Biological organisms involve elaborate infrastructure — organs, bones, and so on
that are themselves (mostly) far-from-equilibrium and self-maintained in constant
overturning of molecular constituents. What renders some parts of the overall far-from- equilibrium system as infrastructure is that infrastructure involves slower time scales of interaction and often higher energy levels than other interactions of the systems. Such infrastructure is required in order to control and to enable the multiple processes — metabolic processes — necessary to self-maintenance and to the creation, selection, and execution of recursively self-maintenant interactions (Ruiz-Mirazo & Moreno, 1998).
The robot is far-from-equilibrium only in the sense that its operations require power; the basic existence of the robot is not far-from-equilibrium, and does not require
self-maintenance or recursive self-maintenance. In that sense, the robot does not have much at stake in its self-maintenant activities, at least relative to the biological case. At issue is the metaphysical significance of those differences for the emergence of function and representation.
Any types of emergent phenomena are likely to have marginal cases, and the robot
example is clearly marginal. But the nature of its marginality is important, I contend, because it illustrates how profoundly far-from-equilibrium conditions, thus self-maintenant
and recursively self-maintenant processes, permeate the emergence of function and representation in paradigm biological cases. Robots can be marginally far-from- equilibrium by design, and, therefore, perhaps marginally self-maintenant and recursively self-maintenant. They do not, however, form an evolutionarily marginal case. Neither function nor representation could have evolved via such a kind of marginality.
There is a fundamental difference between the far-from-equilibrium character of living beings and that of robots that organizes this marginality that robots illustrate into a kind of continuum. Living beings are irreversible far-from-equilibrium systems, while robots, in general, exhibit a reversibility. A robot can be turned off, then on again; if its
battery runs down, nothing in particular is necessarily lost if that battery is later recharged.
The thermodynamics of living systems are not reversible. If the far-from-equilibrium processes that constitute them are halted, they cannot be restarted. This is the deeper
perspective on the significance of far-from-equilibrium infrastructure and metabolism in living systems: it is not just the activities of living systems that require far-from-equilibrium conditions, but their very ontological constitution is far-from-equilibrium, and it begins to
irreversibly move toward equilibrium once that far-from-equilibrium process is sufficiently interrupted (including by the accumulation of errors during normal functioning — aging).