How Physics Fakes Design, and Makes Things Difficult for Theism

Alex Rosenberg

In this chapter, I attempt to accomplish three things: first, to show that physics is sufficient for all appearances of design in nature. Second, that the only way they can emerge consistently with physics is by natural selection. Third, that the particular physical process that drives adaptational evolution makes Darwin’s theory irreconcilable with theism. The standard way most Christian biologists seek to reconcile theism with Darwinism won’t work. It must either misconstrue the theory or treat it as merely useful fiction.

Physics Suffices for Adaptation by Natural Selection

A scientific explanation of adaptation has to be consistent with physics. So, it has to begin with zero adaptation and build up all the rest of the amazing adaptations of biology from the ground up by physics alone. We can’t even leave room for “stupid design,” let alone “intelligent design,” to creep in. Why? Because the process Darwin discovered is a purely physical process operating in a mindless world. If it has to start from even the first, slightest, sliver of design, it is no longer compatible with physics. For physics ruled out purpose, teleology, goals, ends, every economy of means to ends, when Newton overthrew the Aristotelian system.

Logical compatibility with physics is something that science requires from biology. Scientists have long demanded consistency with well-established physics as a requirement on all other theories in science. In fact, the nineteenth-century critics of Darwin’s theory were eager to adopt the standard of consistency with physics as a way to blow the theory out of the water. One of these opponents was Lord Kelvin, of second law fame. Soon after the publication of On the Origin of Species, Kelvin argued that the Darwinian theory of natural selection had to be false. Darwin estimated that at least three hundred million years had been required for natural selection to provide the level of adaptation we see around us. (He was off by three orders of magnitude.) But Kevin thought he could prove that the Sun was no more than twenty million years old. Given the amount of energy it generated and the best theory of heat production available at the time, Kelvin’s theory proposed that the upper limit on the Sun’s age was forty million years. So, there could not been enough time for natural selection to provide adaptations by the process of natural selection. Thus, Kelvin refutes Darwin.

Of course, Kelvin didn’t have the slightest idea what the real source of the Sun’s energy was. It was only after World War II that Hans Bethe won the Nobel Prize in Physics for figuring out that the Sun is a very long-lasting hydrogen-fusion-driven explosion. But in 1870, Kelvin’s objection had to be taken pretty seriously. Explanations of adaptation had to be compatible with the correct physical theory, and in 1870, Kelvin’s was the most well-informed guess about the physics of solar energy. Darwin himself owned up to being very worried about this problem, since he accepted the constraint of consistency with physics as a requirement for any theory of adaptation.

Although Darwinian biologists have not noticed, their theory needs more than mere consistency with physics. It’s not enough to show that Darwinian theory gets it mostly right, except for a few exceptions, a couple of adaptations produced by non-Darwinian processes at work in biology. To close down the wiggle-room, Darwinism needs to show that the only way adaptations can ever happen—even the most trivial and earliest of adaptation—is by natural selection working on zero adaptation. It needs to demonstrate that, given the constraints of physics, adaptation could have no other source than natural selection. We will sketch how such an argument goes, in two stages. First, we’ll show that natural selection doesn’t need any prior adaptation at all to get started: beginning with zero adaptations, it can produce all the rest by physical processes alone. Physics is sufficient for adaptation by natural selection. We need only the second law of thermodynamics to do this. Then, with the same starting point, the second law, we can show that the process Darwin discovered is, necessarily, the only way adaptations can emerge, persist, and be enhanced in a world where the physical facts fix all the facts. That will close down the wiggle room for any alternative source of adaptation in the universe. It will also accomplish the reduction of natural selection to physics.

The second law of thermodynamics tells us that, with very high probability, entropy, the disorder of things, increases over time.

But the biological realm seems to show the opposite of second-law disorder. It reflects persistent orderliness—start out with some mud and seeds, end up with a garden of beautiful flowers. The ever-increasing adaptation of plants’ and animals’ traits to local environments looks like the long-term increase in order and decrease in entropy. So, we have to square the emergence and persistence of adaptation with the second law’s requirement that entropy increases.

You could be excused for thinking that if adaptation is orderly and increases in it are decreases in entropy, then evolution must be impossible. This line of reasoning makes a slight mistake about entropy and magnifies it into a major mistake about evolution. The second law requires that evolution produce a net increase in entropy. Increases in order or its persistence are permitted. But they must be paid for by more increases in disorder elsewhere. Any process of emergence, persistence, or enhancement of adaptation must be accompanied by increases in disorder that are almost always greater than the increases in order. The “almost” is added because increases in entropy are just very, very probable, not absolutely invariable. It won’t be difficult to show that Darwin’s explanation of adaptation—and only Darwin’s explanation—can do this.

Natural selection requires three processes: reproduction, variation, and inheritance. It doesn’t really care how any of these three things get done, just so long as each one goes on for long enough to get some adaptations. Reproduction doesn’t have to be sexual, or even asexual, or even easily recognized by us to be reproduction. Any kind of replication is enough. In chemistry, replication occurs whenever a molecule’s own chemical structure causes the chemical synthesis of another molecule with the same structure—when it makes copies of itself or helps something else make copies of it. This can and does happen several different ways, in the test tube and in nature. The one most directly relevant for evolution on Earth is called template matching—the method DNA uses to make copies of itself.

First step: when atoms bounce around, some bind to one another strongly or weakly, depending on the kind of attraction there is between them—their chemical bond. When the bond is strong, the results are stable molecules. These molecules can only be broken up by forces that are stronger than the bonds. Such forces require more energy than is stored by the stable bond between the atoms in the molecule. Breaking down a stable molecule takes more energy than keeping it together.

