Cycles

Cycles are pretty common in the Universe. Galaxies go through cycles of star formation and decay. Atoms are cycled through stars and back into the interstellar medium. Here on Earth, we have the hydrological cycle, with water evaporating from the oceans or elsewhere, condensing and falling as precipitation, and flowing back to the ocean. Oceanic crust and other materials from the interior of the planet is extruded at mid-ocean rifts and volcanos, cycled across oceans and subducted back into the interior. Oxygen and carbon atoms have their own cycles through rock, ocean, and atmosphere.

Closest to our interest here are the cycles of life. Birth, growth, death and decay take place in a cosmic instant, but these cycles have something crucial in common with all the others - they are all heat engines, powered by taking in energy at low entropy and putting it out at higher entropy. The same of course is true of the cycle that powers your automobile.

All the cycles mentioned are in some regards self-sustaining, as long as we don't run out of low entropy energy. Galaxies are powered mainly by gravitation energy, stars by a combination of gravitation and thermonuclear energy, continental motion by the heat energy released by radioactive decay - itself stored up in mainly gravity powered supernovae.

Most of the power for life comes from the Sun, but some life depends on chemical energy from the Earth's internal heat engine.

Life, we say, is special because it can reproduce and evolve. So can any of those other cycles evolve? They do, of course, but it's a pretty crude sort of evolution, without obvious internal memory, driven just by the changes occuring in their environments.

The metabolism first version or life's origin imagines that, on a primitive planet, chemical cycles could sustain themselves and give rise to descent with modification - the substrate for natural selection. It is an appealing notion, but one that has yet to be demonstrated. The key trick, it seems, is to get that modification in descent parameter tuned correctly. If the cycle is too sensitive to modification, any change will kill it, and natural selection won't lead to any evolution. If the cycle is too insensitive to modification, or too easily modified, selection will be overwhelmed by randomness. An even more critical problem is to actually find any cycle that could fit the most basic qualifications of self-sustainment and reproduction.