\newif\ifOCW %\OCWtrue \documentclass[11pt]{article} \usepackage{amsmath,amssymb,amsthm} \usepackage{hyperref} \hypersetup{colorlinks=true,urlcolor=blue,citecolor=blue,linkcolor=blue} \usepackage{courier} \usepackage{tikz} \usepackage{tikz-cd} \usetikzlibrary{calc,matrix,arrows,decorations.markings} \usepackage{array} \usepackage{color} \usepackage{enumerate} \usepackage{nicefrac} \usepackage{listings} \ifOCW\usepackage{soul}\let\oldhref\href\renewcommand{\href}[2]{\oldhref{#1}{\ul{#2}}}\fi \ifOCW\newcommand{\due}[1]{\hfill\phantom{Due: #1}}\else\newcommand{\due}[1]{\hfill{Due: #1}}\fi \lstset{ basicstyle=\small\ttfamily, keywordstyle=\color{blue}, language=python, xleftmargin=16pt, } \textwidth=5.8in \textheight=9in \topmargin=-0.5in \headheight=0in \headsep=.5in \hoffset -.4in \pagestyle{plain} % growing list of useful macros, use these where appropriate and add to this list as needed \newcommand{\kbar}{\bar{k}} \newcommand{\Fp}{\mathbb{F}_p} \newcommand{\Fpbar}{\overline{\mathbb{F}}_p} \newcommand{\Fq}{\mathbb{F}_q} \newcommand{\Fqbar}{\overline{\mathbb{F}}_q} \newcommand{\Fqm}{\mathbb{F}_{q^m}} \newcommand{\Fqn}{\mathbb{F}_{q^n}} \newcommand{\F}{\mathbb{F}} \newcommand{\Q}{\mathbb{Q}} \newcommand{\Qbar}{\overline{\Q}} \newcommand{\R}{\mathbb{R}} \newcommand{\C}{\mathbb{C}} \renewcommand{\H}{\mathbb{H}} \newcommand{\Z}{\mathbb{Z}} \newcommand{\Zhat}{\widehat\Z} \newcommand{\Zbar}{\overline\Z} \newcommand{\Kbar}{\overline K} \newcommand{\NN}{\mathbb{N}} \newcommand{\Aut}{{\rm Aut}} \newcommand{\Gal}{{\rm Gal}} \newcommand{\SL}{{\rm SL}} \newcommand{\GL}{{\rm GL}} \newcommand{\PGL}{{\rm PGL}} \newcommand{\dy}{\,dy} \newcommand{\dx}{\,dx} \newcommand{\tr}{\operatorname{tr}} \newcommand{\kron}[2]{\bigl(\frac{#1}{#2}\bigr)} \newcommand{\lcm}{\operatorname{lcm}} \newcommand{\Exp}{{\rm E}} \renewcommand{\O}{\mathcal{O}} \newcommand{\OK}{\O_K} \newcommand{\T}{{\rm T}} \newcommand{\N}{{\rm N}} \renewcommand{\Re}{\operatorname{Re}} \renewcommand{\Im}{\operatorname{Im}} \newcommand{\Li}{\operatorname{Li}} \newcommand{\ord}{\operatorname{ord}} \newcommand{\Cl}{\operatorname{Cl}} \newcommand{\disc}{\operatorname{disc}} \newcommand{\p}{\mathfrak{p}} \newcommand{\q}{\mathfrak{q}} \renewcommand{\r}{\mathfrak{r}} \newcommand{\m}{{\mathfrak m}} \renewcommand{\c}{{\mathfrak c}} \renewcommand{\a}{{\mathfrak a}} \newcommand{\Frob}{{\rm Frob}} \newcommand{\Hom}{\operatorname{Hom}} \newcommand{\tor}{{\rm tors}} \newcommand{\I}{\mathcal{I}} \newcommand{\D}{\mathcal{D}} \newcommand{\id}{\operatorname{id}} \newcommand{\A}{\mathcal{A}} \newcommand{\Spl}{\mathrm{Spl}} \renewcommand{\P}{\mathcal{P}} \newcommand{\ab}{{\rm ab}} \newtheorem*{theorem}{Theorem} \theoremstyle{definition} \newtheorem*{definition}{Definition} \newtheorem*{remark}{Remark} \begin{document} \setlength{\unitlength}{1in} \begin{center} \large \textbf{18.785 Number Theory\hfill Fall~2025}\\\vspace{4pt} \textbf{Problem Set \#11\due{12/5/2025}}\\\vspace{-6pt} \normalsize \begin{picture}(5.8,.1) \put(0,0) {\line(1,0){5.8}} \end{picture} \end{center} \subsection*{Description} These problems are related to Lectures 18--20. Your solutions should be written up in latex and submitted as a pdf-file \ifOCW\else to \href{https://www.gradescope.com/courses/1110057}{Gradescope} \fi by midnight on the date due. \medskip \noindent \textbf{Instructions:} Solve any combination of problems that sum to 100 points, or \textbf{alternatively, solve