Second step: occasionally, these relatively stable molecules can be templates for copies of themselves. Their atoms attract other atoms to themselves and to each other, so that the attracted atoms bond together to make another molecule with the same structure. The process is familiar in crystal growth. Start out with a cube of eight atoms in a solution of other atoms of the same kind. They attract another four on each side, and suddenly, the molecule is a three-dimensional cross. As it attracts more and more, the crystal grows from a small cube into a large one. The crystal grows in a solution through “thermodynamic noise”: the increasingly uneven and disorderly distribution of atoms just randomly bouncing around in the solution mandated by the second law. The atoms already in the crystal latch on to other ones in the crystal in the only orientation chemically possible, making the nice shape we can see when they get big enough.

A crystal molecule doesn’t just have to grow bigger and bigger. Instead, the molecule can set up chemical forces that make two or more other unattached atoms, that are just bouncing around, bond with one another, making new copies of the original crystal. Instead of getting bigger, it makes more copies of itself.

The process could involve more steps than just simple one-step replication. It could involve intermediate steps between copies. Think of a cookie that is stale enough to be used as a mold to make a cookie-cutter that takes the stale cookie’s shape. It’s a template—to make a new cookie. Make the new cookie, then throw away the cookie-cutter, let the new cookie go stale, and use it to make a new cookie-cutter. Make lots of copies of the same cookie using one-use cookie-cutters. You get the idea. (In a way, this is just how DNA is copied in the cell, using each cookie-cutter once, except the cookie-cutter is not thrown away. It uses it for something else.)

Physical chemists and organic chemists are discovering more and more about how such complicated structures arise among molecules. They are applying that knowledge in nanotechnology—the engineering of individual molecules. Pretty soon, they’ll be able to induce molecules to build any geometrical shape they choose. Often, the molecule of choice in nanotechnology experiments is DNA. As usual, there are people such as the late Michael Crichton and other commentators on nanotechnology who gain attention, warning us that we will soon be overrun by self-replicating Nano-robots. The reason they claim to be worried is the role of “self-assembly” in building Nano machines.

Chemists building these structures molecule by molecule are remarkable in themselves. What is truly amazing is that the structures assemble themselves. In fact, this is the only way nanotechnology works. There are very few molecules chemists can manipulate one at a time, putting one next to another, and then, gluing them together. All they can do is set up the right soup of molecules milling around randomly (“thermodynamic noise”), and then, wait while the desired structures are built by processes that emerge from the operation of the laws of physics. Just by changing the flexibility of small molecules of DNA in the starting soup at the bottom of the test tube, and changing their concentrations, chemists can produce many different three-dimensional objects, including tetrahedrons, dodecahedrons, and Bucky balls—soccer ball shapes built out of DNA molecules. Of course, what we can do in the lab, unguided nature can do better, given world enough and time.

Replication by template matching is even easier than self-assembly. And it works particularly well under conditions that the second law of thermodynamics encourages: the larger the number of molecules, and more randomly the molecules of different kinds are distributed, the better. These conditions increase the chances that each of the different atoms needed by a template will sooner or later bounce into it to help make a copy. In fact, “works well” is an understatement for how completely template replication exploits the second law.

Let’s assume that the mixture of atoms bouncing around in a test tube or in nature is very disorderly, and getting more so all the time, as the second law requires. As the disorderly distribution of atoms increases, the chances of different atoms landing on the template increase too. Most of the time, an atom bouncing into a template of other atoms is too big or too small or too strongly charged to make a copy-molecule that survives. Even if the new atom bonds to the others, the whole copy may break apart due to differences in size or charge or whatever, sending its constituent atoms off to drift around some more, increasing entropy, of course. In most cases, in the lab and out of it, this disorderly outcome of instability in copying is the rule, not the exception. The exception is, of course, a successful duplicated molecule.

Now, let’s add some variation to the replication. In effect, we are introducing mutation in template copying. Variation is even easier than replication to get going at the level of molecules. It’s imposed on molecules during the process of replication by some obvious chemical facts working together with the second law of thermodynamics.

One look at the columns of the Periodic Table of the Elements is enough to show how disorder makes chemically similar but slightly different molecules. In the table, fluorine is just above chlorine in the same column. They are in the same column because they react with exactly the same elements. Chlorine and sodium atoms bond together and made table salt; that means fluorine and sodium atoms will too (the resulting molecule is a tooth-decay preventer). The reason fluorine and chlorine atoms combine with sodium equally well is that they have the same arrangements of electrons that do the bonding. All that means is that if a chlorine and a fluorine molecule are both bouncing around and bump into the same template, they may both bond the same way with other atoms on the template to make similar, but slightly different molecules. A template with chlorine molecules in it could easily produce a copy-molecule that differs only in having fluorine where chlorine would normally go. Voila variation.

When chemical reactions happen billions of times in small regions of space and time, even a small percentage of exact copies quickly come to number in the millions, as does the percentage of slightly varied copies, with one or two atoms of different elements in place of the original atoms. Most of the time, the outcome of this process is wastage—a molecule that doesn’t replicate itself or falls apart, just as the second law requires. But sometimes—very rarely—variation produces a molecule that is slightly better at replicating, or one that is just a little more stable.

