Searcharxiv⌕ Search

arXiv subjects

Evan M. O'Dorney

Publications and source records attributed to Evan M. O'Dorney.

At least 19 recordsLinked to original sources

Galois groups of integer polynomials with prescribed coefficients

Fix a degree $n$, a nonzero integer constant term, and some further coefficients of a monic integer polynomial, leaving $s$ coefficients to vary in $[-H,H]$. We first prove that, when only the leading and constant coefficients are fixed, the number of polynomials whose Galois group is not $S_n$ is of order $H^{n-2}$ for every $n\ge4$. We then prove a half-dimensional extension: for $n\ge6$ the corresponding exceptional count is of order $H^{s-1}$ whenever $s>n/2+1$, provided the coefficient slice meets the higher-multiplicity loci with a specified dimension bound. This geometric condition holds for every prescribed value if the free coefficients form an initial or final block, and for a nonempty Zariski-open set of prescribed values for arbitrary coefficient positions. It also holds for every value of one additional prescribed interior coefficient. Thus, in degrees seven and eight, any one interior coefficient can be fixed in addition to the leading and constant coefficients. The proof combines a fixed-norm version of Bhargava's primitive-group sieve with sparse Hermite interpolation. A separate imprimitive estimate, using a minimal intermediate field, gives $O(H^{β(n)})$ outside rational partial-product hypersurfaces, where $β(n)=\ell+n/\ell-2\le n/2$ for composite $n$ with smallest prime divisor $\ell$, and has no logarithmic loss. The fixed-constant quartic and quintic endpoints require separate local arguments: exceptional Fourier directions for $A_4$ quartics and a quadratic Gauss sum for quintics. We isolate further even-degree endpoint results, explain the distinct cubic obstruction, and conclude with conditional improvements under the upper-bound form of Malle's conjecture.

math.NT↗

Galois groups of twisted reciprocal polynomials: a uniform asymptotic

We study the Galois group $G_f$ of a random polynomial $f$ in the family of polynomials of degree $2n$ satisfying the twisted reciprocal relation $f(x) = x^{2n}/b^n \cdot f(b/x)$. We use a Euclidean height adapted to this relation. Our main result is an asymptotic theorem of van der Waerden--Bhargava type: for fixed $b \neq 0$ and $n \geq 4$, the number of polynomials of height at most $H$ whose Galois group is not the full hyperoctahedral group $S_2 \wr S_n$ is an explicit constant times $H^n\log H$, with an error of order $H^n$. We determine the dependence of the leading constant on $b$; the leading-order group $G_1$ is of index $2$. This paper is a sequel to a recent paper by Anderson, Bertelli, and the author addressing reciprocal polynomials (i.e. the case $b = 1$).

math.NT↗

Updating Tom Lehrer's "The Elements"

We analyze the poetico-comedic criteria by which Tom Lehrer arranged the chemical elements in his 1959 hit "The Elements," a parody to the tune of Gilbert and Sullivan's "The Modern Major-General" (1879). We compose a new version to the same tune incorporating all 118 currently known elements and fulfilling, so far as possible, the same artistic aims.

math.HO↗

Galois groups of reciprocal polynomials and the van der Waerden-Bhargava theorem

We study the Galois groups $G_f$ of degree $2n$ reciprocal (a.k.a. palindromic) polynomials $f$ of height at most $H$, finding that $G_f$ falls short of the maximal possible group $S_2 \wr S_n$ for a proportion of all $f$ bounded above and below by constant multiples of $H^{-1} \log H$, whether or not $f$ is required to be monic. This answers a 1998 question of Davis-Duke-Sun and extends Bhargava's 2023 resolution of van der Waerden's 1936 conjecture on the corresponding question for general polynomials. Unlike in that setting, the dominant contribution comes not from reducible polynomials but from those $f$ for which $(-1)^n f(1) f(-1)$ is a square, causing $G_f$ to lie in an index-$2$ subgroup.

