Gan--Gross--Prasad cycles and derivatives of $p$-adic $L$-functions

This paper establishes a p-adic analogue of the arithmetic Gan-Gross-Prasad conjectures for unitary groups by constructing a cyclotomic p-adic L-function and proving a formula linking its first derivative to p-adic heights of Selmer classes derived from arithmetic diagonal cycles, thereby advancing the p-adic Beilinson-Bloch-Kato conjecture.

The Gist

Imagine you are trying to solve a massive, cosmic puzzle. On one side of the table, you have Numbers (specifically, complex mathematical functions called $L$-functions that encode deep secrets about prime numbers). On the other side, you have Shapes (geometric objects called cycles living on special surfaces known as Shimura varieties).

For decades, mathematicians have suspected that these two sides are secretly talking to each other. If a number on the left side behaves in a certain way (like having a "derivative" that isn't zero), it should mean there is a specific, non-zero shape on the right side. This is the Gross–Zagier and Beilinson–Bloch–Kato conjecture: a bridge between the world of analysis (calculus) and the world of geometry.

This paper, by Daniel Disegni and Wei Zhang, builds a new, ultra-precise bridge for a specific, high-dimensional version of this puzzle involving Unitary Groups. They don't just build the bridge; they measure the exact weight of the shapes on the bridge using a special "ruler" called a $p$-adic height.

Here is the breakdown of their journey, explained with everyday analogies:

1. The Problem: The "Ghost" Connection

In the old days, mathematicians knew that if you take a specific type of number function (an $L$-function) and look at its slope (derivative) at a specific point, it tells you something about the size of a geometric shape. But this was mostly known for simple, 2-dimensional cases (like elliptic curves, which look like donuts).

The authors wanted to tackle higher dimensions (think of a hyper-donut in 100 dimensions). The problem is that in these high dimensions, the "shapes" (cycles) are incredibly hard to see, and the "numbers" are incredibly hard to calculate.

2. The Tool: The "Relative Trace Formula" (The Magic Mirror)

To connect the numbers and the shapes, the authors use a powerful tool called a Relative Trace Formula.

• The Analogy: Imagine you have a room with two mirrors facing each other. One mirror reflects the "Number World" (spectral side), and the other reflects the "Shape World" (geometric side).
• The Trick: If you shine a light (a specific mathematical test function) into the room, the reflections on both mirrors should look identical. If they match, it proves the connection between the numbers and the shapes exists.
• The Innovation: Previous attempts used "archimedean" mirrors (standard calculus). This paper builds a $p$-adic mirror. This is a mirror made of a different material (using $p$-adic numbers, which are like a different kind of arithmetic based on prime numbers). This allows them to see things that were invisible before.

3. The Journey: Three Major Steps

Step A: Rationality (The "Recipe" Check)

Before building the bridge, they had to make sure the ingredients were real. They proved that the ratios of these complex $L$-values are actually "rational" (they can be described by simple fractions or algebraic numbers).

• Analogy: Before baking a cake, you check if your recipe uses real ingredients. They proved that the "flavor" of these numbers is consistent and predictable, not chaotic.

Step B: The $p$-adic $L$-Function (The "Interpolation Machine")

They constructed a new function, $L_p$, which acts like a universal remote control.

• The Analogy: Imagine you have a radio that can tune into thousands of different stations (different mathematical characters). Usually, you have to tune them one by one. This new function is a "smart radio" that can tune into all of them at once and record their signals into a single, smooth curve.
• Why it matters: This curve allows them to see how the numbers change as you tweak the settings, which is essential for calculating derivatives.

Step C: The Height Formula (The "Weight Scale")

This is the climax. They proved a precise formula linking the slope of their new $p$-adic radio curve to the height (a measure of complexity) of a specific geometric shape called a Gan–Gross–Prasad (GGP) cycle.

• The Analogy: Imagine you have a scale. On one side, you put the "slope" of the number curve. On the other side, you put the "weight" of a geometric shape.
• The Discovery: They proved that these two sides balance perfectly. If the number curve has a non-zero slope, the geometric shape has a non-zero weight. This means the shape actually exists and isn't just a ghost.

4. The "Secret Sauce": How They Did It

The authors didn't just guess; they used a technique called Orbital Integrals.

