Darboux, Moser and Weinstein theorems for prequantum systems
Extracted source text · arXiv 2404.04988
\documentclass[11pt]{amsart}
\setlength{\topmargin}{-.25in}
\setlength{\textwidth}{6.5in}
\setlength{\textheight}{9.0in}
\renewcommand\baselinestretch{1.0}
\setlength{\evensidemargin}{0in}
\setlength{\oddsidemargin}{0in}
\usepackage{amscd,amssymb,palatino}
\usepackage[utf8]{inputenc}
\usepackage[english]{babel}
%\documentclass[openany]{book}
%\special{papersize=8.5in,11in}
\usepackage{caption}
\usepackage{etex}
\usepackage[leqno]{amsmath}%puts equation number on the left
\usepackage{amssymb}
\usepackage{color}
\usepackage{shadow}
\usepackage{epsfig}
\usepackage{epic}
\usepackage{eepic}
\usepackage{graphics}
\usepackage{graphicx}
\usepackage{psfrag}
%\usepackage{alltt}%%%necessary for {matlab} environment to work
%\usepackage{theorem}
\usepackage{calc}
%\usepackage{rotating}
%\usepackage{pspic}
%\usepackage{mycurves}
%\usepackage{multicol}
%\usepackage{multirow}
%\usepackage{sols_man}
%\usepackage{relsize}
%\usepackage{enumitem}
%\usepackage[T1]{fontenc}
%\usepackage{textcomp}
%\usepackage[abs]{overpic}
%\usepackage{pgfplots}
\usepackage{tikz}
\usepackage[dvipsnames,prologue,table]{pstricks}
%\usepackage{pst-grad}
\usepackage{pstcol}
%\usepackage{pst-ghsb}
%\usepackage{pst-slpe,pst-3d}
\usetikzlibrary{arrows,decorations,patterns,positioning,automata,shadows,fit,shapes,calc,decorations.markings,backgrounds,scopes,decorations.text}
%\renewcommand{\rmdefault}{ptm}
%\usepackage[scaled=0.92]{helvet}
%\usepackage{musixtex,harmony}% for \ViPa
%To remove headers/footers from blank page before first chapter
\usepackage{tipa}
\usepackage{fancyhdr}
%\pagestyle{fancy}
\fancyhf{} %reset any previously defined section headings
\fancyhead[RO]{\small\nouppercase\rightmark\quad\
{\large\sf\bfseries\hfill\thepage}}
\fancyhead[LE]{{\large\sf\bfseries\thepage\hfill}\
\quad\small\nouppercase\leftmark}
\renewcommand{\headrulewidth}{0pt} %no rule at top
\renewcommand{\footrulewidth}{0pt} %no rule at bottom
\fancypagestyle{plain}{\fancyhf{}
\fancyfoot[C]{{\large\sf\bfseries\thepage}}}
\setlength{\headsep}{0.18in}
%\makeatletter
%\renewcommand{\@pnumwidth}{1.75em}
%\renewcommand{\@tocrmarg}{2.75em}
%\makeatother
\usepackage{calc}
\setlength\textwidth{6in}
\setlength\textheight{9in}\setlength\oddsidemargin{(\paperwidth-\textwidth)/2 - 1in}
\setlength\evensidemargin{(\paperwidth-\textwidth)/2 - 1in}
\setlength\topmargin{(\paperheight-\textheight
-\headheight-\headsep-\footskip)/2 - 1in}
%\renewcommand{\underbar}{\mbox{\,\vrule height-0.8pt depth1.0ptwidth2pc\,}}
%\newcommand\1{\mbox{\,\vrule height-0.8pt depth1.0pt width2pc\,}}
\newcommand\1{\lower 9pt\hbox{\underbar{}}}
\numberwithin{equation}{section}
\newtheorem {Theorem} {Theorem}
\newtheorem {varTheorem} {Theorem}
%\newenvironment {Theorem'}
% {\begin{varTheorem}{\hspace{-3.5mm}}{\bf '}{\hspace{3.5mm}}}
% {\end{varTheorem}}
\newtheorem {RefTheorem}[equation] {Theorem}
\newtheorem {Assumption}[equation] {Assumption}
\newtheorem {Lemma}[equation] {Lemma}
\newtheorem {Proposition}[equation] {Proposition}
\newtheorem {Conjecture}[equation] {Conjecture}
\newtheorem {Corollary} {Corollary}
\newtheorem {corollary}[equation] {Corollary}
\newtheorem {refproposition} {Proposition}
\theoremstyle{definition}
\newtheorem {Definition}[equation]{Definition}
\newtheorem {Remark}[equation] {Remark}
