Skip to main content

Section C.19 January 2014

Subsection Section I: Group Theory

Activity C.155. Problem 1.

Prove that any group of order \(36\) has a normal subgroup of order \(3\) or \(9\text{.}\)
Solution.
Let \(G\) be a group of order \(36=2^2\cdot 3^2\text{.}\) By Sylow’s Theorems we have \(n_3|4\) and \(n_3\equiv 1\mod{3}\text{,}\) and thus \(n_3=1,4\text{.}\) If \(n_3=1\) then the unique Sylow \(3\)-subgroup is normal by Corollary 6.14, giving us a normal subgroup of order \(9\text{.}\)
Suppose then that \(n_3=4\text{,}\) let \(P\) be one of the subgroups of order \(9\text{,}\) and let \(G\) act on the cosets of \(H\) by left multiplication, inducing the homomorphism \(\rho:G\to S_4\) via the Permutation Representation. This is because there are \(9\) elements in \(P\text{,}\) and thus there are \(4\) left cosets by Lagrange’s Theorem.
As \(|S_4|=24<36=|G|\text{,}\) we see that \(\rho\) cannot be injective by [provisional cross-reference: cite] and thus \(N=\ker(\rho)\) is a non-trivial normal subgroup of \(G\) by [provisional cross-reference: cite].
Recall that the action of \(G\) on its cosets by left multiplication is always a transitive action [provisional cross-reference: cite], meaning there is exactly one orbit, \(\Orb(sP)\text{,}\) for some \(sP\in G/P\text{,}\) which must then have all \(4\) elements of \(G/P\) in it. By [cross-reference to target(s) "cor-orbit-stabilizer" missing or not unique], \(|G|=|\Orb(sP)|\cdot|\Stab(sP)|\text{,}\) and thus \(|\Stab(s)|=9\text{.}\) So there are \(9\) elements in \(G\) that fix \(s\) for any \(sP\in G/P\text{.}\)
Let \(xP \in G/P\) and \(g\in N\text{.}\) Notice that \(g\cdot xP=gxP\text{.}\) As \(g\in N\text{,}\) we know that \(\rho(g)\) yields the identity permutation, and thus that \(gxP=xP\text{.}\)
Thus \(N\subseteq\Stab(xP)\text{,}\) which has order 9.

Activity C.156. Problem 2.

Let \(G\) be a finite group and let \(H\) be a proper subgroup of \(G\) with \([G:H]=h\text{.}\)
  1. Prove that \(H\) has at most \(h\) distinct conjugate sets \(gHg^{-1}\) for \(g\in G\text{.}\)
  2. Prove that \(G\neq \bigcup_{g\in G} gHg^{-1}\text{.}\)
Solution.
  1. Let \(f:G/H\to gHg\inv\) be defined by \(f(gH)=gHg\inv\text{.}\) Suppose \(xH=yH\text{.}\) Notice that \(f(xH)=xHx\inv\) and \(f(H)=yHy\inv\text{,}\) but as \(x=y\) we have equality, and thus \(f\) is well defined. Let \(gHg\inv\) be a conjugate set, and let \(gH\in G/H\text{.}\) Then \(f(gH)=gHg\inv\text{,}\) and so we have surjectivity. As \(|G/H|=|[G:H]|=h\text{,}\) there can be at most \(h\) distinct conjugate sets for \(g\in G\text{.}\)
  2. Let \(G\) act on \(H\) by conjugation. We know that the orbits of this action partition \(H\) by [provisional cross-reference: cite]. However, under this action the orbits are exactly the conjugacy classes of \(H\) by [provisional cross-reference: cite]. There are at most \(h\) conjugacy classes. Each conjugacy class has at most \(|H|\) elements in it, and each one has the identity. As there are at most \(h\) of them then when we add all of their orders we get at most \(h(|H|-1)\text{,}\) which is less than the order of \(G\text{.}\)

Activity C.157. Problem 3 (*).

