Вы находитесь на странице: 1из 5

Math 123, homework 3

February 20, 2019

I have been in the habit of setting “Optional” questions on problem sets, with
questions that are a little more interesting. Problem 7 on this homework is of
this sort (though it is not hard using the ideas from class).

I used to always assign zero credit to optional problems, but I have found that
almost no one attempts them. So this semester I am going to try assigning extra
credit to such problems. The extra credit will be tiny, but a motivation for looking.

Problem 1. Let f 1 and f 2 in F [X ] be polynomials without a common factor in


F [X ]. Let K : F be a splitting extension for f 1 f 2 . Show that f 1 and
f 2 have no common root in K.

(This came up in class. Look back at Problem Set 1 for a closely related question.)

Proof: Done individually

Using the logic from problem 1, PSET 1, we know that f , д having no common
factors (other than 1) in the subring F [X ] means they have no common factors
in K[X ]. Since in K both polynomials split into linear factors, and all these linear
factors are different in each, we see that there are no common roots.

1
Problem 2. Let Q(t) be, as usual, the field of rational functions p(t)/q(t) with
coefficients in Q. Thus Q(t) : Q is a field extension generated by a
single element t, but t is transcendental over Q, not algebraic.

Let α be any element of Q(t). Show that α is algebraic over Q only if α is a


constant (i.e. an element of Q itelf).

Proof: Done individually

Suppose that α = f (t)/q(t) is not constant and we want to show that α is al-
gebraic. Then, ai α i = ai f (t)i /д(t)i = 0. We can then multiply through by
Í Í
д(t)n and have that h(t) = ai f (t)i д(t)n−i = 0, which implies that there is a
Í
polynomial h such that h(t) = 0, which contradicts t not algebraic. So, by the
contrapositive α is algebraic over Q only if α is a constant.
Problem 3. Stewart, Chapter 9. Which of the following extensions are normal?

(a) Q(t) : Q

(b) Q( −5) : Q

(c) Q(α) : Q where α is the real seventh root of 5



(d) Q( 5, α) : Q(α) where α is as in (c)

(e) R( −7) : R

Proof: Done individually

(a) This is not normal. Normal extensions are splitting fields. We cannot get a
splitting field by adjoining a transcendental number t to Q
(b) This is normal, since it is the splitting field of x 2 + 5 over Q
(c) Not normal since it does not contain all factors of x 7 −5 (it does not contain
the complex ones). As a result, x 7 − 5 has a solution in Q(α) but cannot
split into linear factors (since these factors have complex terms).

2
(d) This is normal since it is the splitting field of x 2 − 5 over Q(α) (no solutions
of this polynomial exist in Q(α))

(e) This is the splitting field of x 2 + 7 over R, so it is normal

Problem 4. Let K ⊃ L ⊃ F be finite extensions. Prove or disprove:

(a) If K : F is normal, then K : L is normal.

(b) If K : F is normal, then L : F is normal.

(c) If K : L and L : F are normal, then K : F is normal.

Proof: Done individually

(a) This is true. Since K : F normal there is some д ∈ F [x] for which K is
the splitting field of д over F . Now take this same д ∈ L[x] and factor out
any terms that now have roots in L. This modified д0 gives K: the splitting
field of д0 over L: We can see that this splitting field is inside of K (since
this adjoins all the roots of д0, which are inside of K). At the same time K
cannot be strickly larger than this splitting field: if it were then this field is a
smaller splitting field for д in F [x], which is a contradiction. We conclude
that this splitting field is precisely K : L, which makes K : L normal as
desired.
√ √
(b) Not true. Let F = Q, L = Q( 4 2), K = Q( 4 2, i). We know that L : F not
normal (by the same logic in 3(c)) but we do know that K : L normal, since
this is the splitting field of x 4 + 1 over L.

(c) Take some irreducible f ∈ F [x]. Suppose that it has some root in K. If all
are in L use the fact that L : F normal and we can decompose into linear
factors within L[x] ⊂ L[x]. If some roots are not in L then factor out those
that in L, giving an irreducible ∈ L[x] multiplied by some number of linear
factors. This д has a root in k so use K : L normal and factor into linear
factors, meaning that f has factored into linear factors in K[x]

3
Problem 5. Let K be the splitting field of the polynomial x 4 − 2x 2 − 1 over Q.
What is the degree of K : Q? Find three intermediate fields, E 1 , E 2 ,
E 3 , each of degree 2 over Q. (Verify that no two of the three coincide.)

