Documentation

Atlas.AlgebraicTopologyI.code.Section30

def SymmetricBilinearForms.MatrixCongruent {n : Type u_1} [Fintype n] [DecidableEq n] {R : Type u_2} [CommRing R] (M N : Matrix n n R) :

Two square matrices M, N over a commutative ring are congruent if there exists an invertible matrix A such that A * M * Aᵀ = N. This is the matrix-level equivalence relation underlying isometry of bilinear forms.

Instances For

    The matrix H ⊕ ⟨1⟩ over F₂, formed by the hyperbolic plane and a rank-one form with diagonal entry 1. This is the model form used in Claim 30.7.

    Instances For

      Claim 30.7: the form H ⊕ ⟨1⟩ is congruent over F₂ to the identity form ⟨1⟩ ⊕ ⟨1⟩ ⊕ ⟨1⟩.

      Definition 30.1: a topological manifold is a Hausdorff space that is locally homeomorphic to some Euclidean space at every point (the dimension may a priori depend on the point).

      Instances

        A topological n-manifold: a Hausdorff space equipped with a charted space structure modelled on ℝⁿ. This is the dimension-fixed variant of IsTopologicalManifold.

        Instances
          @[implicit_reducible, instance 100]

          A Hausdorff charted space modelled on ℝⁿ automatically gets the TopologicalManifold n instance.

          @[reducible, inline]

          Singular cohomology with coefficients in F₂ = ZMod 2, as a type.

          Instances For
            @[reducible, inline]
            abbrev HomolZMod2 (X : TopCat) (n : ) :

            Singular homology with coefficients in F₂ = ZMod 2, as a type.

            Instances For
              noncomputable def cupProductPairing (X : TopCat) (dim : ) (μ : HomolZMod2 X dim) (p q : ) (hpq : p + q = dim) :

              The pairing H^p(X; F₂) × H^q(X; F₂) → F₂ induced by cup product with a fixed homology class μ ∈ H_dim(X; F₂), given by (α, β) ↦ ⟨α ⌣ β, μ⟩. Used to formulate Poincaré duality.

              Instances For
                theorem poincareDuality_F2 (dim : ) (M : Type) [TopologicalSpace M] [ChartedSpace (EuclideanSpace (Fin dim)) M] [T2Space M] [CompactSpace M] :
                ∃! μ : HomolZMod2 (TopCat.of M) dim, ∀ (p q : ) (hpq : p + q = dim), (cupProductPairing (TopCat.of M) dim μ p q hpq).IsPerfPair

                Theorem 30.2 (F₂-Poincaré duality). On a closed (compact, Hausdorff) topological dim-manifold there exists a unique fundamental class μ ∈ H_dim(M; F₂) such that, for every p + q = dim, the cup-product pairing H^p(M; F₂) × H^q(M; F₂) → F₂ is a perfect pairing.

                A compact surface: a compact, connected, Hausdorff topological space with a charted-space structure on ℝ² that makes it a C⁰ manifold. This is the geometric setting for the classification of surfaces.

                Instances For

                  The underlying topological space of the connected sum S₁ ♯ S₂ of two compact surfaces: remove an open disk from each surface and glue along the resulting boundary circles.

                  Instances For

                    The connected sum of two compact surfaces is compact.

                    The connected sum of two compact surfaces is connected.

                    The connected sum of two compact surfaces is Hausdorff.

                    A charted space structure on the connected sum modelled on ℝ².

                    Instances For

                      The connected sum S₁ ♯ S₂ of two compact surfaces, packaged as a CompactSurface.

                      Instances For

                        The fundamental class [S] ∈ H_2(S; F₂) of a compact surface, provided by F₂-Poincaré duality.

                        Instances For

                          The intersection form on H¹(S; F₂) of a compact surface S: the symmetric bilinear form (α, β) ↦ ⟨α ⌣ β, [S]⟩.

                          Instances For
                            def SurfacesAndBilinearForms.orthogonalDirectSumForm {V₁ V₂ : Type} [AddCommGroup V₁] [Module (ZMod 2) V₁] [AddCommGroup V₂] [Module (ZMod 2) V₂] (B₁ : LinearMap.BilinForm (ZMod 2) V₁) (B₂ : LinearMap.BilinForm (ZMod 2) V₂) :
                            LinearMap.BilinForm (ZMod 2) (V₁ × V₂)

                            The orthogonal direct sum B₁ ⊥ B₂ of two F₂-bilinear forms on the product space V₁ × V₂, defined by ((x₁, x₂), (y₁, y₂)) ↦ B₁(x₁, y₁) + B₂(x₂, y₂).

                            Instances For

                              The punctured surface S ∖ B: remove a small open disk from S using a chart centred at an arbitrary point.

                              Instances For

                                The continuous inclusion S ∖ B ↪ S of the punctured surface into the original surface.

                                Instances For

                                  The inclusion of the punctured surface into the surface induces an isomorphism on H¹(–; F₂) (removing an open disk does not change ).

