The Geometry of Twistors: Null Planes, the Twistor Norm, and the Robinson Congruence

Intuition Beginner

A point in spacetime is one event: a place at a time. A twistor is a different kind of basic object. Instead of marking where something is, it packages a light ray together with its spin into a single point of a new space. The slogan of twistor theory is that the natural building block of physics is not the event but the ray of light, and a twistor is the algebra of one such ray.

Why trade points for rays? Because light rays are what we actually see, and because the equations of massless physics — light, neutrinos, the gravitational field at large distance — are simplest when written in terms of rays. A twistor records two things about a ray: which direction it points, and how a nearby family of rays twists around it. That twisting is where the name comes from.

The surprise is that this same package also carries a sign, called helicity. A twistor with positive helicity, zero helicity, or negative helicity sits in three separate rooms of twistor space. The boundary room, helicity zero, is exactly the rays of light in ordinary spacetime. The other two rooms hold the "twisting" configurations that have no single point they pass through.

Visual Beginner

The picture to hold is a family of light rays that all twist around a common core, like the threads of a screw or the swirl of water down a drain. A single light ray is one straight line. A twistor that is not on the boundary describes a whole swirling family of rays, none of which is special, all of them winding around together. The amount of winding is fixed by one number attached to the twistor. When that number is zero the swirl collapses to a single straight ray, and the twistor sits on the boundary surface that matches ordinary light rays one-to-one.

Worked example Beginner

Take flat spacetime and the simplest twistor data. A twistor is a list of four complex numbers, written as two pairs: a pair $({\omega }^0,{\omega }^1)$ and a pair $({\pi }_0,{\pi }_1)$. Suppose $({\pi }_0,{\pi }_1)=(1,0)$ and $({\omega }^0,{\omega }^1)=(0,0)$.

We test whether this twistor is "null", meaning it matches a genuine light ray. The test number is built by combining the twistor with its complex conjugate in a fixed pattern. For this choice the test number comes out to $0$, because the $\omega$ part is entirely zero. So this twistor is null: it corresponds to an actual light ray.

Now change one entry. Keep $({\pi }_0,{\pi }_1)=(1,0)$ but set $({\omega }^0,{\omega }^1)=(i,0)$, where $i$ is the square root of $-1$. The same combination now gives a test number equal to $-2$, which is not zero. This twistor is not null. It does not pass through any single ray; it describes a twisting family of rays instead, and the value $-2$ records that it carries negative helicity.

What this tells us: one real number, computed from the four entries, decides whether a twistor is a plain light ray or a twisting bundle, and on which side of the boundary it lands.

Check your understanding Beginner

Formal definition Intermediate+

Work over complexified Minkowski space $\mathbb{CM}$ with the spinor isomorphism ${\mathbb{C}}^4\cong {\mathbb{S}}^A\otimes {\mathbb{S}}^{A^{\prime }}$, soldering a vector index $a$ to an unprimed-primed spinor index pair $A{A}^{\prime }$. Indices are raised and lowered with the alternating spinors ${ϵ}_{AB}$, ${ϵ}_{A^{\prime }{B}^{\prime }}$ in the conventions of 03.02.18 and 03.09.05. Twistor space is the four-complex-dimensional vector space $\mathbb{T}={\mathbb{C}}^4$ whose elements $Z^{\alpha }=({\omega }^A,{\pi }_{A^{\prime }})$ are twistors; its projectivisation is $\mathbb{PT}={\mathbb{CP}}^3$ as introduced in 03.07.14. The incidence relation

$${\omega }^A=i{x}^{A{A}^{\prime }}{\pi }_{A^{\prime }}$$

ties a twistor to a complex spacetime point $x^{A{A}^{\prime }}$.

