arXiv subjects

David Goss

Publications and source records attributed to David Goss.

13 recordsLinked to original sources

arXiv2016

Digit permutations revisited

We discuss here characteristic $p$ $L$-series as well as the group $S_{(q)}$ which appears to act as symmetries of these functions. We explain various actions of $S_{(q)}$ that arise naturally in the theory as well as extensions of these actions. In general such extensions appear to be highly arbitrary but in the case where the zeroes are unramified, the extension is unique (and it is reasonable to expect it is unique only in this case). Having unramified zeroes is the best one could hope for in finite characteristic and appears to be an avatar of the Riemann hypothesis in this setting; see Section [8] for a more detailed discussion.

math.NT↗

arXiv2014

Polynomials of Binomial Type and Lucas' Theorem

We present various constructions of sequences of polynomials satisfying the Binomial Theorem in finite characteristic based on the theory of additive polynomials. Various actions on these constructions are also presented. It is an open question whether we then have accounted for all sequences in finite characteristic which satisfy the Binomial Theorem.

math.NT↗

arXiv2013

A construction of $\mathfrak v$-adic modular forms

The classical theory of $p$-adic (elliptic) modular forms arose in the 1970's from the work of J.-P.\ Serre \cite{se1} who took $p$-adic limits of the $q$-expansions of these forms. It was soon expanded by N.\ Katz \cite{ka1} with a more functorial approach. Since then the theory has grown in a variety of directions. In the late 1970's, the theory of modular forms associated to Drinfeld modules was born in analogy with elliptic modular forms \cite{go1}, \cite{go2}. The associated expansions at $\infty$ are quite complicated and no obvious limits at finite primes ${\mathfrak v}$ were apparent. Recently, however, there has been progress in the $\mathfrak v$-adic theory, \cite{vi1}. Also recently, A.\ Petrov \cite{pe1}, building on previous work of \cite{lo1}, showed that there is an intermediate expansion at $\infty$ called the "$A$-expansion," and he constructed families of cusp forms with such expansions. It is our purpose in this note to show that Petrov's results also lead to interesting ${\mathfrak v}$-adic cusp forms à la Serre. Moreover the existence of these forms allows us to readily conclude a mysterious decomposition of the associated Hecke action.

math.NT↗

arXiv2012

A local field approach to the Riemann Hypothesis

Since the seminal work of Wan, Poonen, and Sheats in the 1990's, we have been searching for the correct general statement of the Riemann Hypothesis ("RH") which appears implicit in their results. Recently, upon viewing the extension $\C/\R$ in light of results derived for the Carlitz module, we were led to view the RH as a statement about ramification which we explore in this short work. We shall see that, combined with some ideas flowing from the proofs of Wan, Poonen, and Sheats, this ramification idea has a good deal of explanatory power in finite characteristic. Indeed, unramified extensions of nonarchimedean local fields are cyclotomic in nature and this fits perfectly with the best possible extension of the result of Wan and Sheats. The notion that the zeroes "lie on a line" seems to be the beginning of the story in finite characteristic, and we show how this fits with having the zeroes be unramified.

math.NT↗

arXiv2011

On the $L$-series of F. Pellarin

The calculation, by L.\ Euler, of the values at positive even integers of the Riemann zeta function, in terms of powers of $π$ and rational numbers, was a watershed event in the history of number theory and classical analysis. Since then many important analogs involving $L$-values and periods have been obtained. In analysis in finite characteristic, a version of Euler's result was given by L.\ Carlitz \cite{ca2} in the 1930's which involved the period of a rank 1 Drinfeld module (the Carlitz module) in place of $π$. In a very original work \cite{pe2}, F.\ Pellarin has quite recently established a "deformation" of Carlitz's result involving certain $L$-series and the deformation of the Carlitz period given in \cite{at1}. Pellarin works only with the values of this $L$-series at positive integral points. We show here how the techniques of \cite{go1} also allow these new $L$-series to be analytically continued -- with associated trivial zeroes -- and interpolated at finite primes.

