Geometric Thermodynamics on Lie Groups

31 concepts10 papers

In classical thermodynamics, temperature is a scalar value read from a thermometer, and entropy is a macroscopic measure of disorder. While these definitions work well for simple gases in stationary boxes, they falter when applied to systems with complex symmetries, relativistic speeds, or abstract statistical manifolds. When a system rotates or moves through space-time, temperature is no longer just a magnitude—it gains a directional dependency.

To resolve this, a modern research direction unifies thermodynamics, statistical mechanics, and information geometry by reframing physical states through the lens of Lie groups. By treating thermodynamic properties as intrinsic geometric features of symmetry groups, this framework provides a fully covariant description of heat, entropy, and probability.

From Scalars to Vectors: Geometric Temperature

The core insight of this geometric framework is that thermodynamic variables should be defined as elements of a Lie algebra and its dual space, rather than as simple scalars.

If a physical or statistical system is invariant under the action of a Lie group $G$ (which represents its symmetries, such as rotations, translations, or gauge transformations), its underlying dynamics are governed by the corresponding Lie algebra $\mathfrak{g}$. In this setting, Geometric Temperature $\beta$ is defined as a vector residing within the Lie algebra $\mathfrak{g}$. For example, in a relativistic setting, this vector manifests as a space-time vector combining both thermal and inertial components.

To pair with this vector, the corresponding thermal observables—such as mean energy or heat—reside in the dual space $\mathfrak{g}^*$. As explained by Frederic BARBARESCO in Koszul Information Geometry and Souriau Geometric Temperature/Capacity of Lie Group Thermodynamics, the natural pairing $\langle \beta, U \rangle$ between the temperature vector $\beta \in \mathfrak{g}$ and the heat/momentum vector $U \in \mathfrak{g}^*$ yields a dimensionless quantity. This directly generalizes the classical thermodynamic term $\frac{E}{kT}$, but ensures that the relationship remains perfectly covariant under any transformation of the coordinate system.

Coadjoint Orbits and Generalized Gibbs States

How do we construct statistical distributions, like the familiar Gibbs distribution, in a way that respects these symmetries? The answer lies in the coadjoint representation of the Lie group on its dual algebra $\mathfrak{g}^*$.

In classical mechanics, a system's phase space can often be mapped to the dual algebra $\mathfrak{g}^*$ via a mathematical tool called the momentum map. The group $G$ acts on this dual space, carving it into distinct geometric surfaces known as coadjoint orbits.

To find the most likely state of a system under these constraints, researchers apply the principle of maximum entropy. As demonstrated by [François Gay-Balmaz] et al. in Lie Group Cohomology and (Multi)Symplectic Integrators: New Geometric Tools for Lie Group Machine Learning Based on Souriau Geometric Statistical Mechanics, leveraging the equivariance of these group actions allows for the definition of generalized Gibbs probability densities directly on Lie groups.

Because these distributions are tied to the geometry of the group itself, they remain invariant under the group's actions. This provides a rigorous method for defining "Gauss densities" on non-Euclidean manifolds, which has historically been an open challenge for statistical modeling on complex spaces (Lie Group Statistics and Lie Group Machine Learning Based on Souriau Lie Groups Thermodynamics & Koszul-Souriau-Fisher Metric).

Entropy as a Topological Feature

One of the most elegant outcomes of this geometrization is that entropy ceases to be an ad-hoc statistical count and becomes a fundamental topological property.

In pure classical mechanics, systems are governed by symplectic geometry, which preserves volume and describes completely reversible, non-dissipative motion. Thermodynamics, however, requires dissipation and the irreversible increase of entropy.

In Jean-Marie Souriau's Symplectic Foliation Model of Sadi Carnot's Thermodynamics, Frederic BARBARESCO demonstrates how Lie group thermodynamics bridges this gap. The dual space $\mathfrak{g}^*$ naturally breaks down into a collection of nested "symplectic leaves" (a foliation). Reversible, conservative dynamics are confined to move entirely within a single symplectic leaf. Irreversible, dissipative processes correspond to movement transversely across different leaves.

Within this geometric picture, entropy emerges as a Casimir function—a mathematical invariant that remains constant along any given leaf but changes value as the system transitions between leaves. This provides a purely geometric origin for the second law of thermodynamics: dissipation is the physical manifestation of a system changing its topological leaf.

Unifying Information Geometry and Physical Heat

In parallel to physics, statisticians use Information Geometry to measure the "distance" between probability distributions using the Fisher Information Metric. Similarly, physicists studying thermodynamic stability use metrics based on the Hessian of internal energy or entropy (such as Weinhold and Ruppeiner geometries) to identify phase transitions, a technique applied to complex systems like black holes by Haotian Liu in Geometric Analysis of Black Hole with Primary Scalar Hair.

Lie group thermodynamics unifies these disparate metrics through Koszul Hessian geometry. By taking the Hessian of the log-partition function (the characteristic function of the Lie group), one derives the Koszul-Souriau-Fisher metric (Geometric Theory of Heat from Souriau Lie Groups Thermodynamics and Koszul Hessian Geometry).

   Thermodynamic Geometry                 Information Geometry
 (Weinhold / Ruppeiner Metrics)         (Fisher Information Metric)
               \                                   /
                \                                 /
                 v                               v
             [ Koszul Hessian Geometry on Lie Groups ]
                                 |
                                 v
                  Koszul-Souriau-Fisher Metric

This metric acts as a grand unification:

  1. From a statistical perspective, it measures information distance on the Lie group manifold.
  2. From a physical perspective, its components explicitly calculate the system's generalized heat capacity.

This geometric interpretation means that physical concepts like thermal capacity and statistical concepts like variance are revealed to be the exact same geometric curvature under different names. This insight forms the bedrock for advanced machine learning algorithms that operate natively on symmetric data structures (Higher Order Geometric Theory of Information and Heat Based on Poly-Symplectic Geometry of Souriau Lie Groups Thermodynamics and Their Contextures).

Extending to Non-Equilibrium Systems

While much of the foundational theory focuses on generalized equilibrium states, the geometric approach naturally extends to non-equilibrium and dissipative kinetics.

To describe systems out of equilibrium, the distinct geometric origins of conservative Hamiltonian mechanics and dissipative gradient dynamics must be reconciled. As explored by [Oğul Esen] et al. in On the role of geometry in statistical mechanics and thermodynamics. I. Geometric perspective, frameworks like the GENERIC formalism achieve this by utilizing cotangent lifts and contact geometry.

By lifting particle-level dissipative trajectories into a higher-dimensional contact manifold, the authors demonstrate how to bridge the gap between individual particle mechanics and macroscopic non-equilibrium thermodynamic laws (Conformal and Contact Kinetic Dynamics and Their Geometrization). This ongoing geometrization ensures that whether a system is in perfect equilibrium or undergoing complex dissipation, its behavior can be understood as a natural trajectory dictated by the geometry of space, time, and symmetry.

History & citationVersion 1 · 2026-10-04

Current · Permanent link

View frozen version 1

Cite this direction

Latest version (a living page):

@misc{lacuna_lie_group_thermodynamics_24925222,
  title = {Lie group thermodynamics},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/lie-group-thermodynamics/rdr_2492522274864a8ea834dbe85771dfba},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-20},
}

This exact version (version 1, frozen text):

@misc{lacuna_lie_group_thermodynamics_24925222_v1,
  title = {Lie group thermodynamics},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/lie-group-thermodynamics/rdr_2492522274864a8ea834dbe85771dfba/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-20},
}

See something wrong? Report a mistake