Skip to main content
QUICK REVIEW

[Paper Review] Galois theory and Diophantine geometry

Minhyong Kim|ArXiv.org|Aug 5, 2009
Algebraic Geometry and Number Theory4 citations
TL;DR

This paper proposes a non-abelian p-adic approach to Diophantine geometry using Galois cohomology and p-adic Hodge theory, showing that rational points on curves of genus ≥2 are finite by proving the image of X(Q) in a De Rham homogeneous space is discrete and compact. The key result establishes finiteness of rational points via the sparseness of zeros of an algebraic p-adic L-function, generalizing classical finiteness theorems through motivic and non-abelian structures.

ABSTRACT

This is an essay to accompany the author's lecture at the introductory workshop on `Nonabelian fundamental groups in arithmetic geometry' at the Newton Institute, Cambridge in July, 2009.

Motivation & Objective

  • To establish a non-abelian p-adic analogue of the Mordell conjecture using Galois cohomology and De Rham fundamental groups.
  • To demonstrate that rational points on a curve of genus ≥2 are finite by analyzing the image of X(Q) in a p-adic homogeneous space.
  • To show that the vanishing of a p-adic L-function's zeros controls the finiteness of Selmer varieties, generalizing the role of L-functions in elliptic curve arithmetic.
  • To provide a motivic framework linking abelian and non-abelian Diophantine problems via the motivic fundamental group.
  • To explore the feasibility of explicit analytic equations for X(Q) ⊂ X(Qp) using computable p-adic logarithms and cohomological functions.

Proposed method

  • Constructs a map κⁿᵘ: X(Q) → H¹_f(G, Uₙ) using unipotent fundamental group representations over Qp.
  • Uses non-abelian p-adic Hodge theory to identify H¹_f(Gₚ, Uₙ) with the quotient Uⁿᴰᴿ / F⁰, a De Rham homogeneous space.
  • Applies the Dieudonné isomorphism D: H¹_f(Gₚ, Uₙ) ≅ Uⁿᴰᴿ / F⁰ to relate Galois cohomology to geometric moduli of flat connections.
  • Defines a non-vanishing convergent power series φ on X(Qp) that vanishes on X(Q), using the composition φ ∘ κⁿᵈʳ/ᶜʳ,ₙ.
  • Leverages the fact that the image of X(Qp) in Uⁿᴰᴿ / F⁰ is space-filling and compact, so its intersection with any proper subvariety is finite.
  • Uses the Iwasawa module structure and Hom estimates over Λ to bound the number of points via O(n²ᵍ⁻¹) growth in dimension.

Experimental results

Research questions

  • RQ1Can the finiteness of rational points on a curve of genus ≥2 be proven via a non-abelian p-adic cohomological method?
  • RQ2To what extent can the De Rham fundamental group and p-adic Hodge theory replace the Jacobian in Diophantine geometry?
  • RQ3Can the vanishing of an algebraic p-adic L-function control the finiteness of Selmer varieties in a way analogous to L-function non-vanishing in elliptic curve theory?
  • RQ4Is the image of X(Q) in the De Rham quotient Uⁿᴰᴿ / F⁰ discrete and compact, ensuring only finitely many rational points?
  • RQ5Can the function φ vanishing on X(Q) be made explicit to yield analytic equations for rational points in X(Qp)?

Key findings

  • The image of X(Q) in Uⁿᴰᴿ / F⁰ is a compact, space-filling subset whose intersection with any proper subvariety is finite.
  • The function φ ∘ κⁿᵈʳ/ᶜʳ,ₙ vanishes on X(Q) and is a non-vanishing convergent power series on each residue disk, hence has only finitely many zeros.
  • The number of rational points is bounded by O(n²ᵍ⁻¹) via Hom estimates over the Iwasawa algebra Λ, where g is the genus.
  • The annihilator ideal L of the Iwasawa module M corresponds to a p-adic L-function whose zero set controls Selmer variety finiteness.
  • The hypothesis dim H¹_f(G, Uₙ) << dim Uⁿᴰᴿ / F⁰ is expected to hold for large n, supported by the Fontaine-Mazur and Bloch-Kato conjectures.
  • In the CM case, the method applies when p is split in the CM field, and the image of G in Aut(H₁(Ẋ, ℤₚ)) is abelian, enabling unconditional finiteness.

Better researchstarts right now

From reading papers to final review, dramatically reduce your research time.

No credit card · Free plan available

This review was created by AI and reviewed by human editors.