math.NT↗

arXiv2011

The Ongoing Binomial Revolution

The Binomial Theorem has long been essential in mathematics. In one form or another it was known to the ancients and, in the hands of Leibniz, Newton, Euler, Galois, and others, it became an essential tool in both algebra and analysis. Indeed, Newton early on developed certain binomial series (see Section \ref{newton}) which played a role in his subsequent work on the calculus. From the work of Leibniz, Galois, Frobenius, and many others, we know of its essential role in algebra. In this paper we rapidly trace the history of the Binomial Theorem, binomial series, and binomial coefficients, with emphasis on their decisive role in function field arithmetic. We also explain conversely how function field arithmetic is now leading to new results in the binomial theory via insights into characteristic $p$ $L$-series.

math.NT↗

arXiv2010

$ζ$-phenomenology

It is well known that Euler experimentally discovered the functional equation of the Riemann zeta function. Indeed he detected the fundamental $s\mapsto 1-s$ invariance of $ζ(s)$ by looking only at special values. In particular, via this functional equation, the permutation group on two letters, $S_2\simeq\Z/(2)$, is realized as a group of symmetries of $ζ(s)$. In this paper, we use the theory of special-values of our characteristic $p$ zeta functions to experimentally detect a natural symmetry group $S_{(q)}$ for these functions of cardinality ${\mathfrak c}=2^{\aleph_0}$ (where $\mathfrak c$ is the cardinality of the continuum); $S_{(q)}$ is a realization of the permutation group on $\{0,1,2...\}$ as homeomorphisms of $\Zp$ stabilizing both the nonpositive and nonnegative integers. We present a number of distinct instances in which $S_{(q)}$ acts (or appears to act) as symmetries of our functions. In particular, we present a natural, but highly mysterious, action of $S_{(q)}$ on a large subset of the domain of our functions that appears to stabilize zeta-zeroes. As of this writing, we do not yet know an overarching formalism that unifies these examples; however, it would seem that this formalism will involve an interplay between the 1-unit group $U_1$ -- playing the role of a "gauge group" -- and $S_{(q)}$. Furthermore, we show that $S_{(q)}$ may be naturally realized as an automorphism group of the convolution algebras of characteristic $p$ valued measures.

math.NT↗

arXiv2006

Zeroes of $L$-series in characteristic $p$

In the classical theory of $L$-series, the exact order (of zero) at a trivial zero is easily computed via the functional equation. In the characteristic $p$ theory, it has long been known that a functional equation of classical $s\mapsto 1-s$ type could not exist. In fact, there exist trivial zeroes whose order of zero is ``too high;'' we call such trivial zeroes ``non-classical.'' This class of trivial zeroes was originally studied by Dinesh Thakur \cite{th2} and quite recently, Javier Diaz-Vargas \cite{dv2}. In the examples computed it was found that these non-classical trivial zeroes were correlated with integers having {\it bounded} sum of $p$-adic coefficients. In this paper we present a general conjecture along these lines and explain how this conjecture fits in with previous work on the zeroes of such characteristic $p$ functions. In particular, a solution to this conjecture might entail finding the ``correct'' functional equations in finite characteristic.

math.NT↗

arXiv2004

Applications of non-Archimedean integration to the $L$-series of $τ$-sheaves

Let $\underline{\mathcal F}$ be a $τ$-sheaf. Building on previous work of Drinfeld, Anderson, Taguchi, and Wan, Böckle and Pink \cite{bp1} develop a cohomology theory for $\underline{\mathcal F}$. In \cite{boc1} Böckle uses this theory to establish the analytic continuation of the $L$-series associated to $\underline{\mathcal F}$ (which is a characteristic $p$ valued ``Dirichlet series'') {\em and} the logarithmic growth of the degrees of its special polynomials. In this paper we shall show that this logarithmic growth is all that is needed to analytically continue the original $L$-series as well as {\em all} associated partial $L$-series. Moreover, we show that the degrees of the special polynomials attached to the partial $L$-series also grow logarithmically. Our tools are Böckle's original results, non-Archimedean integration, and the very strong estimates of Y. Amice \cite{am1}. Along the way, we define certain natural modules associated with non-Archimedean measures (in the characteristic 0 case as well as in characteristic $p$).

math.NT↗

arXiv2003

Can a Drinfeld module be modular?

