Communications in Number Theory and Physics

Volume 6 (2012)

Number 2

Rigid local systems and a question of Wootters

Pages: 223 – 278

DOI: https://dx.doi.org/10.4310/CNTP.2012.v6.n2.a1

Author

Nicholas M. Katz (Department of Mathematics, Princeton University)

Abstract

Recently, we learned from Ron Evans of some fascinatingquestions raised by Wootters \cite{S-S-W}. These questions,which concern exponential sums, arose from hisinvestigations of a particular quantum state with specialproperties, where the underlying vector space is the spaceof functions on the finite field $\F_p:=\Z/p\Z$, $p$ aprime which is $3$ mod $4$. Due to our ignorance of theunderlying physics, we concentrate on the exponential sumsthemselves. In our approach, it costs us nothing to workover an arbitrary finite field $\F_q$ of oddcharacteristic. [Thus $\F_q$ is “the” finite field of $q$elements, $q$ a power of some odd prime $p$.] We alsointroduce a parameter $a \in \F_q^{\times}$. In the Wootterssetup, where $q=p$ is $3$ mod $4$, the parameter $a$ issimply $a= -1$. Ultimately, we end up proving identitiesamong exponential sums, but not at all in a straightforwardway; we need to invoke the theory of Kloosterman sheavesand their rigidity properties, as well as the fundamentalresults of \cite{De-Weil II} and \cite{BBD}. It would beinteresting to find direct proofs of these identities.

Published 19 October 2012