Definition ($\alpha$-plane). A nonzero twistor $Z^{\alpha }=({\omega }^A,{\pi }_{A^{\prime }})$ with ${\pi }_{A^{\prime }}\ne 0$ determines, by the locus of complex points $x$ satisfying the incidence relation, a two-complex-dimensional affine subspace of $\mathbb{CM}$ on which every tangent connecting vector $V^{A{A}^{\prime }}={\lambda }^A{\pi }^{A^{\prime }}$ factorises with the same primed spinor ${\pi }^{A^{\prime }}$. Such a plane is totally null (every tangent vector is null and any two are orthogonal) and self-dual (its tangent bivector ${\lambda }^{[A}{\mu }^{B]}{\pi }^{A^{\prime }}{\pi }^{B^{\prime }}$ is self-dual). It is the $\alpha$-plane $Z_{\alpha }$. Dually, fixing the unprimed spinor ${\lambda }^A$ across a two-plane gives a totally null anti-self-dual $\beta$-plane, the locus associated to a dual twistor $W_{\alpha }=({\lambda }_A,{\mu }^{A^{\prime }})$.

Definition (twistor norm). The complex conjugate of a twistor is a dual twistor ${\bar{Z}}_{\alpha }=({\bar{\pi }}_A,{\bar{\omega }}^{A^{\prime }})$ obtained by complex conjugation, which exchanges primed and unprimed spinors. The Hermitian twistor form is

$${\Sigma }_{\alpha \beta }{Z}^{\alpha }{\bar{Z}}^{\beta }=Z^{\alpha }{\bar{Z}}_{\alpha }={\omega }^A{\bar{\pi }}_A+{\pi }_{A^{\prime }}{\bar{\omega }}^{A^{\prime }}=2\mathrm{Re}\left({\omega }^A{\bar{\pi }}_A\right).$$

In a basis adapted to the splitting $\mathbb{T}={\mathbb{S}}^A\oplus {\mathbb{S}}_{A^{\prime }}$ this form has signature $(+,+,-,-)$, that is signature $(2,2)$. The twistor norm of $Z^{\alpha }$ is the value of this form, a real number. A twistor is null when its norm vanishes. The norm defines the three-way split

$$\mathbb{PT}={\mathbb{PT}}^{+}\bigsqcup \mathbb{PN}\bigsqcup {\mathbb{PT}}^{-},\mathbb{PN}=\{Z^{\alpha }:{\Sigma }_{\alpha \beta }{Z}^{\alpha }{\bar{Z}}^{\beta }=0\},$$

with ${\mathbb{PT}}^{+}$ the locus where the norm is positive and ${\mathbb{PT}}^{-}$ where it is negative; $\mathbb{PN}$ is a real hypersurface of real dimension five inside $\mathbb{PT}$.

Definition (helicity). The helicity carried by a twistor is

$$s=\frac{1}{2}{\Sigma }_{\alpha \beta }{Z}^{\alpha }{\bar{Z}}^{\beta },$$

half the twistor norm. Positive helicity is ${\mathbb{PT}}^{+}$, negative helicity is ${\mathbb{PT}}^{-}$, and helicity zero is the null hypersurface $\mathbb{PN}$.

The form ${\Sigma }_{\alpha \beta }$ is preserved by the group $\mathrm{SU}(2,2)$, the special unitary group of the signature-$(2,2)$ Hermitian form; its action descends to $\mathbb{PSU}(2,2)$ on $\mathbb{PT}$, and this is the (double cover of the) conformal group of compactified Minkowski space as recorded in 03.07.14.

Counterexamples to common slips

• The conjugate of a twistor is not a twistor of the same kind: conjugation maps $\mathbb{T}\to {\mathbb{T}}^{\ast }$ because it swaps primed and unprimed spinors. Writing ${\bar{Z}}^{\alpha }$ as if it had an upper index is the standard slip; the correct object is the dual twistor ${\bar{Z}}_{\alpha }$.
• The twistor norm is not positive-definite. Signature $(2,2)$ is the whole content of the helicity split; a "norm" that could not be negative would collapse ${\mathbb{PT}}^{-}$ and destroy the theory.
• An $\alpha$-plane is null and self-dual; a $\beta$-plane is null and anti-self-dual. Both are totally null two-planes, but they belong to the two distinct families of totally null two-planes in four dimensions, and a single twistor names only one family.