Let \(p\) be an odd prime, \(G\) a group of order \(p(p+1)\text{,}\) and \(P\) a Sylow \(p\)-subgroup of \(G\text{.}\) Assume that \(P\) is not normal in \(G\text{.}\)
  1. Prove that the normalizer of \(P\) in \(G\) is exactly \(P\text{.}\)
  2. Let \(P=\langle t\rangle\) and \(a\) an element of order \(2\) in \(G\) (such an element exists because \(p+1\) is even). Prove that the elements
    \begin{equation*} a, tat\inv,t^2at^{-2}, \dots, t^{p-1}at^{-(p-1)} \end{equation*}
    are all distinct.
Solution.
Let \(p\) be an odd prime, \(G\) a group of order \(p(p+1)\text{,}\) and \(P\) a Sylow \(p\)-subgroup of \(G\) that is not normal in \(P\text{.}\)
  1. First, by Sylow’s Theorems we see that \(n_p=p+1\text{.}\) So there are \((p-1)(p+1)=p^2-1\) elements of order \(p\text{,}\) and \(|G|=p^2+p\text{.}\) So there are \(p+1\) elements left.
    Also from Sylow’s Theorems we know \(G\) acts on \(\Syl_p(G)\) by conjugation. The normalizer of \(P\) is the set \(N_G(P)=\{g\in G|gPg\inv=P\}\text{,}\) which is exactly the stabilizer of \(P\) under this action by [provisional cross-reference: cite].
    Let \(x\in P\text{,}\) and observe \(xPx\inv=P\text{,}\) so \(x\in N_G(P)\text{,}\) and \(P\subseteq N_G(P)\text{.}\)
    The conjugation action on Sylow subgroups is transitive by Part (2) of Sylow’s Theorems, and thus there is exactly one orbit under this action by [provisional cross-reference: cite]. By [cross-reference to target(s) "cor-orbit-stabilizer" missing or not unique] we have \(|G|=|\Orb(x)|\cdot|\Stab(x)|\text{,}\) and thus \(|\Stab(x)|=p+1\) for every \(x\in P\text{.}\) However, \(p\) of these elements are accounted for by \(P\) and the last element is the identity, thus the only elements of \(N_G(P)\) are those in \(P\text{.}\)
  2. Let \(P=\langle t\rangle\text{,}\) \(a\) be an element of order \(2\) in \(G\text{,}\) and \(n<,m\leq p-1\text{.}\) Suppose by way of contradiction that \(t^nat^{-n}=t^mat^{-m}\text{.}\) By strategic multiplication we see \(a=t^kat^{-k}\text{,}\) where \(k=m-n\text{.}\) Thus we can reduce to the case where As \(a=t^kat^{-k}\) for some\(k\leq p-1\text{.}\)
    As \(a\) has order two, it is contained in some Sylow \(2\)-subgroup of \(G\text{.}\) Since \(P=\langle t\rangle\) we have \(t^k=x\) for some \(x\in P\text{.}\) Notice then that \(a=xax\inv\) means that \(x=axa\inv\text{,}\) placing \(a\in N_G(x)\text{,}\) contradicting the fact that \(a\not\in P\text{,}\) as it does not have order \(3\text{.}\) Thus \(a\) is distinct from \(t^kat^{-k}\) for all \(k\leq p-1\text{.}\) If any two powers of

Subsection Section II: Field Theory

Activity C.158. Problem 4 (*).

