Samuël Borza

HOME

A non-smooth theory for sub-Lorentzian geometry

Funded by the European Union

SUBLOR is a research project funded by the European Union under the Horizon Europe programme's Marie Skłodowska-Curie Actions Postdoctoral Fellowships, grant agreement No. 101282277.

The project

The SUBLOR project is devoted to sub-Lorentzian geometry, the non-holonomic analogue of Lorentzian geometry and the Lorentzian counterpart of sub-Riemannian geometry.

A horizontal causal curve tangent to a distribution

A sub-Lorentzian space is a manifold \( M \) together with a family of admissible directions \( \mathcal{D} \subseteq \mathrm{T}(M) \), a Lorentzian inner product \( \langle \cdot, \cdot \rangle \) on those directions, and a choice of time orientation in \( \mathcal{D} \). Typically the number of independent directions available is smaller than the dimension of \( M \), so motion is constrained: not every tangent direction in \( \mathrm{T}_x(M) \) is available.

A curve \( \gamma \) is called horizontal when its velocity stays inside \( \mathcal{D} \). It is causal if \( \langle \dot\gamma(t), \dot\gamma(t)\rangle_{\gamma(t)} \le 0 \), chronological if \( \langle \dot\gamma(t), \dot\gamma(t)\rangle_{\gamma(t)} < 0 \), and it is future-directed when it points consistently along the chosen time orientation. The length of a causal curve is given by the usual Lorentzian formula \[ L(\gamma) := \int_0^T \sqrt{-\langle \dot\gamma(t), \dot\gamma(t) \rangle_{\gamma(t)}} \, \mathrm{d}t. \]

The time-separation function \( \tau(x,y) \) is obtained by taking the supremum of these lengths among all horizontal future-directed causal curves from \( x \) to \( y \): \[ \tau(x, y) := \sup\left\{ L(\gamma) \;\middle|\; \gamma \text{ is horizontal, future-directed causal, from } x \text{ to } y \right\}. \] The associated causal and chronological futures and pasts, denoted by \( J^\pm(x) \) and \( I^\pm(x) \), are defined as in Lorentzian geometry, but using only horizontal causal or timelike curves.

This setup models constrained spacetimes where motion is limited to a subbundle of directions, analogous to sub-Riemannian geometry but in a Lorentzian setting. The figure above illustrates how a causal curve \( \gamma \) must remain tangent to \( \mathcal{D} \), with its velocity \( \dot{\gamma}(t) \) lying in the admissible future-directed cone. Every Lorentzian manifold is a sub-Lorentzian manifold by taking \( \mathcal{D} = \mathrm{T}(M) \), so sub-Lorentzian geometry generalises Lorentzian geometry.

While recent developments have extended Lorentzian geometry to the non-smooth setting and sub-Riemannian geometry has become a well-established and active area of research, sub-Lorentzian geometry remains largely underexplored. The project SUBLOR, mentored by Michael Kunzinger, will address this gap by:

  1. Integrating sub-Lorentzian structures into the framework of Lorentzian length spaces and metric spacetimes.
  2. Characterising the infinitesimal geometry of sub-Lorentzian manifolds via timelike metric tangents.
  3. Developing curvature notions adapted to sub-Lorentzian geometry, including optimal-transport-based conditions and Hamiltonian curvature.

If you would like to hear more about SUBLOR, discuss related research, explore possible collaborations, or get involved, do get in touch!

News

I took part in the Lange Nacht der Forschung, Austria's largest public event dedicated to science, research, and innovation, offering free activities for all ages. Our station featured games, demonstrations, and visual material on Einstein's theory of special and general relativity, and curvature. A personal highlight was Hyperbolica, a non-Euclidean adventure game in which players explore bizarre curved worlds filled with puzzles.

Today marks the launch of the new MSCA research project SUBLOR on the foundations of sub-Lorentzian geometry!