The Setting
Take a natural number and a continuous path () of maps from real lower triangular matrices to real symmetric matrices, starting from a prescribed quadratic map at . When is equivariant under row sign reversals, has a compact zero locus on , and vanishes only at matrices with nonzero rows, must have a lower triangular solution ?
Alongside this matrix continuation problem, for a natural number , the lemma formulates an equivariant continuation principle on under coordinate reflections that fix the -st coordinate, using orbit balls, free action loci, and affine orbit models. Instead of smooth manifold degree theory, it states a finite combinatorial piecewise-linear parity principle on cubical flags under coordinate reflections, together with terminal-zero continuation results on for equivariant maps with affine, locally approximate, or simple initial zero orbits.
Lower Triangular Matrices and Sign Actions
For a natural number , the lower triangular matrix space is the space of real matrices , indexed on rows and columns by , whose entry vanishes whenever .✦ For a natural number , a matrix has nonzero rows if for every row index there exists a column index with .✦
For a vector and a matrix , the row sign action is the lower triangular matrix defined entrywise by (1): (1)✦
The sign conjugation of an real matrix by is the matrix specified by (2): (2)✦
Cubical Flags and Orbit Models
For a natural number , the coordinate reflection group is the group of -tuples of reflection indicators, where leaves a coordinate unreflected and reflects it.✦
Given a group acting on a metric space , a point , and , the group orbit ball is the set .✦ The free action locus of a group acting on a set is the set .✦ Relative to and for an action of a group on a metric space , the orbit ball label is the map that sends each to an element satisfying , and every to .✦
In dimension , the spatial inclusion map is the linear map appending as the -st coordinate, .✦ The spatial projection map for is the linear map dropping the -st coordinate, sending to .✦ Given , a subgroup , a real vector space with a linear representation and a linear isomorphism , together with , , and , the equivariant affine model is the map defined for by (3): (3)✦
A cubical flag in dimension is a triple consisting of a base point , a corner indicator , and a permutation of .✦ For a cubical flag in dimension and , the integer flag vertex is the point whose coordinates are given by (4): (4) The real flag vertex of a cubical flag in dimension at is the point . Given , , a cubical flag in dimension , and weights , the flag convex combination is the vector defined by (5): (5)✦
For a cubical flag in dimension and , the corner flip is the cubical flag obtained from by replacing with .✦ The adjacent cube crossing of a cubical flag in dimension along coordinate is the cubical flag obtained from by replacing with and with .✦ Given and a cubical flag in dimension , the order swap is the cubical flag , where transposes and .✦ When , the facet neighbor flag of a cubical flag in dimension across facet is the cubical flag defined by (6): (6)✦
The bottom seed flag for , a nonzero step , and is the cubical flag in dimension with base , corner indicator , and identity permutation .✦ The mesh grid point for , , and is the point . Given , a subgroup , and an -invariant set of cubical flags in dimension , the invariant flag vertex set is the -invariant subset .✦
The Claim
Theorem 1. In the first three parts, let , let be a subgroup with for all , let be a finite-dimensional real normed vector space with an isometric linear representation and a linear isomorphism , and let . Assume satisfies for all and , every with and lies in the orbit , the slab zero locus is compact, and every zero in this locus belongs to .
If is continuous on , , and the map on has Fréchet derivative at , then (7)✦
If is continuous on , a real radius with satisfies for all distinct , , and every with satisfies for , then (8)✦
If is continuous on , a real radius satisfies for all distinct , and for every with , where , then (9)✦
Let , let be a subgroup with for all , let be a finite -invariant set of cubical flags in dimension , let be a finite-dimensional real normed vector space with a linear representation and a linear isomorphism , and let and satisfy and for all distinct . Assume satisfies for all and , the family is linearly independent in for every and , and is linearly independent in for all and distinct . Assume further that for all , that whenever and satisfy and , that with , and that there exist weights with , for all , , and . Then there exist , with , and weights with for all , , and such that (10)✦
Let , , and with for all . Suppose , , is continuous on , and for every and the matrix is real symmetric and satisfies for all . Assume that for all , that the zero locus is compact, and that every solving for some has nonzero rows. Then (11)✦
In the paper, the result reads: for a smooth -dimensional manifold without boundary and a smooth real vector bundle of rank , if () is a continuous homotopy of sections whose zero locus is compact, and if is smooth and has an odd number of zeros, all transverse, then has a zero.
What Is Checked
For , claims (7), (8), (9), and (10) are stated over finite-dimensional real normed vector spaces with linear representations of subgroups of fixing the last coordinate, orbit balls, free action loci, equivariant affine models, and finite invariant cubical flag complexes with their vertices, mesh grid points, bottom seed flags, and facet neighbors, while claim (11) is stated over the lower triangular matrix space with continuous symmetric-matrix paths equivariant under row sign actions and sign conjugations together with the nonzero-rows condition.