Finite-dimensional normed spaces #
Over a complete nontrivially normed field every linear functional on a finite-dimensional normed
space is continuous, so the continuous dual E →L[𝕜] 𝕜 coincides with the algebraic dual
Module.Dual 𝕜 E. Mathlib records that coincidence as the linear equivalence
LinearMap.toContinuousLinearMap; this file reads off the one consequence of it that dimension
counts need, namely that the continuous dual has the same dimension as the space.
Two complements of the same space in a finite-dimensional space are continuously linearly isomorphic. This allows pointwise choices of complements to be identified with a single model.
Main results #
ContinuousLinearMap.dual_finrank_eq: in finite dimensions the continuous dual has the same dimension as the space.TauCeti.nonempty_continuousLinearEquiv_of_prod_continuousLinearEquiv: complements of the same space in a finite-dimensional space are isomorphic.TauCeti.continuous_det_div_conj: the determinant phasedet A / conj (det A)of a continuous family of continuous linear automorphisms varies continuously.
In finite dimensions the continuous dual E →L[𝕜] 𝕜 has the same dimension as E: every
linear functional on a finite-dimensional space is continuous, so the continuous dual coincides
with the algebraic one.
Over a complete field, two complements of the same space V in a finite-dimensional space W
are continuously linearly isomorphic: both have dimension finrank W - finrank V.
The determinant phase det A / conj (det A) of a continuous family of continuous linear
automorphisms varies continuously. Over ℂ this is the Maslov phase ρ(L, A L) of a maximal
totally real subspace L.