Now, we have replication and variation. What about fitness differences, the last of the three requirements for evolution by natural selection? Fitness is easiest to understand at the level of molecules bouncing around in a world controlled by the second law. Molecules that form from atoms are stable for various durations of time. Some break apart right after forming, as a result of strong atomic forces such as charge. Some break apart because their bonding is too weak to withstand the force of other atoms that bounce into them or even just pass by. Some “fragile” molecules will remain intact for a while. They just happen by chance to avoid bouncing up against other molecules, ones with stronger charges that pull atoms away from their neighbors. Here again, the second law rears its head: as molecules bounce around, any amount of order, structure, pattern almost always gives way to disorder, to entropy. Hardly any molecule is stable for extremely long periods, with the exception of the bonded carbon atoms in a diamond crystal.

There are differences in stability among molecules, owing to the variations that inexact replication permits. Differences in stability have an impact on the replication of different types of molecules. A template molecule produces copies just by random interaction with atoms that bounce into it or pass close enough to be attracted. The longer the original templating molecule stays in one piece—that is, the more stable it is—the more copies it can make. Most of its copies will be just as stable as the original template-molecule since they will be exact duplicates. They will serve as templates for more copies, and so on, multiplying copies of the original.

Of course, just as there are differences in the stability of different molecules, there are differences in their rates of replication. The number of copies of their templates that can be made, and their stability will depend on their “environments”: on the temperature, the local electric and magnetic fields, and the concentration of other atoms and molecules, around them. Consider two molecules that differ from one another along the two dimensions of stability and ease of replication. The first remains intact on average for twice as long as the second; the second templates twice as many copies per unit of time as the first. Over the same period, they will produce exactly the same number of copies. What will the long-run proportions of molecules of the two types be? It will be one to one. As far as producing copies are concerned, the two different molecules will have equal reproductive fitness. And, of course, if their differences in stability and replicability don’t perfectly balance out, then after a time, there are going to be more copies of one type of molecule than of the other.

Molecules randomly bouncing around a region of space, and bonding to form larger molecules, will eventually, randomly, result in a few stable and replicating structures. Their structures will vary, and the variations will have effects on the molecules’ stability and replication rates. These differences, in the ability of molecules to stay intact (stability) and to allow for copies of themselves to form (replicability), will change the proportions of molecules in any region of space. If chemical and physical conditions in that region remain unchanged for long enough, the ratios of the different types of replicating molecules will eventually settle down to a fixed proportion. At that point, all the remaining replicating molecules in the region will be equally fit to survive—whether owing to their stability or replicability or varying combinations of both, they and/or copies of them persist. In other words, a purely physical process has produced molecular adaptation: the appearance, persistence, and enhancement of molecules with chemical and physical properties that enable them to persist and/or replicate, or both. Then, at some point, the chemical environment changes, slightly or greatly: temperatures rise or fall, new and different molecules diffuse through the region, the magnetic field strengthens or weakens. The process of adaptational evolution starts again, thermodynamically filtering for new stable, replicating molecules adapted to the new conditions.

As this process goes on, two other phenomena become inevitable: the size and complexity of the replicating molecules will increase. Eventually, there will start to be molecules that enhance each other’s stability and/or replication through their chemical relations to each other. There are no limits to the repetition of this process, making bigger and more complicated and more diverse molecules. If conditions are favorable, the result will be really big assemblies of stable and/or replicating molecules, for instance, RNA and eventually, DNA sequences and strings of amino acids—i. e., genes and proteins.

The rest is history, that is, natural history. The process we have described begins with zero adaptations and produces the first adaptation by dumb luck, sheer random chance that the second law makes possible. It just had to have happened this way if the physical facts fix all the facts, including the facts of adaptation.

Molecular biologists don’t yet know all the details of this natural history, or even, many of them. Some have been known for a long time. It was in the early 1950s that two scientists—Stanley Miller and Harold Urey—showed how easy it is to get proteins, sugar, lipids, and the building blocks of DNA from simple ingredients available early in the earth’s history. All they did was run an electric current through some water, methane, ammonia, and carbon. Chemists have been designing similar experiments ever since, getting more and more of the building blocks of terrestrial life. Biologists have discovered that the simplest, and oldest of organisms on the planet—the archeabacteria—probably first emerged at least 2.8 billion years ago, and still survive in volcanoes at the bottom of the sea. It is there, in such volcanoes at the base of the ocean under the highest temperatures and greatest pressure, that one finds chemical reactions spewing out boiling lava and producing the largest quantities of entropy on the planet. This is just what the second law requires to drive thermodynamic noise, and through it, to find stable and replicating molecules in a world of random mixing.

How much like the evolution of recognizably biological things—genes, cells, replicating organisms, is the molecular process we have described? Well, recognizably, biological evolution has three distinctive features:

First, natural selection usually finds quick and dirty solutions to immediate and pressing environmental challenges. More often than not, these solutions get locked in. Then, when new problems arise, the solutions to old problems constrain and channel the random search for adaptations that deal with newer problems. The results are jury-rigged solutions permanently entrenched everywhere in nature.

A second feature of biological evolution is the emergence of complexity and diversity in the structure and behavior of organisms. Different environments select among variations in different adaptations, producing diversity even among related lineages of organisms. The longer evolution proceeds, the greater the opportunities for more complicated solutions to continuing design problems.

A third feature of biological evolution is the appearance of cooperative adaptations and competing adaptations. The cooperative ones are sometimes packaged together in the same organism, sometimes separated in two quite distinct ones living symbiotically (like us and the e. coli bacteria in our guts). But natural selection also, and more frequently, produces competition between genes, individuals, lineages, populations, and species. There is always that chance that “arms races” will break out between evolving lineages of traits, in which a new random variation in one trait gets selected for exploiting the other traits it has been cooperating with, or a new variation randomly appears in a trait that enables it to suddenly break up a competitive stalemate it was locked into.

