLfa
;,];,dd8camw;#g,8i/Ùf;,]^8ML]^7P8c;#L³!uv;#7!9.j*d;,l%8kBpI J«ćy|X5uv8MhM8cd#[L5uøW@QB<D tB¯=P5dªg,:cde8pB[
]^5=P5g#g,5uv;#Lf_:rg#:d8rw;,dwg,8.:cd5mfLa7!98.5d;,;,L;#L:rw5mfL7!8MdßQwg,5wuk;,]^8àf;#de8Mw7!;#5Luv;,g#g
Lf57xw9f:cLf8r7!98.:cdml%8cL7x5=Føc[]^5Nic5muk;,g,g8MLfamjb;#L7!98.]^:cl%8rj*g,:cw8riM5m]^7!:cd7!8MaAy
W9fm]x7!98 A;#8cl%:cLL]emfd=Q:w8k=Q5d>r;#]n;#Lfa8M8Mar7!9:725='7P9f8k]^8Mw5La%w:]e8<y
õ >A@QBºeøDEÍ@QBNIGJ
D§øXIØ@QBItJ
D§ø ¢ @CB ¢ D«yW98p:L:g#iM]^;,];#LN7P9f;#]6w:]e8;#]6:;,7?l%5d8
af;#=Pofwmg,7«y ò 8p9:hM8`û'> Ó û<ø.EZf@CB_IJ\D@Cø6IGJ\D![*uv9;,w9%;,]
8Md5.uv98MLNø%E¥`JF5dnBbEJMy
W9f8pw:]e8.øE J;,hM8M]L5%]^5g#mf7P;,5L=P5dB[<uv9;,g#8røNE IpJ;#l%jg#;,8M]XBE IpJz[fuv9;,w9;,hM8M]
>A@IpJcºeøDpEÝIUø ¢ ø ¢ E I*@Qø%IGD@Qø ¢ J
Dkuv9f;#w99:]:af5mg,8%d557zy ò 98cL
BE×J«[7P9f8Md8k;#]FL5.]e5g,m7!;#5L=Q5dFø ° 9f5uv8MhM8Md,[uv8kw:cL%de8Muvd;,7P8k5md8 mf:c7!;#5L%:]
BbE øø X I cø¯I ¤
I cø ¢
365uA[cuv9f8ML`BNEJz[M7!9;,]2w:Lp8nuvd;,7!7P8MLv:]Y¯EIXÖ Ó @QøMIMø ¢ DP[
uv9;,w9p9f:c]2:7!d;,j*g,8nde557
:]Fø ÿ ä 7P9f:c7';,]e[:]xB ìJ«[:cg,g7!9d8c8X]e9f8M8M7!]ª5=<7!98 A;,8Ml%:cLfL`]^mde=Q:cw8w5LhM8Mde8c[]^5
:g#7!95mf9r7!98Mde8X;,]L5¯j*5;,L7n@Jcº«ä¡DW5L`7!98 A;#8Ml%:LLr]^md=P:cw8M[;#='uv8o<g#g;,Lr7P9f8Xj*mfLw7!mde8
³!de8Ml%8Ml%8Md.áNæ Bªºøå Ó @P>WDX;,]¯7!98bof8cg#aØ5=6w5Lf7P;,Lmf5m]`=QmLfw7!;#5Lf]¯5LØ:d:Lw98MaKw5hc8Md
"!#%$ &(')!*'+&-,(.%/0, &1! ´z[*uv8p57P:;#L:¯d:Lw9f8Ma%w5hc8Md7P9f:c7ªd:cl%;,o<8M]xuv;#7!9_:¯7!d;,j*g,8
de557W:c7Jn:La%:paf5mg,8kde557W:c7IpJz[uv9;,w9%;,]F;,La8c8Mar7P9f8kw5hM8MdA;,L_7P9f8/7!9;,da%w:c]^8fy32
465798%: 79@A79BDCFEG BQ79RETS+UWV(X RE 7 8 7 7 :E 8\E B>8 @fE E 7
GD7 ?TYQ g E ?T;-;(<>_h=? = 7iBQE H* 7 =" ?e=1=<IHJ7 =LZUj KNH : MP k O-G H HJ=! 7 ?Ta ;z H)Y =! ?%Z 0 [MF ?\H ][^M`_badc< Y Ye=# = HP ? =
Â?ÃfÄªÅÆ*ÇzÈ é ²
õ A> @QBºeøD_E @CB ¢ J

D§øpIt@QB ¢ J\D§ø ¢ *@QBIGDzy%W9;,]p]^;#7!m:7P;,5LK;#]p:L:cg,55m]¯7!57!98
jd8MhM;,5m]F5L8fy
õ >A@QBºeøDxEZøI`MøI`MB2y ò 89:chc8kû'> Ó û<ø.Etf@CøIJ\DEZY×þ ørE¥`Jtþ BNE¥`Jz[
]^5_>.d:l%;#o<8M]:c7ABNE¥`J ° ;,Laf8M8Ma['>A@¥`JcºeøDE¨@CøpID@Cø/¥GJ\D«[*]^5_>`9f:c]:paf5mg,8/de557
:7n¥`Jcy'W9mf]n;,=x>%;,]6;#ded8Mafmw;,g,8p7P9f8ML;,7ªl%mf]e7?9:hM8p7!98lA;,8Ml%:cLfL]^md=P:cw8p5=2w:]e8¯o<hM8fy
ò 8kw:L%m]^8/7!98kd:7P;,5L:gd557'7!98M5de8Ml7!5w9f8Mwp;,dde8Mamfw;,;#g,;,7Pi ]^;#Lfw8.>`;,]l%5L;,w:cLfa
wm;,wv;#Løc[2uv8`Lf8M8Mab5Lg,i`w98Mw%uv98M7!98Mdn:rw5L]^7!:cL75d:rw5Lf]e7!:cLf77P;,l%8M]/B]e:7P;,]ôf8M]
>ªy<W9f;#]n;#]wg,8M:dg,i¯]e8c8MLLf577!5`8p7!98pw:c]^8pi¯8c8Mj*;,L7!d:w`5=27P9f8v9;,9f8M]e7?a8cde8M8p5=
By
Ë =fiM5màf;#afL 7f8M7'7P9f8Xj*5g,iML5l%;,:cg,]A;#L.7!98X7!9;,da`:cLa.=Q5mfd7!9¯w:]e8c[a5L 7f=P8M8Mg:cma ¡Lf55ai
af;#afo[ n/:La Ë a;,aLf 727!:c8k5=Q='j*5;,L7!]
y
!#"%$'&
(*)qp.-r&
Ë =srb;#]n:cg,8Mde:c;,wX5hM8Mdnst7P9f8MLtrN]^:c7!;#]ô<8M]F]^5l%8vj*5g#icL5l%;,:cgW>s%æ Båú[*]^5r]e;,Lw8.>s%@vu?Dæ B*å [uv8
L5uç7!9:w7 r.;,]2:g#8Md:;#wF5hM8Md?s%v@ u?D«y Ë L= rr;#]ª7Pde:cLf]ew8MLaf8ML7!:cgM5hc8Md?sK7!98cL*ï u ñ ;,]:X7!d:L]^w8MLfa8MLfw8
:c]^;,]F=Q5dn¸Ì5hM8Mds[]e5.iva8Mo<L;,7P;,5Li[ r;#]F:cg,8Mde:c;,wX5hM8Mdns%x@ uªD«y
!#"%$'&
(*)qp.-z
8M7{}|~x@ D ~@xUDk87P9f8;#]^5l%5dj*9f;#]^l 7P:;,Lh cy ò 8uv;#g,g6]e9f5uã7!9:7{Zw:LL57
88Mq7P8cLa8caS7P5K: w5Lf7P;,Lmf5m].=QmLfw7!;#5L 4³!7P9f8]emf]ej*:w87!5j*5g,5i5=7!98w5mfL7!:cg,8
]^8M7 ~v@ D ã;#]v7!98_af;#]^wde8M7!8%7P5j*5g,5mi úćy¯8c7B À ~Ï8%:N]e8mf8MLw87!9:c7nw5LfhM8Md8M]k7P5Pf[
;§y 8fty 6q À l B À E Ay ò 8N9:chc`
8 {6@CB À D%E B À ]^;#LfwA 8 {çofq8M] ~[nm^7 6q À l {6@CB À D_
E Ø:cLfa
{6v@ q6_QÀ l¡BfÀDx E y
,Q'-%#\ ,(..
? G 8 g\ ?\; :E a

Ã EG HP +=6<IHJ= @fE ;+H R E JH ? E Ex:E ?*=9HP]W 77 Z 7 Z=6< Y ?Y EG =! 7 ?Ta
SOLUTION SET # 11
GREGG MUSIKER

√
Problem 1 Determine the irreducible polynomial for α = i + 2 over Q.
There were several ways to do this problem. The basic idea is to find a linear
combination of powers of α that equals zero. Then one needs to explain why the
associated polynomial is irreducible.
√ √
α2 = −1 + 2 −2 + 2 = 1 + 2 −2. Thus (α2 − 1)2 = −8 hence α satisfies
(x2 − 1)2 + 8 = x4 − 2x2 + 9 = f (x). It is probably easiest to prove that this
is irreducible by the theory of field extensions (rather than
√ the tricks from chapter
√
11). Namely, let K be the splitting field √ for f (x).
√ Q( 2, i) contains K. Q(√ 2)
has degree 2 over Q and √ since i 6∈ Q( 2), Q(i, 2) has degree 2 over Q( 2).
Consequently, K = Q( 2, i),[K : Q] = 4 and f (x) is irreducible.
Problem 6 Determine the degree of the splitting field of the following polynomials
over Q.
a) x4 − 1. One can quickly recognize the roots ±1 and/or that x4 = 1 means
the fourth roots of unity will be the roots of this polynomial. Hence x4 − 1 =
(x − 1)(x − i)(x + 1)(x + i) so the splitting field is Q(i) which has degree 2 over Q
since i satisfies the irreducible polynomial x2 + 1.
√
b) x3 − 2. α = 3 2 is clearly a root of x3 − 2. Then after factoring and applying
quadratic formula (if needed) one factors x3 −2 as x3 −2 = (x−α)(x−αζ)(x−αζ 2 )
where ζ is a complex cube root of unity. ζ 2 + ζ + 1 = 0 and ζ 6∈ R hence 6∈ Q(α) so
the splitting field of x3 − 2 has degree 3 · 2 = 6. In fact the splitting field is Q(α, ζ).
√
c) x4 + 1. This polynomial
√ √
has the fourth roots of (−1) as roots thus “± ±i“,
better known as ± 2±i 2
are the four roots of x4 +1. The splitting field will contain
√ 2
both i and 2 and similar to the reasoning in problem 1, we can claim that the
degree of the splitting field is 4.
Problem 8 Let ζ = e2πi/5 .

a) Prove that K = Q(ζ) is a splitting field for the polynomial x5 − 1 over Q, and
determine the degree [K : Q].
ζ 5 − 1 = 0 by inspection, and in fact x5 − 1 = (x − 1)(x4 + x3 + x2 + x + 1) =
(x − 1)(x − ζ)(x − ζ 2 )(x − ζ 3 )(x − ζ 4 ). Thus K is x5 − 1’s splitting field, and since
ζ 6= 1, ζ is the root of an irreducible (cyclotomic polynomial) polynomial of degree
4. Hence [K : Q] = 4.
1
2 GREGG MUSIKER
b) Withouot using Theorem (1.11), prove that K is a Galois extension of Q, and

determine its Galois group.
Let σi send ζ to ζ i . Then σ1 , . . . , σ4 are all automorphisms of K. Since
|Aut(K/Q)| = 4 = [K : Q], K/Q is Galois, and the Galois group is Z/4Z. No-
tice σ24 = σ1 (16) = σ3 (8)σ2 = σ42 thus Gal(K/Q) is of order 4 and has an element
of order 4 thus it cannot be V4 and must be Z/4Z.
√
Problem 12 Determine all automorphisms of the field Q(3 2).
From class we saw that if f (α) = 0, then the √ automorphisms of Q(α) send α to
another root of f (x). This is true if we let α = 3 2. But the other roots are β = αζ
and γ = αζ 2 , both which are not real. Thus an automorphism of Q(α) cannot send
α to β nor γ, thus the only possible automorphism is the identity map.
This highlights the importance of an extension being Galois. Q(α)/Q is not
Galois because Q(α) is not a splitting field for any polynomial in Q. As we saw in
problem 6, α is a root of x3 − 2 whose splitting field has degree 6 over Q.
Problem 13 Let K/F be a finite extension. Prove that the Galois group Gal(K/F )
is a finite group.
To make this easiest to explain, we will assume that F has characteristic zero,
and apply theorem 14.4.1, the primitive element theorem. Thus ∃ α ∈ K such
that K = F (α) since [K : F ] finite (without char zero and the primitive element
theorem, K = F (α1 , . . . , αn ) we just pick α ∈ K which is a root of f (x) ∈ F [x])
and thus α is a root of polynomial f (x) in F [x] of degree [K : F ]. Using the fact
that an automorphism of K (that fixes field F ) must send α to another root of
f (x) ∈ K. But since the number of roots of f (x) which lie in K ≤ [K : F ], the
automorphism group must also be finite. In fact the group will be a subgroup of
the symmetric group S[K:F ] since the automorphisms will permute the roots.
√
Problem 14 Determine all the quadratic number fields Q( d) which contain a
primitive pth root of unity, for some prime p 6= 2.
Let ζ 6= 1√satisfy xp − 1 = 0. [Q(ζ) : Q] = p − 1 since this a cyclotomic extension.
If ζ ∈ Q( d), this implies we can create a tower:
√
Q( d)
Q(ζ)
Q.
√
[Q(√ d) : Q] = 2 and [Q(ζ) : Q] = p − 1 so the only way this tower can occur is if
√
Q( d) = Q(ζ), p = 3. In fact Q( −3) = Q(ζ3 ) but this is the only time that a
quadratic number field contains a pth root of unity for p > 2.
Problem 15 Prove that every Galois extension K/F whose Galois group is the
Klein four group is biquadratic.
By the main theorem, if the Galois group has three subgroups of index 2 (as the
Klein 4 group does) then K contains three subfields containing F which have degree
2 over F . Let two of these subfields be F (β) and F (γ). Since these subfields are
distinct and both have degree 2 over F , β 6∈ F (γ) and γ 6∈ F (β) but β, γ ∈ K.
Thus F (β, γ) has degree 2 over F (β), F (β, γ) has degree 4 over F and K ⊃ F (β, γ)
implies K = F (β, γ). Thus K is the splitting field for (x − β)(x + β)(x − γ)(x + γ)
where (x − β)(x + β) splits in F (β) and (x − γ)(x + γ) splits in F (γ). Thus K/F
is a biquadratic extension. Note: the third subfield of degree 2 over F is F (βγ).