Let $k$ be a global function field with field of constants $\Fr$ and let $\infty$ be a fixed place of $k$. In his habilitation thesis \cite{boc2}, Gebhard Böckle attaches abelian Galois representations to characteristic $p$ valued cusp eigenforms and double cusp eigenforms \cite{go1} such that Hecke eigenvalues correspond to the image of Frobenius elements. In the case where $k=\Fr(T)$ and $\infty$ corresponds to the pole of $T$, it then becomes reasonable to ask whether rank 1 Drinfeld modules over $k$ are themselves ``modular'' in that their Galois representations arise from a cusp or double cusp form. This paper gives an introduction to \cite{boc2} with an emphasis on modularity and closes with some specific questions raised by Böckle's work.

math.NT↗

arXiv2001

The impact of the infinite primes on the Riemann hypothesis for characteristic p L-series

In math.NT/9907019 we proposed an analog of the classical Riemann hypothesis for characteristic p valued L-series based on the work of Wan, Diaz-Vargas, Thakur, Poonen, and Sheats for the zeta function $ζ_{\Fr[θ]}(s)$. During the writing of math.NT/9907019, we made two assumptions that have subsequently proved to be incorrect. The first assumption is that we can ignore the trivial zeroes of characteristic p L-series in formulating our conjectures. Instead, we show here how the trivial zeroes influence nearby zeroes and so lead to counter-examples of the original Riemann hypothesis analog. We then sketch an approach to handling such ``near-trivial'' zeroes via Hensel's and Krasner's Lemmas (whereas classically one uses Gamma-factors). Moreover, we show that $ζ_{\Fr[θ]}(s)$ is not representative of general L-series as, surprisingly, all its zeroes are near-trivial, much as the Artin-Weil zeta-function of $\mathbb{P}^1/\Fr$ is not representative of general complex L-functions of curves over finite fields. Consequently, the ``critical zeroes'' (= all zeroes not effected by the trivial zeroes) of characteristic p L-series now appear to be quite mysterious. The second assumption made while writing math.NT/9907019 is that certain Taylor expansions of classical L-series of number fields would exhibit complicated behavior with respect to their zeroes. We present a simple argument that this is not so, and, at the same time, give a characterization of functional equations.

math.NT↗

arXiv1999

A Riemann Hypothesis for characteristic p L-functions

We propose analogs of the classical Generalized Riemann Hypothesis and the Generalized Simplicity Conjecture for the characteristic p L-series associated to function fields over a finite field. These analogs are based on the use of absolute values. Further we use absolute values to give similar reformulations of the classical conjectures (with, perhaps, finitely many exceptional zeroes). We show how both sets of conjectures behave in remarkably similar ways.

math.NT↗

arXiv1997

Separability, multi-valued operators, and zeroes of L-functions

Let $\k$ be a global function field in 1-variable over a finite extension of $\Fp$, $p$ prime, $\infty$ a fixed place of $\k$, and $\A$ the ring of functions of $\k$ regular outside of $\infty$. Let $E$ be a Drinfeld module or $T$-module. Then, as in \cite{go1}, one can construct associated characteristic $p$ $L$-functions based on the classical model of abelian varieties {\it once} certain auxiliary choices are made. Our purpose in this paper is to show how the well-known concept of ``maximal separable (over the completion $\k_\infty$) subfield'' allows one to construct from such $L$-functions certain separable extensions which are independent of these choices. These fields will then depend only on the isogeny class of the original $T$-module or Drinfeld module and $y\in \Zp$, and should presumably be describable in these terms. Moreover, they give a very useful framework in which to view the ``Riemann hypothesis'' evidence of \cite{w1}, \cite{dv1}, \cite{sh1}. We also establish that an element which is {\it separably} algebraic over $\k_\infty$ can be realized as a ``multi-valued operator'' on general $T$-modules. This is very similar to realizing 1/2 as the multi-valued operator $x\mapsto \sqrt{x}$ on $\C^\ast$. Simple examples show that this result is false for non-separable elements. This result may eventually allow a ``two $T$'s'' interpretation of the above extensions in terms of multi-valued operators on $E$ and certain tensor twists.

math.NT↗