any combination of problems that sum to 200 points to receive credit for two problem sets}. Collaboration is permitted/encouraged, but you must identify your collaborators (including any LLMs you consulted) and any references you consulted outside the course \ifOCW syllabus \else\href{https://math.mit.edu/classes/18.785/2025/syllabus.html}{syllabus}\fi. Include this information after the \textbf{Collaborators/Sources} prompt at the end of the problem set (if there are none, you should enter ``none'', do not leave it blank). Note that each student is expected to write their own solutions; it is fine to discuss problems with others, but your writing must be your own. \subsection*{Problem 1. Higher ramification groups (49 points)} Let $A$ be a complete DVR with finite residue field; its fraction field $K$ is a nonarchimedean local field (Prop. 9.6). Let $L$ be a finite Galois extension of $K$, let $G:=\Gal(L/K)$, and let~$B$ be the integral closure of $A$ in $L$, with maximal ideal $\q=(\pi)$. Fix $\alpha\in B$ so that $B=A[\alpha]$ (via Theorem 10.14), and let $f\in A[x]$ be the minimal polynomial of $\alpha$. The decomposition group $D_\q$ is equal to $G$ (since $\sigma(\q)=\q$ for all $\sigma\in G$), and the inertia subgroup is $I_\q:=\{\sigma\in G:\sigma(x)\equiv x\bmod \q\text{ for all }x\in L\}$ with order equal to the ramification index $e:=e_\q$. For any integer $i\ge -1$ define \[ G_i:= \{\sigma\in G:\sigma(x)\equiv x\bmod \q^{i+1} \text{ for all }x\in B\}, \] so that $G_{-1}=G$ and $G_0$ is the inertia subgroup. The group $G_i$ is the $i$th \emph{ramification group} of $G$ (in the lower numbering). Define $i_G:G\to \Z \cup \{\infty\}$ by $i_G(\sigma):=v_\q(\sigma(\alpha)-\alpha)$. \begin{enumerate}[{\bf(a)}] \item Prove that $G_i=\{\sigma\in G:i_G(\sigma)\ge i+1\}$, show that $G_{i+1}$ is a normal subgroup of~$G_i$, and show that the groups $G_i$ are trivial for all sufficiently large $i$. \end{enumerate} \noindent Recall that the different ideal $\mathcal D:=\mathcal D_{B/A}$ is equal to $(f'(\alpha))$ and satisfies the bounds \[ e-1\le v_\q(\mathcal D) \le e-1+v_\q(e), \] with $e-1=v_\q(\mathcal D)$ if and only if $v_\q(e)=0$, by Proposition 12.23 and Theorem 12.26. \begin{enumerate}[{\bf(a)}] \setcounter{enumi}{1} \item Prove Hilbert's \emph{different formula}: \[ v_\q(\mathcal D) = \sum_{\sigma\ne 1}i_G(\sigma) = \sum_{i\ge 0} (\#G_i-1). \] \end{enumerate} \noindent Let $U_0:= B^\times$ be the unit group of $B$, and for $i > 0$ define \[ U_i:= 1+\q^i = \{x\in U_0:x\equiv 1\bmod \q^i\}. \] \begin{enumerate}[{\bf(a)}] \setcounter{enumi}{2} \item Show that $U_0/U_1\simeq (B/\q)^\times$ and that for $i>0$ we have $U_i/U_{i+1}\simeq \q^i/\q^{i+1}$ isomorphic to the additive group of $B/\q$. Conclude that $L/K$ is tamely ramified if and only if $G_1$ is trivial and totally wildly ramified if and only if $G=G_1$.\footnote{The group $G_1$ is sometimes called the \emph{wild inertia group}.