Problem 1 Let G be a group of automorphisms of a field K. Prove that the fixed
elements K G form a subfield of K.
K G is defined as the set {x ∈ K : g(x) = x for all g ∈ G}. Since K G is a
subset of a field, associativity, commutativity, and distribution will follow as long
as K G is closed under addition, subtraction, multiplication, and division. Note
that being closed under subtraction implies the additive identity is in K G and
likewise for the multiplicative identity. Since g ∈ G is an automorphism of K,
g(x+y) = g(x)+g(y), g(0) = 0, g(1) = 1, and g(xy) = g(x)g(y). Assume x, y ∈ K G .
Then g(x + y) = g(x) + g(y) = x + y and g(xy) = g(x)g(y) = xy. g(−x) + x =
g(−x) + g(x) = g(−x + x) = g(0) = 0 implies g(−x) = −x. g(1/x)g(x) = g(1) = 1
implies g(1/x) = 1/x. Thus K G is a subfield of K.
√ √
−1+ −3
Problem 2 Let α = 3 2, ζ = 2 , β = αζ.
a) Prove that for all c ∈ Q, γ = α + cβ is the root of a sixth-degree polynomial of
the form x6 + ax3 + b.
Note that γ = α(1 + cζ). Thus γ 3 = 2(1 + cγ)3 ∈ Q(ζ). γ 3 6∈ Q unless c = 0.
Assume c 6= 0. Since Q(ζ) has degree 2 over Q (satisfies x2 + x + 1), γ 3 must satisfy
an equation of the form y 2 + ay + b. Consequently, γ satisfies x6 + ax3 + b. If c = 0
then α6 = 4 and α satisfies x6 − 41.
Alternate proof2. The field K = Q(α, β) is the splitting field for x3 − 2 over
Q. This extension has degree 6, its Galois group is S3 , and the three roots of
x3 − 2 are α, β, and βζ = δ. Thus the orbit of γ = α + cβ under the group action
of Gal(K/Q) is the set {α + cβ, β + cα, α + cδ, δ + cα, β + cδ, δ + cβ}. Thus the
irreducible polynomial for γ will divide
g(x) = [x−α(1+cζ)][x−α(ζ+c)][x−α(1+cζ 2 )][x−α(ζ 2 +c)][x−α(ζ+cζ 2 )][x−α(ζ 2 +cζ)].
We expand this expression to get x6 + ax3 + b where a = −4c3 + 6c2 + 6c − 4 and
b = 4(c2 − c + 1)3 .
1Thanks to Philip Zeyliger.

2Thanks to Noam Zeilberger.
4 GREGG MUSIKER
To see this expansion, we note that ζ + ζ 2 = −1 and the constant term is the
product of
α6 [(1 + cζ)(1 + cζ 2 )][(ζ + cζ 2 )(ζ 2 + cζ)][(ζ 2 + c)(ζ + c)] =
α6 [(1 + cζ)(1 + cζ 2 )][ζ(1 + cζ)ζ 2 (1 + cζ 2 )][ζ 2 (1 + cζ)ζ(1 + cζ 2 )]
where each bracketed expression is equal to 1 − c + c2 . One can show using the fact
that (ζ 2 + ζ + 1)x = 0 that the degree five, four, two, and one terms are zero and
the degree three term of the expansion is as above.
Yet another proof. Many people solved this problem using the following method:
γ 6 = c1 + c2 ζ + c3 ζ. Some people noticed that one can rewrite this as γ 6 = d1 + d2 ζ
or d1 + d3 ζ 2 since ζ + ζ 2 = −1. (Here the coefficients are all rational.) Using this
and the fact that γ 3 = e1 + e3 ζ 2 , one finds that γ 6 − d3 /e3 γ 3 is rational. e3 6= 0
unless c = 0 a case which implies γ 6 = 4.
b) Prove that the irreducible polynomial for α + β is a cubic.
Here, I will use Noam’s observation that α + cβ satisfies x6 + ax3 + b where a =
−4c3 + 6c2 + 6c − 4 and b = 4(c2 − c + 1)3 . (Alternatively, one can just show through
explicit computation that α+β satisfies the cubic x3 −2.) Since c = 1, α+β satisfies
x6 + 4x3 + 4 which factors as (x3 − 2)2 .
c) Prove that α − β has degree 6 over Q.
Proving this amounts to proving that α − β is a primitive element for the splitting
field of x3 − 2. Thus letting ai , bi ∈ {α, αζ, αζ 2 }, we form the quotients −a i +α
bj −β for
i 6= j like on page 553. None of these equal −1 so α − β is the primitive element.

