Building upper-unitriangular bases from fixed vectors #
Suppose every nonzero finite-dimensional comodule over a coalgebra with a distinguished element
1 has a nonzero fixed vector, that is, a vector v with coaction v ↦ v ⊗ 1. Then every
finite-dimensional comodule has a basis whose coefficient matrix is upper unitriangular: this is
the weight-vector induction of TauCeti.Algebra.Coalgebra.Comodule.Flag.Triangular with the
weights confined to {1}.
This is the fixed-vector case of the induction common to the Kolchin arguments in Layer 5 of the ReductiveGroups roadmap. Lie–Kolchin supplies eigenlines for solvable groups; for a unipotent group every resulting character is trivial, so the lines are fixed. The theorem here turns that fixed-vector statement into the complete flag needed to embed a faithful representation into an upper-unitriangular group.
Main declarations #
TauCeti.Comodule.HasNonzeroFixedVector: a comodule contains a nonzero vector with coactionv ↦ v ⊗ 1.TauCeti.Comodule.hasNonzeroFixedVector_iff_fixedSubcomodule_ne_bot: that happens exactly when the fixed subcomodule is nonzero.TauCeti.Comodule.exists_basis_coefficientMatrix_isUpperUnitriangular_of_fixed_vectors: if every nonzero finite-dimensional comodule has such a vector, every finite-dimensional comodule admits an upper-unitriangular basis.
References #
- J. C. Jantzen, Representations of Algebraic Groups, I.2.
- T. A. Springer, Linear Algebraic Groups, §2.4.
A comodule has a nonzero fixed vector if some nonzero v has coaction v ⊗ 1.
For the comodule corresponding to a group representation, this says that the represented group
fixes v.
Equations
- TauCeti.Comodule.HasNonzeroFixedVector k H M = ∃ (v : M), v ≠ 0 ∧ TauCeti.Comodule.coact v = v ⊗ₜ[k] 1
Instances For
The defining characterization of a nonzero fixed vector.
A comodule has a nonzero fixed vector exactly when its fixed subcomodule is nonzero.
If every nonzero finite-dimensional H-comodule has a nonzero fixed vector, then every
finite-dimensional H-comodule has a basis with upper-unitriangular coefficient matrix.
The hypothesis is deliberately uniform in the comodule: the induction applies it to successive
quotients. The conclusion allows an arbitrary finite index n; the exhibited basis itself
certifies that n is the dimension of M.