} \item Show that for $i\ge 0$ the map $\phi_i\colon G_i\to U_i$ defined by $\sigma \mapsto \sigma(\pi)/\pi$ induces an injective group homomorphism $G_i/G_{i+1}\hookrightarrow U_i/U_{i+1}$ (that is, the composition of $\phi_i$ with the quotient map $U_i\to U_i/U_{i+1}$ is a homomorphism with kernel $G_{i+1}$). Conclude that (i) $G_0/G_1$ is cyclic of order prime to $p$, (ii) $G_1$ is the unique $p$-Sylow subgroup of $G_0$, (iii) $G_i/G_{i+1}$ an abelian $p$-group for all $i\ge 1$, and (iv) $G=\Gal(L/K)$ is solvable. \item Suppose that $\sigma\in G_i-G_{i+1}$ and $\tau\in G_j-G_{j+1}$ with $1\le i\le j$. Show that $\phi(\sigma\tau)-\phi(\tau\sigma)\equiv (j-i)u\pi^{i+j}\bmod \q^{i+j+1}$ for some $u\in U_0$ and that this implies $\phi(\sigma\tau\sigma^{-1}\tau^{-1})\equiv 1+(j-i)u\pi^{i+j}\bmod \q^{i+j+1}$. Then use this to show $i\equiv j\bmod p$. \end{enumerate} \noindent Let $K=\Q_p$ with $p$ odd, and let $L/\Q_p$ be a totally wildly ramified abelian extension, so in the notation above, $\Gal(L/K)=G=G_0=G_1$, and $\mathcal D$ is the different ideal. \begin{enumerate}[{\bf(a)}] \setcounter{enumi}{5} \item Show that $v_\q(\mathcal D)=2p-2$ if $[L:K]=p$ and $v_\q(\mathcal D)=3p^2-p-2$ if $[L:K]=p^2$. \item Prove that $G$ is cyclic (hint: reduce to $[L:K]=p^2$ then show that if $H\le G$ has order $p$ then $H=G_{p+1}$ by computing the different of $L/L^H$ using \textbf{(b)} and \textbf{(f)}). \end{enumerate} \subsection*{Problem 2. Polar density (49 points)} Let $K$ be a number field and let $S$ be a set of primes of $K$. Recall the \emph{Dirichlet density} \[ d(S):=\lim_{s\to 1^+}\frac{\sum_{\p\in S}\N(\p)^{-s}}{\sum_\p \N(\p)^{-s}} = \lim_{s\to 1^+}\frac{\sum_{\p\in S}\N(\p)^{-s}}{\log\tfrac{1}{s-1}}, \] and the \emph{natural density} \[ \delta(S):=\lim_{x\to\infty}\frac{\#\{\p\in S:\N(\p)\le x\}}{\#\{\p:\N(\p)\le x\}}, \] which are defined whenever these limits exist. As shown on Problem Set 9, if $\delta(S)$ exists then so does $d(S)=\delta(S)$ (you may use this fact even if you did not prove it). \begin{definition} The \emph{partial Dedekind zeta function} associated to $S$ is the complex function \[ \zeta_{K,S}(s):=\prod_{\p\in S}(1-\N(\p)^{-s})^{-1}. \] If for some integer $n\ge 1$ the function $\zeta_{K,S}^n$ extends to a meromorphic function on a neighborhood of $1$, the \emph{polar density} of $S$ is defined by \[ \rho(S):= \frac{m}{n},\qquad m=-\ord_{s=1}\zeta_{K,S}^n(s). \] \end{definition} \begin{enumerate}[{\bf(a)}] \item Show that $\rho(S)$ is well defined (so $m/n$ does not depend on the choice of $n$). \item Prove that $\delta(S)=d(S)=\rho(S)$ whenever $\rho(S)$ exists. \item Show that $\rho(S)=0$ when $S$ is finite and $\rho(S)=1$ when $S$ is cofinite. \item Show that $S\subseteq T$ implies $\rho(S)\le \rho(T)$ whenever both densities exist. \item Let $\mathcal P_1$ denote the set of degree-1 primes of $K$ (those of prime norm). Prove that $\rho(\mathcal P_1)=1$ and $\rho(S\cap\mathcal P_1)=\rho(S)$ whenever $S$ has a polar density. \item Let $L/K$ be a Galois extension and let $\Spl(L/K)$ be the set of primes of $K$ that split completely in $L$. Prove that $\rho(\Spl(L/K))=1/[L:K]$. Conclude that for $G=\Gal(L/K)$ and any normal subgroup $H\subseteq G$, the set $S$ of primes $\p$ of $K$ for which the Frobenius conjugacy class $\Frob_\p$ lies in $H$ has polar density $\rho(S)=\#H/\#G$. \item Prove that if $L/K$ and $M/K$ are Galois extensions of $K$ then $\Spl(M)=\Spl(L)$ if and only if $M=L$. \end{enumerate} \subsection*{Problem 3. The Frobenius density theorem (49 points)} Let $L/K$ be a Galois extension of number fields of finite degree $n$ with Galois group $G:=\Gal(L/K)$. Recall that for each unramified prime $\p$ of $K$, the \emph{Frobenius class} $\Frob_\p$ is the conjugacy class of the Frobenius elements $\sigma_\q$ for $\q|\p$. The \emph{Chebotarev density theorem} states that for any set $C\subseteq G$ stable under conjugation (a union of conjugacy classes), the set of unramified primes $\p$ with $\Frob_\p\subseteq C$ has Dirichlet density $\#C/\#G$.