\newtheorem {Example}[equation] {Example}
\def\cL{\mathcal L}
%Define rounded rectangle starting each chapter
%\setlength{\fboxsep}{6pt}% CLASHING WITH BLUE BOXES padding for \fbox
%\setlength{\fboxrule}{2pt} %Outer border for \fbox
\newcommand{\pr} {\smallskip\noindent{\bf Proof\,\,}}
\newenvironment{Proof} {\pr}{\hspace*{\fill}\qed\\}
%\newenvironment{pf} {\pr}{\hspace*{\fill}\qed\\}
\newenvironment{pf} {\begin{proof}}{\end{proof}}
\begin{document}
\title{Darboux, Moser and Weinstein theorems for prequantum systems}
\author{Eva Miranda}
\thanks{{Both authors are partially supported by the Spanish State Research Agency grant PID2019-103849GB-I00 of MICIU/AEI / 10.13039/501100011033. E. Miranda is supported by the Catalan Institution for Research and Advanced Studies via an ICREA Academia Prize 2021 and by the Alexander Von Humboldt foundation via a Friedrich Wilhelm Bessel Research Award. E. Miranda is also supported by the Spanish State
Research Agency, through the Severo Ochoa and Mar\'{\i}a de Maeztu Program for Centers and Units
of Excellence in R\&D (project CEX2020-001084-M) and partially supported by and by the AGAUR project 2021 SGR 00603. }}
\address{{Department of Mathematics} at Universitat Polit\`{e}cnica de Catalunya and Centre de Recerca Matem\`{a}tica-CRM}
\email{eva.miranda@upc.edu}
\author{Jonathan Weitsman}
\thanks{J. Weitsman was supported in part by a Simons Collaboration Grant \# 579801.}
\address{Department of Mathematics, Northeastern University, Boston, MA 02115}
\email{j.weitsman@neu.edu}
\thanks{\today}
\begin{abstract} We establish analogs of the Darboux, Moser and Weinstein theorems for prequantum systems. We show that two prequantum systems on a manifold with vanishing first cohomology, with symplectic forms defining the same cohomology class and homotopic to each other within that class, differ only by a symplectomorphism and a gauge transformation. As an application, we show that the Bohr-Sommerfeld quantization of prequantum system on a manifold with trivial first cohomology is independent of the choice of the connection.
\end{abstract}
\maketitle
%\thispagestyle{empty}
\section{Introduction}
The Darboux theorem establishes that there are no local invariants in symplectic geometry.
Namely, let $M$ be a smooth manifold and let $\omega, \omega^{\prime}$ be two symplectic forms on $M$, then:
\begin{Theorem}[Darboux \cite{darboux, darboux2}]
{For every point $p\in M$ there exists a neighbourhood $U$ of $p$ and an embedding $\Phi:U \rightarrow M$ isotopic to the inclusion and fixing $p$ such that
$$
\Phi^{*} \omega^{\prime}=\omega|_U \text {. }
$$}
\end{Theorem}
The global invariants of a symplectic manifold include the cohomology class defined by its symplectic form. This gives a complete classification of symplectic $2$-manifolds. In general, there is the following theorem due to Moser.
\begin{Theorem}[Moser \cite{moser}]
Let $M$ be a compact manifold endowed with two symplectic forms $\omega$ and $\omega^{\prime}$. Assume that $[\omega]=\left[\omega^{\prime}\right]$, and that there exists a path $\omega_t$ of symplectic forms such that $[\omega_t ]= [\omega_0]$ for all $t,$ and with $\omega_0=\omega$ and
$\omega_1=\omega^{\prime}$. Then, there exists a diffeomorphism isotopic to the identity
$\Phi:M \rightarrow M$ such that $\Phi^{*} \omega^{\prime}=\omega$.