Each of these three features is found in the nano-evolution of the molecules. And each persists as molecular assemblies grow in size, stability, complexity, and diversity to produce organic life.

Lock-in of quick and dirty solutions: Only a few molecular shapes will ever allow for their own self-assembly and copying (by templating or otherwise). These shapes begin the lock-in that second law processes will have to work with in building the more stable, more replicable, bigger, and more complicated molecules.

Increasing diversity and complexity: Thermodynamic noise constantly makes more and more different environments—different temperatures, different pH, different concentrations of chemicals, different amounts of water or carbon dioxide or nitrogen, or more complicated acids and bases, magnetic fields, and radiation. As a result, there will be a corresponding selection for more and more different molecules. However, they will still be variations on the themes locked in by the earliest stages of molecular evolution.

Cooperation and competition: Some of these molecules will even start to work together, just by luck, having structures that enhance one another’s stability or replicability. Others will explode or dissolve or change their structures when they combine with one another. They will produce new environments that will select for new chemical combinations, ones that cooperate, ones that compete. And so on, up the ladder of complexity and diversity that produces assemblies of molecules so big they become recognizable as genes, viruses, organelles, cells, tissues, organs, organisms, . . . our pets, our parasites, our prey, and our predators . . . and us.

To summarize, molecules bouncing against one another inevitably follow a scenario dictated by the second law. Purely physical and chemical processes in that scenario are all that is needed for the emergence, persistence, and enhancement of adaptation through natural selection at the molecular level. Where and when molecules of some minimal size emerge, there will be some natural selection for chemical structures that confer more stability and replicability than what other chemical structures confer. These chemical properties are adaptations: they have functions for the molecules that exhibit them. They enable the molecules to survive longer than less stable ones in the same chemical bath, and to make more copies of themselves than other ones in the same bath. As a result of molecular natural selection, these molecules are better adapted to their molecular environments than others are.

This is an important outcome for Darwin’s theory. It faces the demand of showing how to get the merest sliver of an adaptation from zero adaptation by purely physical, chemical, thermodynamic processes. More of the same processes can build on that sliver of an adaptation to get more adaptations, and eventually, more robust adaptations. But it can’t cheat. It can’t just assume the existence of that first sliver. The second law makes the first sliver possible. Variation and selection can take it from there. Now, we have to show that this is the only way adaptations—molecular or otherwise—can emerge: by exploiting the second law operating on zero adaptations, to begin with.

Physics Makes Natural Selection the Only Way Adaptations Can Arise

The second law makes the merest sliver of an initial adaptation just barely possible. But it makes no guarantees. For all we know, it might happen only once every 13.7 billion years in an entire universe. If the first adaptation survives long enough, the second law allows for improvements, but only if they are rare, energetically costly, and just dumb luck.

But Darwin needs more from physics than just the possibility of adaptation. Physics won’t fix all the facts, unless the second law’s way of getting adaptations is the only way to get them. Darwinian theory needs to show that blind variation and environmental filtration is the sole route through which life could have emerged in any universe governed by the second law. We have to understand why, in a universe made only by physics, the process that Darwin discovered is the only game in town.

There are only three things we need to keep in mind to do this. First, and as already noted, if we are out to explain how any adaptation at all could ever happen, we can’t help ourselves to some prior adaptation, no matter how slight, puny or insignificant. Second, the second law requires that the emergence and persistence of orderliness of any kind be energetically expensive. An explanation of adaptation’s emergence and persistence is going to have to show that the process is wasteful. The more wasteful, the better as far as the second law is concerned.

The very first adaptation has to be a matter of nothing more than dumb luck. It had to be the random result of a very, very mindless process, such as a monkey hitting the keys of a typewriter randomly and producing a word. Given world enough and time, the bouncing around of things just fortuitously produces adaptations. The second law insists that initial adaptations—no matter how slight, small, or brief—can’t happen very often. And the same has to go for subsequent adaptations that build on these. They will have to be rare and random as well. So, any theory that explains adaptation in general will have to have this feature: that the first adaptation was a fluke, the luck of the draw, just an accident, simply the result of the law of averages. The inevitability of a first, slightest adaptation is ruled out by the second law: to begin with, the second law says that nothing is inevitable, even the heat-death of the universe. More important, the appearance of the merest sliver of an adaptation is an increase in order, and so, at most, improbable.

None of this will be a surprise to Darwinian theory. That’s just what the theory says will happen: variations are random, they are almost always small; in fact, they are almost always molecular; the first ones come out of nowhere, just as a matter of shuffling the cards; mostly, they are maladaptive, and only rarely are they adaptive.

The second law also requires that the process through which later adaptations emerge out of earlier ones be energetically expensive and wasteful: expensive because building more and more order has to cost more and more disorder; wasteful because the early steps—the original adaptations on which later ones are built—will be locked in so that less energetically costly ways of building the later adaptations are not available. Every explanation of adaptation will have to share this feature too. It will have to harness a wasteful process to create order.

Now, one has only to examine natural selection as it has unravelled on earth to see how expensive and wasteful it is. The combination of blind variation and environmental filtration is great at increasing entropy. In fact, the right way to look at the emergence of adaptation on earth is to recognize that it is the most wasteful, most energetically expensive, greatest entropy-producing process that occurs on the planet. The evolution and maintenance of adaptations by natural selection wins the prize for greatest efficiency in carrying out the second law’s mandate to create disorder. Forget design, evolution is a mess. This is a fact about natural selection insufficiently realized and not widely enough publicized in biology.

