Cluster k0498
Eigenspaces and traces for sandwich data over DVRs
This is a large group (most files are restated claims) on matrices and their "eigenspaces" (directions a matrix only stretches) when the entries lie in a discrete valuation ring (a ring with a single prime element, like the p-adic integers). Results include that eigenspaces for different ring maps are independent, that the number of ring maps to the integers equals a rank, that the characteristic polynomial of a left-multiplication is a power, and that a matrix intertwining two representations maps eigenspaces to eigenspaces. Two of the claims restate published theorems (Scholze–Weinstein, Dieudonné–Hodge) as assumptions. So parts rest on assumed external results rather than proofs here.
4 import steps from the main theorem.
Proof status: on 21 Nov 2026 the Lean FRO and the Mathlib Initiative certified the Lean statement and proof (latest check). Names and descriptions on this map are written by Claude from samples of the code and may be wrong.
Prerequisite concepts
Linear algebraValuation rings and lattices over themExterior powers and wedge productsAlgebraic closureSeparable polynomials (no repeated roots) and the derivative testDiscrete valuation rings: numbers with one prime, and reducing mod it
References
Serge Lang, Algebra · ch. XIV (representations; traces) and ch. XVIIIBackground reading suggested by the AI namer · unchecked
Uses
Used by
168 Lean files, main folder Development.AbelianVarietyTheory.Proofs.DivisorProofs.