Key theorem with proof Intermediate+

Theorem (null twistors and null geodesics). A nonzero twistor $Z^{\alpha }=({\omega }^A,{\pi }_{A^{\prime }})$ with ${\pi }_{A^{\prime }}\ne 0$ is null, ${\Sigma }_{\alpha \beta }{Z}^{\alpha }{\bar{Z}}^{\beta }=0$, if and only if its $\alpha$-plane meets the real Minkowski slice $\mathbb{M}\subset \mathbb{CM}$, and then the intersection is a single real null geodesic with tangent direction ${\bar{\pi }}^A{\pi }^{A^{\prime }}$. Up to the projective scaling of $Z^{\alpha }$, this assignment is a bijection between the points of $\mathbb{PN}$ and the oriented null geodesics of $\mathbb{M}$.

Proof. A point $x^{A{A}^{\prime }}$ of the $\alpha$-plane satisfies ${\omega }^A=i{x}^{A{A}^{\prime }}{\pi }_{A^{\prime }}$. Contract with ${\bar{\pi }}_A$:

$${\omega }^A{\bar{\pi }}_A=i{x}^{A{A}^{\prime }}{\pi }_{A^{\prime }}{\bar{\pi }}_A.$$

The point $x^{A{A}^{\prime }}$ is real precisely when $x^{A{A}^{\prime }}={\bar{x}}^{A{A}^{\prime }}$, the Hermitian condition on the $2\times 2$ matrix $x^{A{A}^{\prime }}$. Taking the complex conjugate of the contracted equation and using reality,

$$\overline{{\omega }^A{\bar{\pi }}_A}={\bar{\omega }}^{A^{\prime }}{\pi }_{A^{\prime }}=-i{\bar{x}}^{A^{\prime }A}{\bar{\pi }}_{A^{\prime }}{\pi }_A=-i{x}^{A{A}^{\prime }}{\pi }_{A^{\prime }}{\bar{\pi }}_A,$$

so that ${\omega }^A{\bar{\pi }}_A+{\pi }_{A^{\prime }}{\bar{\omega }}^{A^{\prime }}=i{x}^{A{A}^{\prime }}{\pi }_{A^{\prime }}{\bar{\pi }}_A-i{x}^{A{A}^{\prime }}{\pi }_{A^{\prime }}{\bar{\pi }}_A=0$. Hence existence of a real point on the $\alpha$-plane forces the norm to vanish. Conversely, if the norm vanishes, ${\omega }^A{\bar{\pi }}_A$ is purely imaginary, say ${\omega }^A{\bar{\pi }}_A=ic$ with $c$ real; one solves $i{x}^{A{A}^{\prime }}{\pi }_{A^{\prime }}{\bar{\pi }}_A=ic$ for a Hermitian $x^{A{A}^{\prime }}$ because ${\pi }_{A^{\prime }}{\bar{\pi }}_A$ spans the real null direction ${\bar{\pi }}^A{\pi }^{A^{\prime }}$ and the remaining freedom is real. The solution set is the affine line $x^{A{A}^{\prime }}=x_0^{A{A}^{\prime }}+r{\bar{\pi }}^A{\pi }^{A^{\prime }}$, $r\in \mathbb{R}$: a real null geodesic with tangent ${\ell }^{A{A}^{\prime }}={\bar{\pi }}^A{\pi }^{A^{\prime }}$, which is null because ${\ell }^{A{A}^{\prime }}{\ell }_{A{A}^{\prime }}=({\bar{\pi }}^A{\bar{\pi }}_A)({\pi }^{A^{\prime }}{\pi }_{A^{\prime }})=0$. Rescaling $Z^{\alpha }\mapsto \kappa Z^{\alpha }$ leaves both the geodesic and its orientation fixed, so the map factors through $\mathbb{PN}$, and the recovery of ${\pi }_{A^{\prime }}$ (hence the orientation) from the geodesic makes it a bijection. $\square$

Bridge. This result builds toward the entire correspondence dictionary of twistor theory and appears again in 03.07.14, where the same incidence relation ${\omega }^A=i{x}^{A{A}^{\prime }}{\pi }_{A^{\prime }}$ that here cuts out a null geodesic is the substrate of the contour integral. The foundational reason the Penrose transform lands massless fields on spacetime is exactly that a twistor is a null ray dressed with a spinor, so that integrating over a line in $\mathbb{PT}$ integrates over the rays through a point. This is dual to the point-line correspondence: a spacetime point is a projective line $L_x\subset \mathbb{PT}$, while a null geodesic is a point of $\mathbb{PN}$, and the incidence between them is the central insight that twistor space and spacetime are two readings of one holomorphic geometry. Putting these together, the reality structure ${\Sigma }_{\alpha \beta }$ generalises the bare ${\mathbb{CP}}^3$ of 03.07.11 into a space that knows Lorentzian signature, helicity, and the conformal group at once.