math.NT↗

Reflection theorems of Ohno-Nakagawa type for quartic rings and pairs of $n$-ary quadratic forms

We prove a reflection theorem, conjectured by Nakagawa and Ohno, for the number of quartic rings, or pairs of ternary quadratic forms, with a given cubic resolvent. Over $\mathbb{Z}$, our results are unconditional; we also allow the base to be the ring of integers of a general number field, conditional on some algebraic identities that are Monte Carlo verified. We also establish a reflection theorem for quartic $11111$-forms and $48441$-forms that relates them to the number of $3\times 3$ symmetric matrices with given characteristic polynomial. Along the way, we find elegant new results on Igusa zeta functions of conics and the average value of a quadratic character over a box in a local field. We conjecture that a reflection theorem holds for pairs of $n$-ary quadratic forms for any odd $n$, and we prove this for odd cubefree discriminant. This furnishes a more satisfactory answer for a question raised by Cohen, Diaz y Diaz, and Olivier, namely whether there exist an infinite family of reflection theorems of Ohno-Nakagawa type.

math.NT↗

Composed varieties and reflection theorems of Ohno-Nakagawa type

The Ohno-Nakagawa (O-N) reflection theorem is an unexpectedly simple identity relating the number of $\mathrm{GL}_2 \mathbb{Z}$-classes of binary cubic forms (equivalently, cubic rings) of two different discriminants $D$, $-27D$; it generalizes cubic reciprocity and the Scholz reflection theorem. In this paper, we present a new approach to proving and generalizing this theorem using Fourier analysis on the adelic cohomology $H^1(\mathbb{A}_K, M)$ of a finite Galois module, modeled after the celebrated Fourier analysis on $\mathbb{A}_K$ used in Tate's thesis. This method produces reflection theorems of O-N type from identities of lattice counts over local fields and yields new reflection theorems of O-N type for cubic forms and rings over arbitrary number fields, quartic forms and rings, and also quadratic forms counting by a peculiar invariant, the \emph{superdiscriminant} $a(b^2 - 4ac)$. O-N-type theorems are valuable both for their intrinsic beauty and for analytic applications, which include discriminant-reducing identities, symmetrization of functional equations, and subconvexity for Shintani zeta functions.

math.NT↗

Even-carry polynomials and cohomology of line bundles on the incidence correspondence in positive characteristic

We consider the cohomology groups of line bundles $\mathcal{L}$ on the \emph{incidence correspondence}, that is, a general hypersurface $X \subset \mathbb{P}^{n-1} \times \mathbb{P}^{n-1}$ of degrees $(1,1)$. Whereas the characteristic $0$ situation is completely understood, the cohomology in characteristic $p$ depends in a mysterious way on the base-$p$ digits of the degrees $(d, e)$ of $\mathcal{L}$. Gao and Raicu (following Linyuan Liu) prove a recursive description of the cohomology for $n = 3$, which relates to Nim polynomials when $p = 2$. In this paper, we devise a suitable generalization of Nim polynomials, which we call \emph{even-carry polynomials,} by which we can solve the recurrence of Liu--Gao--Raicu to yield an explicit formula for the cohomology for $n = 3$ and general $p$. We also make some conjectures on the general form of the cohomology for general $n$ and $p$, for which a recurrence relation was recently derived by Kyomuhangi--Marangone--Raicu--Reed.

math.AG↗

Galois groups of reciprocal polynomials II: Twisted reciprocal polynomials

We study the Galois group $G_f$ of a random polynomial $f$ of height at most $H$ in the family of polynomials of degree $2n$ satisfying the twisted reciprocal relation $f(x) = x^{2n}/b^n \cdot f(b/x)$, which arise in a wide variety of applications. Our main result is a theorem of van der Waerden-Bhargava type: the probability that $G_f$ is not the full hyperoctahedral group $S_2 \wr S_n$ is $Θ(H^{-1}\log H)$, independent of $b$, with the leading-order group $G_1$ being of index $2$. This paper is a companion to a recent paper by the authors and Bertelli addressing reciprocal polynomials (i.e. the case $b = 1$).

