🔢 mathematics

Weak Inhomogeneous Poissonian Pair Correlation and Equidistribution

This paper demonstrates that weak inhomogeneous Poissonian pair correlation is a strictly weaker condition than its strong counterpart, does not guarantee equidistribution, and yields mutually independent notions depending on the specific inhomogeneous parameter chosen.

Original authors: Zhiqin Tang, Qing-Long Zhou

Published 2026-07-14

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Zhiqin Tang, Qing-Long Zhou

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you have a giant, endless line of dancers on a circular stage. Mathematicians love to watch how these dancers spread out. Do they fill the whole stage evenly? Do they clump together? Or do they keep a perfect, random distance from their neighbors?

For a long time, mathematicians had a set of rules for how these dancers should behave if they were truly "random" in a specific, high-level way. They called this Poissonian Pair Correlation (PPC). Think of PPC as the "Gold Standard" of randomness: if your dancers have PPC, they are perfectly spaced out, and if you look at the distance between any two dancers, it follows a very specific, predictable pattern.

But here's the twist: What if the dancers aren't just looking at each other, but are also trying to stay a specific distance away from a ghost dancer floating somewhere else on the stage? This is called Inhomogeneous PPC. It's like saying, "I want to be random, but I also want to be exactly $\gamma$ steps away from this invisible friend."

The authors of this paper, Zhiqin Tang and Qing-Long Zhou, decided to test the rules of this game. They asked: "If the dancers follow the 'Gold Standard' rules for the ghost friend, does that mean they are also following the Gold Standard for the regular friends? And does it mean they are spread out evenly across the whole stage?"

The Big Surprise: The Rules Don't Match

In the world of regular randomness (without the ghost friend), there was a nice, tidy ladder of rules. If you had the strongest rule (PPC), you automatically had the weaker ones. It was like a pyramid: if you were at the top, you were definitely at the bottom too.

The authors proved that this pyramid collapses when you add the ghost friend.

They showed that you can have a sequence of dancers that follows the "weak" version of the ghost-friend rule, but fails to follow the strong version. In fact, they proved that the weak version doesn't even guarantee that the dancers are spread out evenly across the stage!

Here is the kicker: In the regular world, if your dancers follow the weak rules, they must be spread out evenly. But in this "ghost friend" world, the authors constructed a specific sequence of dancers that followed the weak ghost rules perfectly, yet they were clumped together in a way that meant they were not evenly distributed. This is a sharp contrast to what everyone expected.

The "Different Ghosts" Experiment

The paper goes even further. Imagine you have two different ghost friends, let's call them Ghost A and Ghost B. They stand at different spots on the stage.

The authors asked: "If my dancers are perfectly arranged to keep a distance from Ghost A, does that mean they are also arranged well for Ghost B?"

The answer is a loud NO.

They proved that for any two different ghost positions (let's say $\gamma_1$ and $\gamma_2$ , both between 0 and 0.5), you can create a dance line that is perfect for Ghost A but a total mess for Ghost B. It turns out that being "good" at one ghost distance doesn't help you at all with another. They are completely independent concepts.

How Did They Prove This?

They didn't just guess; they built these dance lines using a clever mix of math and probability.

For the first surprise (Weak $\neq$ Strong): They imagined a dance line made of random pairs. Half the time, a dancer stands at a random spot; the other half, they stand at that same spot plus the ghost's distance. They showed that while this line looks "weakly" random enough to pass the test, it fails the strict test because the pairs are too predictable.
For the "Not Evenly Distributed" surprise: They used a different trick. They created a line of dancers where the density of dancers changes (some areas are crowded, some are empty). They proved that even with this uneven crowd, the "weak" ghost rule still holds true. This shattered the idea that weak rules always force an even spread.
For the "Different Ghosts" surprise: They built a massive, block-by-block dance routine. They stacked blocks of dancers that were perfect for Ghost A, but because of how they stacked them, the pattern for Ghost B got completely messed up. They proved mathematically that this construction works for any two different ghost positions.

What This Means

The paper doesn't just say "we think this is true." They proved it. They constructed specific examples that act as counter-examples to the old ideas.

They ruled out the idea that weak inhomogeneous rules imply strong inhomogeneous rules.
They ruled out the idea that weak inhomogeneous rules imply the dancers are evenly spread out.
They proved that being good for one ghost distance tells you nothing about being good for another.

So, the next time you see a line of dancers, don't assume that if they look random for one friend, they look random for all friends, or that they are filling the whole stage. In this strange, inhomogeneous world, the rules are much more chaotic and independent than we ever imagined. The authors have shown us that the beautiful, tidy ladder of randomness we thought we knew has broken into many separate, independent rungs.

Technical Summary: Weak Inhomogeneous Poissonian Pair Correlation and Equidistribution

Problem Statement
This paper investigates the relationships between three fundamental concepts in the distribution theory of sequences modulo one: equidistribution, Poissonian pair correlation (PPC), and weak PPC, specifically within the inhomogeneous setting. While the relationships between these properties are well-understood for the homogeneous case (where the shift parameter $\gamma = 0$ ), the behavior of inhomogeneous PPC (where the shift $\gamma \in (0, 1)$ ) and its weak form remains less explored.