!#"%$
&('*),+-)
.0/21436587:9<;>=@?A;>B,CEDGFIH>J!KL36MNKO536J*MP KRQSOJ!536SJ*TMEUWVX?YD[ZRKL\T]^K,5\N_`587aJ*KbJ*SSO36cdNQJ*ef!3NFagh\3
cNJ!KLQSOJ!i#J*M_jM5:kXJ*KlCmDjnporq#sutp]^\J*Q\#J*KvMT5h_KLwGd_jS3^J!MyxRzKOT,5\3l{|_jf!TJ*Kv}4STd~T7Ri#dNKO5
e3, B F
/14365:7:9W;%=?;NBvD4n
;,Cq(F<7R]^36SO3-MNT5hJ!SOS36cNdQJ*ef*3l5\3`M,KLJ*MQ3^J*50J*Kv_î#TMJ*QJ!MN536}436SQdeJ*Q
]^J*5\#QTMKL5_jM5h536Siq
zJ*50]^Tdf*c#\_``3l5Tb\N_`63,Rzh^D[zTS-qp_jKv_STT5z]^\J*Q\J*50cT36KMNT5F
gh\N3-cNJ!KLQSOJ!i#J*M_jM5J*K^Cqô
DjnBCaDjn-o`9Cq4=^st
z(KOT,_`K_èT636z5\N3-{_`f*TJ!K}jSOTd~T:7RJ*Kp B F
!#"%$
&('*),+-
14365:k?E|e3^5\3^c36KLQSOJ!i#J*M_jM5T:7F(:,@5\36My?E#96=zKLT7,J*KJ*SSO36cdNQJ*ef!3T`36S#9<6=
5\dK]^38Q_`M_`KLKOdNi#3b5\_j5v I ,F¡J!MNQ3y¢ #9<6=l£Nl¤?¥Dl_`Mc¦¢ #9<§:=^£(l¤:?¨^<TS|_`M©,STT5v§T
7(z0]^3,ªjMNT]¥5\_`5MNT«SOTT5T7J!KpQTM5_`J*M36cyJ*M¬#9<6=5\dKmKOJ*MQ3«7J!KQdeJ*Q]^3R\N_`635\_`57J!K
J*SS3`cdQJ!ef*3^T636S|#9<6=rF
!#"%$
&('*),+-
14365§®¥e3-5\3lS3`_`f(SOTT5hT7(z_`Mcf*365¯v°>²±¯ e3-5\3lQTi#~f!36³^SOTT5KFG´#3^ªjMNT]5\_`55\N3-cN36}jSO363
¢µxy9<§:=£Gx,¤h?¶¨mKLJ!MNQ37J!K|_`M·J*SSO36cdNQJ*ef!3^QdNeJ!Qz_jMc·5\_j5¯v° ¶±¯ x@9<§:=FNgh\dNK^9W;²C¬¯=9W;²C :±¯ =J!K
J*SS3`cdQJ!ef*3^T636Sxy9§=zKLTR5\3lc3`}jSO363^T¸¹?ºxy9§°O¯=T63`SxJ*KDlo[¨?¶»F
!#"%$
&('*¼,+-½
¾ KOJ*M}R5\N3-~(SOJ!i#J*5J!63^36f*36i#36MN505\36TSO36i#zG_`KLKOdNi#35\N_`5¸¹?¿#9§=_`McyÀN?#9<¯=zNKOT
¸ À ?¶#9<§|°¯=|?#9<¯v°O§:=?¿ À 9§=ÂÁ
14365m7@e3#5\3«i#J!MNJ!i#_jf~Tf!©6MNTi#J!_jf:<TS^§aT636S8mz_`Mc¬f!3`5pÃ¹?ÅÄ^Æ`Ç9<¸8=rF8´#3·ª4MT]È5\_j5p7
KL~(f!J*5KbT636SR¸ÉKLJ!MNQ3¬¸6ÈJ*K8{_`f*TJ!KR_`MNca_`M©~(Tf*©6MNTi#J!_jf]^J!5\a_ySTT5J*M_y{_`f*TJ!Kb36³`536MKLJ*TM
KL~(f!J*5K
F0gh\dKm7#KO~(f*J*5KT636S¸¬ÀÊ¸_`K]^36f!f*z>_`Mc²KOJ*MQ3¸Àh?ºÀË9<§:=_jMc@7:9<;>=mÀ¢ ;¤Ìz%]^3\_j63
5\_`5¸ À J*K5\3lKO~(f*J*55J!MN}bÍ>36f*c#<TS7,T`36S| À zNKOT«¸ À À J*K{_`f*TJ!KF
1j365vÃÀ%?ÎÄ^Æ`Ç9¸ÀÏÀ*=z>_`Mc«MT535\N_`5_jM©36f*36i#36MN5TÃ¶TSÃÀGJ!Kc3`536SOi#J!MN36c#e©\T]ºJ*5_`Q5K
TM§vFÐf!36_jSf*©b_jM©b3`f!36i#3`M5vÑGÀTÃÀÍ>³`3`KpÓÒºÀµz%_`McKLJ!MNQ3,ÑÀ(5_jªj3`Kp§5T#_`MNT5\N36SSOTT5Tv7(z
Ô
ÕÖN×ØÙ(Ú ºÛ Ü
ÑÀjKO5_èJ*f*J!Ý
36Kp¸aFGgh\NdK|ÑÀÞ ßXyÃmz(KLT,]l3^\N_`63^_i#_`~²àu£`ÃÀ>áâÃ¿5_`ª4J*M}RÑÀNãáAÑÀËÞ ßRFgh\NJ!Ki#_`~J!K
J*Mä3`Q5J!63^KLJ!MNQ3bÑÀjJ*K5\3^J!cN36M5J*5©^J*h_`Mc«TMf*©J!ÑÀjÍ>³`36K§hz]^\NJ!Q\#\_`~~(36MK36³`_jQ5f!©^]^\36M«ÑÀËÞ ß¿J!K
5\3lJ!cN36M5J!5©F
!#"%$
&('*¼,+-å
.0/2:¸6J*K|{_`f*TJ*K|5\_j5:¸J*K|5\3^KL~(f!J*55J*M},ÍN3`f!c·TKOTi#37:9<;>=m@#¢ ;¤F%æhdN5KLJ!MNQ3ÈÒ¿ç:z]^3
\N_`63^5\_`5¸ÓJ*Kv5\3^KL~(f!J*55J*M}RÍN36f*c#T7:9W;%=ç-¢ ;¤G5TTz(KOT«¸6ç@J!Kv{_`f*TJ*K
F
/14365Ó?ÈxÈ_`Mcyf*3657:9<;>=^?X;>BlCuDGFgh\36MI7\_`K3`³`_`Q5f*©,TM3,S36_jfhSTT5§F14365çÎ?¥#9§=
_jMcf!3`5¸Èe3-5\N3mKL~(f!J*55J*M}8Í>36f*cRT7F4gh\NdK¸6uJ*K{|_jf!TJ*KOzed5%KOJ*MQ3-5\3-QTi#~(f*36³^STT5K:T
7_`SO3MT50QTM5_`J*M36c#J*Myç:z%ç6J*KMNT50{|_jf!TJ*KlUKLJ*MQ3_b~(Tf!©`MTi#J*_`f]^J*5\#_bSTT50J*M·_8{_`f*TJ!K
36³j53`MKLJ!TM#KL~(f*J!5KOZ`F
èG/é14365-ê?ëxRzçÈ?ì#9í D`=zv_`MNçî?Åçm9%ïí D`=zKOT YÒÓçÓÒ¸ðÒÓ®@F,¡(J*MQ3²_`MN©²cN36}jSO363
D36³`536MNKOJ*TMuJ*K{|_jf!TJ*KLz-]^3ª4MT]ì5\_j5R¸çÎ_`MNcçÓ_`SO3eT5\u{_`f*TJ*K
FñmT]l3`636S*zm5\3
J*SSO36cdNQJ*ef!3R~(Tf!©`MTi#J*_`fUe©8H>J!KL36MNKO536J*MZ87:9<;>=p?X;òlCuD²a#¢ ;¤>\_`K_,STT5J!M¬¸Åed5:cT36K
MNT50KO~f!J*50J*M¸óKOJ*MQ37R\_jK5\3lQTi#~(f*36³^STT5ô ïí DzKOT«¸6J*KMT5{|_jf!TJ*K
F
!#"%$
&('*¼,+$($
<5J!K-5SOd3,5\N_`5¸J!K-5\N3RKL~(f*J!55J*M}²ÍN3`f!cyT_jM@J!SOS36cNdQJ!ef*3QdeJ*Q4Fp1j3`5ÃÅ?óÄ^Æ6ÇË9<¸68=lõ? B z
_`MNcyf!3`5ö2s¥Ãe3_#KOdNe}jSOTd~T:TScN36SlD²U3NF }hF8÷4`°9[Dj=ørZm]lJ*5\QTSSO36KO~TMcJ*M}«J!MN536SOi#36cJ*_`53
Í>36f!c«ç¦?¿¸¬ù8FúT53p5\_j5>KLJ*MQ3bçyJ!K:MT5%{_`f*TJ!KT`36SKOJ*MQ3öûJ*KMT5%MNTSi#_`fGJ*M²Ã,F¡(J*MQ3¸6ç
J*K{|_jf!TJ*K]^J!5\{|_jf!TJ*K}jSOTd~·ö²z]^3m\_j63«¢ ¸A£`çv¤%?¶D]^\J*Q\,i#36_`MNKl¢ ç£jl¤%?¨(F41j3`5§ayçe3-_
~(SOJ!i#J*5J*6336f*36i#36MN5T`36Smz>]lJ*5\·i#J!MNJ!i#_jf%~(Tf!©`MTi#J*_`f7:9W;%=p#¢ ;(¤Fh´#3bªjMNT]º5\_j5v7#cT3`KMT5
KL~(f!J*5T`36SçKOJ*MQ3,ç6J*KMNT50{|_jf!TJ*KLzed55\N_`57cT3`KvKO~f!J*5T63`S¸ÓKOJ*MQ3,¸¿J*K{_`f*TJ!KFNgh\dK
¸ÓJ*Kv5\N3-KL~(f*J!55J!MN}bÍ>36f!c#Th5\3lJ*SS3`cdQJ!ef*3^QdNeJ!Q^7F
ü QTi#i#TMR36SSOTShJ!Mb5\NJ!K~(STef*36iº]^_`K5T^KL_`©ýz4þÂ¸J!K5\3KL~(f!J*55J*M}lÍN3`f!c8T`ÿ ~(Tf!©`MTi#J*_`f*
]^3ädNKO5N\_``35Tl~(ST635\_j5(5\NJ!K~(Tf*©6MTi#J*_`fj\_`K_`M8J*SSO36cdNQJ*ef!3QdeJ*Qv<_jQ5TSËF gh\J*KJ*K_`f*KO3NFúT53
5\_`5e©5\N3~(SJ*i#J*5J*6336f*36i#36MN55\36TSO36i#z
@c3`}jSO363^KOJ*³b{_`f*TJ!K36³j536MNKOJ*TM·J*K|5\3KL~(f*J!55J!MN}RÍ>36f*c
T%_`M·J*SSO36cdNQJ*ef!3^KL36³`5J*Q4F
!#"%$
&('*å,+$
úT53 :QTMN5SO_`S©R5T«~(T~(df*_`Se36f*J!3`*zhTMN3bQ_`MMNT5:KOTf*63b5\J*K~(SOTef!36iþLe©,J!MNKO~(3`Q5J!TM:F ²<5:J!KwGdNJ!53
536cJ*TdKv5Tb}4J*635\N3-KLde}4STdN~(K_jMc#~(SOT63l5\N_`5h5\36SO3-_jS3^MT,T5\N36SKF
1j365Ã?A F J!SOKO5mMT53²5\_j5pKOJ*MQ3y_y5SO_`MNKOJ*5J*63KLde}4STd~öY\N_`K85T¬KL36MNc5\N3²f*36553`S²R5T
36_jQ\@T5\36Sf*36553`S!zJ!5i#dKL5\N_`63#TSOc36S_j5f*36_`KL5 F ü f*KOTz:KLJ!MNQ3ö¹_jQ5KmTM@5\3#KL365l÷4`°D°O¨°qN° Gø
e©8~(3`Si#d5_`5J*TMKLz]^38Q_jM«dNKO385\3KL5_èJ*f!J*Ý
36S WTSOeJ!5<TSOi#df*_85T#QTMQf!dNc385\_j55\3TSOc36SvTöJ!K
ÕÖN×ØÙ(Ú
_«i#dNf!5J*~(f!3«TUKLJ!MNQ3·5\3#KLJ!Ý
3#Tv5\3#TSOeJ!5|Tlz:KL_`©ýzJ*K-_jf!]^_j©6K[Z`F^¡(J*MQ3aÞ B Þ?¹
D6t[z5\N36Mz
]^3«Í>Mc¬5\N_`5,Þ öuÞX÷ °t° (°LD`t(°O¨jt(°Oqjt(°L»6t(°D6tø`F-´#3«]^J*f!f~(SOTQ363`c@e©«Í>McNJ!MN}_jf!f5S_jMKLJ!5J*63
KLde}4STd~K-Ud~5TRQTMädN}j_j5J*TMz_>F ª%F _>F(S36f*_è36f*f*J!MN}RT%5\3lf!365536SKLZTh36_jQ\#TSOc36SýF
Þ ö¦Þ ? GF¬gh\NdK#ö ? õ 6F¬gh\N3TMNf!©36f!3ì#36M5K8TpTSOc36!S ·J*M ,_`S" 3 WQ©6Qf!3`KOzpKLTuöÉJ!K
QTMädN}j_j53^5$ T #O9
D6¨6
q `= %F
Þ ö¦Þp?2tFIgh\N3y¡(©6f*T]Å5\N36TS36i#Kb536f*fdK5\N_`58öÉQTM5_`J*MK,_KLde}4STdN~TpTSOc36! S [zpKO_j©
& ? 9
D6¨6
q `=rFbgh\N3·KL5_jeJ!f*J!Ý
36SlT_`M©²36f*36i#36M5]^J*f!fe3·_KLde}4STdN~TTSOc36SbDGF,¡(J*MQ3'
J*Km}436M3`S_`536cye©_`MN©#ÍN`(3 WQ©`Qf*3#_`Mc5SO_`MKL~(TKLJ!5J*TMz]^3#ª4MT] 5\N_`5v5\3KL5_jeJ!f*J!Ý
36ST-rzTS
J*MKL5_jMQ36z(i#dKL5e3l_cTdNef!3^5S_jMKL~(TKOJ*5J!TMNz(JËF 3NFGJ*50J*K36J*5\36S-9D`¨j=9W)q j=z9Dq4=9<)¨ `=zTSl9D j=Â9<¨`q4=rF
úT]:zv9[D`¨6)q j=Â9D6¨j=9W
q `=v?û9[*¨ j=]^\J*Q\«\N_`KTScN36S5\NS363^_jMc«J*KMT5J!M@ö²zNKLT¬9D`¨j=Â9<)q j=b ö F
ü f!KLTa9[D`¨6)q j=Â9Dq4=9<)¨ `=?9q4¨j*D `=+ #9[D6¨`)q j= %z%KLTyöÅ]lTdNf!c\N_`63,5T«\_``3b5]lT²KOdNe}jSOTd~(K
T:TScN36,S z0]^\J*Q\J*KpJ*i#~(TKOKLJ*ef!3e©,5\N3b¡(©6f*T]¥5\3`TS36i#KFvgh\Nd,K -I? 9D j=Â9<¨`q4=ûö²vMT53
5\_j5 & -². ? - &0/1 ¡(J*MQ3Þ ö¦Þ`2 ? loDz]l3^\N_`63böë3 ? # & %4#5-6%zNKLT,]l3^Q_`M#]^SJ*53
öë3 ? # & ° -7 & ° - ° & -«8 ? - & /1 %
KLTyöÉõ? !k FvúT53,5\N_`5|!k Q_`Me3bSO36_`f*J*Ý[3`c²i#TSO3bQTMNQSO3653`f!©,_`K_KLde}4STd~²T p_`K5\3
_jQ5J!TM#Th5\3lKO©ì#i#365SOJ!3`KvT%_~(36M5_`}4TM#TM#J*5K|636S5J*Q36KF
Þ ö¦Þ? F%´#38ª4MT]Î5\N_`536636SO©}jSOTd~·T0TSOc36S^ pJ*KQ©6Qf*J*Q,UQF ü SO5J*M Õ SOT~(TKOJ*5J!T:M 9F ;F <(Z`F
æhdN5 QTMN5_jJ!MNKMNTR36f*36i#36MN50T%TScN36Sm zKLT,5\N36S3lJ!K|MT,KOdNe}jSOTd~T ThTScN36Sm F
Þ ö¦Þ?ÓD6tFmgh\3¡(©`f!T]û5\N36TS3ì#K536f*fdNK-5\_j5v5\36SO3RJ*K-_«MTSi#_jfKLde}4STdN~TTSc3`, S [zKLT
_jK|dNKOdN_`f>]l3Q_`M²_`KLKOdNi#35\N_`59[D`¨6)q j=@ö F>14365Ã 1 e35\N3lKL5_èJ*f!J*Ý
36SvT05\3^f*365536Slr0MT53
5\_j5Þ Ã 1 Þ`?q(FG´#3lQf*_`J*iÈ5\N_`5Ã 1 J*KvQ©`Qf*J!QGF¡(dN~(~(TKL3lMNT5zKLT«Ã 1 J!KJ!KLTi#TS~(\NJ!Q5TR5\>3 =0f*36J*M
TdS}jSOTd~F|¡(J*MQ3·öìcT36KMNT5QTMN5_jJ!M_5SO_`MNKO~(TKLJ*5J*TMUJ!5QTM5_`J*MK?_ WQ©6Qf*36Zrz:Ã 1 i#dNKO5
QTM5_`J*M¬5\SO363·cTdNef!3²5SO_`MNKO~(TKLJ*5J*TMKLzKOTaÃ 1 ?A÷j`°[9D`¨j=Â9<)q j=r°9DqG=Â9<)¨ `=Â°9*D `=9<¨6qG=øjF^æhd5_`K
e36TS36z9[D`¨6)q j=Â9D6¨j=9W
q `=?¥9[*¨ j=%\N_`KvTSOc36S|¨m]^\J*Q\#cNT36KMT5cJ*6J!cN3,D6t[z_QTM5S_jcJ*Q5J!TM:F
1j3`5¸ @ ? #O9[D`¨6
q `= %rF ¡(J*MQ3 B ¸ A¦Ã ? ÷j`ø,e©@QTi#~_`SJ*M}ITScN36SKbTp36f*36i#36M5KLzp_`MNc
Þ ¸uÞ*Þ Ã 1 Þ`? Þ ö¦Þ z]^3^\_``38¸¬Ã 1 ?¿ö²z(KLT#1 öJ*Kv}j3`M36SO_`536ce©·9[D6¨`)q j=>_jMc#_^}j3`M36SO_`5TST¸aF
úT]û]l3#TMNf!©#\N_`63#5\SO363Q_jKO36K-5TQ\36QªjzKOJ*MQ3#5\N36S3_jS3#KLJ!³·3`f!36i#3`M5KmTTScN36S<TdNSpJ*M
ÊÃ z_jMc5]lTb_`S3-}j3`M36SO_`5TSK:TN3`_`Q\,Q\NTJ!Q3mTÃ 1 FG´#3`f!f*z:9
D6¨`)q `=9D`¨6)q j=?X9)¨ `=9DqG=
]^ò \J*Q\b\1 _`KTScN36ShKOJ*³p_jMc8J*K5\dNKMT5J!M·ö²z`KLTRÃ 1DE ? #O9D6¨`)q j= %F ü f*KOTz9D6qj*¨ j=Â9[D6¨`)q j=Â9*D 6¨`q4=0?
9 qGD6¨4=^@ ¸z]^\NJ!Q\·J!K|e_`c·KLJ*MQ3#¸ÓJ!KMTSOi#_`f*05C \NdKpÃ 1F ? #9Dq4*¨ `= %Fgh\3TMNf!©SO36i#_`J*MJ*M}
~(TKLKLJ!eJ*f!J*5©J!KÃ 1 3 ? #O9D6)¨ qG= %F4143655J*M} & ?¥9
D6¨`)q `=Â° -#?9D`C *¨ 6q4=z5\3`Mz(]^3l\_``3
öëE ? # & ° -7 & ° - ò ° - & ? & B -6%
U5\J*KJ!K_^QTi#~(f!3`53l~(S3`KO36MN5_j5J*TM,KOJ*MQ3-]^3lªjMNTG ] - & - /1 Z`FGgh\NJ!K:J*KvJ!MNc363`c,_^}4STdN~RT>TScN36S
D`t[z0_jKpJ!5J!KpJ*KLTi#TS~(\NJ!Q^5T«5\3,KO36i#J*cJ*S3`Q5v~(SOTcdNQ5 !ò H m]lJ*5\y5\N3R_jQ5J!TMT ò TM
QTi#J!MN}«<SOTi5\3#J*KLTi#TS~(\NJ!KLi ò õ? J9 I( )I= K õ? LNMPO KL36McNJ!MN}²_«}j36MN36S_j5TST ò 5T
5\3l_`dN5Ti#TSO~(\J*KOió;²áA ã ;>B£ á F
Þ ö¦Þ?Î¨`t(F%gh\3¡©6f!T]Î5\36TSO36i#K536f*fNdNK5\N_`5]^38\N_`63TMN38TSvKLJ!³RKOdNe}jSOTd~(KTTSOc36SÍ>63NF
¡(dN~(~(TKL385\_`5:5\36SO3]^36S38KLJ*³RKLde}4STdN~(KTTScN36SÍ>63NF¡(J*MQ3b5\36J*SJ*M536SOKO36Q5J*TMKi#dNKO5e3
5SJ*6J*_`f*z|5\_j5_`QQTdMN5KTSRD6q#T5\3MT6M <J!cN36M5J!5©36f*36i#36M5KF«úT]:z5\3KL5_jeJ!f*J!Ý
36S^T_`MN©
f*365536S|i#dKL5:\_``38TScN36SvKLJ!³jzhKOT#5\_j5_jQQTdNM5K<TS5\3bT5\N36SvÍ>638MTQM WJ*c36MN5J*5©R36f*36i#36M5KT
ö FRæhdN5|5\J*Kî#36_`MNK-5\_j5|36_jQ\¬36f!3ì#36M5J!K^KL5_jeJ!f*J*Ý[3`c¬e©·5\3«KO_ji#3·KOdNe}jSOTd~(z]l\NJ!Q\J*K^_
ÕÖN×ØÙ(Ú GR ;
QTM5S_`cNJ!Q5J*TMKOJ*MQ3R5\3RTMf*©b36f*36i#36MN5T| 5\_`5:KL5_jeJ!f*J*Ý[3`K_`f*fhÍN63Rf!3`5536SKJ*K5\N3bJ*c36MN5J*5©F