\end{Theorem}
This proof based on Moser's method can be adapted when symmetries are present.
Under the conditions above when a compact group acts on $M$ preserving $\omega$ and $\omega^{\prime},$ and where the path $\omega_t$ also consists of invariant forms, we have the following result.
\begin{Theorem}[Weinstein \cite{weinstein}]
There exists an equivariant diffeomorphism $\Phi:M \rightarrow M$ such that $\Phi^{*} \omega^{\prime}=\omega$.
\end{Theorem}
Recall that if $(M, \omega)$ is a symplectic manifold, $\pi: L \rightarrow M$ is a (complex) line bundle, $\nabla$ is a connection on $L$, and $\operatorname{curv}(\nabla)=\omega$, the quadruple $(M, \omega, L, \nabla)$ is a \emph{prequantum system}.
The main purpose of this article is to generalize these three theorems to prequantum systems.
\begin{Theorem} \label{maintheorem1}
Suppose $(M, \omega, L, \nabla)$ and $(M, \omega^{\prime}, L, \nabla^{\prime})$
are prequantum systems. Then for every point $p\in M$ there exists a neighbourhood $U$ of $p,$ an embedding $\Phi:U \rightarrow M$ isotopic to the inclusion and fixing $p,$ and a function $\phi \in C^{\infty}(U, \mathbb C)$ such that\footnote{{For the definition of the pullback of a connection under a smooth map, see e.g. \cite{ms}, p. 292, Lemma 3.}}
$$
\begin{aligned}
& \omega|_U=\Phi^{*} \omega^{\prime} \quad (\Phi^{*} L) = L|_U \\
& \nabla s=(\Phi^{*} \nabla^{\prime})s+ d \phi\otimes s, \, \text{for any} \, s\in \Gamma(L|_U).\end{aligned}
$$
If $L$ is a hermitian line bundle and $\nabla$ is a hermitian connection, we may take $\phi \in \mathcal{C}^{\infty}(M, i\mathbb{R}) $.
\end{Theorem}
For compact manifolds, we have
\begin{Theorem} \label{maintheorem2}
Suppose $(M, \omega, L, \nabla)$ and $(M, \omega^{\prime}, L, \nabla^{\prime})$
are prequantum systems with $M$ a compact manifold. Assume that $[\omega]=\left[{\omega}^{\prime}\right]$, and that there exists a path $\omega_t$ of symplectic forms with $[\omega_t] = [\omega_0]$ for all $t,$ such that $\omega_0=\omega$ and
$\omega_1=\omega^{\prime}$. Assume also that $H^1(M,{\mathbb R})=0.$
Then there exists a a diffeomorphism $\Phi: M\rightarrow M$ isotopic to the identity and a function $\phi \in \mathcal{C}^{\infty}(M ,\mathbb{C})$ such that
$$
\begin{aligned}
& \omega=\Phi^{*} \omega^{\prime} \quad \Phi^{*} L = L \\
& \nabla s=(\Phi^{*} \nabla^{\prime})s+ d \phi\otimes s, \, \text{for any} \, s\in \Gamma(L).
\end{aligned}
$$
If $L$ is a hermitian line bundle and $\nabla$ is a hermitian connection, we may take $\phi \in \mathcal{C}^{\infty}(M, i\mathbb{R}) $.
\end{Theorem}
When there are additional symmetries the following result holds.
\begin{Theorem}\label{maintheorem3} Suppose $G$ is a compact Lie group, and suppose $M$ is a compact $G$-space. Suppose that $(M, \omega, L, \nabla)$ and $(M, \omega^{\prime}, L, \nabla^{\prime})$ are $G$-invariant prequantum systems. Assume that $[\omega]=\left[\omega^{\prime}\right]$, and that there exists a path $\omega_t$ of $G$-invariant symplectic forms with $[\omega_t] = [\omega_0]$ for all $t,$ such that $\omega_0=\omega$ and
$\omega_1=\omega^{\prime}.$ Assume also that $H^1(M,{\mathbb R})=0$.