\footnote{It also has this natural density, but this was proved later.} In this problem you will prove the \emph{Frobenius density theorem}, which says essentially the same thing, but with a different notion of conjugacy. \begin{definition} Two elements $g$ and $h$ of a group $G$ are \emph{quasi-conjugate} if they generate conjugate subgroups $\langle g\rangle$ and $\langle h\rangle$. \end{definition} \begin{enumerate}[{\bf(a)}] \item Show that quasi-conjugacy is an equivalence relation and that each quasi-conjugacy class in a group is a union of conjugacy classes. \item Show that in the symmetric group $S_n$, each quasi-conjugacy class is actually a conjugacy class (so the Frobenius density theorem implies the Chebotarev density theorem in this case), but that this is generally not true for the alternating group~$A_n$. \item Suppose $G$ is cyclic. For each $d|n$, let $S_d$ be the set of primes $\p$ of $K$ for which the primes $\q|\p$ have inertia degree $f_\q=d$. Prove that the set $S_d$ has polar density $\rho(S_d)=\phi(d)/[L\!:\!K]$ and conclude that infinitely many primes of $K$ are inert in $L$. \end{enumerate} \noindent Fix $\sigma\in G$, let $K'=L^\sigma$ be its fixed field, let $H=\langle\sigma\rangle\subseteq G$, and let $d=\#H$. Recall that in any number field, a \emph{degree-1 prime} is a prime whose absolute norm is prime. For each prime $\p$ of $K$ (resp. $K'$) that is unramified in $L$, let $\overline\Frob_\p$ denote the quasi-conjugacy class in~$G$ (resp. $H$) that contains the conjugacy class $\Frob_\p$. \begin{enumerate}[{\bf(a)}] \setcounter{enumi}{3} \item Let $S'$ be the set of degree-1 primes $\p'$ of $K'$ for which $\p=\p'\cap \O_K$ is unramified in $L$ and for which $\sigma\in \overline\Frob_{\p'}$. Prove that $S'$ has polar density $\rho(S')=\phi(d)/d$. \item Let $S$ be the set of unramified degree-1 primes $\p$ of $K$ for which $\sigma\in \overline\Frob_\p$. Show that map $\p'\mapsto \p'\cap \O_K$ defines a surjective map $\pi\colon S'\to S$. \item Show that the fibers of $\pi$ all have cardinality $[K':K]/c$, where $c$ is the number of distinct conjugates of $H$ in $G$. \item Show that $S$ has polar density \[ \rho(S)=\frac{c\phi(d)}{[L\!:\!K]}. \] \item Prove that for any set $C\subseteq G$ stable under quasi-conjugation the set of unramified primes $\p$ of $K$ with $\overline\Frob_\p\subseteq C$ has polar density $\#C/\#G$. \end{enumerate} \subsection*{Problem 4. The Hilbert symbol (98 points)} Let $K$ be a local field whose characteristic is not $2$. \begin{definition} The \emph{local Hilbert symbol} is the map $(\cdot,\cdot)\colon K^\times/K^{\times 2}\times K^\times/K^{\times 2}\to \{\pm 1\}$ \[ (a,b) := \begin{cases} 1 &\text{if $ax^2+by^2=1$ has a solution in $K$},\\ -1 &\text{otherwise.} \end{cases} \] Here and throughout this problem $a,b\in K^\times $ are understood to represent elements of $K^\times/K^{\times 2}$ whenever the context requires it. \end{definition} \begin{enumerate}[{\bf(a)}] \item Prove that the Hilbert symbol satisfies: \begin{enumerate}[{\bf(i)}] \item $(a,b)=(b,a)$ (symmetry); \item $(a,bc)=(a,b)(a,c)$ and $(ab,c)=(a,c)(b,c)$ (bilinearity); \item For any $a\in K^\times$, if $(a,b)=1$ for all $b\in K^\times$ then $a\in K^{\times 2}$ (nondegeneracy). \item $(a,1-a)=1$ (for $a\ne 1$) and $(a,-a)=1$ (Steinberg relations). \end{enumerate} \item In part (a) where (if anywhere) did you use the fact that $K$ is a local field? Determine which of (i)-(iv) hold for all fields whose characteristic is not 2, and for those that do not, give explicit counterexamples. \item Prove that for $a\not\in K^{\times 2}$ we have $(a,b)=1$ if and only if $b\in \N_{K(\sqrt{a})/K}(K(\sqrt{a})^\times)$. \item Let $L/K$ be an abelian extension. Let $r_{L/K}\colon K^\times/\N_{L/K}(L^\times)\to\Gal(L/K)$ be the isomorphism given by Artin reciprocity, and $\langle\cdot,\cdot\rangle := \Gal(\Kbar/K)\times K^\times/K^{\times 2}\to \{\pm 1\}$ the Kummer pairing $\langle \sigma,a\rangle := \sigma(\sqrt{a})/\sqrt{a}$ (which can be applied to $\sigma\in \Gal(L/K)$ whenever $L\subset\Kbar$ contains $\sqrt{a}$). Prove that the Hilbert symbol satisfies \[ (a,b) = \left \langle r_{K(\sqrt{b})/K}(a),b\right\rangle. \] \item For $a,b,c\in K^\times$ prove $ax^2+by^2=c$ has a solution if and only if $(-ab,c)=(a,b)$. \item For $a,b\in K^\times$ define the quaternion algebra $H_{a,b}$ as the $K$-algebra $K(i,j)$ with $i^2=a$, $j^2=b$, $ij=-ji$. Show that $(a,b)=1$ if and only if $H_{a,b}\simeq \mathrm{M}_2(K)$, the $2\times 2$ matrix algebra over $K$ (such quaternion algebras are said to \emph{split}). Then show that $H_{a,b}\simeq H_{a,c}$ if and only if $[b]=[c]$ in $K^\times/N_{K(\sqrt{a})/K}(K(\sqrt{a})^\times)$ and deduce that the isomorphism class of $H_{a,b}$ depends only on the Hilbert symbol $(a,b)$. \item Show that for archimedean $K$ we have $(a,b)=-1$ if and only if $K\simeq \R$ and $a,b<0$. \item Suppose that $K$ is nonarchimedean with residue field of odd cardinality $q$. Let $\O$ be its valuation ring, $\pi$ a uniformizer for $\O$. Define the residue symbol \[ \left(\frac{a}{\pi}\right):=\begin{cases} 1 & \text{if }\bar a\in \F_q^{\times 2},\\ -1 & \text{if }\bar a\not\in \F_q^{\times 2}, \end{cases} \] where $\F_q:=\O/(\pi)$ is the residue field (which does not depend on the choice of $\pi$). For $a,b\in K^\times$, let $a=u_a\pi^\alpha$, $b=u_b\pi^\beta$ with $u_a,u_b\in \O^\times$. Prove the \emph{reciprocity law}: \[ (a,b) = (-1)^{\alpha\beta(q-1)/2}\left(\frac{u_a}{\pi}\right)^{\beta}\left(\frac{u_b}{\pi}\right)^\alpha. \] \item (\textbf{optional}) Now let $K$ be a global field of characteristic not $2$, and for each place $v$ of $K$ let $(a,b)_v$ denote the Hilbert symbol of the completion of $K$ at $v$. Prove the \emph{product formula}, which states that \[ \prod_v(a,b)_v=1 \] for all $a,b\in K^\times$ (and in particular, $(a,b)_v=1$ for all but finitely many places $v$). (Hint: Use the global Artin homomorphism (see Definition 28.2 in \href{https://math.mit.edu/classes/18.785/2021fa/LectureNotes28.pdf}{these notes})). \end{enumerate} \noindent Useful references for this problem if you get stuck are \cite[Ch.\,III]{serre73} and \cite[\S 5.6]{voight}. \subsection*{Problem 5. Profinite groups (98 points)} Recall that a topological space is \emph{totally disconnected} if every pair of distinct points can be separated by open neighborhoods that partition the space; totally disconnected spaces are obviously Hausdorff. \begin{enumerate}[{\bf(a)}] \item Show that products and inverse limits of totally disconnected compact topological spaces are totally disconnected and compact. Conclude that every profinite group is a totally disconnected compact group. \end{enumerate} Let $G$ be a totally