Using results from Wednesdays class or otherwise, confirm moreover that Ei is


not isomorphic to E j for i , j.

Proof: Done individually


p √ p √ p √ p √
We can see that the roots of this polynomial are ± 1 + 2, ± 1 − 2 = ± 1 + 2, ±i 1 + 2.
Let u = x 2 and then solve u 2 −p2u −√1. Solve p in √ terms of u and then reduce into
x. pSo the splitting field is Q(p 1+ 2, i 1 + 2). We see this p is√equivalent to
√ √ p √ p √
Q( 1 + 2, i). We seepthat i 1 + p2/ 1 + 2 =pi, so i ∈ Q( 1 p + 2, i 1 + 2).
√ √ √ √
At the same time, i · 1 + 2 = i 1 + 2, so i 1 + 2 ∈ Q( 1 + 2, i). This
shows the extensions are equal. We can now use the tower law to compute the
p √
degree: |Q( 1 p + 2)| = 4 (since
p the degree of thepirreducible polynomial is 4),
√ √ √
and then |Q(i, 1 + 2) : Q( 1 + 2)| = 2, since 1 + 2 doesn’t contain i and
its minimal polynomial is x 2 + 1. So, the degree is 8.
p √ p √ √ p √
Note that i ∈ Q(i, 1 + 2). Also, we see that ( 1 + p 2)2 −1 = 2 ∈ Q(i, 1 + 2).By
√ √
multiplying these together we also get i 2 ∈ Q(i, 1 + 2). From this,pwe can
√ √ √
also see three intermediate fields E 1 , E 2 , E 3 = Q( 2), Q(i), Q(i 2) ⊂ Q(i, 1 + 2).
Each have degree 2 since their respective minimal polynomials √ are x 2 − 2, x√
2 +

1, x 2 + 2. They don’t coincide since E 1 , E 2 don’t have i 2, E 2 , E 3 don’t have 2,


and E 1 , E 3 don’t have i

These intermediate fields do not admit isomorphisms of field extensions between


them. We’ll show this on all cases:

• E 2  E 1 : suppose this to be true via isomorphism γ . Now since γ restricted


to Q is the identity, we see that γ (2) √ = 2. Since γ surjective,
√ there must
be an element η of E 2 with γ (η) = 2. It cannot be 2 since E 2 does not
have this element nor is it a rational number (since all the rationals are
sent to themselves). So √ η is some complex number. Then, we observe that
η 2 − 2 , 0, since η , 2, but we see that γ (η 2 − 2) = γ (η)2 − 2 = 2 − 2 = 0.

4
This means the kernel is nontrivial and this map isn’t injective, which is a
contradiction.
• E
√2  E 3 : with similar logic we suppose γ exists and note γ (−2) = −2. Since

−2 doesn’t exist in E 2 , there is some other number η with γ (η) = −2.
Now, η 2 +2 , 0, but we see that γ (η 2 +2) = 0, which contradicts injectivity.
• E 1  E 3 : this proof is entirely similar to the first two.

The next two questions are from Cohn’s ‘Algebra’, vol. 2.

Problem 6. An equation is called reciprocal if whenever α is a root, 1/α is also a


root. Show that if an irreducible polynomial over Q has a complex
root of modulus 1, then the equation is reciprocal.

Hint. Given the minimal polynomial of α, how do you write down the minimal
polynomial of 1/α?

Proof: Done in collaboration with Philip Nicol

Suppose that there is a root α of modulus 1, then we know that α = 1/α (basic
complex numbers fact). So if we have some irreducible polynomial f , f (α) =
Í i
qi α = 0 =⇒ α −n qi α i = 0 =⇒
Í i−n
0 =⇒ qi α = 0. Now take
Í
the conjugate on both sides and know that conjugation respects addition and
qi (α)i−n = 0 =⇒
Í n−i
multiplication. So, qi α i−n = 0 =⇒ qi α = 0 =⇒
Í Í
Í i Í n−i
qi α = qi α . This now shows us that qi = qn−i .

So, if we have qi x i = 0 we also have qi x n−i = 0. Multiply through by x −n


Í Í
and we have that qi x = 0 =⇒ 1/x is a root.
Í −i

Problem 7. Extra credit. Let f be an irreducible polynomial over F . Show that in


a normal extension N of F , the polynomial f factorizes (if at all) into
factors that are all of the same degree. (You may assume the normal extension is
finite.)

Peter Kronheimer

Вам также может понравиться