• The Analogy: Think of a drum. If you hit it, it vibrates in specific patterns (frequencies). The authors hit the "drum" of the number world and the "drum" of the shape world with the same stick (a test function).
• The Challenge: In high dimensions, the drum skin is very complex. They had to invent a special stick (a "Gaussian" test function) that vibrates the drum in just the right way to isolate the specific frequency they care about, filtering out all the noise.
• The "Almost Unramified" Condition: To make the math work, they had to assume the shapes they were studying were "almost smooth" (a technical condition called almost unramified). It's like saying, "We can only measure the weight of the object if it's not covered in too much mud." They proved that even with this limitation, the bridge holds.

5. The Big Picture: Why Should We Care?

This paper is a massive leap forward in the Beilinson–Bloch–Kato Conjecture.

• The Conjecture: This is one of the "Holy Grails" of number theory. It predicts that the number of solutions to certain equations (related to the Selmer group) is exactly equal to the order of the zero of an $L$-function.
• The Impact: By proving this formula for $p$-adic numbers, the authors have provided a new way to count these solutions. They showed that if the $L$-function behaves a certain way, there is guaranteed to be at least one non-trivial geometric solution.
• The Metaphor: It's like proving that if the wind blows in a specific direction (the $L$-function derivative), there must be a tree growing in a specific spot (the geometric cycle). This gives mathematicians a powerful new tool to hunt for solutions to equations that have stumped them for centuries.

Summary

Disegni and Zhang built a $p$-adic bridge connecting the world of numbers to the world of shapes. They used a magic mirror (trace formula) to show that the slope of a number curve perfectly predicts the weight of a geometric shape. This confirms a deep, hidden relationship in the universe of mathematics, bringing us one step closer to solving the mysteries of prime numbers and algebraic geometry.

Technical Summary

This paper, "Gan–Gross–Prasad Cycles and Derivatives of p-adic L-functions" by Daniel Disegni and Wei Zhang, establishes a precise formula relating the first derivative of a $p$-adic $L$-function to the $p$-adic heights of arithmetic cycles on unitary Shimura varieties. This work formulates and proves a $p$-adic analogue of the Arithmetic Gan–Gross–Prasad (GGP) conjecture for unitary groups, providing a significant step toward the $p$-adic Beilinson–Bloch–Kato (BBK) conjecture for Rankin–Selberg motives.

1. The Problem

The classical Gross–Zagier formula relates the height of Heegner points (arithmetic cycles) to the derivative of a complex $L$-function at the central critical point. The Gross–Zagier–Perrin-Riou conjecture and the Arithmetic GGP conjecture generalize this to higher-dimensional Shimura varieties and automorphic representations of unitary groups.

While the complex version (Ichino–Ikeda conjecture) has been largely resolved, the $p$-adic analogue remains elusive. The central problem addressed here is:

• Can one construct a $p$-adic $L$-function $L_p(M_\Pi)$ for a conjugate-selfdual cuspidal automorphic representation $\Pi$ of $GL_n \times GL_{n+1}$ over a CM field?
• Does the derivative of this $p$-adic $L$-function at the trivial character relate to the $p$-adic height of specific algebraic cycles (Gan–Gross–Prasad cycles) on unitary Shimura varieties?
• Can this relationship be used to prove cases of the $p$-adic Beilinson–Bloch–Kato conjecture, specifically linking the order of vanishing of the $L$-function to the dimension of the Selmer group?

2. Methodology

The authors employ a relative trace formula (RTF) approach, comparing two distinct distributions with $p$-adic coefficients. This strategy, inspired by Jacquet–Rallis and previous work by Wei Zhang, involves:

A. The Analytic Side (The $p$-adic $L$-function)

• Construction: They construct a $p$-adic $L$-function $L_p(M_\Pi)$ by defining a $p$-adic analytic relative-trace distribution $I$.
• Rationality: They first prove the rationality of twisted Rankin–Selberg $L$-values (Theorem A) using "Gaussian" test functions and the Jacquet–Rallis RTF in rational coefficients.
• Interpolation: Using $p$-ordinary representations, they construct a distribution on the Hecke algebra that interpolates these $L$-values. The spectral expansion of this distribution yields the $p$-adic $L$-function.
• Derivative: They analyze the derivative of this distribution, $\partial I$, which isolates the term corresponding to the first derivative of the $L$-function.

B. The Arithmetic Side (The Height Distribution)

• Cycles: They define Gan–Gross–Prasad cycles arising from the embedding of a Shimura variety associated with $U(n)$ into one associated with $U(n) \times U(n+1)$.
• Heights: They define a distribution $J$ encoding the $p$-adic heights of these cycles using Nekovář's theory of $p$-adic heights and biextensions.
• Arithmetic RTF: They prove an arithmetic relative-trace formula for $J$, expressing it as a sum over local arithmetic intersection numbers on integral models of Shimura varieties (specifically, Rapoport–Zink spaces).