Examples are obvious. A female leopard frog will lay up to six thousand eggs at a time—each carrying exactly half of all the order required for an almost perfect duplicate offspring. Yet, out of those six thousand, the frog will produce only—on average—two surviving offspring. Some fish are even more inefficient, laying millions of eggs at one time just to make two more fish. Compared to that, a dandelion is miserly with seeds. It will spread only a thousand seeds and produce, on average, one copy of itself. But the human male competes well with profligate fishes. He ejaculates millions of sperm, full of information about how to build a person, almost all capable of fertilizing an egg, and yet, 999,999 sperm out of 1,000,000 fail to do it. A high proportion of most organisms go through an entire life cycle, building and maintaining order, and then leaving no offspring at all. Insofar as Darwinian processes make reproduction central to adaptation, they rely on a wonderfully wasteful process. It’s hard to think of a better way to waste energy than to produce lots of energetically expensive copies of something, and then, to destroy all of them except for the minimum number of copies that you need to do it all over again.

How about heredity? Another amazingly entropy-increasing process! Molecular biologists know that DNA copy fidelity is very high, and therefore, low in information loss, compared to, say, RNA copy fidelity. But think of the costs of this much high fidelity. In every cell, there is a vast and complex apparatus whose sole function is to ensure copy fidelity: it cuts out hairpin turns when the very sticky DNA sticks to itself; it proofreads all the copies, removes mutations, breaks up molecules that could cause mutations, and so on. In Homo sapiens, at least sixteen enzymes—polymerases—have so far been discovered whose sole functions are to repair different kinds of errors that thermodynamic processes produce in DNA sequences. The costs of high-fidelity heredity—both to build the equipment that protects and corrects the DNA sequences, and to operate it—are very great, just what the second law requires.

Now, let’s look at this energy waste on an even larger scale. The evolution of adaptation reflects environmental change over the earth’s history. The vast diversification of flora and fauna is also the result of differences between local environments in different places on earth. From long before continental drifts and long after global warming, environments on the earth change over time and increase entropy as they do so. What is more, once natural selection kicks in, flora and fauna remake their environments in ways that further accelerate entropy-increase. When nature started selecting molecules for stability and replicability, it began producing arms races more wasteful than anything the Americans and Soviets could ever have dreamt up. From that time on, there has been a succession of move and countermove in adaptational space made at every level of organization. It has happened within and between every descending lineage of molecules, genes, cells, and organisms. Each line of descent has forever searched for ways to exploit its collaborators and its competitors’ adaptations. All that jockeying for position is wasted when one organism, family, lineage, species is made extinct by the actions of another organism, family, lineage, and so on. This is what Darwin called the struggle for survival.

Add sexual reproduction to the struggle for survival, and it’s impossible to avoid the conclusion that Darwinian selection must be nature’s favorite way of obeying the second law. Natural selection invests energy in the cumulative packaging of coadapted traits in individual organisms just in order to break them apart in meiosis—the division of the sex cells. Then, it extinguishes the traits and their carriers in a persistent spiral of creative destruction. Think about this for a moment: 99 percent of the species that have been extant on this planet are now extinct. That is a lot of order relentlessly turned into entropy!

It’s well-known that every major change and many minor ones in the environment condemn hitherto fit creatures to death, and their lineages to extinction. As environments change, yesterday’s adaptation becomes tomorrow’s maladaptation. In fact, it looks like three different cataclysmic events have repeatedly killed off most of the life forms on earth. The dinosaur extinction 65 million years ago, owing to an asteroid collision on the Yucatan peninsula, is well-established. There are no dinosaur bones in any of the younger layers of stone around the world, but there is a layer of iridium—an element found in much higher concentrations in asteroids than on earth—spread evenly around vast parts of the earth centered on the Yucatan in the layers of rock 65 million years old. In that layer, iridium is 1,000 times more concentrated than elsewhere above or below it in the earth’s crust. At a stroke, or at least, over only a few years, all the vast numbers of dinosaur species, which had been around adapting to their environment beautifully for 225 million years, just disappeared. That’s what made the world safe for the little shrew-like mammals from which we are descended. The fossil record reveals a bigger extinction event 500 million years ago on earth and an even more massive extinction after that, 225 million years ago: the Permian-Triassic extinction in which 3/4ths of all ocean-living genera and almost 100 percent of ocean-dwelling species, along with 75 percent of land species, became extinct. This is order-destroying waste on a world-historical scale.

Long before all this, it was the build up of oxygen in the oceans and the atmosphere that killed off almost everything on earth! Build up of oxygen? How could oxygen be a poison? Remember, yesterday’s adaptation can be today’s maladaptation. Life started in the oceans with anaerobic bacteria—ones that don’t need oxygen. In fact, they produce oxygen as a waste product, the way we produce carbon dioxide. Just as the plants clean up our mess by converting carbon dioxide into oxygen and water, the environment cleaned up all that oxygen pollution by molecular action, binding the oxygen to iron and other metals. At some point, the amount of oxygen waste produced by the anaerobic creatures exceeded the absorption capacity of the environment. As a result, they all began to be poisoned by increasing levels of oxygen around them. 2.4 billion years ago, these bacteria were all almost completely wiped out, making enough space for the aerobic bacteria, the ones that live on oxygen and produce carbon dioxide as waste. We evolved from these bacteria.