Exercises Intermediate+

Lean formalization Intermediate+

Mathlib does not provide twistor space, two-component spinor calculus over Lorentzian signature, the signature-$(2,2)$ Hermitian twistor form, or the group $\mathrm{SU}(2,2)$ acting on ${\mathbb{CP}}^3$. No formal statement of the null-twistor / null-geodesic correspondence can be written against current Mathlib; the unit ships with lean_status: none and the gap recorded in frontmatter as an upstream-contribution target. A first formalisation milestone would be the abstract index calculus of ${ϵ}_{AB}$, ${ϵ}_{A^{\prime }{B}^{\prime }}$ and the soldering ${\mathbb{C}}^4\cong {\mathbb{S}}^A\otimes {\mathbb{S}}^{A^{\prime }}$, on top of which the Hermitian form ${\Sigma }_{\alpha \beta }$ and its signature could be defined.

Advanced results Master

The geometry above acquires its full force once the reality structure is read against the projective and Grassmannian pictures.

The Robinson congruence. Fix a non-null twistor $Z^{\alpha }$, so its norm $s=\frac{1}{2}{\Sigma }_{\alpha \beta }{Z}^{\alpha }{\bar{Z}}^{\beta }\ne 0$. Its $\alpha$-plane meets $\mathbb{M}$ in no real point, so $Z^{\alpha }$ is not a single null geodesic. Instead, consider every null twistor $X^{\alpha }\in \mathbb{PN}$ incident to $Z^{\alpha }$, that is $X^{\alpha }{\bar{Z}}_{\alpha }=0$. Each such $X^{\alpha }$ is a real null geodesic, and the family of these geodesics fills $\mathbb{M}$ as a shear-free null geodesic congruence whose rays twist around one another with rotation proportional to $s$. This is the Robinson congruence: spatially, at a fixed time, the rays appear tangent to a family of nested circles, the stereographic image of the Hopf fibration of $S^3$. The congruence twists left or right according to $\mathrm{sign}(s)$, and it is the geometric object that gives twistor theory its name. By the Kerr theorem, every analytic shear-free null geodesic congruence in $\mathbb{M}$ arises this way from a holomorphic surface in $\mathbb{PT}$, of which the Robinson congruence (a single twistor, hence a single point of $\mathbb{PT}$) is the simplest case.

The Klein quadric. Complexified compactified Minkowski space ${\mathbb{CM}}^{\#}$ is the Grassmannian $\mathrm{Gr}(2,4)$ of two-planes in $\mathbb{T}={\mathbb{C}}^4$: a spacetime point is the projective line $L_x\subset \mathbb{PT}$ it determines, and a line in ${\mathbb{CP}}^3$ is a two-plane in ${\mathbb{C}}^4$. Under the Plücker embedding $\mathrm{Gr}(2,4)\hookrightarrow \mathbb{P}({\Lambda }^2\mathbb{T})={\mathbb{CP}}^5$, this Grassmannian is cut out by a single quadratic equation, the Klein quadric $Q\subset {\mathbb{CP}}^5$, whose defining bilinear form is the wedge pairing ${\Lambda }^2\mathbb{T}\otimes {\Lambda }^2\mathbb{T}\to {\Lambda }^4\mathbb{T}\cong \mathbb{C}$. The conformal metric of ${\mathbb{CM}}^{\#}$ is exactly this quadratic form: two points are null-separated iff their lines meet in $\mathbb{PT}$ iff their Plücker bivectors are orthogonal under the wedge pairing.