math.NT↗

The structure of the double discriminant

For a polynomial $f(x) = \sum_{i=0}^n a_i x^i$, we study the double discriminant $DD_{n,k} = \operatorname{disc}_{a_k} \operatorname{disc}_x f(x)$. This object has been well studied in algebraic geometry, but has been brought to recent prominence in number theory by its key role in the proof of the Bhargava--van der Waerden theorem. We bridge the knowledge gap for this object by proving an explicit factorization: $DD_{n,k}$ is the product of a square, a cube, and possibly a linear monomial. Our proof is entirely algebraic. We also investigate other aspects of this factorization.

math.NT↗

Galois groups of random integer matrices

We study the number $M_n(T)$ be the number of integer $n\times n$ matrices $A$ with entries bounded in absolute value by $T$ such that the Galois group of characteristic polynomial of $A$ is not the full symmetric group $S_n$. One knows $M_n(T) \gg T^{n^2 - n + 1} \log T$, which we conjecture is sharp. We first use the large sieve to get $M_n(T) \ll T^{n^2 - 1/2}\log T$. Using Fourier analysis and the geometric sieve, as in Bhargava's proof of van der Waerden's conjecture, we improve this bound for some classes of $A$.

math.NT↗

Étale algebras and the Kummer theory of finite Galois modules

Galois cohomology groups $H^i(K,M)$ are widely used in algebraic number theory, in such contexts as Selmer groups of elliptic curves, Brauer groups of fields, class field theory, and Iwasawa theory. The standard construction of these groups involves maps out of the absolute Galois group $G_K$, which in many cases of interest (e.g. $K = \mathbb{Q}$) is too large for computation or even for gaining an intuitive grasp. However, for finite $M$, an element of $H^i(K,M)$ can be described by a finite amount of data. For the important case $i = 1$, the appropriate object is an étale algebra over $K$ (a finite product of separable field extensions) whose Galois group is a subgroup of the semidirect product $\operatorname{Hol} M = M \rtimes \operatorname{Aut} M$ (often called the \emph{holomorph} of $M$), equipped with a little bit of combinatorial data. Although the correspondence between $H^1$ and field extensions is in widespread use, it includes some combinatorial and Galois-theoretic details that seem never to have been written down. In this short quasi-expository paper, we fill in this gap and explain how the étale algebra perspective illuminates some common uses of $H^1$, including parametrizing cubic and quartic algebras as well as computing the Tate pairing on Galois coclasses of local fields.

math.NT↗

Raney Transducers and the Lowest Point of the $p$-Lagrange spectrum

It is well known that the golden ratio $ϕ$ is the ''most irrational'' number in the sense that its best rational approximations $s/t$ have error $\sim 1/(\sqrt{5} t^2)$ and this constant $\sqrt{5}$ is as low as possible. Given a prime $p$, how can we characterize the reals $x$ such that $x$ and $p x$ are both ''very irrational''? This is tantamount to finding the lowest point of the $p$-Lagrange spectrum $\mathcal{L}_p$ as previously defined by the third author. We describe an algorithm using Raney transducers that computes $\min \mathcal{L}_p$ if it terminates, which we conjecture it always does. We verify that $\min \mathcal{L}_p$ is the square root of a rational number for primes $p < 2000$. Mysteriously, the highest values of $\min \mathcal{L}_p$ occur for the Heegner primes $67$, $3$, and $163$, and for all $p$, the continued fractions of the corresponding very irrational numbers $x$ and $p x$ are in one of three symmetric relations.