Let \(F\) be a field and \(F^*\) its group of units.
  1. Prove that any finite subgroup of \(F^*\) is cyclic.
  2. Suppose that \(F\) is algebraically closed that has characteristic \(p>0\text{.}\) For any positive integer \(n\text{,}\) prove that \(F^*\) has a subgroup of order \(n\) if and only if \(p\) does not divide \(n\text{.}\)
Solution.
  1. Let \(G\) be a finite subgroup of \(F^*\text{.}\) Let \(|G|=n\text{.}\)
    Let \(k\) be the LCM of all orders of elements in \(G\text{.}\) Then \(g^k=1\) and thus \(g\) is a root of the polynomial \(x^k-1\) for all \(g\in G\text{.}\) By Lagrange’s Theorem every element divides \(n\text{,}\) and so we have \(k\leq n\text{.}\) However, by the Factor Theorem CITEX the polynomial \(x^k-1\) can have at most \(k\) roots, and we have \(n\) distinct elements, and thus we have \(k=n\text{.}\) Thus there must exist an element of order \(n\) in \(G\text{,}\) making \(G\) cyclic, as desired.
  2. Let \(n\in\N\) and suppose that \(F\) has characteristic \(p>0\) and is algebraically closed.
    First, suppose by way of contradiction that \(F^*\) has a subgroup of order \(n\text{,}\) \(G\text{,}\) and \(p|n\text{.}\) From Part (a) \(G\) is cyclic and generated by some \(x\) such that \(|x|=n\text{.}\) However, as \(p|n\) we see that \(x^p=0\text{,}\) given that we are in an additive group. This contradicts the fact that \(n\) is the smallest number such that \(x^n=0\text{.}\)
    Now we proceed via the contrapositive. Suppose \(F^*\) does not have a subgroup of order \(n\text{.}\) Then there cannot exist a unit \(x\) such that \(|x|=n\text{.}\) Consider the polynomial \(x^n-1\text{.}\) As \(F\) is algebraically closed there exists some root \(r\in F\text{.}\) (how do we know this isn’t 1???) Notice that this means \(r^n=1\) and so \(r\) is a unit in \(F\text{.}\) As \(r\) cannot have order \(n\text{,}\) it must have an order that divides \(n\text{.}\) We also know that \(r^p=1\) as we are in a field of characteristic \(p\text{.}\) Thus \(|r|\) either divides \(p\) or is \(p\text{.}\) Either there is a non-identity root of \(x^n-1=(x-1)^n\text{.}\) Note that in this case \(n>p\) as we are

Activity C.159. Problem 5 (*).

Let \(E\) be a subfield of \(\C\) and suppose every element of \(E\) is a root of a polynomial of degree \(10\) in \(\Q[x]\text{.}\) Prove that \([E:\Q]\leq 10\text{.}\) (Note: \(E\) is not assumed to be a finite extension of \(\Q\text{.}\))
Solution.
Coming Soon!

Activity C.160. Problem 6 (*).

Let \(\omega\) be a primitive \(25\)th root of unity.
  1. Find \([\Q(\omega):\Q]\) and generator(s) for \(\Gal(\Q(\omega)/\Q)\text{.}\)
  2. Draw the subfield lattice for \(\Q(\omega)\) and indicate the degrees of each extension. (You do not have to find generators for each of the subfields)
Solution.
Let \(\omega\) be a primitive \(25\)th root of unity.
As \(\omega\) is a \(25\)th root of unity we see that \(\omega^{25}=1\) and thus that it is the root of the polynomial \(f(x)=x^{25}-1\in\Q[x]\text{.}\) We can factor out an \((x-1)\) from \(f(x)\) to see \(f(x)=(x-1)(x^{24}+x^{23}+\dots+x+1)\text{,}\) and thus \(\omega\) is a root of \(g(x)=(x^{24}+x^{23}+\dots+x+1)\) as well. Notice that the content of \(g(x)=1\text{,}\) and thus by Gauss’s Lemma \(g(x)\) is irreducible in \(\Q[x]\) if and only if it is irreducible in \(\Z[x]\text{.}\)

Subsection Section III: Rings, Modules, and Linear Algebra

Activity C.161. Problem 7.

