More Options

Does Evolution Have a Thermodynamics Problem?

Jason Rosenhouse

May 19, 2006

Writing in a recent issue of the politically conservative magazine The American Spectator, mathematician Granville Sewell revived the argument that conventional evolutionary theory runs afoul of the second law of thermodynamics. When he later made available an online lecture promoting the same argument, ID advocate William Dembski endorsed it enthusiastically at his blog.

This is highly significant. You see, the thermodynamics argument is one of the very worst creationists have ever used. The argument is wrong, of course. But more than that it is wrong in a way that betrays an extreme simple-mindedness about science in general and physics in particular. Consequently, among scientists the thermodynamics argument has become a symbol for the sort of mind-numbing ignorance that is the stock-in-trade of creationists.

It was precisely this level of silliness that ID proponents were keen to avoid. They, after all, were supposed to be serious scientists, not Bible thumping extremists. Now we have William Dembski, who provides most of the tiny amount of intellectual oomph the ID folks can claim, supporting this ridiculous argument. In so doing he has completely sacrificed any pretensions to seriousness ID might once have entertained.

The basic creationist argument is this: The second law states that a spontaneous, natural process can only lead to an increase in the entropy of a system. Entropy is roughly a measure of disorder or complexity. So the second law implies that natural processes can only cause things to become more disordered and less complex over time. But evolution asserts that natural processes have caused organisms to grow more complex over time. This is a contradiction, and since no one is inclined to abandon the second law, evolution must not be correct.

I first encountered this argument before I had devoted any time to learning the basic principles of thermodynamics. Even then, it struck me as suspicious. For one thing, I was aware that the second law of thermodynamics had been formulated well before Darwin did his work. Accepting the creationist argument as valid would require that several generations of scientists had simply overlooked an elementary contradiction between evolution and thermodynamics. That seemed unlikely.

Furthermore, the second law appeared to play only a rhetorical role in the creationist argument. After all, you don’t need fancy principles of thermodynamics to argue that the growth in complexity of organisms over time is something that requires a special sort of explanation. It is a simple fact of everyday life that without maintenance things tend to break down and fall apart.

Everyone agrees that the growth in complexity that evolutionists claim took place over the course of natural history requires an explanation. And biologists have one. Many generations of natural selection acting on random genetic variations can cause the average complexity of organisms to increase. This is not theoretical. Natural selection has demonstrated its ability, in both the field and the lab, to increase the level of order and complexity in organisms. The same principle is at work in artificial life experiments, and in the use of evolutionary algorithms in engineering problems. Granted, the experiments I am referring to tend to show relatively small increases in order, but that is enough to establish that no principle of thermodynamics prohibits known evolutionary mechanisms from increasing biological complexity.

Sewell himself inadvertently concedes this. When it comes time for him to explain why natural selection is not an adequate explanation for the growth of biological complexity, thermodynamics goes out the window. Instead he simply parrots the irreducible complexity argument of which ID folks are so fond. For example, in a discussion of the insect-catching apparatus of the aquatic bladderwort he writes:

The development of any major new feature presents similar problems, and according to Lehigh University biochemist Michael Behe, who describes several spectacular examples in detail in Darwin’s Black Box (Free Press, 1996), the world of microbiology is especially loaded with such examples of “irreducible complexity.”

It seems that until the trigger hair, the door, and the pressurized chamber were all in place, and the ability to digest insects, and to reset the trap to be able to catch more than one insect, had been developed, none of the individual components of this carnivorous trap would have been of any use. What is the selective advantage of an incomplete pressurized chamber?

Recall that Darwin’s Black Box is the book in which Michael Behe introduced the irreducible complexity argument in its modern form.

Sewell goes on like this at considerable length in an attempt to persuade the reader that natural selection is incapable in principle of explaining the growth of biological complexity over time. The arguments he makes in this regard are entirely incorrect biologically, but of more relevance to the present discussion is that they have nothing to do with thermodynamics.

Let me make things even simpler. Things that are thermodynamically impossible do not occur. But natural selection does occur, and is undeniably capable in principle of explaining increases in biological complexity. Therefore, there is no principle of thermodynamics that says that evolution is an incorrect theory.

But let’s push this a bit further. Depending on the particular textbook from which you learn your thermodynamics, you may find several different statements all labeled as the second law. There are certain physical situations in which it can be useful to think of the second law as a statement about order and disorder, and some books present it in that context. In other situations it might be presented as a statement regarding the efficiency of an engine in converting heat into work. Popular level treatments will sometimes describe the second law by the simple statement that heat always travels from a hot body to a cold body. There are other formulations as well.

