Fermat's principle of least time has been shown to reproduce the relativistic laws of light-ray refraction at a uniformly moving interface between vacuum and a homogeneous, isotropic, nondispersive medium, in the two geometries where the medium velocity is parallel or perpendicular to the interface, including a predicted 'backward refraction' regime with threshold speeds κ1=(1+n2)−1/2\kappa_{1}=(1+n^{2})^{-1/2} and κ2=1/n\kappa_{2}=1/n. The source paper leaves open the extension of this least-time analysis to (a) dispersive media, (b) general three-dimensional incidence geometries, and (c) interfaces between two uniformly moving non-vacuum media with different rest-frame refractive indices. A 2025–2026 line of work on Fresnel drag in dispersive and anisotropic moving media covers much of (a) and the vectorial aspect of (b) within an equivalent group-velocity (generalized Snell) framework, but the dispersive ray laws in the Fermat effective-index formulation — in particular for perpendicular motion and the backward-refraction thresholds — and the case of two uniformly moving media remain without a published treatment, and no experiment has exercised the backward-refraction regime.

Background

Fermat's principle of least time states that the path taken by a light ray between two points is the one that minimizes the travel time. In optics, a 'ray' is the path of light energy, which travels at the group velocity (the velocity of a wave packet), and this in general differs in direction from the wave normal, which advances at the phase velocity (the velocity of surfaces of constant phase). In a medium at rest that is homogeneous (uniform throughout), isotropic (properties independent of direction), and nondispersive (refractive index nn independent of frequency), the two directions coincide and Fermat's principle yields the ordinary Snell law of refraction at an interface.

When the medium moves, several new effects appear. Light dragging (Fresnel or Fizeau drag, confirmed in Fizeau's 1851 running-water experiment) is the partial entrainment of light by a moving medium: the light speed and direction in the laboratory frame depend on the medium velocity. Relativistically, a medium that is isotropic in its own rest frame becomes optically anisotropic — direction-dependent — in a frame where it moves uniformly, because Lorentz velocity addition tilts the photon velocity depending on its direction relative to the medium velocity vv. Gjurchinovski and Skeparovski (2007) quantified this with an effective refractive index nef(φ)n_{ef}(\varphi), the ratio of the vacuum light speed to the photon speed in the laboratory frame, depending on the angle φ\varphi between the ray and the medium velocity, and inserted it into Fermat's principle to derive ray refraction formulas for a photon incident from vacuum onto a uniformly moving material half-space, for motion parallel to the interface and for motion perpendicular to it. For perpendicular motion against the incident light they identified a 'backward refraction' regime, in which the refracted ray bends so far that its tangential component reverses: it first appears (at sufficiently oblique incidence) once the medium speed exceeds κ1=(1+n2)−1/2\kappa_{1}=(1+n^{2})^{-1/2}, and occurs for all incidence angles once the speed exceeds κ2=1/n\kappa_{2}=1/n (speeds in units of cc). They also explained why earlier plane-wave electrodynamic treatments (Tai, Yeh, and successors), which track the propagation vector rather than the ray, give different angular formulas: those formulas describe refraction of the wave normal (phase velocity), while Fermat's principle tracks the ray (group velocity); the two coincide only when the medium is at rest. This reconciliation is already given inside the 2007 paper.

Independent of the least-time approach, modern work treats moving media through Lorentz transformations applied to the dispersion relation (the relation between wave frequency and wavevector) in the medium's rest frame, yielding 'generalized Snell laws' for both wavevector and group velocity. Along this route, Langlois, Braud and Gueroult (2025) derived transmission conditions for a wave refracted at the interface of an anisotropic, dispersive medium in uniform motion and applied them to a magnetized plasma (an ionized gas with an imposed magnetic field, in which wave propagation depends on direction and frequency); Langlois and Gueroult (2026) extended the method to nonuniform velocity fields using gradient-index (GRIN) optics — the classic technique of discretizing an inhomogeneous medium into thin uniform layers and applying refraction laws between them — and recovered, for nondispersive isotropic media, the results of the Gordon-metric approach, which describes a moving dielectric by an effective spacetime metric exploiting the analogy between moving media and gravitational fields. Notably, that work states that within the covariant, Fermat-type (Gordon-metric) formulation a description valid for general dispersive isotropic media is still missing, and a 2026 review (Gueroult, Braud and Langlois) concludes that motion effects on waves in plasmas remain largely unexplored. The open question is whether and how the original Fermat least-time construction extends beyond its original nondispersive, vacuum-interface, two-dimensional setting, and what remains to be worked out given these partial advances.

Problem Statement

Extend the Fermat principle of least time, as applied to a light ray (photon path) refracted at a uniformly moving interface, beyond the nondispersive, vacuum-to-medium, two-dimensional cases solved in the source paper. Specifically, derive the ray refraction laws — through the direction-dependent effective refractive index of the moving medium and its induced optical anisotropy — for: (a) media that are dispersive in their rest frames, i.e. with rest-frame refractive index nˉ(ω′)\bar n(\omega') depending on the rest-frame frequency, including how dispersion modifies the effective index nef(φ)n_{ef}(\varphi) and the fate of the backward-refraction regime and its threshold speeds κ1=(1+n2)−1/2\kappa_{1}=(1+n^{2})^{-1/2} and κ2=1/n\kappa_{2}=1/n; (b) general three-dimensional geometries in which the plane of incidence need not be normal to the interface nor parallel to the medium velocity; and (c) an interface between two uniformly moving non-vacuum media having different refractive indices in their rest frames and possibly different velocities, the planar interface between them moving uniformly with a prescribed velocity. The common objective is a least-time formulation of ray refraction at uniformly moving boundaries that is closed in the same sense as the source's two-dimensional nondispersive formulas: explicit refraction laws, recoverable as ordinary Snell's law when the media are at rest, and a delineation of any new ray-path regimes induced by motion, such as backward refraction. All three parts (a), (b) and (c) are required, and each is to be obtained within this least-time, effective-index formulation; that an equivalent group-velocity framework already covers parts of (a) and (b) does not by itself discharge them.

References

  1. A. Gjurchinovski and A. Skeparovski, 'Fermat's principle of least time in the presence of uniformly moving boundaries and media', European Journal of Physics 28, 933–951 (2007). DOI: 10.1088/0143-0807/28/5/017; arXiv:physics/0702100, https://arxiv.org/abs/physics/0702100
  2. A. Gjurchinovski and K. Trencevski, 'Huygens' construction in a dispersive medium moving at a constant velocity' (2007), arXiv:0711.0585, https://arxiv.org/abs/0711.0585
  3. J. Langlois, A. Braud and R. Gueroult, 'Fresnel drag in a moving magnetised plasma', Journal of Plasma Physics 91, E47 (2025). DOI: 10.1017/S0022377825000182, https://doi.org/10.1017/S0022377825000182
  4. J. Langlois and R. Gueroult, 'Light drag in nonuniformly moving anisotropic media through the lens of gradient-index optics', Physical Review E 113, 045205 (2026). DOI: 10.1103/8gx6-rnrf; arXiv:2512.01718, https://arxiv.org/abs/2512.01718
  5. R. Gueroult, A. Braud and J. Langlois, 'Wave drag in moving plasmas: Recent developments and prospects', Physics of Plasmas 33, 062107 (2026). DOI: 10.1063/5.0329641, https://doi.org/10.1063/5.0329641
  6. A. Gjurchinovski, 'Aberration of light in a uniformly moving optical medium', American Journal of Physics 72, 934 (2004). DOI: 10.1119/1.1648683, https://doi.org/10.1119/1.1648683