gh\NdK5\N36S3J*KTMNf!©-TM3pKLde}4STdN~¸ TNÍN633`f!36i#3`M5KON]l3mQ_`M,_`KLKLdi#3p5\_j50¸ J*K:}j3`M36SO_`536c
e©²9
D6¨`)q `=rF
úT]z|]^3«\_j63TS·t#KOdNe}jSOTd~(KTTScN36Sb¨[m]^3²i#dKL5\_``3 t#e36Q_`dKL3]l3²Q_jMMT5
\N_`63_·MNTSi#_`fKOdNe}jSOTd~yTT SOc36S-¨be36Q_`dKL3#5\_`5]lTdNf!c©6J!3`f!c@_jM@36f*36i#36MN5vTTScN36SR F
ñpT]^36`36S!z5\3`S3m_jS3lTMf*©,DpTo S U ?¶D`t36f*36i#36MN5KT>TSOc36S¨pJ!M' Lz(KLT#öói#dKL5hQTMN5_jJ!M#_jf!fNT
5\36iFNM·~(_`SO5J*QdNf!_jS!zv9
D6¨j=@öaFæhdN59[D6¨4=Â9[D6¨`)q j=Â9<¨4D
=? 9
D
q 6¨j=@ ¸z(_8QTM5S_`cNJ!Q5J*TMF
gh\NdKv5\36SO3-J*KvMNTR5S_jMKLJ!5J!`3lKLde}4STdN~ThTSc3`S¨jtF
Þ ö¦Þ?qjt(F8gh\N3·KO5_èJ*f*J!Ý
36S8Ã ÒÓ T|5\N3·f!3`5536VS ,\_`K^TSOc36XS W[]^3·Qf*_`J*i¹5\_j5pÃ _`Q5K
5S_jMKLJ!5J!`36f!©bTMI÷jj°LD(°O¨(°Oq(øjF¡(dN~(~(TKL38ò MT5rF>gh\dNKpVÃ QTM5_`J*MKMT36f*36i#36MN5T0TSOc36S<TdS*zKLT

