The weight decomposition of Symᵈ(ℂ²) under the maximal torus of SU(2) #
TauCeti/RepresentationTheory/SU2/SymmetricPower.lean computes the character of the symmetric
power Symᵈ(ℂ²) of the standard representation of SU(2) on the maximal torus as the weight
string z^{-d} + z^{2-d} + ⋯ + z^d. A character is a trace, so it records the weights only with
multiplicity; this file builds the decomposition behind it.
The monomial basis of Symᵈ(ℂ²), reindexed by Fin (d + 1) through
TauCeti.symFinTwoEquiv, is a basis of weight vectors: diag(z, z⁻¹) acts on the i-th one
by the scalar z^{2i - d}. The d + 1 weights 2i - d are pairwise distinct, so each weight
occurs with multiplicity exactly one, and the character above,
TauCeti.SU2.character_symPower_torusHom_zpow, is their sum.
Two consequences make this the form the highest-weight classification uses.
- A torus-stable subspace is spanned by weight vectors. Nothing forces this for a single
operator with repeated eigenvalues; here the eigenvalues are distinct, which is exactly the
hypothesis of the general coordinate argument
TauCeti/LinearAlgebra/Eigenspace/DiagonalBasis.lean. Feeding it a torus element whose integer powers are pairwise distinct —exp(i), by the irrationality ofπ— separates thed + 1weights at once. That element is a device of the proof and stays private to this file. - The weight spaces are lines. A nonzero vector on which the whole torus acts through a
single character
z ↦ z^mis a multiple of one basis vector, andmis one of thed + 1weights.
Main definitions #
TauCeti.SU2.weightBasis: the weight basis ofSymᵈ(ℂ²), indexed byFin (d + 1).TauCeti.SU2.weight: the weight2i - dof thei-th weight vector.
Main results #
TauCeti.SU2.symPower_torusHom_weightBasis:diag(z, z⁻¹)acts on thei-th weight vector byz^{2i - d}, andTauCeti.SU2.weight_injective: the weights are pairwise distinct.TauCeti.SU2.weightBasis_mem_of_repr_ne_zeroandTauCeti.SU2.eq_span_weightBasis_mem: a torus-stable subspace contains every weight vector occurring in one of its elements, and is spanned by the weight vectors it contains.TauCeti.SU2.exists_weight_eq_of_forall_torusHom_smul: a nonzero vector on which the torus acts through a single characterz ↦ z^mis a multiple of a single weight vector, andmis that vector's weight.
References #
This is the "weight/string decomposition" milestone of the SU(2) engine case of
TauCetiRoadmap/RepresentationTheory/CompactGroups/README.md, which asks for the weights
{d, d-2, …, -d} of Symᵈ(ℂ²) "each with multiplicity one, computed from the diagonal action",
as the input to the highest-weight argument.
- D. Bump, Lie Groups, 2nd ed., Springer GTM 225 (2013), Chapter 3.
- T. Bröcker, T. tom Dieck, Representations of Compact Lie Groups, Springer GTM 98 (1985), Chapter II, §5.
The weight basis and the weights #
The weight basis of Symᵈ(ℂ²): the monomial basis of the symmetric power, whose index
i : Fin (d + 1) counts the factors equal to the first standard basis vector of ℂ². Its
elements are eigenvectors of the maximal torus, by
TauCeti.SU2.symPower_torusHom_weightBasis.
Equations
Instances For
The i-th weight vector is the monomial basis vector of Symᵈ(ℂ²) with i factors equal to
the first standard basis vector of ℂ² and d - i equal to the second.
Every weight vector is a pure symmetric tensor of standard basis vectors: some ordering of the unordered tuple indexing it lists its factors.
The d + 1 weights 2i - d are pairwise distinct. Together with
TauCeti.SU2.symPower_torusHom_weightBasis this is what makes each weight of Symᵈ(ℂ²) occur
with multiplicity one, in the form TauCeti.SU2.exists_weight_eq_of_forall_torusHom_smul.
The torus acts diagonally in the weight basis #
The maximal torus is diagonal in the weight basis: diag(z, z⁻¹) multiplies the i-th
weight vector by z^{2i - d}. This is the weight decomposition of Symᵈ(ℂ²), of which the
character TauCeti.SU2.character_symPower_torusHom is the trace.
The coordinates of a vector in the weight basis are scaled by the weights: the torus is
diagonal, so it multiplies the i-th coordinate by z^{2i - d}.
A torus element with pairwise distinct powers #
Torus-stable subspaces are spanned by weight vectors #
A torus-stable subspace contains every weight vector that occurs in one of its elements.
The torus is diagonal in the weight basis and the weights weight d i are pairwise distinct, so
a single torus element already separates the coordinates of a vector.
The weight spaces are lines #
A vector transforming under a single character of the torus is a weight vector. A nonzero
w on which the whole maximal torus acts through z ↦ z^m is a multiple of a single weight
vector, and m is that vector's weight, one of the d + 1 integers 2i - d.