Then there exists a $G$-equivariant diffeomorphism $\Phi: M\rightarrow M$ isotopic to the identity and a $G$-invariant function $\phi \in \mathcal{C}^{\infty}(M ,\mathbb{C})$ such that
$$
\begin{aligned}
& \omega=\Phi^{*} \omega^{\prime} \quad \Phi^{*} L = L \\
& \nabla s=(\Phi^{*} \nabla^{\prime})s+ d \phi\otimes s, \, \text{for any} \, s\in \Gamma(L).
\end{aligned}
$$
If $L$ is a hermitian line bundle and $\nabla$ is a hermitian connection, we may take $\phi \in \mathcal{C}^{\infty}(M, i\mathbb{R}) $.
\end{Theorem}
\begin{Remark}
The function $\phi$ is called a \emph{gauge transformation}. We have thus proven that two prequantum systems, with associated symplectic forms lying in the same cohomology class and homotopic to each other, differ only by a symplectomorphism and a gauge transformation. Observe that it is not possible to dispense with the gauge transformation: Suppose $\nabla$ is the trivial connection on the trivialized bundle $L = M \times {\mathbb C},$ and $\nabla^\prime$ is some other connection on $L.$ Then for any section $s \in \Gamma(M,L),$ we have $\nabla^\prime s = \nabla s +d\phi\otimes s,$ where $\phi$ is somewhere non-constant. And no symplectomorphism will transform $\phi$ into an everywhere constant function (since diffeomorphisms take constant functions to constant functions).
\end{Remark}
As an application of this theorem, we prove that for manifolds with trivial first cohomology, the Bohr-Sommerfeld quantization of a prequantum system $(M,\omega,L,\nabla)$ does not depend on the choice of the connection $\nabla;$ in other words if $(M,\omega,L,\nabla^\prime)$ is another prequantum system associated with the same manifold, line bundle, and symplectic form, its Bohr-Sommerfeld quantization coincides with that of $(M,\omega,L,\nabla).$
If $H^{1}(M) \neq 0$ a similar result holds if the holonomy representations of $\nabla$ and $\nabla^{\prime}$ coincide.
{\bf Acknowledgement:} We would like to thank Peter Crooks for helpful discussions on this paper.
\section{Proof of the main theorems}
We first prove Theorem \ref{maintheorem1}.
\begin{proof}
By the Darboux theorem, we know that there exists a neighborhood $U$ of $p,$ which we may take to be diffeomorphic to an open ball, and {an embedding $\phi: U \rightarrow M$
isotopic to the inclusion preserving $p$ such that
$$
{\Phi^{^*} \omega^{\prime}=\omega|_U .}
$$
Since} $\Phi$ is isotopic to the inclusion we also have
$$
\Phi^{*} L=L|_U
$$
It remains to compare the two connections and consider the difference
{
$$
\nabla s-(\Phi^{*} \nabla^{\prime}) s \, \text{for any} \, s\in\Gamma(L|_U).
$$}
Since the space of connections on a line bundle on $U$ is an affine space modeled on $\Omega^{1}\left(U, \mathbb C\right) ,$ we know that
$$
\nabla s-(\Phi^{*} \nabla^{\prime})s= \xi\otimes s
$$
for some $\xi \in \Omega^{1}\left(U, \mathbb C\right)$.
Since $\operatorname{curv}\nabla=\operatorname{curv}\Phi^{*} \left(\nabla^{\prime}\right)$, it follows
that $d \xi=0$. Since $H^{1}(U,{\mathbb R})=0$,
$
\xi=d \phi$ for some $\phi \in \mathcal{C}^{\infty}(U,\mathbb{C}).
$
If the line bundle $L$ is hermitian and the connection is unitary, we may repeat this argument to show that we may take $\phi \in \mathcal{C}^{\infty}(U,i\mathbb{R}).$
\end{proof}
We next prove Theorem \ref{maintheorem2}.
\begin{proof} The existence of the diffeomorphism is guaranteed by Moser's theorem.
