The World of SPLectrum
Home > Positioning > Subjects > Phase transitions
Phase Transitions
Phase transitions are matter changing state: water boiling, iron magnetising, an alloy ordering. Physics distinguishes two kinds. First-order transitions — boiling among them — proceed by latent heat and coexistence: two phases side by side, an interface, an abrupt jump. Continuous transitions are stranger and became the field’s central subject: as a critical point is approached — the temperature at which a ferromagnet loses its magnetism, the point where liquid and gas cease to differ — fluctuations grow to every scale, and quantities diverge in power-law fashion, characterised by critical exponents. An order parameter — magnetisation, density difference — goes continuously to zero, and the transition is marked by a change of symmetry rather than a jump. The field built around these phenomena, phase transitions and critical phenomena, produced what is widely regarded as among twentieth-century physics’ most consequential findings about how levels of description relate.
The field’s development
The canonical object is the Ising model — spins on a lattice, each interacting with its neighbours — and the founding exact result is Onsager’s solution of the two-dimensional case (1944), which settled that a genuine transition can arise at all from a finite-range interacting model. The general framework came from Landau: order parameters and symmetry as the organising concepts, with mean-field theory as the workhorse approximation. Mean-field’s quantitative failure is the field’s productive crisis — it predicts critical exponents, and experiment returns different ones — and the steps out of the crisis set up everything below: Widom’s scaling hypothesis (1965) showed the exponents obey exact relations among themselves, as if governed by a single underlying scale-invariance, before anyone could say why.
Universality
The discovery is called universality, and it is an experimental fact before it is a doctrine. Measure the critical exponents of a liquid-gas system at its critical point; measure those of a uniaxial ferromagnet — a different material, different forces, different constituents. The numbers are identical. Systems with little in common at the micro-level fall into a small number of universality classes, and membership is fixed by remarkably little: the spatial dimensionality, the symmetry of the order parameter, and the range of the interactions, since long-range forces change the class. Everything else about the microscopic details — the lattice structure, the strength of the couplings, the chemistry — makes no difference to the critical behaviour. Where physics expected the macro to depend on the micro, it found, at criticality, a quantified and measurable independence.
The mechanism — the renormalisation group
Universality was explained, not just observed. Leo Kadanoff’s block-spin picture (1966) made Widom’s scaling intuitive: average a magnet’s spins over blocks, treat each block as a new spin, and repeat — the system viewed at successively coarser scales. Kenneth Wilson turned the picture into machinery (1971; Nobel Prize 1982): under this repeated coarse-graining, systems trace a flow, and at criticality the flow runs to a fixed point. Different microscopic systems flow to the same fixed point, and the critical behaviour is a property of the fixed point, not of the starting system — which is the mechanism of universality. The micro-details that distinguish the systems are, in the technical vocabulary, irrelevant operators: contributions that shrink away under the flow. The word is exact — the scheme does not merely ignore the details; it shows they do not matter to the critical behaviour, and says which ones.
What the renormalisation group delivers, in the vocabulary of levels: an account of macro-autonomy. That the same macro-behaviour arises across different micro-constitutions is not asserted as a mystery of levels but exhibited — with the analysis showing why a description in order-parameter language succeeds without tracking the constituents. Symmetry breaking, the companion concept (Anderson’s “More is Different” made it the emblem of level-autonomy), enters the same story: the ordered phase has less symmetry than the laws that govern it, and its description needs vocabulary — the order parameter — that the micro-level does not use.
The wider family
“Phase transition” reaches past equilibrium criticality. Kosterlitz–Thouless transitions are topological — order changing without an order parameter in the usual sense (Nobel Prize 2016). Quantum phase transitions occur at zero temperature, driven by quantum rather than thermal fluctuations. Non-equilibrium transitions — Prigogine’s dissipative structures the classic case — arise in driven systems far from equilibrium. And the vocabulary has an extensive life beyond physics: percolation thresholds, epidemic transitions, abrupt transitions in networks, the criticality hypothesis in neuroscience, and Bak’s self-organised criticality — systems that tune themselves to the critical state.
The philosophical dispute
The field hosts a sharply defined philosophical debate, because here both sides argue over mathematics they share.
Robert Batterman (The Devil in the Details, 2002, and the universality papers since) reads the physics as vindicating a distinctive, non-reductive form of explanation: asymptotic reasoning. Universality is explained by what survives as information is systematically excluded, and the explanation essentially involves singular limits — the thermodynamic limit in which the number of constituents goes to infinity, where alone the mathematics of the transition (non-analyticity) strictly lives. On his reading, neither the reductionist framing nor the classical emergentist one captures what limit-explanations do; his own framing of the field’s lesson is that universality explains precisely by showing which micro-details never needed deriving.
Jeremy Butterfield (“Less is Different: Emergence and Reduction Reconciled,” 2011) answers that the drama is unnecessary: emergence — novel and robust behaviour at the limit — and reduction are compatible, the infinite limit a convenience rather than a metaphysical necessity, with the relevant behaviour already approximated at large finite size. A body of work alongside — Callender, Menon, Palacios, Wu — presses the same line against the essential-idealisation reading: real systems are finite, real transitions happen, and the physics does not need the actual infinity the strong reading leans on. The dispute — whether limit-idealisations are essential to the explanation or eliminable conveniences — is live and technical.
Persons
Anderson — symmetry breaking, “More is Different”; Prigogine — dissipative structures; Bak — self-organised criticality. The field’s other principals are external: Kenneth Wilson (the renormalisation group), Leo Kadanoff (block spins), Michael Fisher (critical phenomena), Robert Batterman and Jeremy Butterfield (the philosophical dispute), Robert Laughlin (emergent protectorates).
See also: Complex adaptive systems · Anderson · Bak · Prigogine · Bedau · Kim · Emergence