10.2 The Main Theorem of Galois Theory

1.
a. img is the splitting field of x5 − 1 over img, and img is separable as char img. By Example 4 §10.1, img where σ(u) = u2 (because img. The lattices are:
img
Thus Himg is the only intermediate field, img is Galois (as H img G), and img. We have σ2(u) = u4 = u−1 so σ2(u + u4) = (u + u4). Hence img so, since img, img. Of course img
c. img splits (x2 + 1)(x2 − 3) over img, so it is a Galois extension. Clearly x2 + 1 is the minimal polynomial of i, and has roots {i, − i} in E ; and x2 − 3 is the minimal polynomial of img and has roots img It follows by Theorem 1 that img where G = gal(E : Q) . Construct img such that img then extend σ0 to img where σ(i) = − i. Hence σ img G and img Similarly, construct img such that τ0(i) = i ; then extend τ0 to img where img Hence τ img G and τ(i) = i. It follows easily that σ2 = ε and τ2 = ε, so img Now consider img Since img and img it follows that img If H0 = img σ img , H1 =img τ img and H2 = img στ img, the lattices are
img
Here each img is Galois (because Hi img G) and img Hence img means img; then img because img. Similarly img Finally, img so img
e. img, img; so img by Theorem 6 §10.1. Then E is the splitting field of x4 − 2 = (xu)(x + u)(xiu)(x + iu) so img is Galois. By Exercise 13 §10.1, img where o(σ) = 4, o(τ) = 2, στσ = τ and img The lattice diagrams are:
img
Here imgσ2img , img σ2, τ img , img σ img and imgσ2, τσimg are normal in G (because imgσ2 img = Z(G)), so img, img, img and img are Galois.
1. img: Hence σ2(i) = i; σ(u2) = − u2 so

img

Thus img. Also img and so we are done because img. Now img by Theorem 6 §10.1.
2. img: We have img; and img
3. img: We have τ(u2) = u2 and σ2(u2) = u2 (in (1)) so img. But img and img
4. img: Check img and img
5. img: Clearly img and img as x4 − 2 is irreducible.
6. img: We have img as τσ(iu) = − u and τσ(u) = − iu. This is equality as x4 + 8 is the minimum polynomial of img
7. img: Compute img; img has minimum polynomial x4 − 2.
8. img: We have img because τσ3(u) = iu and τσ3(iu) = u. The minimum polynomial of img is x4 + 8.
2.
a. If G = gal(E : F) then |G| = p2 implies G is abelian, so img or GCp × Cp by Theorem 7, §8.2. If img, the lattices are:
img
Here imgσ2 img imgF is Galois because imgσ2 img img G, and [img σ2 img img : F] = 2 . If GCp × Cp = img σ, τ img the lattices are:
img
Each of imgσ img imgF and imgτ img imgF are Galois because [img σ img img : F] = img and [img τ img img : F].
3. By Example 6 §10.1, img where σ : EF is the Frobenius automorphism given by σ(u) = up. By the Dedekind-Artin theorem [E : Gimg] = |G| = n. But img so, since img, img. Thus img is a Galois extension by Lemma 4.
Now the (inverted) lattice of subgroups of G = img σ img , when o(σ) = 12, is shown below at the right. Write Ek = img σk img img = {u img E img σk(u) = u} for each divisor k of n. Then the subfield lattice is shown below on the left.
img
Note that

img

so |Ek| = pk, and EkGF(pk).
4. a. img, σi img X, img. Thus u img H is fixed by all τ img H if and only if σ(u) = u for all σ img X.
5. a. Let img. Then

img

If char F = 2, this always holds and K = EG = E. Thus t img K so m = xt. If char F ≠ 2, write h(t) = f(t)g(− t). Then h(t) = h(− t) so h(t) = k(t2) for some polynomial k (because char F ≠ 2) . Thus f(t)g(− t) = k(t2) . Similarly and g(t)g(− t) = l(t2) for some polynomial l. Thus img, from which K = F(t2) . It follows that m = x2t2. Note that this shows [E : K] = 2 = |G| in this case, as the Dedekind-Artin theorem guarantees.
7. a. K′ is a subgroup of the abelian group G, so it is normal. Then the main theorem applies.
9.
a. If HHimg is onto, and K is an intermediate field, then K = Himg for a subgroup H, so Kimg′ = Himgimg = H′ = K. Thus K is closed. Conversely, if all intermediate fields K are closed then K = Kimg is the image of K′, and the map is onto.
11. Write G = gal(E : F) = {σ img autE img σ fixes F}. Since KF we have

img

Thus:
EK is Galois img K is the set of elements of E fixed by gal(E : K)
img K = {u img E img σ(u) = u for all σ img K′}
img K = Kimg
img K is closed in EF .
13. If [E : K] = 6 then img, so K′ is a subgroup of A4 of order 6. There is no such subgroup (Exercise 34 §2.6).
15. a. By its definition, K img L is the smallest intermediate field containing both K and L. Since the Galois connection KK′ is order reversing, (K img L)′ is the largest subgroup of G = gal(E : F) contained in both K′ and L′; that is (K img L)′ = K′ ∩ L′.
17. If K = σ(K1) let img. If img write img, u img K1. Then

img

so σKσ−1K′. On the other hand, if μ img K′ then img for all img, so μ[σ(u)] = σ(u) for all u img K1. Hence img, so img. This proves img
Conversely, assume img, that is img. If u img K1 then μ(u) = u for all img, so σ−1λσ(u) = u for all λ img K′, that is λ[σ(u)] = σ(u) for all λ img K′, that is σ(u) img Kimg. Since EF is Galois, Kimg = K, so this shows σ(u) img K, u img σ−1(K). Hence K1σ−1(K), so σ(K1) ⊆ K. Similarly Kσ(K1).
19. a. Let E = F(u1, u2, . . ., um) where X = {u1, u2, . . ., um} is the set of distinct roots of f in E. If σ img G then σ(ui) img X for all i so σ : XX is one-to-one (hence onto). Thus σ induces a permutation of these roots:

img

Now the map img is a group homomorphism GSm because

img

for all σ, τ in G. Moreover img means σ(ui) = ui for all i, so σ = ε in G by Theorem 3 §10.1. Thus img is an embedding.
20.
a. If img then, given

img

But τσ traces all of G as σ does, so img. This is true for all τ img G, so img. The proof that T(u) is in F is analogous.
c. If KF is Galois and H = gal(K : F), then |H| = n by Theorem 3, Corollary 3. Now p splits in K and has distinct roots u1, u2, . . ., um (Theorem 3). Hence p = (xu1)(xu2) img (xun). For each i there exists σi img H such that σi(u) = ui, so p = ∏ σimgG(xσ(u)). Thus the coefficients of xn−1 and 1 are, respectfully

img

21. If img define img for τ img G. Thus fτ = ∏ σimgG[xτσ(u)] = f because τσ runs through G as σ does. It follows that img for all τ img G, whence img for each j. Thus f img F[x]. Since f(u) = 0, p|f where p is the minimal polynomial of u over F. Write f = pmg where p and g are relatively prime in F[x]. If g ≠ 1, let q be an irreducible factor of g. Since q|f, 0 = q[σ(u)] = σ[q(u)] for some σ img G, so q(u) = 0. Since p(u) = 0 this is a contradiction because gcd(p, q) = 1 in F[x]. In fact, 1 = pr + qs gives 1 = 0 when we substitute x = u.
..................Content has been hidden....................

You can't read the all page of ebook, please click here login for view all page.
Reset
3.131.13.37