Tensor multiplicities of irreducible modules #
Let L be a finite-dimensional Lie algebra with non-degenerate Killing form over an algebraically
closed field of characteristic zero, H a Cartan subalgebra and b a base of its root system.
The tensor multiplicity TauCeti.tensorMultiplicity b lam mu nu is the dimension of the space
of morphisms L(nu) →ₗ⁅K,L⁆ L(lam) ⊗ L(mu). For dominant integral lam and mu the tensor
product is a finite-dimensional L-module, so Weyl's theorem decomposes it into irreducibles, and
at a nu with L(nu) nonzero, hence irreducible, that dimension is the number of copies of
L(nu) in the decomposition, the structure constant c^nu_{lam mu}. The definition itself, and
the lemmas reading irreducibility off a nonzero multiplicity, ask only for a triangularizable
Cartan subalgebra; algebraic closure enters with the character identity, which calls on Weyl's
complete reducibility theorem.
The theorem about it is the character identity
ch L(lam) · ch L(mu) = ∑_nu c^nu_{lam mu} · ch L(nu),
which combines two facts already in place: formal characters are multiplicative on tensor products
(TauCeti.formalCharacter_tensor), and the character of a finite-dimensional module is the
multiplicity-weighted sum of the irreducible characters
(TauCeti.formalCharacter_eq_finsum_isotypicMultiplicity_smul). It is what makes the tensor
multiplicities computable from characters, and the identity a Pieri or Littlewood-Richardson rule
evaluates.
Every L(nu) is irreducible, the Verma module M(nu) being nonzero
(TauCeti.vermaGenerator_ne_zero), so the identity is a statement about honest irreducibles, and
it never degenerates to 0 = 0: some c^nu_{lam mu} is nonzero
(TauCeti.exists_tensorMultiplicity_ne_zero), because L(lam) ⊗ L(mu) is a nonzero
finite-dimensional module.
Main definitions #
TauCeti.tensorMultiplicity: the multiplicityc^nu_{lam mu}ofL(nu)inL(lam) ⊗ L(mu).
Main results #
TauCeti.irreducibleFormalCharacter_mul_eq_finsum_tensorMultiplicity_smul: the character identitych L(lam) · ch L(mu) = ∑_nu c^nu_{lam mu} · ch L(nu).TauCeti.exists_tensorMultiplicity_ne_zero: some tensor multiplicity is nonzero, so the character identity is not a statement about zero modules.
References #
This is the tensor-multiplicity item of the milestone "tensor multiplicities and the minuscule
Pieri rule" in the Layer 6 decomposition toolkit of
TauCetiRoadmap/RepresentationTheory/LieHighestWeight/README.md, which asks for c^ν_{λμ} "with
the character identity ch L(λ) · ch L(μ) = Σ_ν c^ν_{λμ} ch L(ν) through formalCharacter_tensor".
The minuscule Pieri rule itself, which evaluates these multiplicities, awaits the Weyl character
formula.
- J. E. Humphreys, Introduction to Lie Algebras and Representation Theory, GTM 9, Ch. VI, §24.
The tensor multiplicity c^nu_{lam mu}: the multiplicity of L(nu) in
L(lam) ⊗ L(mu) in the sense of LieModule.isotypicMultiplicity, that is the dimension of the
space of morphisms L(nu) →ₗ⁅K,L⁆ L(lam) ⊗ L(mu).
It is the number of copies of L(nu) in a decomposition of the tensor product into irreducibles
when that reading is available: lam and mu dominant integral, so that the tensor product is
finite-dimensional and completely reducible, L(nu) being irreducible.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The tensor multiplicity is the multiplicity of L(nu) in the tensor product.
The character identity #
The character of L(lam) is the character of a decomposition into irreducibles, so from here on
the field is algebraically closed: that is what Weyl's complete reducibility theorem is available
over in TauCeti/Algebra/Lie/HighestWeight/Decomposition.lean.
The character identity for a tensor product of highest weight modules: the product of the
characters of L(lam) and L(mu) is the sum of the characters of the L(nu), weighted by the
tensor multiplicities.
Formal characters are multiplicative on tensor products, so the left-hand side is the character of
L(lam) ⊗ L(mu); that module is finite-dimensional, so its character is the
multiplicity-weighted sum of the irreducible characters.
The character identity is never a statement about zero modules. L(lam) ⊗ L(mu) is a
nonzero finite-dimensional module, so at least one tensor multiplicity is nonzero, and the sum
ch L(lam) · ch L(mu) = ∑_nu c^nu_{lam mu} · ch L(nu) has a term that survives.