Prove that \(\Z[\sqrt{-5}]\) is not a UFD. Justify all details.
Solution.
First, notice that \(2\cdot 3=6=(1+\sqrt{-5})(1-\sqrt{-5})\text{.}\) Define a function
\begin{equation*} N:\Z[\sqrt{-5}]\to\Z, N(a+b \sqrt{-5})=a^2+5b^2. \end{equation*}
Let \(\alpha,\beta\in\Z[\sqrt{-5}]\) and behold
\begin{equation*} \begin{align} N(\alpha\beta) &=N((a+b \sqrt{-5})(c+d \sqrt{-5}))\\ &=N(ac-5bd+ad \sqrt{-5}+bc \sqrt{-5})\\ &=(ac-5bd)^2+5(ad+bc)^2\\ &=((ac)^2-10acbd+25(bd)^2)+5((ad)^2+2abcd+(bc^2))\\ &=a^2c^2+25b^2d^2+5a^2d^2+5b^2c^2\\ &=(a^2c^2+5a^2d^2)+(5b^2c^2+25b^2d^2)\\ &=a^2(c^2+5d^2)+5b^2(c^2+5d^2)\\ N(\alpha)N(\beta)&=(a^2+5b^2)(c^2+5d^2).\end{align} \end{equation*}
So there’s that done.
Suppose by way of contradiction there exist \(\a,\b\) such that \(2=\a\b\text{.}\) Then \(\a=a+b\sqrt{-5}\) and \(\b=c+d\sqrt{-5}\text{,}\) and
\begin{equation*} N(2)=4=N(\a\b)=N(\a)N(\b)=(a^2+5b^2)(c^2+5d^2). \end{equation*}
Thus \((a^2+5b^2)=\pm1,\pm2,\) or \(\pm4\text{,}\) as these are the only integer divisors of \(4\text{.}\) However, there do not exist integers \(a,b\) such that this is true. Thus \(2\) is irreducible in \(\Z[\sqrt{-5}]\text{.}\)
Suppose by way of contradiction that \(2\) is prime in \(\Z[\sqrt{-5}]\text{.}\) Note that \(2|6=(1+\sqrt{-5})(1-\sqrt{-5})\text{.}\) Thus \(2\) divides one of these factors.
First, suppose there exists some \(a+b\sqrt{-5}\) such that \(2(a+b\sqrt{-5})=(1\pm\sqrt{-5})\text{.}\) Thus \(2a+2b\sqrt{-5}=1\pm\sqrt{-5}\text{,}\) and so \(2a=\pm1\text{.}\) However, \(\pm\frac12\) is not an integer, and thus \(2\) cannot divide either of these factors. Thus \(2\) is not prime in \(\Z[\sqrt{-5}]\text{.}\) By Part (a), this is not a UFD.

Activity C.162. Problem 8 (*).

Let \(R\) be a commutative ring with identity and \(M\) an \(R\)-module. Let \(f\) : \(M \rightarrow M\) be a surjective \(R\)-module homomorphism.
  1. Give an cxample of such an \(R, M\text{,}\) and \(f\) such that \(f\) is not an isomorphism.
  2. Prove that if \(M\) is Noctherian then \(f\) is an isomorphism.
Solution.
Coming Soon!

Activity C.163. Problem 9.

Let \(F\) be a field and \(A\) a square matrix with entries from \(F\text{.}\) Prove that \(A\) is similar to its transpose.
Solution.
Let \(L\) be the algebraic closure of \(F\text{.}\) Thus \(A\) has a Jordan Canonical Form in \(L\text{.}\) For each Jordan block \(J_i\) in the JCF of \(A\text{,}\) let \(B_i\) denote the transpose of the identity matrix, and notice that \(B_iJ_iB_i\inv=J_i^T\text{.}\) As this is the case for every Jordan block, we see that the JCF of \(A\text{,}\) \(J\text{,}\) is similar to its transpose. As the \(A\) is similar to \(J\text{,}\) \(A^T\) is similar to \(J^T\text{,}\) and \(J\) is similar to \(J^T\text{,}\) we see that \(A\sim A^T\) in \(L\) by transitivity.
Suppose \(A\) and \(B\) are similar in \(\Mat_{n \times n}(L)\text{.}\) As \(A\) and \(B\) have entries in \(F\text{,}\) then they are both in \(\Mat_{n \times n}(F)\text{.}\) Thus there exist matrices \(C,D\in\Mat_{n \times n}(F)\) in RCF such that \(A\) is similar to \(C\) and that \(B\) is similar to \(D\text{.}\) However, \(A\) is similar to \(C\) and that \(B\) is similar to \(D\) in \(\Mat_{n \times n}(L)\) as well. Notice \(C\) and \(D\) are still in RCF. However, as the RCF is unique, this means that \(C=D\) in \(\Mat_{n \times n}(L)\text{,}\) making them equal in \(\Mat_{n \times n}(F)\) as well. Thus \(A\) is similar to \(B\text{,}\) as similarity is transitive.
This yields \(A\sim A^T\) in \(F\text{.}\) # Jan 15