Evolution and the second law
One of the best things I've read about evolution and the second law of thermodynamics is an essay called "Thermodynamics and Evolution" by John W. Patterson in Scientists Confront Creationism, edited by Laurie R. Godfrey.
Briefly, a prominent creationist named Henry Morris argued that the second law of thermodynamics means that evolution is impossible. (Patterson shows that Morris did not actually come up with this argument; among other things, R. E. D. Clark and E. H. Betts anticipated the argument. But, Patterson notes, Morris made the argument popular in a book called The Genesis Flood and in his subsequent writing.)
Patterson offers useful analogies that highlight the distinction between local and global increases in entropy. The second law requires that global entropy increases as a result of any thermodynamic process, but not that the entropy of any particular object or system increase. In fact, common processes like metabolism do produce decreases of entropy in some places (inside the body) while producing increases of entropy elsewhere (like in the body's waste products and in the waste heat the body releases into the environment). Patterson cites a discussion from a 1978 textbook by Ira N. Levine in support of this description of what the body does; Levine specifically attacks the notion that "life processes violate the second law", which Patterson thinks is a necessary (if inconvenient) consequence of creationist reasoning about the second law.
One example Patterson does not use is a computer. Anyone who has been in the market for a computer power supply, or who has used a recent laptop, will be aware that modern computers consume a lot of power and generate a lot of waste heat. (They also generate thermal noise and a great deal of waste electromagnetic radiation -- not all of which is quite as disordered as we might wish, given that sometimes useful information about the internal state of the computer or the computer's display can be recovered from it!)
One reason that computers do consume so much power and generate so much waste heat is precisely that they are constantly producing extremely low entropy within their internal circuitry. The state of the computer's processing circuits is highly ordered and remains so over a long period of time even as external inputs are applied to them. (There are actually some rules about how much power must be consumed -- as a matter of physical necessity -- in order to perform certain computations accurately. These rules are derived in the thermodynamics of computation, something I don't understand very well, but a field that's been developed very extensively since the 1970s. Actual computers consume a lot more power than they theoretically would need to, but actual engines are less efficient than an ideal heat engine, too, so it's no surprise that we can't build machines at the ideal of thermodynamic efficiency.)
But a computer is an increasingly familiar machine that generates a huge amount of entropy immediately outside itself (at its exhaust vents) in the process of producing an extremely low entropy within itself. So is a metabolizing organism -- an example Patterson does present. Many people using recent computers are also aware that computers behave erratically when their heat sinks or other cooling systems fail; in that case, they can't move their waste heat outside to dissipate it into the environment, and indeed, their internal entropy increases beyond appropriate engineering tolerances, and they begin to make errors. It is also true that an organism that can't dissipate waste heat effectively is likely to suffer illness as a result of an unreasonably high internal temperature. We need to rely on the outside environment as a kind of enormous heat sink into which we constantly dump entropy. If we couldn't do that, we would die.
Once we have done this, other organisms -- using an external energy source and producing their own waste products -- organize some of our waste products into a useful form again. As lots of people know, this activity couldn't continue without an immense and constant energy input from the sun. Within the sun's power, individual organisms couldn't get the energy they need to keep organizing themselves and to keep disorder out. (I realize that the thermodynamics here are more complicated than what I have described.)
One oddity is that Henry Morris seems to recognize the concept that an open thermodynamic system may experience a decrease in entropy -- but he introduces a spurious requirement that the system "must also contain a 'program' to direct" that decrease. Many people answer Morris by talking about the difference between closed and open systems, but Morris seems to try to add an additional criterion of the "program".
In the case of the computer, for instance, the "program" must consist in the layout of the circuitry, for presumably applying wall current directly to any random circuit has a decent chance of increasing its entropy. And in the case of the body, the "program" must consist in those structures that enable metabolic processes to take place.
Since Patterson does not address Morris's concept of a "program", some people may conclude that Patterson has not actually rebutted Morris. After all, Patterson stresses the open system/closed system distinction, but Frank Steiger notes that Morris -- at least in his The Remarkable Birth of Planet Earth -- has added a "program" criterion. But as Steiger observes, thermodynamics does not actually contain such a criterion, and it is definitely not a part of the second law, which limits itself to a conclusion about the total entropy in the universe or in some other closed system.
If there were such a thing as Morris's "program" principle, it would have to come from somewhere other than the second law of thermodynamics. Patterson, for example, describes a machine called a hydraulic ram pump, which "can get low-lying water to pump part o itself uphill" -- an example of "an uphill or backward process [that] can spontaneously occur in nature by being [...] coupled to a more dominant downhill process [in such a way that] the downhill process can actually drive the other process 'backward' or in the so-called uphill direction". This means that, in the presence of a hydraulic ram with no external power source, a portion of a stream of water flowing downhill can provide the means of making another portion of the stream flow uphill. Presumably, if Morris tried to answer this, he would argue that the hydraulic ram is an instance of a "program" and that the hydraulic ram itself could not arise by chance, which is to say spontaneously.
Steiger calls that "misleading" and "arbitrar[y]" because many systems do exhibit behavior that is analogous to the behavior of the hydraulic ram. (Steiger doesn't actually discuss the hydraulic ram example; instead, he uses an example shared with Patterson: the formation of extraordinarily highly ordered snowflakes spontaneously in the atmosphere. It's Morris's program criterion that Steiger sees as "misleading" and "arbitrar[y]" because it isn't actually a principle of physics but simply a way of speaking that can easily lead to incorrect intuitions about which processes are actually possible.)
Patterson's discussion of coupling one process to another -- as in the case of the hydraulic ram -- is extremely illuminating. It reminds me of a remarkable quotation from Tommaso Toffoli, which is given in several different forms.
In a sense, nature has been continually computing the "next state" of the universe for billions of years; all we have to do -- and, actually, all we can do -- is "hitch a ride" on this huge, ongoing Great Computation.
or
We never perform a computation ourselves, we just hitch a ride on the great Computation that is going on already.
In a similar sense, we might say that, in order to do anything at all, we just hitch a ride on the entropy flows that are already going on around us; we couple things we care about to them, so that the mighty flow of increasing entropy will momentarily help decrease the entropy within us and ours.
