The adjoint action on the tangent space #
The points of a Hopf algebra act on the counit-valued derivations — the tangent
vectors at the identity — by convolution conjugation: Ad g d = g ⋆ d ⋆ g⁻¹ in the
convolution semiring of linear maps, the differential of the conjugation
c_g x = g ⋆ x ⋆ g⁻¹ of the group of points. Conjugation is a semiring automorphism of
the whole convolution semiring; this file shows it restricts to the derivations, and
packages the restriction, one coefficient algebra B at a time, as a representation of
the B-points on the B-valued tangent vectors (adRepresentation). For commutative
A these are the coefficientwise layers of the adjoint action of the corresponding
affine group scheme on its Lie functor; their compatibility as B varies
(postcomposition naturality) is later infrastructure, not packaged here.
Closure is composition-level, by the exterior-product calculus of Tangent.Basic: an
algebra-map point satisfies g ∘ mul = g ⊠ g, a derivation satisfies
d ∘ mul = e ⊠ d + d ⊠ e for the convolution unit e, and the conjugates collapse
by g ⋆ e ⋆ g⁻¹ = e, leaving the Leibniz form for g ⋆ d ⋆ g⁻¹. No antipode
computation appears; inverses come from the group of points.
Main declarations #
Derivation.adDerivation: the conjugate of a tangent vector by a point.Derivation.toConv_coe_adDerivation: that conjugate as the convolution productg ⋆ d ⋆ g⁻¹, the form the algebraic manipulations use.Derivation.snd_conjugate_tangentPoint: conjugation on dual-number tangent points induces the adjoint action on their infinitesimal coefficients.Derivation.adRepresentation: the adjoint action, as a representation of the convolution group of points on the tangent space.
Compatibility of the action with the Lie bracket needs the Lie structure and so lives in
Tangent.Lie.Adjoint.Basic, keeping this module free of it.
The action stays on the Lie functor B ↦ Derivation R A (CounitAlgebra R A B); at
each B it is a genuine representation. Identifying the functor's value at B with
B ⊗ Lie(G)(R) — the classical fixed-module G → GL(Lie G) — needs a
finite-projectivity hypothesis on the conormal module and is not attempted here.
References #
- J. S. Milne, Algebraic Groups (2017), §14 (the adjoint representation).
- J. C. Jantzen, Representations of Algebraic Groups, I.7.18.
The conjugate of a tangent vector by a point: the adjoint action
Ad g d = g ⋆ d ⋆ g⁻¹, the differential of conjugation on the group of points.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The adjoint action is the convolution conjugate, on underlying linear maps.
The adjoint action, valuewise: the conjugate evaluated at an element of the bialgebra.
The conjugate of a tangent vector, in convolution form: Ad g d is the product
g ⋆ d ⋆ g⁻¹ in the convolution algebra of linear maps. This is the form in which the
adjoint action is manipulated algebraically, and the bracket compatibility in
Tangent.Lie.Adjoint.Basic is proved from it.
The adjoint action of the group of points on the tangent space, as a
B-linear representation: conjugation is a semiring automorphism of the convolution
semiring, it restricts to the derivations, and scalars of the coefficient algebra pass
through the conjugation because B is commutative.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Conjugating a tangent point by a constant dual-number point induces the adjoint action on its infinitesimal coefficient.