C. The Comparison (The Core Proof)

• Matching: The authors construct specific "test" Hecke measures that match spectrally and geometrically between the linear group side ($GL_n \times GL_{n+1}$) and the unitary group side.
• Fundamental Lemma & Transfer: They utilize the Fundamental Lemma (Yun, Beuzart-Plessis) and the Arithmetic Fundamental Lemma (AFL) and Arithmetic Transfer (AT) conjectures (proven by Z. Zhang, Luo, Mihatsch, and others) to equate the local terms of the two distributions.
• Global Comparison: By comparing the global geometric expansions of the analytic derivative $\partial I$ and the arithmetic height distribution $J$, they derive the main formula.

3. Key Contributions and Results

Theorem A: Rationality of $L$-values

They prove the rationality of the ratio of twisted Rankin–Selberg $L$-values for trivial-weight hermitian cuspidal automorphic representations. This establishes that these values lie in a specific number field, a prerequisite for $p$-adic interpolation.

Theorem B: Existence of $p$-adic $L$-functions

They construct a unique cyclotomic $p$-adic $L$-function $L_p(M_\Pi)$ that interpolates the algebraic parts of the twisted $L$-values for $p$-ordinary representations. This is achieved via a new method using $p$-adic relative-trace formulas.

Theorem D: The $p$-adic Arithmetic GGP Formula

This is the paper's central result. Under specific local assumptions (unramified extensions, split primes at $p$, and conditions on conductors), they prove a precise formula:
$$h_\pi(Z_\pi(\phi), Z_{\pi^\vee}(\phi')) = e_p(M_\Pi)^{-1} \cdot \frac{1}{4} \partial L_p(M_\Pi) \cdot \alpha(\phi, \phi')$$
where:

• $h_\pi$ is the $p$-adic height pairing.
• $Z_\pi$ are the GGP cycles.
• $\partial L_p(M_\Pi)$ is the derivative of the $p$-adic $L$-function.
• $\alpha$ is a canonical invariant functional.

This formula generalizes the $p$-adic Gross–Zagier formula to higher dimensions and arbitrary unitary groups.

Theorem C: Applications to the $p$-adic Beilinson–Bloch–Kato Conjecture

As a consequence of Theorem D, they prove a non-vanishing criterion for the Selmer group:
$$\text{ord}_{\chi=1} L_p(M_\Pi) = 1 \implies \dim_L H^1_f(F, \rho_\Pi) \ge 1$$
Furthermore, under additional "admissibility" conditions (using recent bounds by Liu–Tian–Xiao–Zhang), they prove equality:
$$\text{ord}_{\chi=1} L_p(M_\Pi) = 1 \implies \dim_L H^1_f(F, \rho_\Pi) = 1$$
This confirms the $p$-adic BBK conjecture for a broad class of motives attached to Rankin–Selberg products, showing that the Selmer group is generated by the class of an algebraic cycle.

4. Significance

• Resolution of a Major Conjecture: This work provides the first proof of the $p$-adic arithmetic GGP conjecture for unitary groups beyond the case of Heegner points (dimension 1).
• New Methodology: The construction of $p$-adic $L$-functions via relative-trace formulas in $p$-adic coefficients is a novel technique that avoids reliance on modular symbols or $p$-adic families of automorphic forms in the traditional sense.
• Bridge between Geometry and Analysis: It rigorously connects the analytic derivative of $L$-functions to the arithmetic geometry of Shimura varieties (heights of cycles) in the $p$-adic setting.
• Selmer Group Bounds: The results provide concrete evidence for the $p$-adic BBK conjecture, linking the analytic rank of $L$-functions to the algebraic rank of Selmer groups for higher-dimensional Galois representations, a major open problem in number theory.
• Technical Advances: The paper resolves deep technical issues regarding the vanishing of absolute cohomology of integral models and the explicit computation of $p$-adic orbital integrals for "almost unramified" representations, which are crucial for the arithmetic transfer conjectures.

In summary, Disegni and Zhang have successfully extended the Gross–Zagier philosophy to the $p$-adic world for general unitary groups, establishing a profound link between $p$-adic $L$-functions, arithmetic cycles, and Galois cohomology.