We again have $\Phi^{*} L=L$, and again there exists $\xi \in \Omega^{1}\left(M, \mathbb C\right)$ so that
$$
\nabla s-(\Phi^{*} \nabla^{\prime})s= \xi\otimes s \, \text{for any} \, s\in\Gamma(L).
$$
Again $d \xi=0$, so since $H^1(M,{\mathbb R}) = 0,$ $\xi=d \phi$ for some $\phi,$ as needed. If the line bundle $L$ is hermitian and the connection is unitary, we may repeat the argument to show that we may take $\phi \in \mathcal{C}^{\infty}(M,i\mathbb{R}).$
\end{proof}
We now proceed to the proof of Theorem \ref{maintheorem3}.
\begin{proof} The existence of a $G$-equivariant diffeomorphism $\Phi$ is given by Weinstein's theorem.
Since $H^1(M,{\mathbb R})=0,$ again there exists a function $\phi$ such that
\begin{equation}\label{eqn1}
\nabla s-(\Phi^{*} \nabla^{\prime})s= d\phi\otimes s ~ \text{for any} \, s\in\Gamma(L).
\end{equation}
Let $\bar{\phi}$ be the average of $\phi$ over $G$ using the Haar measure $\mu$, i.e., $$\bar{\phi}:=\int_G \alpha_g^*\phi~ d\mu(g),$$ where we have denoted by $\alpha_g$ the diffeomorphism giving the action of $g \in G$ on $M.$
Then since $\nabla$ and $\nabla^{\prime}$ are $G$-invariant, we obtain
$$
\nabla s-(\Phi^{*} \nabla^{\prime})s= d \bar{\phi}\otimes s
$$
by averaging both sides of equation (\ref{eqn1}).
Again, if the line bundle $L$ is hermitian and the connection is unitary, we may repeat this argument to show that we may take $\phi \in \mathcal{C}^{\infty}(M,i\mathbb{R})^G.$
\end{proof}
\begin{Remark}\label{remark}(The case where $H^1(M,{\mathbb R}) \neq 0$)
Suppose we are given two connections $\nabla$ and $\nabla^{\prime}$ on $L$ with the same curvature. Then for any section $s \in \Gamma(M,L),$ we have $\nabla s=\nabla^{\prime} s + \alpha \otimes s$ for some $\alpha \in \Omega^1(M, {\mathbb C})$ with $d \alpha = 0$.
If $H^1(M,{\mathbb C}) \neq 0,$ we cannot conclude that $\alpha = d\phi$ for some function $\phi.$ However, if $\nabla$ and $\nabla^{\prime}$ also have the same holonomy representation, it follows that the holonomy representation of $\alpha,$ thought as a connection on the trivialized bundle $L\otimes L^*=\mathbb C$, is trivial.
That means that for every oriented curve $C$ in $M$, $\int_C \alpha = 2 \pi n$, where $n \in \mathbb Z$. This does not imply that $\alpha$ is exact; but there exists a map $t: M \to S^1$ with $\alpha = t^*( d \theta)$, where $d\theta$ is the translation invariant one form on $S^1$.\footnote{Morally $\alpha = t^{-1} dt,$ where $t : M \to S^1$ is a (possibly homotopically nontrivial) element of the gauge group ${\rm Map}(M,S^1).$}
\end{Remark}
\section{Application to Quantization in a Real Polarization}
Let $(M, \omega, L, \nabla)$ be a prequantum system. Recall that a Lagrangian submanifold $\Lambda$
is {\em integral} if $\left.(L, \nabla)\right|_{\Lambda}$ is trivial as a bundle with connection.
\begin{Corollary}\label{cor1} Let $M$ be a compact manifold with $H^1(M,{\mathbb R}) = 0$ and let $(M, \omega, L, \nabla)$ and $(M, \omega, L, \nabla^{\prime})$ be two prequantum systems.
A Lagrangian submanifold $\Lambda\subset M$ is integral for $(M, \omega, L, \nabla)$ if and only if it is integral for $(M, \omega, L, \nabla^{\prime})$.