The conformal group. Linear transformations of $\mathbb{T}$ preserving ${\Sigma }_{\alpha \beta }$ are $\mathrm{U}(2,2)$; the special unitary subgroup $\mathrm{SU}(2,2)$ acts on ${\Lambda }^2\mathbb{T}\cong {\mathbb{C}}^6$ preserving the wedge form of signature $(2,4)$, giving the isomorphism

$$\mathrm{SU}(2,2)\cong \mathrm{Spin}(2,4),$$

the double cover of the connected conformal group ${\mathrm{SO}}^{\uparrow }(2,4)/\{\pm 1\}$ of compactified Minkowski space. Real compactified Minkowski space ${\mathbb{M}}^{\#}$ is the real slice of $Q$ fixed by the reality structure ${\Sigma }_{\alpha \beta }$, an $S^1\times S^3/{\mathbb{Z}}_2$ on which the conformal group acts by its defining action; the points of ${\mathbb{M}}^{\#}$ are exactly the projective lines lying inside the real hypersurface $\mathbb{PN}$.

Synthesis. The bridge is that one Hermitian form does all the work: the foundational reason a twistor carries helicity, a reality structure, and a conformal symmetry simultaneously is that ${\Sigma }_{\alpha \beta }$ of signature $(2,2)$ is exactly the datum that splits $\mathbb{PT}$ into helicity chambers, cuts out $\mathbb{PN}$ as the real light rays, and (on ${\Lambda }^2\mathbb{T}$) becomes the conformal metric of spacetime. This is exactly the statement that the point-line correspondence is dual to the null-geodesic correspondence: putting these together, a spacetime point is a line in $\mathbb{PT}$ and a light ray is a point of $\mathbb{PN}$, and the Klein quadric makes the incidence between them the conformal metric. The central insight generalises the bare incidence relation of 03.07.14 into the full Penrose dictionary, and it builds toward both the curved deformations of 03.07.11 and the asymptotic null geometry where this same shear-free congruence structure reappears at infinity. What looked in 03.07.14 like a coordinate $4$-tuple is, with ${\Sigma }_{\alpha \beta }$ restored, the complete geometry of a light ray.

Full proof set Master

Proposition 1 (the null hypersurface is the real-geodesic locus). The map sending a null twistor (mod scale) to its real null geodesic is a real-analytic bijection $\mathbb{PN}\to \{\text{oriented null geodesics of }\mathbb{M}\}$.

Proof. Injectivity and the existence of the geodesic are the Key theorem above. For surjectivity, let $\gamma$ be an oriented null geodesic with point $x_0\in \mathbb{M}$ and future-pointing null tangent ${\ell }^{A{A}^{\prime }}$. Since $\ell$ is real and null, it factorises as ${\ell }^{A{A}^{\prime }}={\bar{\pi }}^A{\pi }^{A^{\prime }}$ for some ${\pi }_{A^{\prime }}$, unique up to a phase ${\pi }_{A^{\prime }}\mapsto e^{i\theta }{\pi }_{A^{\prime }}$. Set ${\omega }^A=i{x}_0^{A{A}^{\prime }}{\pi }_{A^{\prime }}$. Then $Z^{\alpha }=({\omega }^A,{\pi }_{A^{\prime }})$ is a twistor whose $\alpha$-plane contains $x_0$, hence by the Key theorem its norm vanishes, so $Z^{\alpha }\in \mathbb{PN}$, and its associated geodesic is $\{x_0+r\ell :r\in \mathbb{R}\}=\gamma$ with the given orientation. The phase freedom in ${\pi }_{A^{\prime }}$ and the additive freedom $x_0\mapsto x_0+r\ell$ together rescale $Z^{\alpha }$ projectively, so the preimage is a single point of $\mathbb{PN}$. Real-analyticity of both directions is immediate from the algebraic formulae. $\square$

Proposition 2 (helicity is $\mathrm{SU}(2,2)$-invariant and conjugation-odd). Under $g\in \mathrm{SU}(2,2)$, $s(gZ)=s(Z)$; under the operation exchanging ${\mathbb{PT}}^{+}$ and ${\mathbb{PT}}^{-}$ induced by $Z^{\alpha }\mapsto {\bar{Z}}_{\alpha }$ (dualisation), $s\mapsto -s$.