                                  The induced linear isomorphism H¹(S ∖ B; F₂) ≃ₗ H¹(S; F₂).

                                  Instances For

                                    Inclusion of the first punctured factor S₁ ∖ B₁ into the connected sum S₁ ♯ S₂.

                                    Instances For

                                      Inclusion of the second punctured factor S₂ ∖ B₂ into the connected sum S₁ ♯ S₂.

                                      Instances For

                                        Mayer–Vietoris restriction map on for the connected sum: send a cohomology class to its pair of restrictions to the two punctured factors.

                                        Instances For

                                          The Mayer–Vietoris connecting homomorphism in degree zero, H⁰(S¹; F₂) ≅ F₂ → H¹(S₁ ♯ S₂; F₂).

                                          Instances For

                                            Exactness of the Mayer–Vietoris sequence at the H¹(S₁ ♯ S₂) spot: the image of the connecting homomorphism δ⁰ equals the kernel of the restriction map.

                                            The Mayer–Vietoris restriction map in degree zero, which on F₂ × F₂ → F₂ simply adds the two components.

                                            Instances For

                                              Exactness at the H⁰(S¹; F₂) spot of the Mayer–Vietoris sequence: the image of H⁰(S₁∖B) ⊕ H⁰(S₂∖B) → H⁰(S¹) equals the kernel of δ⁰.

                                              The degree-zero restriction map F₂ × F₂ → F₂ (addition) is surjective.

                                              Since the preceding restriction map is surjective, the connecting homomorphism δ⁰ : H⁰(S¹) → H¹(S₁ ♯ S₂) is the zero map.

                                              Because δ⁰ = 0, exactness at H¹(S₁ ♯ S₂) forces the restriction map to be injective.

                                              The Mayer–Vietoris map H¹(S₁∖B) × H¹(S₂∖B) → H¹(S¹) given by restricting to the boundary circle .

                                              Instances For

                                                Exactness at the H¹(S₁∖B) × H¹(S₂∖B) spot of the Mayer–Vietoris sequence: an element of the product is in the image of the restriction map iff it is annihilated by the boundary-restriction map.

                                                The Mayer–Vietoris connecting homomorphism in degree one, H¹(S¹; F₂) → H²(S₁ ♯ S₂; F₂).

                                                Instances For

                                                  Exactness of the Mayer–Vietoris sequence at the H¹(S¹; F₂) spot: the image of γ equals the kernel of δ¹.

                                                  The degree-one connecting homomorphism δ¹ : H¹(S¹) → H²(S₁ ♯ S₂) is injective (it is the dual to the inclusion of the fundamental class).

                                                  Since δ¹ is injective, exactness at H¹(S¹) forces the boundary-restriction map γ to vanish identically.

                                                  Since γ = 0, exactness at the product term implies the restriction map H¹(S₁ ♯ S₂) → H¹(S₁∖B) × H¹(S₂∖B) is surjective.

                                                  Combine injectivity and surjectivity: the restriction map is bijective.

                                                  Mayer–Vietoris isomorphism (punctured version): H¹(S₁ ♯ S₂; F₂) ≃ H¹(S₁∖B; F₂) × H¹(S₂∖B; F₂).

                                                  Instances For

                                                    Proposition 30.8: the Mayer–Vietoris isomorphism H¹(S₁ ♯ S₂; F₂) ≃ H¹(S₁; F₂) × H¹(S₂; F₂), obtained by combining the punctured-surface isomorphism with the isomorphisms H¹(Sᵢ∖B; F₂) ≃ H¹(Sᵢ; F₂).

                                                    Instances For

                                                      Compatibility of the intersection form with the Mayer–Vietoris decomposition: the intersection form on S₁ ♯ S₂ decomposes as the sum of the pulled-back intersection forms on S₁ and S₂.

                                                      Under the Mayer–Vietoris isomorphism, the intersection form of S₁ ♯ S₂ is the orthogonal direct sum of the intersection forms of S₁ and S₂.

                                                      Existential form of Proposition 30.8: there exists a linear isomorphism H¹(S₁ ♯ S₂; F₂) ≃ H¹(S₁; F₂) × H¹(S₂; F₂) under which the intersection form of S₁ ♯ S₂ equals the orthogonal direct sum of those of S₁, S₂.

                                                      Two compact surfaces are homeomorphic if there exists a homeomorphism between their underlying topological spaces.

                                                      Instances For

                                                        The set of homeomorphism classes of compact surfaces, as a quotient of CompactSurface by the homeomorphism relation.

                                                        Instances For

                                                          The canonical map sending a compact surface to its homeomorphism class.

                                                          Instances For
                                                            @[implicit_reducible]

                                                            The connected sum endows the set of homeomorphism classes of compact surfaces with the structure of a commutative monoid.

                                                            A nondegenerate symmetric bilinear form over F₂ packaged with its underlying finite-dimensional F₂-vector space.

                                                            Instances For

                                                              Two nondegenerate symmetric bilinear forms over F₂ are isometric if there is a linear isomorphism of the underlying vector spaces that takes one form to the other.

