1 Problem 1

√ Z[x]
Z[ −14]/(p) = = (Z/p[x])/(x2 + 14)
(p, x2 + 14)
Now we count the number of solutions of x2 + 14 in Z/p:
1. p = 2 : x2 + 14 = x2 , there is only one solution of this polynomial in Z/2
which is 1. So p = 2 ramifies
2. p = 3 : x2 + 14 = x2 − 1, this has two solutions 1, 2 in Z/3. So it splits.
3. p = 5 : x2 + 14 = x2 − 1 this has two solutions 1, 4 in Z/5. So it splits.
4. p = 7 : x2 + 14 = x2 this has one solution 0 in Z/7. So it ramifies.
5. p = 11 : x2 + 14 = x2 + 3. Testing x = 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 gives
3, 4, 7, 12 = 1, 19 = 8, 28 = 6, 6, 8, 1, 7, 4 respectively, none of which is a
solution in Z/11. So 11 is inert.
6. p = 13 : x2 + 14 = x2 − 64 this has two solutions 8, −8 in Z/13. So it
splits.

2 Problem 2
√
Let d−1
4 = d′ and ω = 1+2 d , then by direct calculation we have that ω 2 = ω +d′
Consider two cases:
1. If d′ is even (d ≡ 1 mod 8), then 2 is not prime because 2 + ω, 1 + ω ∈
/ (2)
but (2 + ω)(1 + ω) = 2 + d′ + 4ω ∈ (2)
√
2. If d′ is odd, take an arbitrary pair x + yω, z + tω ∈ Z[ 1+2 d ] such that
(x + yω)(z + tω) ∈ (2).
In other words, there are integers a, b such that:

2a + 2bω = (x + yω)(z + tω) = (xz + ytd′ ) + (yz + tx + yt)ω

This forces 0 ≡ xz + ytd′ ≡ xz + yt mod 2 and 0 ≡ yz + tx + yt mod 2.

If x is even then our conditions are equivalent to 0 ≡ yt mod 2 and 0 ≡
yz + yt mod 2, suggesting that both yt and yz are even. If y is even then
x + yω ∈ (2), otherwise both z, t are even and z + tω ∈ (2).
A similar reasoning holds when z is even. Assume that both x, z are
odd. Then our conditions are equivalent to 0 ≡ 1 + yt mod 2 and 0 ≡
y + t + yt mod 2. The first condition forces both y, t to be odd, which is
incompatible with the second. So this case cannot hold.
√
From the cases above it can be seen that if x + yω, z + tω ∈ Z[ 1+2 d ] such
that (x + yω)(z + tω) ∈ (2) then at least one element is in (2), indicating
that 2 is a prime.

1
3 Problem 3
Let n1 , ..., nk be a set of elements that generate N , and m1 + N, ..., ml + N be
the set of cosets that generate M/N .
Take an arbitrary element m ∈ M . It has to be inside some coset a1 (m1 +
N ) + ... + al (ml + N ) = (a1 m1 + ... + al ml ) + N, ai ∈ R because such cosets
generate M/N .
This means m − (a1 m1 + ... + al ml ) ∈ N . This allows us to write it as
b1 n1 + ... + bk nk , bj ∈ R, making m = a1 m1 + ... + al ml + b1 n1 + ... + bk nk .
This representation of m shows that M is finitely generated by the set
n1 , ..., nk , m1 , ..., ml .

4 Problem 4
1. Direct
√ calculations show that the matrix of the linear map mα for α =
a + b d is:  
a db
b a
2 2
This has determinant√ N (α) = a − db and trace T r(α) = 2a. Then the
element α := a − b d satisfies αα = N (α). By the multiplicativity of
linear map determinants, we also have multiplicativity of the norm.
√
2. Expanding (α − a)2 = (b d)2 shows that α is a solution of the polynomial
P (x) = x2 − T r(α)x + N (α)
If T r(α), N (α) are both integers then P (x) is a monic integral polynomial
admitting α as a solution, proving that α is an algebraic integer.
Suppose α is an algebraic integer that is the solution of its minimal poly-
nomial Q(x) ∈ Z[x] that is monic. In that case, it is also the solution of
the remainder polynomial r(x) obtained by dividing Q(x) by P (x) in Q[x].
If r(x) is not the zero polynomial then it violates the smallest degree role
of Q(x), so r(x) = 0 and so P (x)|Q(x) in Q[x].
That is Q(x) = R(x)P (x), R(x) ∈ Q[x]. Comparing the coefficient of
the highest degree, since that of Q(x) and P (x) are both 1, so is that
of R(x). Write P (x) = p(x)/q1 , R(x) = r(x)/q2 where p(x), r(x) are
primitive polynomials with integer coefficients and q1 , q2 ∈ Z+ . But then
p(x)r(x)/(q1 q2 ) = Q(x) ∈ Z[x], and p(x)r(x) is a primitive polynomial
(the product of two primitives), so q1 q2 = 1 is the only way for the left-
hand side to have integer coefficients. Then P (x) = p(x)/q1 = p(x) ∈ Z[x],
implying its coefficients, including T r(α), N (α), are integers
3. For the forward implication,
√ note that N (α) ∈ Z, if αβ = 1 for some
algebraic integer β ∈ Q[ d]. Taking the norm of both sides, using the fact
that norm is a multiplicative function: N (α)N (β) = N (1) = 1. Because
both norms are integers (following from the previous part and that α, β
are both algebraic integers), they are units in Z.