Proof. By definition $\mathrm{SU}(2,2)$ preserves ${\Sigma }_{\alpha \beta }$: ${\Sigma }_{\alpha \beta }(gZ{)}^{\alpha }\overline{(gZ)}{}^{\beta }={\Sigma }_{\gamma \delta }{Z}^{\gamma }{\bar{Z}}^{\delta }$ for all $g$, hence $s$ is invariant. For the second claim, the dual twistor ${\bar{Z}}_{\alpha }$ paired with its own conjugate gives the form on ${\mathbb{T}}^{\ast }$, which is ${\Sigma }^{\alpha \beta }$, the inverse Hermitian form; its signature is again $(2,2)$ but with the positive and negative subspaces interchanged relative to the identification $\mathbb{T}\cong {\mathbb{T}}^{\ast }$, so the sign of the evaluated norm reverses, $s(\bar{Z})=-s(Z)$. Concretely ${\Sigma }^{\alpha \beta }{\bar{Z}}_{\alpha }{Z}_{\beta }=\overline{{\Sigma }_{\alpha \beta }{Z}^{\alpha }{\bar{Z}}^{\beta }}$ with the role of ${\mathbb{PT}}^{+}$ and ${\mathbb{PT}}^{-}$ exchanged. $\square$

Proposition 3 (Klein-quadric metric). Two points $x,y\in {\mathbb{CM}}^{\#}$ with twistor lines $L_x,L_y$ and Plücker bivectors $P_x,P_y\in {\Lambda }^2\mathbb{T}$ are null-separated if and only if $P_x\wedge P_y=0$ in ${\Lambda }^4\mathbb{T}$.

Proof. Two two-planes ${\Pi }_x,{\Pi }_y\subset \mathbb{T}$ share a one-dimensional subspace iff their lines $L_x,L_y\subset \mathbb{PT}$ meet at a point. Decomposable bivectors $P_x=u\wedge v$, $P_y=p\wedge q$ satisfy $P_x\wedge P_y=u\wedge v\wedge p\wedge q$, which is the volume of the four vectors and vanishes iff they are linearly dependent, i.e. iff ${\Pi }_x\cap {\Pi }_y\ne \{0\}$. The wedge form $P_x\wedge P_y$ is the conformal metric on $\mathrm{Gr}(2,4)$ by definition of the Plücker quadric, and meeting lines are exactly the null-separated points of ${\mathbb{CM}}^{\#}$. Hence $P_x\wedge P_y=0⟺x,y$ null-separated. $\square$

These three propositions are the algebraic skeleton; the Robinson-congruence statement and the Kerr theorem are quoted as advanced results and are proved in Penrose-Rindler Vol. 2 Ch. 7 ^{[Penrose-Rindler 1986]}.

Connections Master

• The incidence relation ${\omega }^A=i{x}^{A{A}^{\prime }}{\pi }_{A^{\prime }}$ and the projective space $\mathbb{PT}={\mathbb{CP}}^3$ used throughout are introduced in 03.07.14 purely as the substrate of the Penrose contour integral; this unit supplies the reality structure ${\Sigma }_{\alpha \beta }$, the ${\mathbb{PT}}^{+}/\mathbb{PN}/{\mathbb{PT}}^{-}$ split, and the null-geodesic geometry that the transform silently assumes. The helicity $s=\frac{1}{2}{\Sigma }_{\alpha \beta }{Z}^{\alpha }{\bar{Z}}^{\beta }$ defined here is the same $h$ that labels the line bundles $\mathcal{O}(-2h-2)$ there.

• The two-component spinor calculus — ${ϵ}_{AB}$, the soldering ${\mathbb{C}}^4\cong {\mathbb{S}}^A\otimes {\mathbb{S}}^{A^{\prime }}$, raising and lowering, the abstract-index conventions — is consolidated in 03.02.41, which this unit uses at every step. That unit is the algebraic toolkit; this one is its first geometric payoff in Lorentzian signature.

• The principal null directions and the Weyl-spinor factorisation of 03.02.18 are the local-curvature analogue of the global ray geometry here: a shear-free null geodesic congruence (the Robinson congruence is the flat example) is, by Goldberg-Sachs, tied to algebraic speciality of the Weyl spinor, linking $\mathbb{PN}$ to the Petrov types.