disconnected compact group, let $\widehat G:= \varprojlim G/N$ be its profinite completion (so $N$ varies over finite index open normal subgroups of $G$ ordered by containment), and let $\phi\colon G\to \widehat G$ be the natural map that sends each $g\in G$ to its images in the finite quotients $G/N$. \begin{enumerate}[{\bf(a)}] \setcounter{enumi}{1} \item Show that every open subgroup of $G$ has finite index and contains an open normal subgroup (which necessarily also has finite index). \item Show that $\phi(G)$ is both dense in $\widehat G$ and closed, hence equal to $\widehat G$; thus $\phi$ is surjective. \item Show that to prove that $\phi$ is injective it suffices to show that the intersection of all open subgroups of $G$ is trivial. Then show that for every $g\in G-\{1\}$ there is a neighborhood $U$ of $1$ that is both open and closed and does not contain $g$, and it is enough to show that every such $U$ contains an open subgroup $H$. \item Let $U$ be a neighborhood of $1$ that is both open and closed. Show that $U$ contains an open neighborhood of $1$ that is closed under multiplication and inversion, hence a subgroup (this requires some work; you will need to use the fact that the multiplication map $G\times G\to G$ is continuous and that $U$ is compact). \item Show that $\phi$ is a continuous open map, hence a homeomorphism. Conclude that $G$ is isomorphic to its profinite completion, and in particular, a profinite group. \item Show that for a profinite group $G$ the following are equivalent: (i) the topology of $G$ is induced by a metric, (ii) $G\simeq \varprojlim G_n$, with $n\in \Z_{\ge 1}$, the $G_n$ finite, and $G_{n+1}\to G_n$ surjective, (iii) the number of open subgroups of $G$ is countable. \item Show that the equivalent conditions (i)-(iii) in (g) imply that $G$ contains a countable dense subset (so $G$ is \emph{separable} as a topological space), and give an example showing that the converse does not hold. \item Let $p$ be prime, let $\Z_p:=\varprojlim_n \Z/p^n\Z$, and let $G_p:=\prod_n \Z/p^n\Z$. Show that $G_p$ and $\Z_p\times (G_p/\Z_p)$ are isomorphic as groups. Are they isomorphic as topological groups? \item Show that $\widehat \Z^\times \simeq \widehat\Z\times \prod_n\Z/n\Z$ as topological groups. \end{enumerate} \noindent A useful reference for this problem if you get stuck is Ribes and Zalesskii~\cite{RZ}. \subsection*{Problem 6. Arithmetically equivalent number fields (98 points)} Two number fields are said to be \emph{arithmetically equivalent} if their Dedekind zeta functions coincide. Isomorphic number fields obviously have the same zeta functions, but as proved by Gassmann \cite{gassmann}, the converse need not hold. A particularly simple example is given by the fields $\Q(\sqrt[8]{a})$ and $\Q(\sqrt[8]{16a})$; as shown by Perlis \cite{perlis}, for $a\in \Z$ not square and not twice a square, these fields are arithmetically equivalent but nonisomorphic (they have the same Galois closure, which is obtained by adjoining a primitive 8th root of unity). \begin{enumerate}[{\bf (a)}] \item Show that arithmetically equivalent number fields must have the same Galois closure (up to isomorphism). Conclude that if $K/\Q$ is Galois then $K'$ is arithmetically equivalent to $K$ if and only if $K'\simeq K$. \end{enumerate} In view of (a) we now consider two non-Galois extensions $K$ and $K'$ of $\Q$ with Galois closure $L/\Q$. Let $G:=\Gal(L/\Q)$ and put $H:=\Gal(L/K)$ and $H':=\Gal(L/K')$. \begin{enumerate}[{\bf (a)}] \setcounter{enumi}{1} \item Show that if $K$ and $K'$ are arithmetically equivalent number fields then for every prime $p$ there is a bijection between the primes of $K$ lying above $p$ and the primes of $K'$ lying above $p$ that preserves inertia degrees. Conclude that $K$ and $K'$ must have the same degree. \item Given a cyclic $C\subseteq G$, we can partition $G$ into double cosets $H\sigma_1C,\ldots,H\sigma_gC$. When $H$ is not normal these cosets need not have the same size; each will have cardinality $f_i\cdot \#H$ for some integer $f_i\ge 1$. Assume the $f_i$ are in increasing order and call the tuple $(f_1,\ldots,f_g)$ the \emph{coset type} of the pair $(H,C)$. Prove that if $p$ is a prime that is unramified in $L$ and $C$ is any decomposition group of $p$ in $G$, then the coset type of $(H,C)$ is equal to the tuple of inertia degrees of the primes of $K$ lying above $p$ when arranged in increasing order. Conclude that if $K$ and $K'$ are arithmetically equivalent then $(H,C)$ and $(H',C)$ have the same coset type for every cyclic subgroup $C$ of $G$. \item Two subgroups $H$ and $H'$ of a finite group $G$ are said to be \emph{Gassmann equivalent} if for conjugacy class $c$ of elements in $G$ the sets $c\cap H$ and $c\cap H'$ have the same cardinality. Show that this holds if and only if for every cyclic group $C$ the coset types of $(H,C)$ and $(H',C)$ coincide. \item Suppose $H$ and $H'$ are Gassmann equivalent. Show that $K$ and $K'$ have the same number of real and complex places (hint: consider the ``decomposition group" of the prime $p=\infty$ in $G$). \end{enumerate} \noindent Like the Riemann zeta function, the Dedekind zeta function has a meromorphic continuation to $\C$ that satisfies a functional equation. Define $\Gamma_\R(s):=\pi^{-s/2}\Gamma(s/2)$ and $\Gamma_\C(s):=2(2\pi)^{-s}\Gamma(s)$, and let \[ Z_K(s):=|D_K|^{s/2}\Gamma_\R(s)^{n_1}\Gamma_\C(s)^{n_2}\zeta_K(s), \] where $n_1$ and $n_2$ are the number of real and complex places of $K$, respectively. Then \[ Z_K(s)=Z_K(1-s) \] as meromorphic functions on $\C$ (you may assume this in what follows). \begin{enumerate}[{\bf (a)}] \setcounter{enumi}{5} \item Let $f(s)=f_1(s)/f_2(s)$ be a ratio of Euler products over a finite set of primes that extends to a meromorphic function on $\C$ that satisfies a functional equation $f(s)=g(s)f(1-s)$ for some meromorphic function $g(s)$ whose zeros and poles do not coincide with any zero or pole of $f$. Prove that $g(s)=1$, then use this and the functional equations for $\zeta_K(s)$ and $\zeta_{K'}(s)$ to prove that if $H$ and $H'$ are Gassmann equivalent then $K$ and $K'$ are arithmetically equivalent. \item A \emph{Gassmann triple} (or \emph{Gassmann-Sunada triple}) is a triple $(G,X,Y)$ in which $G$ is a group that acts faithfully on sets $X$ and $Y$ such that every element of $G$ fixes the same number of elements in $X$ and $Y$ but $X$ and $Y$ are not isomorphic as $G$-sets. Show that $K$ and~$K'$ are arithmetically equivalent but not isomorphic if and only if $(G,G/H,G/H')$ is a Gassmann triple, where $G/H$ and $G/H'$ denote sets of cosets (for consistency with part (c), use right cosets and put the $G$-action on the right). \item Show that if $K$ and $K'$ are arithmetically equivalent then they have the same normal core (largest subfield that is a normal extension of $\Q$). Use this to