Signature and definiteness of integral lattices #
This file defines the radical and signature (n₊, n₀, n₋) of an integral symmetric lattice and
the standard definiteness predicates. The indices of inertia are Mathlib's sigPos
and sigNeg; the null index is the dimension of the kernel of the bilinear form.
Positive- and negative-semidefiniteness are expressed using Mathlib's bilinear-form predicates. The characteristic theorems relate every predicate both to the signature and to the usual elementwise inequalities. In particular, an indefinite lattice has vectors of both signs, and a degenerate lattice has a nonzero vector in its radical. The final section proves that an integral-lattice isometry transports the radical and preserves all three inertia indices and every definiteness predicate.
References #
- W. Ebeling, Lattices and Codes, Chapter 1.
TauCetiRoadmap/IntegralLattices/README.md, Layer 1.
Main definitions #
TauCeti.IntegralLattice.radical: the kernel of the rational bilinear form.TauCeti.IntegralLattice.signature: the positive, null, and negative indices.TauCeti.IntegralLattice.IsPosSemidefand related definiteness predicates.TauCeti.IntegralLattice.isPosDef_ofGramMatrix_iff: a lattice presented by a Gram matrix is positive definite exactly when the matrix is, read overℚ.TauCeti.IntegralLattice.Isometry.map_radicaland theIsometry.sigPos_eq,sigNull_eq, andsigNeg_eqtheorems: isometry invariance of the signature.
The radical of an integral lattice is the kernel of its rational bilinear form.
Equations
- L.radical = LinearMap.ker L.form
Instances For
A vector lies in the radical of an integral lattice if and only if it annihilates all vectors under the rational bilinear form.
The radical of the quadratic form associated to a lattice is its bilinear radical.
The positive index of inertia of an integral lattice.
Equations
Instances For
The dimension of the radical of an integral lattice.
Equations
- L.sigNull = Module.finrank ℚ ↥L.radical
Instances For
The negative index of inertia of an integral lattice.
Equations
Instances For
The positive, null, and negative indices exhaust the rank of the lattice.
An integral lattice is positive-definite when its quadratic form is positive on every nonzero vector.
Equations
Instances For
An integral lattice is positive-semidefinite when its symmetric bilinear form is nonnegative.
Equations
Instances For
An integral lattice is negative-definite when the negative of its quadratic form is positive-definite.
Instances For
An integral lattice is negative-semidefinite when the negative of its symmetric bilinear form is positive-semidefinite.
Equations
- L.IsNegSemidef = (-L.form).IsPosSemidef
Instances For
An integral lattice is degenerate when its radical is nontrivial.
Equations
- L.IsDegenerate = (L.radical ≠ ⊥)
Instances For
An integral lattice is indefinite when both its positive and negative indices are nonzero.
Instances For
The radical of an integral lattice is trivial if and only if its null index vanishes.
Positive-definiteness has its usual elementwise characterization.
A lattice presented by a Gram matrix is positive definite exactly when the matrix is,
read over ℚ. The form of ofGramMatrix b G hG is the bilinear form of G in the coordinates
of b, so this is Mathlib's comparison of a quadratic map with the matrix of its bilinear form
in a basis.
Positive-semidefiniteness has its usual elementwise characterization.
Negative-definiteness has its usual elementwise characterization.
Negative-semidefiniteness has its usual elementwise characterization.
Positive-semidefiniteness is equivalent to the vanishing of the negative index.
Negative-semidefiniteness is equivalent to the vanishing of the positive index.
Positive-definiteness is positive-semidefiniteness together with nondegeneracy.
A positive-definite lattice is positive-semidefinite.
Positive-definiteness is equivalent to zero null and negative indices.
Negative-definiteness is negative-semidefiniteness together with nondegeneracy.
Negative-definiteness is equivalent to zero positive and null indices.
Degeneracy is equivalent to a positive null index.
A lattice is degenerate exactly when its bilinear form is not nondegenerate.
A lattice is degenerate exactly when its radical contains a nonzero vector.
Indefiniteness is equivalent to the failure of both semidefiniteness conditions.
An indefinite lattice has, and is characterized by, vectors of both signs.
An integral-lattice isometry maps the source radical onto the target radical.
Isometric integral lattices have the same positive index.
Isometric integral lattices have the same null index.
Isometric integral lattices have the same negative index.
Isometric integral lattices have the same signature.
Nondegeneracy of the ambient form is invariant under integral-lattice isometry.
The nondegeneracy mixin is invariant under integral-lattice isometry.
Transport the nondegeneracy mixin along an integral-lattice isometry.
Positive-definiteness is invariant under integral-lattice isometry.
Positive-semidefiniteness is invariant under integral-lattice isometry.
Negative-definiteness is invariant under integral-lattice isometry.
Negative-semidefiniteness is invariant under integral-lattice isometry.
Degeneracy is invariant under integral-lattice isometry.
Indefiniteness is invariant under integral-lattice isometry.