Can any other process produce entropy as fast and on such a scale as natural selection? Just try to think of a better way of wasting energy than this: build a lot of complicated devices out of simpler things, and then, destroy all of them except the few you need to build more such devices. Leaving everything alone, not building anything, won’t increase entropy nearly as fast. Building very stable things, such as diamond crystals, will really slow it down, but building adaptations will use up prodigious amounts of energy. Adaptations are complicated devices: they don’t fall apart spontaneously; they repair themselves when they break down. They persistently get more complicated, and so, use even more energy to build and maintain themselves. Any long-term increase in the number of adapted devices without increased energy consumption would make a mockery of the second law. If such devices are ever to appear, besides being rare, they had better not persist and multiply, unless by doing so, they inevitably generate more energy wastage than there would have been without them. This is the very process Darwin envisioned: in Tennyson’s words, “nature red in tooth and claw.”

Kelvin had the wrong end of the stick when he argued that there has not been enough time for natural selection to produce the adaptation, the order, we see on earth. What really needs to be explained is the fact that adaptation here on earth produces so much disorder. The second law that does exactly this, by allowing adaptations, but only on the condition that their appearance increases entropy. Any process competing with natural selection as the source of adaptations has to produce adaptations from non-adaptations, and every one of the adaptations it produces will have to be rare, expensive, and wasteful, We’ll see that this requirement—that building and maintaining orderliness always has to cost more than it saves—rules out all of natural selection’s competitors as the source of adaptation, diversity, and complexity.

Could there be a process that produces adaptations that is less wasteful than the particular way in which Darwinian natural selection unrolled on earth? Probably. How wasteful any process producing adaptation can be depends on the starting materials, and on how much variation emerges in the adaptations built from them. But every one of these possible processes has to rely on dumb luck to produce the first sliver of an adaptation. In that respect, they would still just be more instances of the general process Darwin discovered—blind variation and environmental filtration. A process that explained every later adaptation, by more dumb-luck shuffling and filtering of the earlier adaptations, would still be Darwinian natural selection. It will be Darwinian natural selection even if the process was so quick and so efficient as to suggest that the deck was stacked. So long as the deck wasn’t stacked to produce some prearranged outcome, it’s still just blind variation and environmental filtration. Any deck stacking—a process of adaptational evolution that started with some unexplained adaptation already in the cards—is ruled out by physics.

The Second Law, Natural Selection, and Theism

The attentive reader will have noticed that at the beginning of the last section, I said only three things were required to show that, given the laws of physics, the only way adaptations could have emerged is by natural selection. The first two are the requirement that the first and all subsequent adaptations be random and rare events, the second that the process by which adaptations persist and improve be energetically expensive. But the third requirement we need to show that natural selection is the only game in town for building adaptations is a rather deeper and less widely noticed feature of the second law.

No matter what brings it about, the process of adaptation is different from the more basic physical and chemical processes in nature. They are all “time symmetrical.” Adaptation is not. But the only way a time-asymmetrical process can happen is by harnessing the second law.

A time-symmetrical process is one that is physically reversible. One example of such time-reversible process is well-known: any set of ricochets on a billiard table can be reproduced in exactly the opposite order. Here are some more examples: hydrogen and oxygen can combine to produce water, but water can also release hydrogen and oxygen. Even the spreading circular waves made when a drop of liquid falls into a pool can be reversed to move inward and expel the drop upward from the surface. No matter in what order the basic chemical and physical processes go, they can go the reverse order too.

The second law creates all asymmetrical processes and gives them their direction in time. Now, the evolution of adaptations is a thoroughly asymmetrical process. Take a time-lapse film of a standard natural selection experiment. Grow bacteria in a petri dish. Drop some antibiotic material into the dish. Watch the bacterial slime shrink until a certain point, when it starts growing again as the antibiotic resistant strains of the bacteria are selected for.

Now, try reversing the time-lapse video of the process of bacterial selection for resistance. What you will see just can’t happen. You will watch the population of the most resistant strain diminish during the time the antibiotic is present. After a certain point, you will see the spread of the bacteria that can’t resist the antibiotic, until the drops of the antibiotic leave the petri dish altogether. But that sequence is impossible. It’s the evolution of maladaptation, the emergence, survival, and spread of the less fit.

There is only one physical process available to drive asymmetrical adaptational evolution. That is the entropy increase required by the second law of thermodynamics. Therefore, the second law must be the driving force in adaptational evolution. Every process of adaptational evolution—whether it’s the one Darwin discovered, or any other—has to be based on second law entropy increase.

The physical facts—the starting conditions at the Big Bang, plus the laws of physics—fix all the other facts, including the chemical and biological ones. All the laws of physics except the second law work backwards and forwards. So, every one-way process in the universe must be driven by the second law. That includes the expansion of the universe, the build up of the chemical elements, the agglomeration of matter into dust motes, the appearance of stars, galaxies, solar systems, planets, and all other one-way processes. And that will eventually include, on one or more of these planets, the emergence of things with even the slightest, merest sliver of an adaptation. We can put it even more simply. In a universe of things moving around and interacting on paths that could go in either direction, the only way any one-way patterns can emerge is by chance, here and there, when conditions just happen to be uneven enough to give the patterns a start.

