1. Physics did not discover that everything is unstable, but that stability did not exhaust the real
Classical physics first advanced by isolating regularities, invariants and systems simple enough to be described mathematically. The mechanics of Galileo and Newton did not, however, assume that all systems tend toward equilibrium. An ideal mechanical system can oscillate, rotate or pursue its motion indefinitely without converging toward a state of rest. Nor does the principle of least action describe a general tendency toward stability: it makes it possible to determine the trajectory followed by a system among the evolutions compatible with its constraints.
Non-equilibrium thermodynamics established that open systems, traversed by flows of matter or energy, can maintain or produce organised structures. These dissipative organisations are not motionless states. They exist thanks to the transformations they sustain and disappear when the flows that support them cease. Fluctuations can favour a bifurcation or select one configuration among several possibilities, without constituting the sole cause of all organisation. A continuous supply of energy can thus maintain structures that would not be accessible at thermodynamic equilibrium.
Chaotic dynamics likewise showed that a deterministic law does not guarantee a durable prediction. In the model studied by Edward Lorenz in 1963, very close initial states can produce rapidly diverging trajectories, even though the evolution remains entirely governed by deterministic equations. Chaos is neither an absence of law nor a universal instability. It denotes a particular regime in which the dynamics remains bounded while being aperiodic and extremely sensitive to initial conditions.
Contemporary physics leads to a more precise conclusion: the real reduces neither to stability nor to instability. It contains stable states, metastable states, transitions, oscillations, chaotic regimes and organisations sustained far from equilibrium. The central question then bears on the conditions that allow certain configurations to maintain themselves within this dynamic diversity.
2. Organised systems channel transformations
Organisation does not begin with the biological living. Long before the appearance of the first cells, matter had already produced systems capable of temporarily preserving a structure thanks to the interactions, flows and constraints of their environment.
A star such as the Sun offers a major example. It persists thanks to a dynamic regime in which gravity tends to compress matter while internal pressure opposes collapse. That pressure is bound up with the extreme temperatures of the stellar plasma and with the energy released by nuclear fusion in the core. The resulting hydrostatic equilibrium is not an immobility. The Sun continually transforms part of its matter, transports energy from its core to its surface, radiates into space, progressively loses mass, and acts on its environment through its radiation, its magnetic field and its solar wind.
The Sun remains an organised physical system and not an organism. It possesses no cellular metabolism, no hereditary reproduction and no Darwinian evolution. The comparison bears solely on dynamic persistence under constraints. It must not efface the differences of mechanism and of autonomy.
From stars to active planets, from geochemical cycles to organisms, the Universe produces different forms of dynamic persistence. They rest neither on the same mechanisms nor on the same capacities. Some structures last thanks to the physical interactions that constitute them. Others sustain more complex cycles, exchanges or regulations. The biological living crosses a particular threshold by integrating an active boundary, a metabolism, repair mechanisms, a transmissible memory, reproduction, adaptation and evolution.
3. The biological living filters, damps or exploits fluctuations
The biological living belongs to this older history of organised matter, but it introduces a new form of persistence. It does not merely maintain a structure under the effect of the physical interactions that constitute it. It actively sustains the conditions of its own viability through metabolism, regulation, repair, memory and reproduction.
Homeostasis does not correspond to the immobility of an organism around a perfectly fixed point. It denotes a set of regulations that keep certain variables within ranges compatible with life. Temperature, chemical concentrations, pressure, energy availability and hormonal activity fluctuate continually. Feedback loops correct some variations, tolerate others, and may modify the reference values according to the history or the environment of the system. Homeostatic mechanisms thus produce dynamic regimes, sometimes relatively constant, sometimes oscillatory.
The articulation between robustness and variation then becomes essential. An organism too sensitive to every perturbation would rapidly lose its coherence. An entirely rigid organism could no longer respond to a changing environment. Robustness protects certain functions despite perturbations, while plasticity allows the functioning or the form to be modified when conditions require it. Evolvability preserves, at the scale of lineages, the possibility of producing variations liable to be selected.