show that~$K$ and $K'$ contain the same roots of unity and conclude that the unit groups $\O_K^\times$ and $\O_{K'}^\times$ are isomorphic as abelian groups. \item Show that if $K$ and $K'$ are arithmetically equivalent then $|\disc\O_K|=|\disc \O_{K'}|$ and conclude that $h_KR_K=h_{K'}R_{K'}$, where $h_K:=\#\Cl\O_K$ is the class number and $R_K$ is the regulator of $K$ (and similarly for $K'$). \end{enumerate} \noindent It is natural to ask whether the class numbers and regulators must also match. This is not the case; the arithmetically equivalent fields $K:=\Q(\sqrt[8]{-15})$ and $K':=\Q(\sqrt[8]{-240})$ have class numbers 8 and 4, respectively (you do not need to prove this). \begin{enumerate}[{\bf(a)}] \setcounter{enumi}{9} \item You showed in (b) that in arithmetically equivalent fields the multisets of inertia degrees above any prime $p$ must match (including at ramified primes); it is natural to ask whether the same is true of the ramification indices. Show that this is not the case by showing that the polynomials $x^8-97$ and $x^8-16\cdot 97$ (which you may assume define arithmetically equivalent fields $K$ and $K'$), do not have the same factorization pattern in $\Q_2[x]$. Conclude that the different ideals $\D_{K/\Q}$ and $\D_{K'/\Q}$ do not necessarily have the same factorization pattern, even though the discriminant ideals $D_{K/\Q}:=\N_{K/\Q}(\D_{K/\Q})$ and $D_{K'/\Q}:=\N_{K'/\Q}(\D_{K'/\Q})$ do. \end{enumerate} \noindent A useful reference for this problem if you get stuck is Perlis' paper \cite{perlis}. \subsection*{Problem 7. Survey (2 points)} Complete the following survey by rating each problem you attempted on a scale of 1 to~10 according to how interesting you found it (1 = ``mind-numbing," 10 = ``mind-blowing"), and how difficult you found it (1 = ``trivial," 10 = ``brutal"). Also estimate the amount of time you spent on each problem to the nearest half hour. \begin{center} \begin{tabular}{l|r|r|r|} & Interest & Difficulty & Time Spent\\\hline Problem 1 & & & \\\hline Problem 2 & & & \\\hline Problem 3 & & & \\\hline Problem 4 & & & \\\hline Problem 5 & & & \\\hline Problem 6 & & & \\\hline \end{tabular} \end{center} \noindent Please feel free to record any additional comments you have on the problem sets and the lectures, in particular, ways in which they might be improved. \medskip \noindent \textbf{Collaborators/Sources} \begin{thebibliography}{99} \bibitem{gassmann} F. Gassmann, \href{http://link.springer.com/article/10.1007/BF01283860}{\textit{Bemerkungen zu der vorstehenden Arbeit von Hurwitz}} (comments on \textit{\"Uber Beziehungen zwischen den Primidealen eines algebraischen K\"orpers und den Substitutionen seiner Gruppe}, by Hurwitz) Math. Z. \textbf{25} (1926), 655-665. \bibitem{perlis} R. Perlis, \href{http://www.sciencedirect.com/science/article/pii/0022314X77900701}{\textit{On the equation $\zeta_K(s)=\zeta_{K'(s)}$}}, J. Number Theory \textbf{9} (1977), 342--360. \bibitem{RZ} L. Ribes and P. Zalesskii, \href{http://link.springer.com/book/10.1007/978-3-642-01642-4}{\textit{Profinite groups}}, 2nd edition, Springer, 2010. \bibitem{serre73} J.-P. Serre, \href{https://link.springer.com/book/10.1007/978-1-4684-9884-4}{\textit{A course in arithmetic}}, Springer, 1973. \bibitem{voight} J. Voight, \href{https://link.springer.com/book/10.1007/978-3-030-56694-4}{\textit{Quaternion algebras}}, Springer, 2021. \end{thebibliography} \end{document}