Twistor Space

A twistor Zα = (ωA, π&Adot;) ∈ ℂ4 encodes a null line in complexified Minkowski space via the incidence relation ωA = ixA&Adot;π&Adot;. The projective version PT = ℂP³ is projective twistor space. Penrose's insight was that the conformal structure of spacetime, and the massless field equations, have a natural and transparent encoding in the holomorphic geometry of PT.

A massless field of helicity h in the anti-self-dual sector corresponds to a cohomology class in H¹(PT, 𝒪(−2h−2)). For gravitons (h = −2): H¹(PT, 𝒪(−6)). This is the Penrose transform.

The Grassmannian Gr(2,4)

The space of oriented null lines in M⁴ is parametrised by the Grassmannian:

Gr(2,4) = U(4) / (U(2) × U(2))The compact homogeneous space of 2-planes in ℂ⁴. Every twistor line Lx ≅ ℂP¹ corresponds to a fibre over a point of Gr(2,4).
Theorem — Haar Self-Duality on Gr(2,4)

The orthogonal complement map ⊥: Gr(2,4) → Gr(2,4), Λ ➦ Λ, preserves the Haar measure:

Gr) = dμGr(Λ)Proved via U(4)-equivariance of the Hodge star on the Plücker embedding.

This single measure-theoretic fact is the origin of the shadow symmetry of the celestial CFT.

The Googly Problem

The Penrose–Ward correspondence produces only the anti-self-dual (ASD) sector: Ward bundles E → PT encode solutions with F+ = 0. The self-dual (SD) sector — positive helicity, the googly modes — requires the dual twistor space PT*, which has no canonical identification with PT. This is the googly problem: Penrose named it in 1991 and it remained open for fifty years.

Theorem — Googly Resolution (Toupin 2026)

The shadow bundle Ê = T(E) with transition functions

˜g}ij(Z̅) = gij(Z)̅−T

encodes the SD sector. Both helicity sectors are unified in the Yang–Mills twistor datum (E, Ê, φ), where φ is a shadow pairing satisfying the Knizhnik–Zamolodchikov equations at the shared conformal boundary. Dual twistor space PT* is not an independent structure — it is the T-image of PT.

The Three-Step Chain: Gr(2,4) to Shadow Transform

Step 1. The Hodge star ☆: Λ24 → Λ24 acts on the Plücker embedding of Gr(2,4) as the complement map Λ ➦ Λ, preserving Haar measure.

Step 2. Via the Penrose correspondence, this induces the antipodal map ι: [Zα] ➦ [−Z̅α] on ℂP¹. The spinor normalisation ⟨λ,˜λ⟩ = 1 forces energy inversion ω ➦ ω−1 under the swap λA ⇔ ˜λ&Adot;.

Step 3. Under ω ➦ ω−1 the Mellin transform sends Δ ➦ 2−Δ. This is the shadow transform. Since T is antiunitary, it acts on the Mellin weight by conjugation: T(ωΔ−1) = ωΔ̅−1 = ω(2−Δ)−1 on the principal series. Hence shadow = T.

Zitterbewegung and the 4π Periodicity

Every massive Dirac fermion oscillates between the two T-sectors at the Compton frequency ω = 2mc²/ħ with amplitude λC/4π. This oscillation is Schrödinger's zitterbewegung, reinterpreted as T-boundary traversal. The 4π periodicity of spin-½ representations follows topologically: the spinor bundle over M⁴+ℐ⁰ M⁴ has π1(Spin(3) ×T Spin(3)) ≅ ℤ4, since T² = −1 on spin-½ gives a non-split extension.

Read the Paper on Zenodo ↗  Celestial QG ↗