Certain biological architectures protect essential functions while maintaining margins of transformation. At the scale of populations, genetic and phenotypic diversity distributes the possible responses to variable environments. This diversification results from evolutionary mechanisms and presupposes no conscious search for instability.
The major transitions of evolution shed further light. Entities previously capable of reproducing or acting relatively autonomously were integrated into new levels of organisation: molecules brought together in systems of transmission, chromosomes, eukaryotic cells, multicellular organisms or animal societies. These transitions required new forms of coordination, of information transmission and of limitation of internal conflict.
The eukaryotic cell and multicellularity illustrate the integration of entities into a wider level of organisation. New coordinations appear, with risks of conflict or proliferation. The persistence of the higher level depends on mechanisms that limit these internal dynamics.
Biological stability may be defined as an active and transformable persistence. It depends on exchanges with the environment, on a metabolism, on repair mechanisms, on regulation and on capacities of adaptation.
4. The analytical path was necessary in order to understand complexity
The human mind cannot grasp all the relations of a complex system simultaneously. Science therefore advanced by isolating objects, controlling experimental conditions and seeking invariants. This analytical reduction made it possible to define with precision mass, energy, entropy, symmetries, feedbacks, attractors and bifurcations.
The difficulty appears when this method of decomposition is taken for a complete description of the real. Knowing the constituents of a system does not suffice to explain its organisation. The properties observed also depend on their configuration, their interactions, their environment and their history.
The science of complex systems does not overturn analytical physics. It extends it by reintroducing the relations that analysis had provisionally isolated. It seeks to understand how identifiable mechanisms combine to produce collective behaviours, regimes of stability, transitions and new capacities.
Analysis and reconstruction are complementary. The first identifies the constituents and the mechanisms. The second examines how their interactions produce an organisation, a regime of stability or a transition.
This reconstruction does not consist in attributing an intention to matter. To say that a system « seeks » to persist can only be a metaphor as long as no mechanism of representation, decision or anticipation is present. A star does not choose its hydrostatic equilibrium. It adopts a regime determined by physical interactions and by the initial conditions that made its formation possible.
Organisms modify their behaviour, their physiology or their environment according to internal and external signals. Cognition, learning and representation add levels at which certain consequences can be anticipated and several actions selected.
5. Distinguishing regimes of persistence
A rigorous demonstration requires distinguishing several notions that are often confused. A fluctuation is a variation around a value or a regime. A dynamic instability appears when a small perturbation carries the system away from its initial state or trajectory. Metastability denotes a state capable of persisting for a long time while remaining liable to tip over into another state. A non-equilibrium steady regime preserves certain macroscopic properties thanks to continuous flows.
Robustness corresponds to the conservation of a structure or a function despite perturbations. Regulation modifies certain internal processes in order to maintain conditions compatible with the continuity of the system. Resilience denotes its capacity to recover a viable functioning after a perturbation. Biological adaptation corresponds to characters or modifications that improve the viability of an organism or a population in a given environment.
These properties are not interchangeable. A crystal can be stable without being regulated. A flame can maintain a form thanks to a flow without possessing biological autonomy. A star can display physical feedbacks without having hereditary mechanisms. A cell, for its part, combines an active boundary, metabolic flows, internal regulation, repair and a transmissible molecular memory.
The Universe contains stable, metastable and unstable regimes as well as systems maintained out of equilibrium. Material structures appear and persist when their interactions and their environment make certain regimes accessible and sufficiently durable.
Physical systems can temporarily maintain certain properties thanks to their interactions and their flows. Biological systems add an organisational autonomy. They maintain a boundary, select exchanges, transform resources, repair certain degradations and produce a part of their own components.
Reproduction and heredity extend this persistence beyond the individual. Evolution transforms populations across generations. In some organisms, learning and memory also modify responses over the course of a lifetime.
This historical continuity does not efface the thresholds. Physical organisation, biological autonomy, Darwinian evolution and representation rest on different mechanisms. Their comparison becomes useful when it specifies these differences instead of presenting the living as a merely higher degree of complexity.