Penrose’s Googly Problem: Twistors and the Missing Right Hand

In cricket, a googly is a leg-break delivery bowled with the same action as an off-break — it spins the "wrong" way. Roger Penrose borrowed the term in the 1960s for an analogous problem in his twistor programme: twistor space naturally represents fields with anti-self-dual (left-handed) helicity, but self-dual (right-handed) fields are described awkwardly by the dual twistor space T^* rather than T itself. For electromagnetism, this asymmetry is an aesthetic annoyance. For gravity, it becomes a fundamental obstacle: full general relativity contains both self-dual and anti-self-dual curvature, and gluing the two twistor descriptions together non-linearly has never been achieved in a satisfactory form.

The deeper issue is that twistor space is inherently chiral. A twistor Z^α = (ω^A, π_A′) consists of a pair of two-component spinors: ω^A has undotted (unprimed) spinor index and π_A′ has dotted (primed) spinor index. Undotted spinors transform under the left-handed SL(2,ℂ) factor of the complexified Lorentz group; dotted spinors under the right-handed factor. The incidence relation ω^A = ix^AA′π_A′ ties the spatial coordinate x to both spinor types, but the geometric content of twistor space — holomorphic line bundles, cohomology classes, Penrose transforms — is driven by the complex structure of ℙ(T), which is controlled by the undotted spinors. Right-handed fields live in the cohomology of ℙ(T^*), a different space, and there is no canonical isomorphism between ℙ(T) and ℙ(T^*) compatible with the Penrose transform.

What the Problem Actually Requires

To include right-handed gravitons in the twistor description, one would need to define a non-linear graviton construction that couples both helicities. The non-linear graviton construction (Penrose, 1976) deforms the complex structure of regions of twistor space to encode anti-self-dual vacuum metrics — metrics with vanishing self-dual Weyl curvature W⁺_ABCD. The googly problem asks how to extend this to metrics with non-vanishing W⁺. Partial answers have been proposed — ambitwistor strings, the Metsaev-Tseytlin action, various non-local constructions — but none achieves the geometric naturalness of the original non-linear graviton construction.

The T-Image Resolution

The Shadow Framework identifies the googly problem as a manifestation of the two-sided structure of spacetime. The forward-time sector M⁴₊ and its T-image M⁴₋ are related by a time reversal at the conformal boundary ℐ⁰. In the twistor language, this T-reflection swaps the dotted and undotted spinor indices:

The "googly field" is not a field in the same spacetime sector as its left-handed counterpart. It is the T-image: a field in M⁴₋ viewed from M⁴₊ through the conformal boundary. Right-handed gravitons are the mirror images of left-handed gravitons across the T-boundary. This is why they cannot be described by ℙ(T): they live in the dual twistor space ℙ(T^*) because ℙ(T^*) is the twistor space of M⁴₋, the mirror sector.

Implications for Scattering Amplitudes

This resolution has immediate consequences for the theory of scattering amplitudes. In the modern amplitude programme, the Parke-Taylor formula for maximally helicity-violating (MHV) gluon scattering is naturally a holomorphic object in twistor space. The googly MHV amplitudes — those with the opposite helicity assignment — are equally simple but live in dual twistor space. Witten's twistor string theory reproduces the correct split between ℙ(T) and ℙ(T^*) for the two helicity sectors, which the Shadow Framework identifies as the scattering between the two sides of the T-boundary.

The googly problem is not an obstacle to twistor theory. It is twistor theory correctly identifying that right-handed fields belong to a different side of the universe — the T-image sector reached only through the conformal boundary at null infinity.

This perspective explains why ambitwistor strings work: they describe scattering on the conformal boundary itself, where both ℙ(T) and ℙ(T^*) make simultaneous contributions through the boundary operator algebra. The boundary is the meeting point of the two twistor spaces, and the full S-matrix is a bilinear pairing between them — exactly the shadow pairing of the BMS algebra at Δ = 1.