Documentation

LeanPool.HasseMinkowski.Main

Assembly of the Hasse–Minkowski principle over ℚ (WP6) #

This file assembles the local–global principle for quadratic forms over ℚ from the rank-by-rank inputs proved in the earlier layers:

The two diagonal ingredients RankFourDiagonalHM (RankFour.lean) and RankFiveLeDiagonalHM (the WP5.3 induction, HighRank.lean) enter as explicit hypotheses of the conditional hasseMinkowski_of and meyer_of; the unconditional hasseMinkowski and meyer below specialise them with rankFourDiagonalHM and rankFiveLeDiagonalHM rankFourDiagonalHM.

Main results #

WP6.1 — a degenerate form is isotropic #

In characteristic different from 2 a nonzero vector of the radical satisfies Q x = 0 (it is orthogonal to itself, and associated Q x x = 2 * Q x). This is the "degenerate Q" case of the Hasse–Minkowski principle.

WP6.2 — the rank-by-rank dispatch on the diagonalized form #

A nondegenerate form is equivalent to a weighted sum of squares ⟨w₀, …, w_{n-1}⟩ with nonzero rational weights. The local hypotheses of h4 and h5 are exactly the isotropy of this diagonal form at every place, so we dispatch on n = finrank ℚ V: ranks 0, 1, 2, 3 are the proved layer results, rank 4 is h4, and rank ≥ 5 is h5.

WP6.3 — Meyer's theorem #

An indefinite real form of rank ≥ 5 is isotropic over ℝ; over each ℚ_[p] every form in at least five variables is isotropic (isotropic_weightedSumSquares_of_five_le), and these local data feed WP6.2.

The unconditional theorems #

With RankFourDiagonalHM proved in RankFour.lean and its rank-≥ 5 counterpart in HighRank.lean, the conditional theorems above specialise to the targets: