Polarization data for quadratic spaces #
This file packages the decomposition used to construct spinor modules: two isotropic subspaces in a left- and right-nondegenerate polar pairing and an orthogonal remainder embedded in the scalar line. It also records two elementary consequences: the polar form vanishes identically on each isotropic summand, and a basis of the first summand has a Kronecker-dual family of vectors in the second.
The decomposition also counts dimensions. Over any commutative ring the remainder embeds in the
scalar line, so is at most a line; over a field the two isotropic summands are moreover dual to
each other, so equidimensional. Hence dim V = 2 · dim W + dim line with dim line ≤ 1, and the
parity of dim V decides which, giving dim W = l both in dimension 2l (type Dₗ) and in
dimension 2l + 1 (type Bₗ).
Main definitions #
TauCeti.SpinPolarizationDatarecords the decomposition and its consumer-facing coordinates.TauCeti.SpinPolarizationData.dualVector: given a basis of the first isotropic summand, the vector of the second summand matching a coordinate functional of that basis.TauCeti.SpinPolarizationData.dualBasis: the resulting basis of the second isotropic summand.
Main results #
TauCeti.SpinPolarizationData.polar_W_eq_zeroandTauCeti.SpinPolarizationData.polar_W'_eq_zero: the polar form vanishes on each isotropic summand.TauCeti.SpinPolarizationData.isOrtho_W_lineandTauCeti.SpinPolarizationData.isOrtho_line_W': the orthogonal remainder is orthogonal to each isotropic summand.TauCeti.SpinPolarizationData.polar_dualVector: the dual vectors pair with the chosen basis by the Kronecker delta.TauCeti.SpinPolarizationData.isOrtho_basis_dualVector: off the diagonal a basis vector is orthogonal to a dual vector.TauCeti.SpinPolarizationData.dualBasis_apply: the dual basis hasdualVectoras its underlying family.TauCeti.SpinPolarizationData.finrank_line_le_one: over any commutative ring the remainder is at most a line.TauCeti.SpinPolarizationData.finrank_W'_eq_finrank_WandTauCeti.SpinPolarizationData.finrank_eq_two_mul_finrank_W_add_finrank_line: over a field the isotropic summands are equidimensional, and a finite-dimensional polarized space has dimension2 · dim W + dim line.TauCeti.SpinPolarizationData.line_eq_bot_of_even_finrank,TauCeti.SpinPolarizationData.even_finrank_of_line_eq_botandTauCeti.SpinPolarizationData.finrank_W_eq_of_finrank_eq_two_mul: in even dimension the remainder vanishes, and conversely, and the isotropic summand has half the dimension.TauCeti.SpinPolarizationData.finrank_W_eq_of_finrank_eq_two_mul_add_oneandTauCeti.SpinPolarizationData.finrank_line_eq_one_of_finrank_eq_two_mul_add_one: in dimension2l + 1the isotropic summand again has dimensionl, and the remainder is exactly a line.TauCeti.SpinPolarizationData.lineCoordinate_surjective_of_ne_bot: over a field the coordinate of a nonzero remainder takes every scalar value.TauCeti.SpinPolarizationData.nondegenerateandTauCeti.SpinPolarizationData.nondegenerate_of_line_eq_bot: a polarized quadratic form is nondegenerate, as soon as2is a regular scalar of a reduced ring in general and unconditionally when there is no remainder.
A polarization of a quadratic space (V, Q): a decomposition V = W ⊕ W' ⊕ line into two
isotropic submodules, paired by the polar form so that W' is identified with the dual of W and
no nonzero vector of W pairs to zero with all of W', and an orthogonal remainder embedded in
the scalar line by a coordinate whose square is Q. It is the data from which the
exterior model ⋀·W of a spin representation is built.
- W : Submodule K V
The isotropic subspace used for exterior multiplication.
- W' : Submodule K V
The complementary isotropic subspace used for contraction.
- line : Submodule K V
The orthogonal remainder, embedded in the scalar line by
lineCoordinate. The direct-sum coordinates of the quadratic space.
- decompositionEquiv_apply (x : (↥self.W × ↥self.W') × ↥self.line) : self.decompositionEquiv x = ↑x.1.1 + ↑x.1.2 + ↑x.2
The decomposition equivalence adds the three components in the ambient module.
The first summand is isotropic.
The second summand is isotropic.
The polar form identifies the second summand with the dual of the first.
- pairingEquiv_apply (y : ↥self.W') (x : ↥self.W) : (self.pairingEquiv y) x = QuadraticMap.polar ⇑Q ↑x ↑y
The pairing equivalence is evaluation by the polar form.
- pairing_separatingLeft (x : ↥self.W) : (∀ (y : ↥self.W'), QuadraticMap.polar ⇑Q ↑x ↑y = 0) → x = 0
The polar pairing has trivial left radical.
The coordinate on the at-most-one-dimensional remainder.
- lineCoordinate_injective : Function.Injective ⇑self.lineCoordinate
The remainder embeds in the scalar line through its coordinate.
The quadratic form on the remainder is the square of its coordinate.
The remainder is orthogonal to the first isotropic summand.
The remainder is orthogonal to the second isotropic summand.
Instances For
Recover the three components of a vector assembled from polarization coordinates.
A vector of the first isotropic summand has only that coordinate.
A vector of the second isotropic summand has only that coordinate.
A vector of the orthogonal remainder has only that coordinate.
The polar pairing has trivial right radical.
The polar form vanishes on each isotropic summand #
The polar form vanishes on the first isotropic summand of a polarization.
The polar form vanishes on the second isotropic summand of a polarization.
The remainder is orthogonal to both isotropic summands #
A vector of the first isotropic summand is orthogonal to the remainder.
The remainder is orthogonal to the second isotropic summand.
The dual isotropic vectors of a basis #
The vector of the second isotropic summand W' dual to the i-th basis vector of W: the
polarization pairing identifies W' with the dual of W, and this is the vector matching the
i-th coordinate functional.
Equations
- P.dualVector b i = P.pairingEquiv.symm (b.coord i)
Instances For
The dual vector of the i-th basis vector pairs with W as the i-th coordinate
functional.
The basis of the second isotropic summand dual to b under the polarization pairing.
Instances For
The dual basis consists of the dual vectors.
The dual vectors pair with the basis of W by the Kronecker delta.
A basis vector of W pairs with its own dual vector to 1.
Off the diagonal a basis vector of W is orthogonal to a dual vector.
A polarization without an orthogonal remainder has nondegenerate quadratic form.
A polarization has nondegenerate quadratic form, whatever its orthogonal remainder, over a
reduced ring in which 2 is a regular scalar. So over such a ring nondegeneracy never needs to be
assumed alongside polarization data. When the remainder vanishes,
TauCeti.SpinPolarizationData.nondegenerate_of_line_eq_bot gives the same conclusion with neither
hypothesis.
The orthogonal remainder of a polarization is at most a line, over any commutative ring.
The dimensions of the three summands #
A polarization is a decomposition V = W ⊕ W' ⊕ L in which the polar form identifies W' with
the dual of W and L sits inside the scalar line. The first fact makes the two isotropic summands
equidimensional and the second, finrank_line_le_one above, bounds the remainder by one
dimension, so the dimension of V determines the dimension of W up to the parity of
finrank V.
The two isotropic summands of a polarization have the same dimension.
A nonzero orthogonal remainder is coordinatized onto the scalars: its coordinate takes every value in the base field.
The dimension of a polarized quadratic space: twice the dimension of the isotropic summand
W, plus the dimension of the remainder. Together with finrank_line_le_one this pins
finrank W to finrank V / 2.
In even dimension a polarization has no remainder. This is the hypothesis under which the
exterior parity of ⋀·W splits the spin representation into its two half-spin
representations.
A polarization with no remainder has even dimension. This is the converse of
SpinPolarizationData.line_eq_bot_of_even_finrank, so the even-dimensional theory can be stated
with the single hypothesis P.line = ⊥ that the parity splitting of the spinor module needs.
In even dimension the isotropic summand has half the dimension. The isotropic summand W
of a polarization of a 2l-dimensional space has dimension l. This is the type Dₗ case, where
the spinor module ⋀·W has dimension 2ˡ.
In odd dimension the isotropic summand again has half the dimension, rounded down. The
isotropic summand W of a polarization of a 2l + 1-dimensional space has dimension l, the odd
dimension being taken up by the remainder. This is the type Bₗ case, where the spinor module
⋀·W again has dimension 2ˡ but the parity splitting is not one of representations.
In odd dimension the remainder is exactly a line. It is spanned by an anisotropic vector,
whose action mixes the two exterior parities of the spinor module, which is why they are not
subrepresentations in the type Bₗ case.