\end{Corollary}
In particular, suppose we are given a real polarization of $M$, that is, a foliation by Lagrangian submanifolds. Consider from now on prequantum systems $(M, \omega, L, \nabla)$ where where $L$ is hermitian and the connection $\nabla$ is unitary.
The Bohr-Sommerfeld quantization of $(M, \omega, L, \nabla)$ associated to this polarization is given by the vector space generated by the integral leaves of the foliation, or equivalently by the covariant constant sections of $(L,\nabla)$ on those leaves.
{In \cite{guilleminsternberg} Guillemin and Sternberg consider the case where the foliation is given by a map $\pi : M \to B,$ where $B$ is a simply connected subset of $ {\mathbb R}^n$ and the map $F: M\longrightarrow B$ is the moment map for an integrable system away from the singularities of the foliation. The Bohr-Sommerfeld quantization is then determined by the integral points of the image of the moment map $F$.}
Then Corollary 1 implies
\begin{Corollary}\label{cor2}
Let $(M, \omega)$ be a symplectic manifold with $H^1(M,{\mathbb R})=0.$ Let $L$ be a complex line bundle on $M$ with $c_1(L) = [\omega].$ Choose a connection $\nabla$ on $L$ so that $(M,\omega,L,\nabla)$ is a prequantum system. Suppose we are given a real polarization of $M.$ Then the quantization of $(M, \omega, L, \nabla)$ with this real polarization is independent of the choice of $\nabla.$
\end{Corollary}
\begin{Remark} Note that in the case where the foliation is given as in \cite{guilleminsternberg} by an integrable system, the generic leaves are tori, on which $(L,\nabla)$ may have complicated holonomy representations. We have shown that the global condition $H^1(M,{\mathbb R}) = 0 $ guarantees these holonomy representations are independent of $\nabla.$\end{Remark}
\begin{Remark} For complex polarizations, an analog of Corollary \ref{cor2} holds, since the index of the Dolbeault $\bar\partial$ operator on $(M, L)$ depends only on $M$ and $c_{1}(L)=[\omega]$, not on the connection; this is due to the Riemann-Roch theorem
\begin{equation}\label{eqn:index}
\operatorname{ind}(\bar\partial)=\int_{M} Td(T M) e^{[\omega]}.
\end{equation}\end{Remark}
Note that the right-hand side of equation (\ref{eqn:index}) depends only on the class $[\omega]$ and not on $\nabla$.
\begin{Remark}
In the case where $H^1(M,{\mathbb R}) \neq 0,$ any two connections $\nabla$ and $\nabla^{\prime}$ with the same curvature still give the same quantization if, in addition, the trivial bundle $L \otimes L^*$, equipped with the flat connection arising from $\nabla$ and $\nabla^{\prime}$, is trivial as a bundle with connection.
\end{Remark}
\begin{thebibliography}{ABCD}
\bibitem{darboux} G. Darboux, \emph{Sur le problème de Pfaff}, Bull. sci. math. astron. (2) 6 (1882), 14-36.
\bibitem{darboux2} G. Darboux, \emph{Sur le problème de Pfaff}, Bull. sci. math. astron. (2) 6 (1882), 49-62.
\bibitem{guilleminsternberg} V. Guillemin and S. Sternberg,
Journal of Functional Analysis
Volume 52, Issue 1,1983, 106-128.
\bibitem{kobayashi}
S. Kobayashi, Differential Geometry of Complex Vector Bundles. Princeton University Press, 1987.
\bibitem{ms} J. Milnor and J. Stasheff, Characteristic Classes. Princeton University Press, 1974.
\bibitem{moser}
J. Moser, \emph{On the volume elements on a manifold.}
Trans. Amer. Math. Soc., 120 (1965), pp. 286-294.
\bibitem{weinstein} A. Weinstein,
\emph{Symplectic manifolds and their Lagrangian submanifolds},
Advances in Mathematics,
Volume 6, Issue 3,
1971,
Pages 329-346.
\end{thebibliography}
\end{document}