J*5NJ*K:i#_`cN3pd~R36M5J!SO36f*©pTKLJ!MN}jf*3m_jMc,cTdNef!3p5S_jMKL~(TKLJ!5J!TMNK

F4æhdN5%¡(©6f*T]¿5\N36TS©536f*f!K:dK:5\_j5
ö¹\N_`Kl_·KOdNe}jSOTd~@TTSc3`YS z:KOT]^3#Q_jM_jKOKLdi#3·5\_j59[D6¨`)q j=b¿ö²5\36SO36<TSO3·Ã Q_jM
QTM5_`J*M¬MNTyKOJ*M}4f!35S_`MNKO~TKOJ*5J*TMUe3`Q_`dNKO3²$ _ WQ©6Qf!3²_`MNc¬_5SO_`MNKO~(TKLJ*5J*TM}436M3`S_`536KR Z`F
gh\NdK|Ã QTM5_`J*MK5\3pJ*c3`M5J!5©m_jMc,KO36`36MRcNTdef*3m5S_jMKL~(TKOJ*5J!TMNKOzG_lQTM5S_jcJ*Q5J!TMRKOJ*MQ3b ò
TMNf!©8\N_`K|5\NS363cNTdef*35SO_`MKL~(TKLJ!5J*TMKF>gh\dKmVÃ J*K_R5SO_`MNKOJ*5J*63KLde}4STdN~·T ò F>æh© ü SO5J*Mz
Ô ;F 9F Ô(Ô z5\N3mTMf*©l5SO_`MKLJ*5J*63mKLde}4STdN~RT TNTSOc36ZS WJ!Kk Fæhd5k QTM5_`J*MK5]^T8KOJ*M}4f*3
5S_jMKL~(TKOJ*5J!TMNKRJU [NJ*~(~(J*M}²_èTdN5v5\N3RcNJ!_j}jTMN_`ò fT:5\N3,KLwGd_jS36Zrò z_·QTMN5ò SO_`cJ*Q5J*TMFgh\N36S36TS3
5\36SO3-J*KvMNTR5S_jMKLJ!5J!`3lKLde}4STdN~#T%TSc3`Svq4tF
Þ ö¦Þ?¹»`t(F#gh\3²TMf*©KOdNe}jSOTd~ITm TTScN36Sb»`t#J!K \ e:© ]TSc_j6M WNñ ^f!cN36S!z|KOJ*MQ$ 3 \ J!K
KLJ!i#~f!3NF
Þ ö¦Þ`?X[D`t[zGJËF 3NF%öì?¶ F
gh\36SO36<TSO3p5\N3-~TKOKLJ!ef*3^KOdNe}jSOTd~(K_`S3lJ*KOTi#TSO~(\J*Q5T °Ok ° ò ° °
_`Mc z ]^J!5\#5\3
\
36i#e36cNcJ*M}4Kv}4J!63`M·_èT63>F H
!#"%$
&('*å,+-)
gh\J*K<Tf*f*T]lKSTiì5\3~(S3`6J!TdNK^~(STef*36i#zvKLJ!MNQ35\N3«TMNf!©5S_`MNKOJ*5J!63²KOdNe}jSOTd~(KTl]^\TKL3
TScN36S:J!K_8i#df*5J*~(f*3^T¨_jSD3 \ _`Mc F
!#"%$
&('*å,+!_
` 36Q_`f*fvU ü SO5J*M«gh\N36TS36i Ô F ;F <(Zm5\N_`5_`M36f!3ì#36M5|§@Q_jM²e38QTMKL5SOdQ536ce©8SOdf*36S|_jMcQTi#~(_jKOK
J*h_`Mc#TMNf!©J*h5\36SO3-J*Kv_5T]l3`SThÍ>36f*c#36³`536MNKOJ*TMK
x? 1 Ò ÒEooo(Ò0a UÔZ
]^J!5\u¢ 0 bp£%0 b /'1 ¤?ûDlTS|36_jQ\ô_jMc¬c a? 0 a /1 9<§:=F|úT53b5\_`5J*5J*NK TdMN36Q36KLKL_`SJ*f*©R5SdN3R5\_j5
36_jQ\0 bÒ¿x@9<§:=z(TS5\_j5¢ 0 a£4x,¤h?¿qfF e
orqrsts,qvu,wyxzxqvxz{ Ö w| xwL }~ j wc}wyu l ~ l qvu m {
gihkj)l mn m p
ÕÖN×ØÙ(Ú < Û
úT]¶f!365§®X\_j63mi#J*MJ*i#_`f~(Tf!©`MTi#J*_`fh7:9<;>=@xy¢ ;(¤4T>c36}4S3`3^q

