Ilya Prigogine (1977), Nobel Lecture, 8 December 1977 (Les Prix Nobel, Chemistry 1977, pp. 263–285); reprinted in Science 201(4358):777–785, 1978 — DOI-10.1126/science.201.4358.777
Key takeaways
- A “dissipative structure” is a new, ordered state of an open system that appears when a steady state far from equilibrium becomes unstable: a fluctuation that would die out near equilibrium is amplified and then held in place by a continuous exchange of energy (and matter) with the surroundings.
- The minimum-entropy-production theorem is not the law of these structures. It holds only for steady states close enough to equilibrium that the flow–force relations are linear with constant coefficients. Far from equilibrium the lecture says behaviour can be “even opposite” to it, and Bénard convection raises entropy production.
- The only general far-from-equilibrium criterion given is a sufficient condition for stability. Instability needs specific kinetics: in chemistry, autocatalytic steps, which are necessary but not sufficient.
Standing in the record
Filed on 2026-10-02 at the owner’s request, as the read statement of the dissipative-structure theory in the record’s new coverage of Prigogine’s programme and its canonical systems. Its two monographs, Glansdorff and Prigogine’s LIT-537 and Nicolis and Prigogine’s LIT-524, are filed Deferred beside it, and the lecture is the reading that stands in for them: it cites the first for the stability theory (chs. II, V, VII) and the second for the Brusselator (ch. VII).
It bears on two documents the record already holds.
- The demarcation dispute in LIT-192. Nahas and Sachs report a dispute over separating organisms from candle flames and tornadoes. The lecture gives the physicist’s side of that line: what makes something a dissipative structure is an instability of an open system held far from equilibrium, with the new state kept going by flows across its boundary. Nothing in that definition separates a convection cell from a cell. The lecture does name “biochemical cycles involving oscillatory enzymes” and a town or “a living system” as having “functional order” of this kind (§1). It does not argue that being a dissipative structure is sufficient for life or for goal-directedness, and it offers no criterion that would draw that line.
- The “Dissipative Self-Organizing Systems” cluster in LIT-211. The 21 definitions in that cluster rest on the thermodynamic picture stated here. The lecture is a fair test of what such a definition can carry: a far-from- equilibrium, flux-maintained, fluctuation-born order, which the Brusselator and Bénard convection satisfy as well as any organism.
The scope of the minimum-entropy-production theorem is the point most often misreported downstream, and the lecture is explicit (p. 266): the theorem “requires even more restrictive conditions than the linear relations” and is valid in a “strictly” linear theory. The parallel filings of the critics and of the later entropy-production principles are outside this entry.
No instruction for machine-learning practice; nothing here belongs in the anthology.