Home > Positioning > Subjects > Thermodynamics

Thermodynamics

Thermodynamics began in 1824 with a practical question — how much work can be got from a heat engine — and produced an answer whose form turned out to be general: every transformation has a cost, and the cost is not negotiable by ingenuity. Its laws were established before anyone knew whether matter is made of atoms, and they survived the discovery that it is, untouched. That indifference to substrate is the field’s most distinctive feature, and the reason the same formalism is now applied to gases, magnets, living cells, computation, and black holes.

It is also a field that argues with itself, at the foundations, in ways that have not settled. What entropy is, whether thermodynamics is a theory in its own right or statistical mechanics awaiting complete reduction, why processes run one way when the underlying laws run both — these are open questions with named positions on several sides. This bundle gathers what the field says and where it disagrees.

What the field is about

There is no single formulation, and the differences are not cosmetic.

The oldest framing is constraint: thermodynamics states what cannot happen. Einstein put it as a theory that “seeks by analytical means to deduce necessary conditions, which separate events have to satisfy, from the universally experienced fact that perpetual motion is impossible” — a principle theory, in the distinction he drew in 1919, starting from empirically discovered general constraints rather than constructing phenomena from a hypothesised microscopic scheme. On that reading the substrate-indifference is no accident: a theory that never commits to a substrate cannot be embarrassed by one.

A second framing is equilibrium and the macroscopic: thermodynamics describes the behaviour of systems in terms of pressure, temperature and volume, with equilibrium as the state an isolated system reaches when left alone. This is the textbook-adjacent statement and the most neutral.

A third takes the subject to be the direction of time — the one place physics distinguishes past from future. The identification is common and is also more complicated than it looks: Jos Uffink’s “Bluff Your Way in the Second Law of Thermodynamics” (2001) is largely devoted to separating time-reversal non-invariance from irreversibility, which are routinely run together.

A fourth locates the subject in the relation between macroscopic and microscopic descriptions — though that is a claim about statistical mechanics, and treating the two as one field already takes a side in the dispute below.

A fifth, more recent: control. David Wallace’s “Thermodynamics as Control Theory” (2014) reads the subject as the theory of what an agent with limited control over a system’s degrees of freedom can and cannot make it do — indexing thermodynamics not to a scale but to a degree of access.

One finding cuts across all of these, and Uffink is its source: there is no single second law. The formulations of Clausius, Kelvin, Planck, Carathéodory, and Lieb and Yngvason are not straightforwardly equivalent to one another. Speaking of “the” second law conceals a family.

The bundle

The core

The laws and the reduction — the four laws and why there is no single second law; Boltzmann and Gibbs; what the H-theorem does not prove; and the dispute over whether thermodynamics reduces to statistical mechanics.
What entropy is — four quantities sharing a name, and the readings that divide over what they measure: objective, epistemic, relative to a physical coupling, and relative to the means of manipulation available — with Gibbs’s formalism, which most of the field computes in, distinguished from any of them.

Regimes

Away from equilibrium — Onsager and the near-equilibrium regime, the several competing non-equilibrium frameworks, Prigogine’s programme in its three very different parts, and the stochastic thermodynamics that now covers small systems.

Readings

Information and computation — Maxwell’s demon, Szilard’s engine, Landauer’s principle and its experimental confirmation, and the unresolved argument over whether the principle does the foundational work claimed for it.
The quantum regime — a family of second laws rather than one, erasure at negative cost, thermalisation without a bath, and the unresolved question of what work means in quantum mechanics.
Black holes — the laws of black hole mechanics, Bekenstein entropy and Hawking radiation, and whether the correspondence is literal or analogical.

The working subject

The working subject — chemical thermodynamics and the free energies, the engineering cycles the field began with, and bioenergetics: the settled trunk where nearly all use lies, largely untouched by the disputes above.

A note on the arrangement. These divisions are one map among possible ones, not the field’s own taxonomy — there is no consensus taxonomy. Two axes cross here: a foundational one running from phenomenological thermodynamics to statistical mechanics, and a regime one running from equilibrium to the small-system limit. The readings above are not a further rung on either but cross-cutting reinterpretations applied at several points at once, and the working subject is largely independent of both.

Principals

Internal: Carnot’s successors Boltzmann — the statistical reading of entropy — Gibbs — ensembles and the free energies — and Maxwell, whose demon has outlived most of the physics of its era; Prigogine on dissipative structures; Landauer and Bennett on the thermodynamic cost of computation; Shannon, whose entropy shares a formula with Gibbs’s; Einstein on what kind of theory this is; Hawking on black holes; Rovelli on entropy relative to a physical coupling.

External: Sadi Carnot (1824, the founding analysis), Rudolf Clausius (who coined “entropy” in 1865), William Thomson, Lord Kelvin, Lars Onsager (the reciprocal relations), Leó Szilárd (the 1929 engine), E. T. Jaynes (the information-theoretic reading), Udo Seifert and Christopher Jarzynski (stochastic thermodynamics), Jacob Bekenstein (black hole entropy), and among philosophers of physics Jos Uffink, John Norton, Craig Callender and Robert Batterman.

See also: Entropy — the word across its other homes · Phase transitions · Emergence · Relational quantum mechanics · Complex adaptive systems