The central questions addressed are:

Does weak inhomogeneous PPC imply inhomogeneous PPC?
Does weak inhomogeneous PPC imply equidistribution? (This contrasts with the homogeneous case, where weak PPC implies equidistribution).
Are weak inhomogeneous PPC properties for distinct shift parameters $\gamma_1$ and $\gamma_2$ independent of one another?

Methodology
The authors employ a combination of probabilistic constructions and asymptotic analysis to establish their results. The primary methodology involves:

Probabilistic Constructions: The paper constructs specific sequences using independent and identically distributed (i.i.d.) random variables.
- For Theorem 1.1, a sequence is constructed by interleaving i.i.d. uniform variables $y_n$ with shifted versions $y_n + \gamma$ . This construction is shown to satisfy weak inhomogeneous PPC almost surely while failing the stronger inhomogeneous PPC condition.
- For Theorem 1.2, the authors utilize random variables distributed according to a specific non-uniform density function $g(x)$ . By leveraging Lemma 3.1, which relates the pair correlation limit to the autocorrelation integral of the density ( $\int g(x)g(x+\gamma)dx$ ), they construct a density where this integral equals 1 (satisfying weak PPC) but the density is not identically 1 (implying non-equidistribution).
- For Theorem 1.3, a more complex inductive construction is used. The sequence is built in blocks where specific shifts ( $\gamma_2$ ) are repeated to create "clumping" that violates weak $\gamma_2$ -PPC, while the randomness of the underlying blocks ensures that weak $\gamma_1$ -PPC holds for a distinct $\gamma_1$ .
Analytical Tools:
- Integral Characterization: The proofs rely on Lemma 2.1, which establishes an equivalence between the weak PPC limit and the convergence of an integral involving local counting functions.
- Concentration Inequalities: Bernstein's inequality and the Borel–Cantelli lemma are used to prove that the constructed random sequences satisfy the required limits almost surely.
- Variance Estimates: Chebyshev's inequality and variance calculations are used to extend convergence from subsequences to the full sequence limit.

Key Contributions and Results

Separation of Weak and Strong Inhomogeneous PPC (Theorem 1.1):
The authors prove that for any $\gamma \in (0, 1/2]$ and $0 < \delta < 1$ , there exists a sequence satisfying weak $\gamma$ -PPC that does not satisfy $\gamma$ -PPC. This demonstrates that the weak form is strictly weaker than the strong form in the inhomogeneous setting, mirroring the known result for the homogeneous case.
Failure of Equidistribution Implication (Theorem 1.2):
A significant finding is that weak inhomogeneous PPC does not imply equidistribution. The paper constructs a sequence with weak $\gamma$ -PPC that is not equidistributed. This stands in "sharp contrast" to the homogeneous case, where Steinerberger proved that weak PPC implies equidistribution. The construction relies on finding a density $g$ such that $\int_0^1 g(x)g(x+\gamma)dx = 1$ while $g \not\equiv 1$ .
Independence of Parameters (Theorem 1.3):
The paper proves that weak inhomogeneous PPC properties for distinct parameters are mutually independent. Specifically, for distinct $\gamma_1, \gamma_2 \in (0, 1/2]$ , there exists a sequence that has weak $\gamma_1$ -PPC but fails to have weak $\gamma_2$ -PPC. This implies that the notion of weak inhomogeneous PPC is parameter-dependent and does not collapse into a single universal property.

Significance and Claims
The paper positions itself as providing a "partial contribution" to understanding the relational diagram of equidistribution, PPC, and weak PPC in the inhomogeneous setting.

Refining the Relational Diagram: The results update the known hierarchy (Figure 2 in the paper). While the homogeneous case (Figure 1) shows a clear chain of implications (PPC $\implies$ Weak PPC $\implies$ Equidistribution), the inhomogeneous case breaks the link between Weak PPC and Equidistribution.
Open Questions: The authors modestly note that their work does not fully resolve all relationships. Specifically, it remains an open question whether $\gamma$ -PPC is strictly stronger than weak $\gamma$ -PPC (the authors conjecture they are incomparable) and whether an equidistributed sequence exists that fails to be weak $\gamma$ -PPC.
Theoretical Impact: The work highlights that introducing an inhomogeneous shift $\gamma$ fundamentally alters the landscape of pair correlation statistics, decoupling properties that are tightly linked in the homogeneous case.

The paper does not propose experimental applications or future directions beyond the mathematical resolution of these specific distributional properties. The significance lies entirely in the rigorous classification of these statistical properties for deterministic and random sequences.

The Big Surprise: The Rules Don't Match

The "Different Ghosts" Experiment

How Did They Prove This?

What This Means

Technical Summary: Weak Inhomogeneous Poissonian Pair Correlation and Equidistribution

More like this