math.NT↗

Diophantine approximation on conics

Given a conic $\mathcal{C}$ over $\mathbb{Q}$, it is natural to ask what real points on $\mathcal{C}$ are most difficult to approximate by rational points of low height. For the analogous problem on the real line (for which the least approximable number is the golden ratio, by Hurwitz's theorem), the approximabilities comprise the classically studied Lagrange and Markoff spectra, but work by Cha-Kim and Cha-Chapman-Gelb-Weiss shows that the spectra of conics can vary. We provide notions of approximability, Lagrange spectrum, and Markoff spectrum valid for a general $\mathcal{C}$ and prove that their behavior is exhausted by the special family of conics $\mathcal{C}_n : XZ = nY^2$, which has symmetry by the modular group $Γ_0(n)$ and whose Markoff spectrum was studied in a different guise by A. Schmidt and Vulakh. The proof proceeds by using the Gross-Lucianovic bijection to relate a conic to a quaternionic subring of $\operatorname{Mat}^{2\times 2}(\mathbb{Z})$ and classifying invariant lattices in its $2$-dimensional representation.

math.NT↗

Renaissance canons with asymmetric schemes

By a "scheme" of a musical canon, we mean the time and pitch displacement of each entering voice. When the time displacements are unequal, achieving consonant sonorities is especially challenging. Using a first-species theoretical model, we quantify the flexibility of schemes that Renaissance composers used or could have used. We craft an algorithm to compute this flexibility value precisely (finding in the process that it is an algebraic integer). We find that Palestrina consistently selected some of the most flexible schemes, more so than his predecessors, but that he by no means exhausted all feasible schemes. To add support to the model, we present two new compositions within the limits of the style utilizing unexplored canonic schemes. In the Online Supplement (attached via Papers with Code), we provide MIDI realizations of the musical examples and Sage code used in the numerical computations.

math.NT↗

Large orbits on Markoff-type K3 surfaces over finite fields

We study the surface $\mathcal{W}_k : x^2 + y^2 + z^2 + x^2 y^2 z^2 = k x y z$ in $(\mathbb{P}^1)^3$, a tri-involutive K3 (TIK3) surface. We explain a phenomenon noticed by Fuchs, Litman, Silverman, and Tran: over a finite field of order $\equiv 1$ mod $8$, the points of $\mathcal{W}_4$ do not form a single large orbit under the group $Γ$ generated by the three involutions fixing two variables and a few other obvious symmetries, but rather admit a partition into two $Γ$-invariant subsets of roughly equal size. The phenomenon is traced to an explicit double cover of the surface.

math.NT↗

Reflection theorems for number rings

This is the original version of my Ph.D. thesis. The main results have been divided up between papers arXiv:2111.09784 and arXiv:2204.10924. This paper has been kept on the arXiv to preserve some proofs of elementary lemmas that will be removed from the peer-reviewed version.

math.NT↗

Quintic algebras over Dedekind domains and their sextic resolvents

Bhargava parametrized quintic rings over $\mathbb{Z}$ by quadruples of $5\times 5$ alternating matrices. We extend the construction to work similarly over any Dedekind domain $R$. No assumptions are needed on the characteristic of $R$. The resolvent consists of a pair of locally free modules $L$, $M$ with two multilinear maps between them; we can view $L$ as $Q/R$, for $Q$ the quintic ring, and $M$ as $S/R$, where $S$ is a sextic resolvent ring. As in Bhargava's treatment, any quintic ring has a resolvent ring, and for a maximal ring, the resolvent is unique. We hope that this work will enable the removal of the condition that the characteristic be different from $2$ in Bhargava-Shankar-Wang's proof of Linnik's conjecture on the asymptotic distribution of discriminants of relative extensions.

math.NT↗

Visibly irreducible polynomials over finite fields

H. Lenstra has pointed out that a cubic polynomial of the form (x-a)(x-b)(x-c) + r(x-d)(x-e), where {a,b,c,d,e} is some permutation of {0,1,2,3,4}, is irreducible modulo 5 because every possible linear factor divides one summand but not the other. We classify polynomials over finite fields that admit an irreducibility proof with this structure.

math.NT↗