Stokes vectors and Minkowski spacetime: Structural parallels

Abstract: The Stokes formalism of polarization physics has astounding structural parallels with the formalism used for relativity theory in Minkowski spacetime. The structure and symmetry properties of the Mueller matrices are the same as those for the matrix representations of the electromagnetic tensor and the Lorentz transformation operator. The absorption terms $\eta_k$ in the Mueller matrix correspond to the electric field components $E_k$ in the electromagnetic tensor and the Lorentz boost terms $\gamma_k$ in the Lorentz transformation matrix, while the anomalous dispersion terms $\rho_k$ correspond to the magnetic field components $B_k$ and the spatial rotation angles $\phi_k$. In a Minkowski-type space spanned by the Stokes $I,Q,U,V$ parameters, the Stokes vector for 100 % polarized light is a null vector living on the surface of null cones, like the energy-momentum vector of massless particles in ordinary Minkowski space. Stokes vectors for partially polarized light live inside the null cones like the momentum vectors for massive particles. In this description the depolarization of Stokes vectors appears as a "mass'' term, which has its origin in a symmetry breaking caused by the incoherent superposition of uncorrelated fields or wave packets, without the need to refer to a ubiquitous Higgs field as is done in particle physics. The rotational symmetry of Stokes vectors and Mueller matrices is that of spin-2 objects, in contrast to the spin-1 nature of the electromagnetic field. The reason for this difference is that the Stokes objects have substructure: they are formed from bilinear tensor products between spin-1 objects, the Jones vectors and Jones matrices. The governing physics takes place at the substructure level.