These rare one-way patterns will eventually just peter out into nothing. Trust the second law. Consider the one-way process that built our solar system and maintains it. It may last for several billion years. But eventually, the nice pattern will be destroyed by asteroids, or comets, or the explosion of the Sun, or the merging of the Milky Way with other galaxies, whichever comes first. That’s entropy increase in action on a cosmic scale. On the local scale, entropy increase will occasionally and randomly result in adaptational evolution. And that is the only way adaptations can emerge in a universe where all the facts are fixed by the physical facts.

Because entropy increase is a one-way street, the second law is also going to prevent any adaptation-building process from retracing its steps and finding a better way to skin the cat. Once a local adaption appears, it can’t be taken apart and put together in different, more efficient, less entropy-increasing ways. The only way to do that is to start all over independently. Adaptation-building has to produce local equilibria in stability and replication that get locked in, built into the woodwork, like a knot, and have to be worked around in the creation of new adaptations. Natural selection is famous for producing such examples of inferior design, Rube-Goldberg complexity, and traits that could only have been “grandfathered in.” It’s not just the oft-cited example of the blind spot in the mammalian eye resulting from the optic nerve’s coming right through the retina. An even more obvious case is the crossing of the trachea and the digestive system at the pharynx. Convenient, only if you like choking to death. We think of the giraffe’s neck as an adaptation par excellence. But the nerve that travels from its brain to its larynx has to go all the way down the neck, under the aorta and back up—a twenty-foot detour it doesn’t need to take.

Any adaptation-creating process has to produce sub-optimal alternatives all the time. It has to do this not just to ensure entropy increase, but also to honor the one-way direction the second law insists on. Perhaps our most powerful adaptation is the fact that our brains are very large. They have enabled us to get to the top of the carnivorous food chain everywhere on earth. But this is only the result of a piece of atrocious design. The mammalian birth canal is too narrow to allow large-brained babies to pass through. This bit of locked-in bad design meant that the emergence of human intelligence had to await random changes that made the skull small and flexible at birth, and thus, delayed brain growth till after birth. This is where the large fontanel—the space separating the three parts of the infant’s skull—comes in. Now, the kid has room to get through the birth canal and has a skull that will immediately afterward expand to allow a big brain to grow inside it. But brain growth after birth introduced another huge problem: the longest period of childhood dependence of any species on the planet. All this maladaptation, corner-cutting, jury-rigging, is required by the second law of any process that produces adaptations.

What does all this come to? The only way a recipe for building adaptations can get along with the second law is by employing it. The only recipe that can do that is the process that Darwin discovered: dumb luck variation, one-way filtering, and a very expensive copying machine.

The process of natural selection is a matter of probabilities in the same way the spread of gas molecules in the up-stroke of a bicycle pump is a matter of probabilities, and for the same reason: the process of entropy increase is probable, not certain. This will be no surprise, given the role of the second law in generating evolutionary asymmetries. But even before this role became clear, the nature of evolutionary probabilities was well understood. The theory tells us that fitness differences between organisms will probably lead to differences in their reproductive successes. The Darwinian mechanism cannot guarantee the reproductive success of the fitter of two organisms, lineages, or populations. Therefore, it doesn’t guarantee the evolution of adaptation. If it did claim that fitness differences guaranteed reproductive success, Darwin’s theory would be false. Biologists are well aware of several quite rare circumstances under which, just through bad luck, the fittest among competing creatures do not survive at all, let alone have more offspring than the less fit ones. Most of these circumstances fall under the category of “drift.”

It is the role of low probability events in evolution that makes it incompatible with Christian theism.

One feature is arguably indispensable to the theism of all the Abrahamic religions—Judaism, Christianity, and Islam. They all say, “God created man in his own image. In God’s image he created him; male and female he created them” (Genesis 1:27). One thing they agree on is we exist and have our features, owing to God’s design, based on an idea—an image, a design—of what he wanted. How literally to take this claim remains a matter of debate among the various sects, creeds, churches, divinity schools, madrassas, and biblical exegetes of the Abrahamic religions. What theism can’t give up is the credo that he created us sapient creatures, and that he did so intentionally. Theism cannot accept the notion that we are a side effect, a by-product, an accident or coincidence, an unintended outcome.

Now, the standard reconciliation of Darwinism and theism goes like this. Everything else extant in zoology and botany might well be such a by-product, a side effect, and an unintended, but foreseen outcome of God’s handiwork in creating us. Elephants and e. coli are by-products since they were produced, along with us, by natural selection, and natural selection was the instrument, the means, the technique God used to produce his one intended outcome—us. A more ecumenical view might have it that all God’s creatures, great and small, are equally intended outcomes, along with us, and all were produced by the device of blind variation and natural selection, just as the omniscient lord knew they and we would be so produced.

The trouble with this reconciliation is that it does not take Darwin’s theory seriously. It just pretends to do so. Darwin’s theory tells us that we, and every other creature that roams the earth or the deep, are, at best, improbable outcomes of natural selection. It is, therefore, a highly unreliable means of making us or anything that looks like us. Indeed, it is an unreliable way of making intelligent life of any sort, or for that matter, making life of any sort. Any omnipotent deity or even a very powerful agency who decided to employ natural selection to produce us would be disappointed many more than 99 times out of 100 attempts.^[1] An omnipotent agency that employed natural selection to make Homo sapiens would certainly not be an omniscient one. He would have had to be rather dim. Even weak-minded creatures such as us know that blind variation and environmental filtration is not a reliable way of making anything that you have a very definite design for already.