                                                              Instances For

                                                                The set of isometry classes of nondegenerate symmetric bilinear forms over F₂.

                                                                Instances For
                                                                  @[implicit_reducible]

                                                                  The orthogonal direct sum endows the set of isometry classes of nondegenerate symmetric F₂-bilinear forms with the structure of a commutative monoid.

                                                                  The canonical map sending a nondegenerate symmetric F₂-bilinear form to its isometry class.

                                                                  Instances For

                                                                    The intersection form on H¹(S; F₂) of a compact surface is symmetric (over characteristic 2 the cup product is graded-commutative without sign).

                                                                    The intersection form on H¹(S; F₂) of a compact surface is nondegenerate, as a consequence of Poincaré duality.

                                                                    The first F₂-cohomology of a compact surface is finite-dimensional.

                                                                    Package the intersection form of a compact surface as a nondegenerate symmetric F₂-bilinear form (on the finite-dimensional space H¹(S; F₂)).

                                                                    Instances For

                                                                      The map sending a compact surface to the isometry class of its intersection form.

                                                                      Instances For

                                                                        Homeomorphic compact surfaces have isometric intersection forms, so the map descends to a function on surface classes.

                                                                        The induced map on homeomorphism classes of surfaces: send a class [S] to the isometry class of its intersection form on H¹(S; F₂).

                                                                        Instances For

                                                                          Theorem 30.9: the intersection form yields a monoid isomorphism between the monoid of homeomorphism classes of compact surfaces (under connected sum) and the monoid of isometry classes of nondegenerate symmetric F₂-bilinear forms (under orthogonal direct sum).

                                                                          Instances For

                                                                            For a reflexive bilinear form B, the restriction B|_W is nondegenerate iff W and its orthogonal complement intersect only at zero.

                                                                            theorem BilinearForm.restrict_orthogonal_nondegenerate {K : Type u_1} {V : Type u_2} [Field K] [AddCommGroup V] [Module K V] [FiniteDimensional K V] (B : LinearMap.BilinForm K V) (hB : B.Nondegenerate) (hB_refl : B.IsRefl) {W : Submodule K V} (hW : (B.restrict W).Nondegenerate) :

                                                                            If B is nondegenerate and reflexive and B|_W is nondegenerate, then the restriction of B to the orthogonal complement W^⊥ is also nondegenerate.

                                                                            noncomputable def BilinearForm.orthogonalProdEquiv {K : Type u_1} {V : Type u_2} [Field K] [AddCommGroup V] [Module K V] [FiniteDimensional K V] (B : LinearMap.BilinForm K V) (hB_refl : B.IsRefl) {W : Submodule K V} (hW : (B.restrict W).Nondegenerate) :
                                                                            (W × (B.orthogonal W)) ≃ₗ[K] V

                                                                            When B|_W is nondegenerate, V decomposes as the internal direct sum W ⊕ W^⊥, giving a linear equivalence W × W^⊥ ≃ₗ V.

                                                                            Instances For
                                                                              theorem BilinearForm.orthogonalProdEquiv_respects_form {K : Type u_1} {V : Type u_2} [Field K] [AddCommGroup V] [Module K V] [FiniteDimensional K V] (B : LinearMap.BilinForm K V) (hB_refl : B.IsRefl) {W : Submodule K V} (hW : (B.restrict W).Nondegenerate) (x y : W × (B.orthogonal W)) :
                                                                              (B ((orthogonalProdEquiv B hB_refl hW) x)) ((orthogonalProdEquiv B hB_refl hW) y) = (B x.1) y.1 + (B x.2) y.2

                                                                              Under the direct-sum decomposition V ≃ W × W^⊥, the form B becomes the orthogonal direct sum of B|_W and B|_{W^⊥}; the cross terms vanish because they lie in the orthogonal complement.

                                                                              theorem BilinearForm.restrict_nondegenerate_characterization {K : Type u_1} {V : Type u_2} [Field K] [AddCommGroup V] [Module K V] [FiniteDimensional K V] (B : LinearMap.BilinForm K V) (hB : B.Nondegenerate) (hB_refl : B.IsRefl) (W : Submodule K V) :
                                                                              ((B.restrict W).Nondegenerate Disjoint W (B.orthogonal W)) ((B.restrict W).Nondegenerate(B.restrict (B.orthogonal W)).Nondegenerate) ((B.restrict W).Nondegenerate∃ (e : (W × (B.orthogonal W)) ≃ₗ[K] V), ∀ (x y : W × (B.orthogonal W)), (B (e x)) (e y) = (B x.1) y.1 + (B x.2) y.2)

                                                                              Lemma 30.5 (packaged form): a triple characterisation of nondegeneracy of the restriction B|_W of a nondegenerate reflexive bilinear form, combining the disjointness criterion, the nondegeneracy of the restriction to the orthogonal complement, and the orthogonal-direct-sum decomposition of V together with the corresponding factorisation of B.