Also of importance is the nature of the system to which the second law applies. Some texts require that the system under consideration be isolated from its surroundings, so that neither matter nor energy is crossing the boundary. Typically such restrictions are made to simplify the presentation, not to reflect any fundamental requirement of the physics of the situation. By contrast, others will allow even the most general situations, in which anything, either matter or energy, is allowed to cross the boundary.

In my view, however, the second law is best viewed as a purely mathematical statement. It says that the change in entropy of a system in going between two states must be larger than a certain mathematical quantity (the integral of dQ over T, for those who know some calculus and some thermo notation). The technical details of what this means need not detain us here. This is the formulation that is most relevant to determining whether or not evolution runs afoul of the second law.

If you make the added assumption that your system is completely isolated from the outside world, then the integral I mentioned ends up having the value zero, and the second law tells us that the change in entropy must be positive. In other words, the entropy must increase in this situation. Creationists of old tended to ignore this assumption, and argued simply that the second law rules out any possibility of natural forces causing order to increase. As a result, scientists generally replied that the Earth is not an isolated system, since we receive copious amounts of energy from the Sun. That is certainly an important observation, and it does, indeed, refute some primitive versions of the second law argument.

The second law still applies when energy is crossing the boundary of the system, and in this case it says that the change in entropy must be equal to or larger than the mathematical function I mentioned previously. Entropy can, indeed, decrease in this situation, but the second law still makes a definite statement about the magnitude of that decrease. Sewell understands this, and gives a tolerable, if highly non-technical, description of this fact. However, formulating the second law mathematically makes it clear that Sewell cannot merely assert that some process (evolution by natural selection in this case) violates the second law. There is a very clear test to pass to show that a given process really has a second law problem.

You see, any claim that evolution violates the second law must be backed up with a calculation. Sewell believes that the second law is a problem for evolution? Very well. Let him evaluate the integral I mentioned and show that the change in entropy has been smaller than it should be. Anything short of that is no longer an argument based on thermodynamics. It is just ye olde argument from personal incredulity, in which Sewell is expressing nothing more than his own disbelief that biological complexity could have evolved naturally. Since every formulation of the second law allows for local increases of order and complexity, the mere observation of such increases does not constitute an apparent violation of thermodynamic principles.

The reason Sewell will not carry out this calculation is that he cannot. No one can. Entropy calculations are always carried out in the context of a reversible process, and no one has the faintest idea how to describe a reversible process for assembling an organism from its component substances. That is why serious scientists do not try to apply the second law to biological processes in the simple-minded ways ID folks prefer.

Sewell, however, has another trick up his sleeve. He prefers to formulate the second law as a statement about probability. In his American Spectator article he writes:

Natural forces, such as corrosion, erosion, fire and explosions, do not create order, they destroy it. The second law is all about probability, it uses probability at the microscopic level to predict macroscopic change: the reason carbon distributes itself more and more uniformly in an insulated solid is, that is what the laws of probability predict when diffusion alone is operative.

To see what he is talking about, consider that the historically earliest formulations of the second law, and indeed of all the basic principles of thermodynamics, were made without any reference to the fact that matter is made up of smaller particles. Eventually scientists realized that viewing matter as a large ensemble of individual molecules could shed some additional light on basic thermodynamic principles. This realization led eventually to the development of statistical mechanics; “mechanics” because it was devoted to explaining the motions of particles, “statistical” because it was concerned not with the trajectory of any individual particle, but rather with the average behavior of a large number of particles. Sewell’s statement likening the second law to an assertion about probability only makes sense within this context.

The basic idea is this: Given a box that is filled with gas and has been sitting, untouched, for some time, we expect the gas molecules to be distributed roughly evenly throughout the box. We would be very surprised to find all of the gas on one side of the box with empty space on the other side. We can explain this in terms of probability: There are vastly more configurations in which the molecules are distributed roughly evenly than there are where all the molecules are on one side of the box. So other things being equal, we can say that it is vastly more likely that we will encounter one of the even distributions. The distributions in which the gas is evenly distributed can plausibly be said to be less ordered, and therefore have higher entropy, then the highly uneven distributions. This permits a probabilistic interpretation of the second law.

