The universe evolves by particles transitioning between modes of motion, no more and no less. Such transitions occur at a mean rate governed by quantum mechanics. Regardless of the interpretation of quantum mechanics, transitions occur randomly within a defined system of particles and between the system and its environment. Evolution is deterministic (unitary) only between wave function transition events, aka quantum jumps, induced by interaction.

This statement is a crucial departure from standard theory. Diffusion cannot be derived in classical mechanics or unitary quantum theory. Deterministic mechanics produces laminar motion in which coherence recurs - half of the systems in a randomly prepared ensemble would show increasing coherence rather than the observed decreasing trend. Imposing stochastic forces does not simulate collapse consistently.

Physics experiments are designed to limit interaction such that deliberate measurement is the primary cause of collapse. "Irreversible" or "spontaneous" trends manifest when the resolution of a measurement is greater than the system correlation time and length. In other words, physics "emerges" in the limit of negligible diffusion, i.e. long correlation time, but diffusion does not emerge in deterministic Lagrangian mechanics as system complexity increases.

Open quantum system (OQS) theory has developed along with our recent experimental ability to control microsystems. If the microsystem has only a few degrees of freedom, then it is reasonable to assume dissipation is due to coupling with its environment. Alternative approaches deduce master equations for the evolution of the system density matrix by some assumed stochastic mechanism in the environment [wiki][Review] or for evolution of the microsystem wave function by quantum jumping [Review]. The new theory generalizes the latter approach for complex systems in which each particle acts as both a microsystem and the environment of its neighbors.

The aggregate of all particle mode transitions resolve into macroscopic flow and diffusive flux. All particles diffuse toward states of local stability, establishing well-defined phases of matter with distinct thermodynamic characteristics such as heat capacity, compressibility, diffusivity, marginal particle energy, etc. The local environment is as important as composition of a particle in determining thermodynamics. Therefore, particles should be distinguished by their material phase and elemental constitution, and identified as members of distinct species. All particle dynamics may be viewed as transition among their modes or transformation into other species. For example, transformations among particles of different composition are chemical reactions, and evaporation-condensation involves transformation between gas and liquid phases of the same composition.

All thermodynamic properties may be derived from the mode spectrum available to each particle and the quantum transition rates among them. Far out of equilibrium, conditions are highly local and variable and few particles share the same spectrum. Computing system behavior in this case would require enormous resources and extreme precision. Diffusion tends to stabilize local conditions, the mode spectrum and distribution. Quasi-equilibrium is the regime when the spectrum is stable enough that the mean particle mode distribution, averaged over a measurement period, approximates the limiting equilibrium case of zero net mean flows of particles, momentum and energy among systems. Near equilibrium flows may be approximated by assuming that adjacent systems are locally thermalized.

Macroscopic here refers to the time-average of any property of the present system (not an ensemble average), which depends on the choice of measurement period. Thermodynamic equations balance change in these mean values. A property value at any instant fluctuates randomly about this trend. Fluctuations do not enter thermodynamic analysis beyond initiating phase transitions and random walk. A measurement period much shorter than the system correlation time produces erratic mean values while much longer obscures dynamics. The distinction between macroscopic and microscopic motion, and the collective effect called friction, is clear in large systems when measurement resolution is larger than the correlation length and time. Empirical "natural laws" describing bulk behavior lose validity for smaller mesoscopic systems.

Practical analysis requires independent kinetic equations for the rate of exchange between species and among modes within a species. The latter produces a Maxwellian distribution in quasi-equilibrium, independent of mode transition rates. The former does depend on mode transition rates. Clausius and Gibbs both made a simplifying foundational assumption in standard theory that has caused endless confusion since by neglecting these rates.

"All systems diffuse toward a steady state'' replaces the Second Law ("Total entropy never decreases", "Heat never flows from colder to warmer by itself", etc.) as a general description of local evolution, even during external or active disruption.