z_jMcf*365ÃÎ?EÄ^Æ6ÇË970jx·=rF
ÃJ!K5\NdK_p5SO_`MNKOJ*5J*63KLde}4STdN~lT0 ò z4KOTlJ*5J*KTM3T5\3}4STdN~(Kf*J!KL536c8J*M ü SO5J*M Ô ;F 9F Ô(Ô z4M_ji#36f*©
ò ° \ ò °Ok ò ° ò °6TS5\V3 =0f*36J*M«TdS}4STdN~F01j3`5v¸`xXe385\3KL~(f*J!55J!MN},Í>36f*c«T7(KLd~(~(TKL385\_j5
7:9W;%= ? òb 1 9<;«C c§ b=J*M¸z(]^\3`S3b§¬?§ 1 F
Ð Ù 0 <F Ô F § i 6L 6 PG! Ã?k ò ° ò '
ÕÖN×6× F
UL? Z ´#3@]^J!f*fKO\NT]Å5\N_`5lJ!8Ã ? ò TS \ ò 5\N36M¿§J*KRMNT5-QTMKL5SOdQ5J*ef!3>F¡(dN~(~(TKL3@5\3
QTM5S_`SO©ýz|5\_j5^§J*KbQTMKL5SOdQ5J*ef!3>F@14365^?A0a /'1 zKLT0aº?A#9§=rFgh\NdK85\3i#J*MJ*i#_`f
~(Tf*©6MNTi#J!_jfÑ<TS§¿T636S#Ó\_`K,c3`}jSO363y5]^T8KLd~~(TKO3]lJ*5\Td5-f!TKLK,Tp}j3`M36SO_`f*J!5©5\_j5
Ñh9W;%=-?9W;C§ 1 =9W;C§ =F|gh\dNKpÑ%9tj=m? § 1 § mz0_jMcKOJ*MQ3«7:9<;>=L[Ñh9W;>=8mz0]^3bª4MT]
5\_j50§ B § ò ,FGgh\dNK¯?¶§ 1 § § B § ò ,FGúT]:zG_`KJ!M ü SO5J*M,~F Û 9 zG5\N3pTSeJ*5>T>¯dNMc3`S
\
ò ÃJ*K
X
÷[¯ 1 ?§ 1 § ¦§ B § ò °I¯ ?¶§ 1 § B § § ò °I¯ B ?¶§ 1 § ò § § B ø

UKLJ!MNQ3²9D`¨6qG= \ ò Z`FG´#3^KL\Tdf*c·MNT53l5\N_`50MNTR5]^T¯ b _`SO3l5\3lKL_ì#36z(KLJ*MQ3TS:J!MNKO5_`MNQ36z
§ 1 § § B § ò ?¶§ 1 § B § § ò § 1 9§ C § B =v?¶§ ò 9<§ C § B =
]^\J*Q\²J!K_RQTM5S_jcJ*Q5J!TM²KOJ*MQ3#7·\_jK|cNJ!KL5J!MNQ5STT5KÛKLJ!MNQ3·x \_jK|Q\_`SO_`Q536SOJ!KL5J!Q-Ý[3`ST_`MNc
7J!KJ!SOS36cNdQJ*ef*36Z`FGgh\3`S36TS3^¯«\N_`KvcN36}jSO363^5\NS363lT636SxFGæhd5|xy9W¯=Òmz(KLT
¢ ¥£x,¤>?¥¢ £4xy9<¯=Ë¤ý¢µxy9W¯=£4x¤
J*K_i#dNf!5J!~f!3^Th5\S3`36zG]l\NJ!Q\#QTMN5SO_`cJ*Q5Kv5\3l<_jQ505\_j5¢ £4x,¤J!K|_~(T]^36S:TDRUKLJ*MQ3R¿J!K
J*M·_5T]l3`SU Ô ZThKLw4dN_`SO3lSOTT5KLZ`F4gh\dNK§J*KvMNT50QTMNKO5SdNQ5J!ef*3NF
U ?Z ¡(d~~(TKO3m5\N_`5ÃÎ?¶k ò ° ò z4TZS FG5NKLdÍNQ36K5TÍ>McR_m~(SOT~(36S0J*M536Si#3`cJ*_`5336³`536MKLJ*TM
çe365]^3636MxÎ_`MNc@xy9§=zKLJ!MNQ3^5\N36M
x çxy9<§:=

]^Tdf*ce3-_^5T]^36ST>c3`}jSO363-5]^T836³j536MNKOJ*TMKFGúT]:zJ!Ã? ò TS¿5\N36M·¸¹?ºxy9§=hJ*Kv5\3

KL~(f*J!55J*M}bÍN36f*czG_jMc_`MN©-TSOc36S5]^T8KLde}4STdN~RTÃ]^J*f!f\N_`63l_-Í>³`3`cRÍ>36f*c·çy]^\J*Q\KO_j5J*KLÍN36K
TdNSS36wGdJ*S3ì#36M5K
F
<|ÃÓ? k 5\N36M¢ ¸2£0xy9§=ý¤v?ÈD[z0KLTyÃÀ?ÈÄ^Æ6Ç9¸jxy9<§:==vJ*Kp_KLde}4STd~T:TScN36S|5]^T
KLJ!MNQ3ÃÀ%~(36Sò i#dN53`K5\N3«§ b _jMc@Í>³`36K§¶?§ 1 zÃÀ>i#dKL5e3}436M3`S_`536ce©#_#5S_`MNKO~TKOJ*5J*TMz
KL_`©²ÃÀv ? #9§ °O§ = %F´#3]^Tdf*cyf!J*ª43,5T«Í>Mc_#KLde}4STdN~öìT:TScN36Sqb5\_j5QTM5_`J*MKÃÀµz
KLT@5\_`5-ç?ìB ¸¬ùò J*K_ÍN36f*cIQTM5_`J*M3`c¬J*Mxy9<§:=]^J!5\¢ ¸ £vç¤-? q ? ¢ x@9<§:=£vç¤-?
DF,´#36f!f*zvJ*89§ °O§ =«EÃû5\N36M¿9§ °O§ =«EÃÅUSO36i#36i#e36Sm5\N3·36i#e36cNcJ*M}Tk J!MN5Ta
QTMKLJ!KL5J*M}RT%5B \N3-_jò Q5J!TM#TM#5\3l636SO1 5J*Q36 KTh5\3lKL©6i#i#365SJ*36KT%_KOwGd_jS36ZrzKOT,5\N3-KLò de}4STdN~ ò
öë?º÷4`°[9[Dj=r°[9<¨6qG=r°[9[Dj=Â9¨6q4=rø]^J*f!fNcTFgh\NdK§J!K|QTMKL5SdQ5J*ef*3NF
❒
*'
TJ¡ ¡

Math 122, Solution Set No. 1: 1 Chapter 1.1 Problem 16

Uploaded by

Copyright:

Available Formats

You might also like

Math 122, Solution Set No. 1: 1 Chapter 1.1 Problem 16

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Math 122, Solution Set No. 1: 1 Chapter 1.1 Problem 16

Uploaded by

Copyright:

Available Formats

Math 122, Solution Set No.

