Green's function on the Tate curve

The Big Picture

Imagine describing the shape of a donut to someone whose world runs on completely different arithmetic, where “closeness” is defined not by size but by how many times a prime number divides the difference between two values. That’s the setting of p-adic string theory, and this paper solves a problem that’s been open since the subject began in the late 1980s.

String theory posits that the fundamental constituents of nature are tiny vibrating strings. As each string moves through spacetime, it sweeps out a surface called its worldsheet, much like a moving loop traces out a tube. At the simplest level, worldsheets are spheres or disks. One step up in complexity, the surface closes back on itself to form a torus: the donut shape.

Physicists have handled ordinary worldsheets for decades. But a parallel version of string theory is built on p-adic numbers, an alternative number system (one for each prime $p$) where arithmetic is governed by divisibility rather than magnitude. In the p-adic world, simpler worldsheets are well understood. The torus has remained out of reach.

An Huang, Rebecca Rohrlich, Yaojia Sun, and Eric Whyman tackle that problem here. They define a mathematical operator governing how fields spread across a p-adic torus and solve for the function describing field interactions. A purely number-theoretic object from arithmetic geometry turns out to be the large-$p$ limit of a p-adic quantum field theory correlation function.

Key Insight: The Néron local height function on the Tate curve, a fundamental object in number theory, is the Green’s function of a p-adic Laplacian. This gives it a direct physical interpretation as a quantum field theory correlator.

How It Works

Start with the Tate curve, the p-adic version of a torus, written as $E_q = \mathbb{Q}_p^\times / q^\mathbb{Z}$. This is p-adic numbers modulo a lattice, the p-adic analogue of constructing a flat torus by gluing opposite edges of a plane together. Here $q$ is a p-adic number with $|q| < 1$, and it controls the torus’s shape the way a modular parameter does in ordinary geometry.

The central object is the Green’s function, a field theory’s version of an impulse response. It solves $DG = \delta - 1/V$, where $D$ is the Laplacian, $\delta$ is a perfectly localized disturbance, and $V$ is the volume of the fundamental domain. If the Laplacian describes how a field propagates across the torus, the Green’s function captures the response to a point source. In string theory, this is the two-point correlation function of the worldsheet conformal field theory: it encodes how two points on the string interact.

The authors define the p-adic Laplacian on the Tate curve via an integral operator:

$$D\phi(x) := \int_E \frac{\phi(z) - \phi(x)}{|z-x|^2} |x|, dz$$

This is a non-Archimedean Laplacian: it uses p-adic norms, which measure divisibility rather than ordinary size. On any small patch of the p-adic torus, this operator coincides with the well-known Vladimirov derivative on simpler geometries, just as any piece of a torus locally looks like a flat plane.

To solve the resulting equation, the authors use a filtration strategy:

• Restrict to functions depending on only the first $k$ p-adic digits, reducing the infinite-dimensional problem to finite-dimensional matrix equations at each level $k$
• This filtration corresponds geometrically to truncating the Bruhat-Tits tree $T_p/\Gamma$, a combinatorial tree encoding p-adic geometry, at radius $k$
• Take the limit $k \to \infty$ to recover the full Green’s function

The result is an explicit closed-form formula. On an ordinary torus, the Green’s function involves logarithms and Jacobi theta functions. Here it involves p-adic norms and sums over the tree $T_p/\Gamma$: the same architecture, translated into p-adic language.

The punchline comes in the large-$p$ limit. When the j-invariant of the Tate curve has odd valuation (a condition on how “degenerate” the torus is), the Green’s function converges as $p \to \infty$ to the Néron local height function. Arithmetic geometers use this to measure heights of points on elliptic curves, and it connects to questions around the Birch and Swinnerton-Dyer conjecture.

Why It Matters

Number theorists have studied Néron local height functions since the 1960s as abstract algebraic tools for measuring arithmetic complexity. Physicists have studied Green’s functions as concrete objects governing particle propagation. This paper shows they are the same thing, at least in the p-adic world. The Néron height isn’t abstract bookkeeping; it’s a correlator encoding how information propagates across a p-adic string worldsheet.

Scattering amplitudes in string theory are built from Green’s functions, so this result makes p-adic string amplitudes at genus one computable for the first time. The paper points toward connections with the adelic product formula: just as Tate’s thesis showed that products of p-adic and Archimedean zeta functions over all primes satisfy elegant identities, similar product formulas may hold for the torus, stitching each prime’s version of string theory into a single adelic picture.

The Bruhat-Tits tree filtration gives a concrete computational method for non-Archimedean geometry that could extend to higher-genus worldsheets, pushing toward a full p-adic perturbative string theory.

Bottom Line: Huang, Rohrlich, Sun, and Whyman find the Green’s function on the p-adic Tate curve, solving an open problem in p-adic string theory and showing that a number-theoretic height function is secretly a quantum field theory correlator.