The weights of a Verma module and their multiplicities #
Let L be a finite-dimensional Lie algebra with non-degenerate Killing form over a field K of
characteristic zero, H a splitting Cartan subalgebra and b a base of its root system. The
Verma module M(lam) is a free U(n⁻)-module of rank one on its canonical generator v_lam
(TauCeti.universalEnvelopingEquivVermaModule), so a Poincaré--Birkhoff--Witt basis of U(n⁻)
gives a basis of M(lam): the vectors
f_{i₁} ⋯ f_{iₖ} · v_lam, i₁ ≤ ⋯ ≤ iₖ,
for an ordered basis (f_i) of the negative nilradical n⁻. When every f_i is an
H-eigenvector of weight w_i, the vector attached to the exponent n is an H-eigenvector of
weight lam + ∑ᵢ nᵢ wᵢ, so the weight space M(lam)_mu is spanned by the basis vectors of that
weight and its dimension is the number of exponents n with lam + ∑ᵢ nᵢ wᵢ = mu
(TauCeti.finrank_weightSpace_vermaModule_eq_natCard).
The negative nilradical has such a basis made of root vectors, one for each positive root α,
lying in the line L_{-α} (TauCeti.negativeNilradicalBasis). With it the count becomes the
number of ways of writing lam - mu as a sum of positive roots with multiplicity, which is the
Kostant partition function: the weight multiplicities of a Verma module are
dim M(lam)_mu = P(lam - mu)
(TauCeti.finrank_weightSpace_vermaModule). In particular the weights of M(lam) are exactly the
elements of lam minus the monoid generated by the positive roots
(TauCeti.weightSpace_vermaModule_ne_bot_iff), and M(lam) is the direct sum of its weight
spaces (TauCeti.isInternal_weightSpace_vermaModule).
Main definitions #
TauCeti.vermaBasis: the basis ofM(lam)obtained by applying the ordered monomials in a basis ofn⁻to the canonical generator.
Main results #
TauCeti.vermaBasis_mem_weightSpace: for a basis ofn⁻made of weight vectors, every vector of the associated basis ofM(lam)is a weight vector, of the weight read off its exponent.TauCeti.weightSpace_vermaModule_eq_spanandTauCeti.finrank_weightSpace_vermaModule_eq_natCard: a weight space ofM(lam)is spanned by the basis vectors of that weight, so its dimension counts the exponents of that weight.TauCeti.finrank_weightSpace_vermaModule:dim M(lam)_mu = P(lam - mu), the Kostant partition function oflam - mu.TauCeti.weightSpace_vermaModule_ne_bot_iff:muis a weight ofM(lam)exactly whenlam - muis a sum of positive roots.TauCeti.iSup_weightSpace_vermaModule_eq_topandTauCeti.isInternal_weightSpace_vermaModule: the Verma module is the sum, indeed the internal direct sum, of its weight spaces.
References #
- J. E. Humphreys, Introduction to Lie Algebras and Representation Theory, GTM 9, §20.3 and §24.1.
- J. E. Humphreys, Representations of Semisimple Lie Algebras in the BGG Category
O, §1.3.
The Poincaré--Birkhoff--Witt basis of a Verma module #
The Poincaré--Birkhoff--Witt basis of the Verma module attached to an ordered basis B of
the negative nilradical: the vector of exponent n is the ordered monomial of
Module.Basis.pbwBasis in B, with exponent n, applied to the canonical generator
v_lam (TauCeti.vermaBasis_apply). It is a basis because M(lam) is a free U(n⁻)-module of
rank one on v_lam.
Equations
- TauCeti.vermaBasis b lam B = B.pbwBasis.map (TauCeti.universalEnvelopingEquivVermaModule b lam)
Instances For
The basis vector of exponent n is the ordered monomial of exponent n in B, applied to the
canonical generator.
The basis vectors of a Verma module are weight vectors. If the Cartan subalgebra acts on
each vector B i of the basis of n⁻ through w i, then the basis vector of exponent n has
weight lam + ∑ᵢ nᵢ w i.
A weight space of a Verma module is spanned by the basis vectors of that weight. If the
Cartan subalgebra acts on each vector B i of the basis of n⁻ through w i, the weight space
M(lam)_mu is spanned by the basis vectors whose exponent n satisfies lam + ∑ᵢ nᵢ w i = mu.
The weight multiplicities of a Verma module count exponents. If the Cartan subalgebra acts
on each vector B i of the basis of n⁻ through w i, then the weight space M(lam)_mu has
dimension the number of exponents n with lam + ∑ᵢ nᵢ w i = mu. When there are infinitely many
such exponents both sides are 0.
The weights of a Verma module. If the Cartan subalgebra acts on each vector B i of the
basis of n⁻ through w i, then mu is a weight of M(lam) exactly when it is
lam + ∑ᵢ nᵢ w i for some exponent n.
The multiplicities are given by the Kostant partition function #
The weight multiplicities of a Verma module are given by the Kostant partition function:
the weight space M(lam)_mu has dimension P(lam - mu), the number of ways of writing lam - mu
as a sum of positive roots with multiplicity.
The weights of a Verma module: mu is a weight of M(lam) exactly when lam - mu is a
sum of positive roots with multiplicity.
A Verma module is the sum of its weight spaces.
The weight-space decomposition of a Verma module: M(lam) is the internal direct sum of
its weight spaces.