• The Euclidean instanton picture of 03.07.11 is the signature-$(4,0)$ sibling: there $\mathbb{PN}$ is empty, the relevant group is $\mathrm{SU}(4)=\mathrm{Spin}(6)$, and twistor lines fibre $S^4$ by the Hopf map. The Lorentzian reality structure of this unit is what distinguishes the physical twistor theory from its Euclidean computational cousin.

• The spinor bundle of 03.09.05 is the global geometric setting in which ${\omega }^A$, ${\pi }_{A^{\prime }}$, and the conjugation $Z^{\alpha }\mapsto {\bar{Z}}_{\alpha }$ live as sections, and the conformal-weight machinery there is what lets the twistor norm and helicity be defined invariantly on ${\mathbb{CM}}^{\#}$.

Historical & philosophical context Master

Twistor theory was introduced by Roger Penrose in 1967 ^{[Penrose 1967]}, where the twistor algebra, the incidence relation, and the Hermitian norm of signature $(2,2)$ first appear together with the helicity classification into the three regions ${\mathbb{PT}}^{+}$, $\mathbb{PN}$, ${\mathbb{PT}}^{-}$. The motivating idea was to take the light ray, not the spacetime event, as primitive, and to encode the conformal and spinorial structure of Minkowski space in a complex projective geometry. The Robinson congruence, named for Ivor Robinson, supplied the geometric image — a twisting family of null rays — from which Penrose coined the word "twistor"; the picture and its role are developed at length in the Penrose-MacCallum review ^{[Penrose-MacCallum 1972]}.

The identification of compactified complexified Minkowski space with the Grassmannian $\mathrm{Gr}(2,4)$ and the Klein quadric, and of the conformal group with $\mathrm{SU}(2,2)\cong \mathrm{Spin}(2,4)$, places the construction inside nineteenth-century line geometry: the Klein correspondence between lines in ${\mathbb{CP}}^3$ and points of a quadric in ${\mathbb{CP}}^5$ is exactly the point-line duality twistor theory exploits, a connection drawn out in Ward and Wells ^{[Ward-Wells 1990]}. The full geometric treatment over Lorentzian signature, including the $\alpha$-plane / $\beta$-plane decomposition and the Robinson congruence, is the subject of Penrose and Rindler's second volume ^{[Penrose-Rindler 1986]}, the canonical reference for this unit.

Bibliography Master

@article{Penrose1967,
  author  = {Penrose, Roger},
  title   = {Twistor Algebra},
  journal = {Journal of Mathematical Physics},
  volume  = {8},
  number  = {2},
  pages   = {345--366},
  year    = {1967},
  doi     = {10.1063/1.1705200}
}

@article{PenroseMacCallum1972,
  author  = {Penrose, Roger and MacCallum, M. A. H.},
  title   = {Twistor Theory: An Approach to the Quantisation of Fields and Space-Time},
  journal = {Physics Reports},
  volume  = {6},
  number  = {4},
  pages   = {241--316},
  year    = {1972},
  doi     = {10.1016/0370-1573(73)90008-2}
}

@book{PenroseRindlerVol2,
  author    = {Penrose, Roger and Rindler, Wolfgang},
  title     = {Spinors and Space-Time, Volume 2: Spinor and Twistor Methods in Space-Time Geometry},
  publisher = {Cambridge University Press},
  year      = {1986},
  doi       = {10.1017/CBO9780511524486}
}

@book{WardWells1990,
  author    = {Ward, R. S. and Wells, R. O.},
  title     = {Twistor Geometry and Field Theory},
  publisher = {Cambridge University Press},
  series    = {Cambridge Monographs on Mathematical Physics},
  year      = {1990},
  doi       = {10.1017/CBO9780511524493}
}

@book{HuggettTod1994,
  author    = {Huggett, S. A. and Tod, K. P.},
  title     = {An Introduction to Twistor Theory},
  edition   = {2},
  publisher = {Cambridge University Press},
  series    = {London Mathematical Society Student Texts 4},
  year      = {1994},
  doi       = {10.1017/CBO9780511624018}
}

@article{Penrose1977programme,
  author  = {Penrose, Roger},
  title   = {The Twistor Programme},
  journal = {Reports on Mathematical Physics},
  volume  = {12},
  number  = {1},
  pages   = {65--76},
  year    = {1977},
  doi     = {10.1016/0034-4877(77)90047-7}
}