2 Chapter 2.1 Problem 1(b)

3 Chapter 2.1 Problem 8

4 Chapter 2.2 Problem 2

6 Chapter 2.2 Problem 4

7 Chapter 2.2 Problem 14

8 Chapter 2.2 Problem 20

(ab)r = ar br = (am )z (bn )y = ez ey = e

Since (ab)r = e, The order of ab must divide r..

9 Chapter 2.3 Problem 1

ϕ(x + y) = 2x+y = 2x 2y = ϕ(x)ϕ(y)

So ϕ is a homomorphism. Furthermore, 2x = 1 iff x = 0. So Ker(ϕ) = 0, i.e. ϕ is injective.

10 Chapter 2.3 Problem 7

CAC −1 = DAD−1 ⇒ (D−1 C)A = A(D−1 C)

11 Chapter 2.3 Problem 12

xy = (y −1 x−1 )−1 = ϕ(y −1 x−1 ) = ϕ(y −1 )ϕ(x−1 ) = yx

and so G is abelian. If G is abelian, then

ϕ(xy) = (xy)−1 = y −1 x−1 = x−1 y −1 = ϕ(x)ϕ(y)

and so ϕ is an bijective homomorphism, i.e. an automorphism. Thus ϕ is an automorphism

(123)(12)(23)(12)−1 (123)−1 = (12)

(12)(123)(23)(123)−1 (12)−1 = (23).

(c). Let ϕ : G → R∗ be the map defined by ϕ(A) = a11 /a22 . We have

So ϕ is a homomorphism. Because its kernel is exactly the set H, H ≤ G and H is

So ϕ is a homomorphism. Because it’s kernel is exactly the set H, H ≤ G and H is

The characteristic polynomial of D is easily calculated to be xn . By the Cayley-

p(P AP −1 ) = (P AP −1 )n + an−1 (P AP −1 )n−1 + ... + a0 I =

P An P −1 + an−1 P An−1 P −1 + ... + a0 I =

||D − A|| = ||P −1 (C + E)P − A|| = ||P −1 CP + P −1 EP − A||

≤ ||P −1 CP − A|| + ||P −1 EP || = 0 + ||P −1 ||||E||||P ||

Assuming G is simple, ker ϕ = G or {e}. If it’s G, then conjugation by G fixes

Assuming G is simple, ker ϕ = G or {e}. If it’s G, then conjugation by G fixes

(b, c) It is easy to check that the basis mij is orthonormal:

Thus the signature is (2, 1).

a < f, g > +b < h, g >

We will use this matrix to transform the quadric.

Note (3, 0, 1)P = (3, −6, 1).

X 0T (P T AP )X + (BP )X + c = 0 = 5x02 + 20y 02 − 4z 02 + 3x0 + 2y 0 − 2z 0 − 6 = 0

We complete the square:

(a + b21/3 + c41/3 )(d + e21/3 + f 41/3 ) =

(ad + 2bf + 2ed) + (bd + ae + 2cf )21/3 + (cd + af + be)41/3 ,

(a0 + a1 t + ... + an tn )(b0 + b1 t + ... + bn tn )

1 Chapter 11, Section 1

2 Chapter 11, Section 2

3 Chapter 11, Section 3

f (x)po (x) + g(x)qo (x) = 1

and thus the gcd of f(x) and g(x) is 1.

f (x)po (x) + g(x)qo (x) = 1

2. Artin §11.4 #11

d) We have f (x) = x3 + 6x2 + 7 ≡ x3 + x2 + 2 (mod 5). As above, f has no root in F5 , so

e) This f (x) = x5 − 3x4 + 3 is irreducible by Eisenstien’s criterion with p = 3.

3. Artin §11.4 #11

First we prove the following useful fact.

L EMMA 3.1. If M is any n × n complex matrix with M k = In for some k, then M is

1 Taken from Noam Zeilberger’s solutions

|v|2 > 0. Now let

Now, for any v, w ∈ V , hM v, M wi = hv, wi. Thus M is unitary, so M M ∗ = In =

4. Artin §11.4 #12

a) (1 − 3i)(1 + 3i) = 10 = 2 · 5 = (1 − i)(1 + i)(1 − 2i)(1 + 2i). After some investigation,

2°46;?ÝCO`9#B6;49V9:J2»9K¸COK7%J9EK?C/9#K¸K`96;E%?Óæ)jÃÐ¸ÓÐÑAql ç5jhN»MNAqlWÈSmQN:E4:ÝE%C/9#K¸CF6;M

? K9PCOØ^J68BF25CvG'K@Í3R3È'257%7Q)2L±;?JK¥9:J689DhkDE%CD6eCOØ6;B2LG'K@Í0E%I}68?J@bK?7MeEIÙÍ°ÀÞkªjÃÂ/Ä^lQ

½(3¾4SDk5$IAR5_ÿM&NrAjNOF@C(DEFGH&3UDVÙ×·eiNLD136M5IG36J!Çb¼DJBÇb×rP*ÅF5_ b´×£NOM&SGT317

Q-FGK-K-DNO35$IAR5 \M&NApVsJL363;BC(DEFGH&3Re_NOD] k3cIA<m63ApC(DEFGH&3cM&NODC(DJK-IGM&NOCµ7 9y; = PY2<365

Û·DSmR36JNO36Hg.eiNOFKK-DNL3U5IGAR5 (b qsatnMN{5IG3K-JMSTM&KARH*M&EG36ARHi436S36JLAR5$36E@gUARSp36H&3RC(36S5y]kIGM&TI

ÅF5_7{qWª¡O-¡|||¡O<t{b>W`x7{qw<¡L¡|||¡L4tubNLMST3YWob¼MS£;j« ªekIGM&TI\TDSG5JLAREMT5$Nr5IG3VsART5·5$IA<5

I)NO59EO526 =6 1"Q2F$=6 N13a13E

|mv 0)13F$13>"=?N>3J+-R o?2} o:S~U ` q i eHR}EOQÅ8A=ET2AQ>313259=?N]V=46HQNOEO~8AQZW59<959N3R\2Q2A5Q]µÅ À PÅÏûPÅ

EP&Q S À Î i a Ë B 2 Ë ^K 2 û X"

Ë =?¸Ì9f:c]xw9:cde:cw7P8Mde;#]^7!;#wv7!98ML%7!98k:c5hM8vj*de5amfw7Wuv;#g,g8JEIpJMy

õ Q@ í V ºëD2d8Mjde8M]^8ML7!]7P9f8/wg#:]e]nuv;#7!9bè V :p7!d:L]^j5]^;#7!;#5L%:Labè TWV :.

õ A> @QBºeøD_E @CB ¢ J

,Q'-%#\ ,(..

? G 8 g\ ?\; :E a

QTM5S_`cNJ!Q5JTMKOJMQ3R5\3RTMf©b36f36i#36MN5T| 5\_`5:KL5_jeJ!fJÝ[3`K_`ffhÍN63Rf!3`5536SKJK5\N3bJc36MN5J5©F

J5NJK:i#_`cN3pd~R36M5J!SO36f©pTKLJ!MN}jf3m_jMc,cTdNef!3p5S_jMKL~(TKLJ!5J!TMNK

úT]¶f!365§®X\_j63mi#JMJi#_`f~(Tf!©`MTi#J*_`fh7:9<;>=@xy¢ ;(¤4T>c36}4S3`3^q

÷[¯ 1 ?§ 1 § ¦§ B § ò °I¯ ?¶§ 1 § B § § ò °I¯ B ?¶§ 1 § ò § § B ø

]^Tdfce3-_^5T]^36ST>c3`}jSO363-5]^T836³j536MNKOJTMKFGúT]:zJ!Ã? ò TS¿5\N36M·¸¹?ºxy9§=hJ*Kv5\3