2
 For the reverse implication, suppose that N (α) is a unit in Z, meaning it is
1 or −1. Note that α has the same minimal polynomial as α (in notation
of the previous part, it is P (x)), so it is an algebraic integer as well. This
also extends to its additive inverse. Then we have that:
α
α =1
N (α)

which means α is a unit in the ring of algebraic integers.

√ √
4. Integers in Q[ −3 are Z[ω], ω = 1+ 2 −3 . A simple calculation shows
ω 2 = ω + 1.
Take some α = a + bω from this ring, then a simple calculation shows that
T r(α) = 2a + b, N (α) = a2 + ab + b2 are both integers because a, b are.
This means that α is an algebraic
√ integer, or that our ring is a subring of
the algebraic integers inside Q[ −3].
√
If α = a + bω is a unit in the integers of Q[ −3], then it is at least a
unit in the set of algebraic integers. By the previous part, this amounts
to N (α) = a2 + ab + b2 is a unit in Z, since it is nonnegative, it can only
be 1.
Since 1 = N (α) ≥ 0.75a2 , 0.75b2 , |a|, |b| < 2. Exhausting the remaining
possibilities gives the values ±ω, ±1, ±(1 − ω). These are all units because
1 = (±1)2 = (±ω)(±(ω − 1)).

5 Problem 5
Each endomorphism ϕ : M → M is uniquely determined by ϕ(1) because ϕ(r) =
ϕ(r.1) = rϕ(1), ∀r ∈ M .
And so the map that sends r ∈ R to the endomorphism ϕ : M → M
satisfying ϕ(1) = r is an isomorphism between R and HomR (M, M ).

6 Problem 6
P
1. View KG as the set { ar · r|ar ∈ K, r ∈ G.
r∈G
P P
Take an element f = αi ( r) in KG, where we group together the
r∈Ai
formal variables with the same coefficient, ie: αi ̸= αj , ∀i ̸= j

3
 Then:

f is in the center of KG
↔gf g −1 = f, ∀g ∈ G
X X X X
↔ αi ( r) = αi ( grg −1 ), ∀g ∈ G
r∈Ai r∈Ai
X X
−1
↔ r= grg , ∀g ∈ G, ∀i
r∈Ai r∈Ai

because the coefficients αi between groups differ. The last line is equivalent
to the fact that each group Ai is closed under conjugation by g. This
means that Ai ’s are unions of conjugacy classes. Translating this back to
the function perspective, this means that f is a function that takes on the
constant value αi on the conjugacy classes in Ai , ∀i.
2. Let v1 , v2 , ..., vn be an orthonormal basis of Rn , note that the coefficient
at the vi component of a vector v is ⟨v, vi ⟩.
Take a matrix f in the center of Mn (R), and denote the matrix ω(j,i) , 1 ≤
i, j ≤ n as the matrix with all entries being zero, except for the one at
(j, i) where the entry is 1.
Viewing these matrices as representing linear transformations in the basis
v1 , ..., vn , it can be seen that ω(j,i) represents the linear transformation
that turns vi into vj and all other basis vectors to 0.
This implies f.ω(j,i) represents the linear transformation that turns vi into
f (vj ) and all other basis vectors to 0.
The matrix above is the same as ω(j,i) .f , which transforms v to f (v) then
to ⟨f (v), vi ⟩vj (ω(j,i) filters out the vi component then switches to vj ) for
every vector v.
Matching the images of the corresponding basis vectors, we obtain 0 =
⟨f (vk ), vi ⟩vj , ∀k ̸= i, and f (vj ) = ⟨f (vi ), vi ⟩vj (starting from the vector
vi ). The first condition shows us that f (vk ), k ̸= i, has no vi component.
Let both k, i vary, this means that any basis vector vi going through f
cannot have a nonzero coefficient in any basis vector other than vi .
The second condition shows that f (vj ) = ⟨f (vi ), vi ⟩vj , ∀i, j. This means
that f scales the basis vectors vi by the same factor. In other words, f is
a scaling map, whose matrix is a diagonal matrix with all entries on the
diagonal being the same.
It can be easily verified that such matrices commute with every real-valued
n-by-n matrix. To sum up, the center of Mn (R) is {c.I|c ∈ R, I is the identity matrix}.

7 Problem 7
Consider a nonzero R-module homomorphism ϕ : M → N .

4
 Ker(ϕ) is a submodule of M , an irreducible module, so Ker(ϕ) is either M
or {0}. The first case implies ϕ is a zero map, so it cannot hold. The second
case implies that ϕ is injective.
Im(ϕ) is a a submodule of N , an irreducible module, so Im(ϕ) is either N
or {0}. The second case implies ϕ is a zero map, so it cannot hold. The first
case implies that ϕ is surjective.
Combine the two facts above, we have that ϕ is a bijection, coupling with
homomorphism implies isomorphism.