In theoretical and mathematical physics, twistor theory is a mathematical theory mapping the geometric objects of conventional 3+1 spacetime (Minkowski space) into geometric objects in a 4 dimensional space with metric signature (2,2). This space is called twistor space, and its complex valued coordinates are called "twistors."
Twistor theory was first proposed by Roger Penrose in 1967,^{[1]} as a possible path to a theory of quantum gravity. The twistor approach is especially natural for solving the equations of motion of massless fields of arbitrary spin.
In 2003, Edward Witten^{[2]} proposed to marry twistor and string theory by embedding the topological B model of string theory in twistor space. His objective was to model certain YangMills amplitudes. The resulting model has come to be known as twistor string theory.
Contents 
Twistor theory is unique to 4D Minkowski space and the (2,2) metric signature, and does not generalize to other dimensions or metric signatures. At the heart of twistor theory lies the isomorphism between the conformal group Spin(4,2) and SU(2,2), which is the group of linear transformations of determinant 1 over a four dimensional complex vector space. These transformations leave invariant a Hermitian norm of signature (2,2).
, , and are all homogeneous spaces of the conformal group.
admits a conformal metric (i.e., an equivalence class of metric tensors under Weyl rescalings) with signature (+++−). Straight null rays map to straight null rays under a conformal transformation and there is a unique canonical isomorphism between null rays in and points in respecting the conformal group.
In , it is the case that positive and negative frequency solutions cannot be locally separated. However, this is possible in twistor space.
For many years after Penrose's foundational 1967 paper, twistor theory progressed slowly, in part because of mathematical challenges. Twistor theory also seemed unrelated to ideas in mainstream physics. While twistor theory appeared to say something about quantum gravity, its potential contributions to understanding the other fundamental interactions and particle physics were less obvious.
Witten (2003) proposed a connection between string theory and twistor geometry, called twistor string theory. Witten (2004)^{[2]} built on this insight to propose a way to do string theory in twistor space, whose dimensionality is necessarily the same as that of 3+1 Minkowski spacetime. Hence twistor string theory is a possible way to eliminate the need for more than 3 spatial dimensions when doing (super)string theory. Although Witten has said that "I think twistor string theory is something that only partly works," his work has given new life to the twistor research programme. For example, twistor string theory may simplify calculating scattering amplitudes from Feynman diagrams.

