A conjecture for arithmetic spherical maximal functions
K Hughes · 2026
Read A conjecture for arithmetic spherical maximal functions by K Hughes (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-22712.txt.
Introduction
Hughes puts forward a conjecture on boundedness for maximal functions over a sparse family of discrete spherical averages. Better bounds beyond the full discrete spherical maximal range have stayed open for 24 years. The paper states a theorem backing the proposed claim. The note is short and conjectural, not a full answer, and it sets a precise target for later work on these averages.
Catalog record
| CREATOR | K Hughes |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 5103 |
| LICENSE | http://creativecommons.org/licenses/by-sa/4.0/ |
| RIGHTS | CC BY SA |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.22712
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A conjecture for arithmetic spherical maximal functions
K Hughes
Abstract
. For 24 years, it has been an open problem to obtain improved bounds, for the maximal function over a sparse sequence of discrete spherical averages, going beyond the range for the full discrete spherical maximal function. I formulate a conjecture to characterize the boundedness of such maximal functions and state a theorem in support of it.
We begin with some notation so that we may precisely formulate the problem.
Let
d\geq 4
.
Define the arithmetic spherical averages
A_{\lambda}f(x)=\frac{1}{\#\{y\in\mathbb{Z}^{d}:|y|^{2}=\lambda\}}\sum_{y\in\mathbb{Z}^{d}:|y|^{2}=\lambda}f(x-y)
for
\lambda\in\mathbb{N}
,
f\in\ell^{1}(\mathbb{Z}^{d})
and
x\in\mathbb{Z}^{d}
.
For
\Lambda\subset\mathbb{N}
, define the arithmetic spherical maximal function
M_{\Lambda}f(x):=\sup_{\lambda\in\Lambda}|A_{\lambda}f(x)|.
In
[
14
]
, it was proved that
M_{\mathbb{N}}
is bounded on
\ell^{p}(\mathbb{Z}^{d})
if and only if
d\geq 5
and
p>\frac{d}{d-2}
. The restricted weak-type endpoint result was given in
[
10
]
. Let
\Lambda\subseteq\mathbb{N}
. Since
M_{\mathbb{N}}f(x)\geq M_{\Lambda}f(x)
for all
f\in\ell^{p}(\mathbb{Z}^{d})
and
x\in\mathbb{Z}^{d}
,
[
14
]
implies that
M_{\Lambda}
is bounded on
\ell^{p}(\mathbb{Z}^{d})
when
d\geq 5
and
p>\frac{d}{d-2}
.
A natural question is:
Is this range of
\ell^{p}(\mathbb{Z}^{d})
spaces sharp for a given sequence
\Lambda
?
By analogy with the continuous Euclidean lacunary spherical maximal function
[
3
,
4
]
, the folklore conjecture was that, for each lacunary sequence
\Lambda\subset\mathbb{N}
,
M_{\Lambda}
is bounded on
\ell^{p}(\mathbb{Z}^{d})
for all
p>1
and
d\geq 5
. I also conjectured this is true when
d=4
for sequences avoiding the 2-adic obstruction when
d=4
.
The first progress towards this question was made in
[
8
]
. Subsequently,
[
6
,
5
]
showed that there are infinite sequences that are bounded on
\ell^{p}(\mathbb{Z}^{d})
for
d\geq 5
and all
p>1
. At my behest,
[
12
]
combined my ideas and methods introduced in
[
8
]
and
[
9
]
to prove that – for
any
lacunary sequence
\Lambda\subset\mathbb{N}
–
M_{\Lambda}
is bounded on
\ell^{p}(\mathbb{Z}^{d})
when
p>\frac{d-2}{d-3}
and
d\geq 5
. Such maximal functions have been studied further in
[
7
,
1
,
15
,
13
,
17
,
2
]
.
While writing up
[
8
]
, Prof. Jacek Zienkiewicz showed me a probabilistic construction proving the existence of lacunary sequences
\Lambda\subset\mathbb{N}
whose associated arithmetic spherical maximal function
M_{\Lambda}
is unbounded near
\ell^{1}(\mathbb{Z}^{d})
for each
d\geq 5
.
His construction disproved the folklore conjecture.
I quickly adapted the construction to show that, for each
p<\frac{d}{d-1}
and
d\geq 4
, there exist arbitrarily fast growing sequences, say
\Lambda
, in
\mathbb{N}
such that
M_{\Lambda}
is unbounded on
\ell^{p}(\mathbb{Z}^{d})
. This showed that intuition from the continuous setting did not apply; see
[
7
]
.
Yet, the intuition from the continuous setting can apply and the folklore conjecture can be true for averages defined by objects similar to spheres, but with substantially less arithmetic, such as the equation
\lambda=\sum_{i=1}^{d}\lfloor x_{i}^{c}\rfloor
where
c
is not integral; see
[
11
]
.
Given the mixture of positive and negative results, the above question becomes:
For a given sequence
\Lambda\subset\mathbb{N}
, can we characterize the range of
\ell^{p}(\mathbb{Z}^{d})
spaces on which
M_{\Lambda}
is bounded?
The point of this note is to conjecture a characterization.
To set this up: Let
\mathfrak{P}
denote the set of rational primes.
For each prime
\mathfrak{p}\in\mathfrak{P}
, let
\delta_{\mathfrak{p}}(\Lambda)
denote the
\mathfrak{p}
-adic upper Minkowski dimension of
\Lambda
.
Let
\delta_{\infty}(\Lambda)
denote the smallest parameter
\delta\geq 0
such that there exists a constant
C_{\delta}
– independent of
N\in\mathbb{N}
– with
\#(\Lambda\cap[N,2N])\leq C_{\delta}N^{\delta}
uniformly for all
N\in\mathbb{N}
.
These dimensional quantities are analogous to those in
[
16
]
.
Define the exponent
\eta(\Lambda):=\max\Big\{\sup_{\mathfrak{p}\in\mathfrak{P}}\{1+\frac{\delta_{\mathfrak{p}}(\Lambda)}{d-1}\},1+\frac{2\delta_{\infty}(\Lambda)}{d-2}\Big\}.
We have the following unboundedness result (from 2015-6) whose proof I will defer to another paper (though its statement makes the proof obvious).
Theorem 1
.
Let
\Lambda\subset\mathbb{N}
.
M_{\Lambda}
is unbounded on
\ell^{p}(\mathbb{Z}^{d})
for all
p<\eta(\Lambda)
.
Finally, I conjecture the following.
Conjecture 1
.
Let
\Lambda\subset\mathbb{N}
.
Let
d\geq 5
.
M_{\Lambda}
is bounded on
\ell^{p}(\mathbb{Z}^{d})
for all
p>\eta(\Lambda)
.
Remark 1
.
In particular, Conjecture
1
states that any lacunary arithmetic spherical maximal function would be bounded on
\ell^{p}(\mathbb{Z}^{d})
for
p>\frac{d}{d-1}
and
d\geq 5
.
Remark 2
.
Theorem
1
and Conjecture
1
generalize to non-degenerate positive integral quadratic forms and to higher degree homogeneous integral forms.