Role of Relative Stability in Self-Repair and Self-Maintenance

Abstract

This chapter opens with a discussion of relative stability, from the point of view of the physics of information storage. Bistable elements that store information are of two kinds: (1) static devices with performance resembling that of particles occupying either one of several energy (potential) wells in a field; (2) dynamic systems consisting of structures that by energy conversions and dissipations hold one of several possible steady states. The latter, dynamic type is examined here.

The relative stability of a locally stable state of a dynamic system depends strongly on the character of the noisealong paths between two (or more such states). No amount of study of the possible terminal states, by themselves, will suffice to discover the more probable one. To do that, we must take account of the fluctuations in the unlikely intervening states between those that are locally stable. This account does not require that a distinction be made between “internal” and “external” noise.

Because of fluctuations, a multistable system may be buffeted or agitated out of the noisy regions of its state space, then “condense” into a low-noise “cold trap” region, where it remains. Simplistic schemes based only on “information” or “entropy production” cannot suffice as accounts of relative dynamical stability in the presence of fluctuations nor of the constructive aspects of noise for self-organizing systems such as life.

“Self-organization” is in some respects an unfortunate term, because it hints at an extraphysical “élan vital”, which we, as physical reductionists, should drive out of our explanations. Even worse, the term “self-organizing” is sometimes applied to Bénard cells in hydrodynamic fields exposed to strong thermal gradients, but considering that under the constraints these “structures” are the only physically allowed pattern of behavior, do they really represent “self-organization” any more than does an electron circulating about an atom in a quantum state? Surely the notion of “self-organization” has often been used too grandiosely.

Experience with technological machinery shows that it always tends to wear out. true, a scheme of maintenance can sustain its operations for long or very long epochs. But can a technological machine be made self-maintaining, so that it persists with form and function intact, without the intervention of people acting as deus ex machina? How does life accomplish this without help from humans? Two schemes of maintenance or adaptation are examined: (1) slow and careful modidfication of parts or blueprints and (2) reproduction, variation, and descent with modification. the scheme by living systems to ensure persistence (especially at the species level of organization) is of the latter kind. The role of random mutations in that scheme resembles that of nucleation in the physical case of first-order phase transitions. The fluctuation leads from one locally stable state to another, different stable state. Mutations do notresemble physical second-order phase transitions that require that a symmetry-breaking threshold exceeded. After the threshold is exceeded, the more symmetrical original state is no longer stable—anyfluctuation then drives the system away from it. Though such bifurcations are of great interest in the mathematical field of qualitative dynamics, it is not clear that they pertain to the unfolding of the evolution of the terrestetrial biosphere.

The chapter closes with two related questions: (1) Will a system close to but not at (thermal) equilibrium evolve, and if so, does the level of complexity that can be achieved depend on the degree of departure from equilibrium? (2) Does simplicity of environment (e.g., uniformity) prevent evolution. It is suggested that complex systems require for their continuation or origination large deviations from equilibrium, and nonuniform, rich environments. —The Editor