Sewell’s insistence that the second law is fundamentally a probabilistic statement will come as news to most physicists. Statistical mechanics offers one way of looking at the second law, but it is neither more fundamental nor better than the classical view. It is merely different, superior in certain situations but not as effective in others. In his book Understanding Thermodynamics, H. C. Van Ness described the situation this way:

Statistical mechanics adds to thermodynamics on its theoretical side, as a means for or as an aid to the calculation of properties. The other half of thermodynamics, the applied half, benefits only from a wider availability of the data needed in the solution of engineering problems. Although statistical mechanics is based on the presumed reality of atoms and molecules, it does not provide, any more than does thermodynamics, a detailed description of atomic and molecular behavior and of atomic and molecular interactions. However, it does provide, as thermodynamics does not, the means by which thermodynamic properties may be calculated whenever detailed descriptions of atomic and molecular behavior are provided from other studies, either theoretical or experimental. Thus statistical mechanics adds something very useful to thermodynamics, but it neither explains thermodynamics nor replaces it. (Emphasis added)

Returning now to Sewell’s argument, we find him summing up his thinking here with the following formulation:

In these simple examples, I assumed nothing but heat conduction or diffusion was going on, but for more general situations, I offered the tautology that “if an increase in order is extremely improbable when a system is closed, it is still extremely improbable when the system is open, unless something is entering which makes it not extremely improbable.” (Emphasis in original)

To which I reply, “Yes, of course. But so what?”

Accepting for the moment Sewell’s idiosyncratic terminology, we can say that if we take the Earth by itself as our system, then there is definitely something entering to make an increase in biological complexity more likely. The solar energy received by the Earth fuels the chemical reactions that allow living organisms to survive and reproduce. This cycle of survival and reproduction ultimately leads to natural selection, which can, in turn, lead to increases in biological complexity. Minus that energy living organisms would quickly go extinct and evolution would not occur.

Nor can Sewell retreat to the question of life’s origin, for then he must confront the fact that the various sources of energy bathing the early Earth would have fueled chemical reactions that are believed to have led to the first primitive life forms. Once again, it is for him to back up his claims about probability with something more substantive than his own beliefs.

In his online lecture Sewell describes evolution as being a “film running backwards” by which he means that we see complexity increasing in apparent violation of the second law (just as in a movie run backwards you might see the shattered pieces of a broken coffee cup reassemble themselves into a functional vessel). If he wants us to take this claim seriously, he needs to follow the dictates of his own theorizing. Does evolution require us to believe that something incredibly improbable has occurred in the course of natural history? Let him carry out the probability calculation that shows that to be the case. Then let him explain what significance his calculation is supposed to have. (Improbable things happen all the time, after all). Once he has done that, he will have an actual argument, and we can revisit this subject at that time. Without such a calculation, he has only a lot of polysyllabic bluster.

Sewell will have no more luck carrying out these probability calculations than he had with the prior (entropy) calculation. And that is because these sorts of probabilities are effectively impossible to calculate. The probability of any particular set of outcomes of several billion years of evolution depends on far more variables than can possibly be included in a practical calculation. Probability theory finds many applications in biology, but this is not one of them. There is a reason real scientists do not talk about probability calculations in Sewell’s haphazard manner.

The pattern in Sewell’s arguments is now rapidly becoming clear. When he wants to impress us with the rigor and sophistication of his arguments, he talks about entropy and order and probability and the history of thermodynamics. But when it comes time to apply any of this to evolution he retreats to simple-minded assertions about films running backwards and atoms arranging themselves into microchips. The reason he does this is that, in reality, thermodynamics and probability play no role at all in his argument. More precisely, they play no scientific role. They do play a strong rhetorical role, however, since the casual use of technical scientific jargon is an effective means of confusing lay people.

Sewell closes his essay as follows:

The development of life may have only violated one law of science, but that was the one Sir Arthur Eddington called the “supreme” law of Nature, and it has violated that in a most spectacular way. At least that is my opinion, but perhaps I am wrong. Perhaps it only seems extremely improbable, but really isn’t, that, under the right conditions, the influx of stellar energy into a planet could cause atoms to rearrange themselves into nuclear power plants and spaceships and computers. But one would think that at least this would be considered an open question, and those who argue that it really is extremely improbable, and thus contrary to the basic principle underlying the second law, would be given a measure of respect, and taken seriously by their colleagues, but we aren’t. (Emphasis in original).

Knowledgeable people will not show any respect for Sewell’s argument, because he has produced virtually no argument at all. He describes it as his opinion that evolution violates the second law. This is not the sort of thing about which scientists are supposed to have opinions. We have ample evidence that evolution happened and that natural selection was the driving force of it. Biologists find evolutionary thinking to be very helpful in their research. If Sewell believes that it runs afoul of the second law nevertheless, then he needs to carry out the calculations that show that to be case. Otherwise he has only an opinion based on nothing.

These sorts of considerations should be obvious to anyone with a modicum of mathematical or scientific training. That they are not obvious to Sewell is another reason his quest for respect will be in vain.

Jason Rosenhouse

Jason Rosenhouse is the author of EvolutionBlog, providing commentary on developments in the endless dispute between evolution and creationism.