Theism, or at any rate, a theism that makes the Abrahamic religions sensible is just not compatible with Darwin’s theory. You can’t believe both. But if you are a true believer in theism, you can come close to reconciling them, perhaps close enough to deceive yourself. The trick is to believe that God is powerful enough, smart enough, and present at enough places and times to guide the process of evolution so that it looks like random variation and natural selection to anyone not in on the secret. Suppose God employed a method to obtain us—Homo sapiens—that was so complex, so subtle, and so difficult to uncover, that the closest we could come to figuring it out was Darwin’s theory of blind variation and environmental filtration. You might even go further and suppose that the closest any finite intelligence could come to uncovering God’s method was Darwin’s theory. You might even suppose that God designed our cognitive capacities so that the only scientific theory of evolution we could ever conceive of would be Darwin’s. It would be, as Kant supposed of Newton’s theory, the only game in town, not because it is true, but because its falsity is unthinkable by us once we see the evidence for it.

These suppositions would readily explain to a theist why so many smart scientists believe that Darwin’s theory is true. For that matter, it would also explain why the smartest among them are atheists as well, since they recognize that theism and Darwinism are incompatible. But this stratagem won’t reconcile theism and Darwinism.

In fact, one clear implication of this stratagem is that Darwinism is false. Darwinian theory tells us that evolutionary outcomes are the result of probabilities—chance reproduction, chance variation, chance environmental events, operating on low population numbers. But the theist thinks that the process God chose was fiercely complicated and absolutely certain to have produced us. As such, the theist is committed to Darwin’s theory being wrong, false, mistaken, erroneous, incorrect. Darwinian theory is false, but just looks correct to smart people trying to figure out what caused adaptation. Saying it’s false is no way to reconcile the truth of a theory with theism. Of course, the ways of God may be so far beyond our ken, that the smartest and most knowledgeable of us mistakenly think adaptation is the result of natural selection. But that wouldn’t reconcile theism and Darwinism. At most, it shows how to be a theist and to treat Darwin’s theory as a useful fiction, one that gets the mechanism of evolution completely wrong.

William Lane Craig’s Response

It’s a challenge to know how to respond to Alex Rosenberg’s views. In the case of Professor Maudlin’s argument, there were enough relevant points to comment on that one could ignore the red herrings and still have plenty to talk about. But in Dr. Rosenberg’s case, the entire argument is off-topic. Had the Greer-Heard Forum this year been on the topic of God and evolutionary biology or the existence of God in light of Darwinian evolution, then Dr. Rosenberg’s arguments would make a relevant contribution to the discussion. But that’s not the topic. The topic is the existence of God in light of contemporary cosmology, and Dr. Rosenberg’s arguments have nothing to say about the subject of cosmology.

So, do I just sit down now, or do I follow the hounds in pursuit of the red herring? Since our quarry is nowhere in sight, I think that I may as well join the hounds for a short distance.

First, it’s evident that Rosenberg fails in his first stated objective, to prove that physics is sufficient for all appearances of design in nature. For he assumes that all appearances of design in nature are biological in nature. But that is to ignore the fine-tuning of the universe, which sets the table for biological evolution. In particular, his appeal to the second law of thermodynamics takes no cognizance of one of the most extraordinary instances of fine-tuning, namely, the initial low entropy condition of the universe.

Second, he fails to establish his second stated objective, that natural selection is the only way consistent with physics from which the appearances of design can emerge. For he never proves that natural selection is sufficient for the appearances of biological design. Rosenberg’s entire paper reads like a Darwinian, or, better, naturalistic, confession of faith. Especially breathtaking is the following transition:

Eventually there will start to be molecules that enhance each other’s stability and/or replication through their chemical relations to each other. There are no limits to the repetition of this process, making bigger and more complicated and more diverse molecules. If conditions are favorable, the result will be really big assemblies of stable and/or replicating molecules, for instance, RNA and eventually DNA sequences and strings of amino acids—i. e., genes and proteins.

The rest is history, that is, natural history.

I don’t think we would be uncharitable to ask for a little more detail here. I’ve read that the odds of getting even a single functioning protein molecule by chance is approx. 1:10¹⁶⁴. That’s a trillion, trillion, trillion, trillion, trillion, trillion, trillion times smaller than the odds of finding a single specified particle among all the possible particles in the universe.^[2] And that says nothing of the odds of evolving a single living cell, much less multicellular organisms.

So, what is the evidence that natural selection operating on random variations accounts for observed biological complexity? I think it’s worth emphasizing just how extraordinary an extrapolation Darwinism involves. Many of us probably think that if random mutation and natural selection can explain, for example, the evolution of the horse, then that surely shows the power of the Darwinian mechanisms. In fact, evolution within a single kind such as this is trivial, compared to the vast range of life. You might think, well, if we could show that random mutation and natural selection could explain, say, how a bat and a whale evolved from a common ancestor, that would really show the power of these mechanisms! Sorry! Think again.

Consider the following figure from Sean Carroll (a different Sean Carroll!).

Figure 1: The various phyla of multicellular animals.

The figure shows the major groups or phyla of the animal kingdom. Notice that a bat and a whale are both mammals, which is just one of the subcategories under the chordates. Even the evolution of a bat and a whale from a common ancestor is an utter triviality, compared to the vast range of the animal kingdom.

It would do nothing to explain, for example, how a bat and a sea urchin evolved from a common ancestor, not to speak of a bat and a sponge. Darwinism represents an extrapolation of gargantuan proportions. Indeed, it represents an enormous leap of faith in the efficacy of the Darwinian mechanisms.

Now, if that extrapolation takes your breath away, take a look at this!

Figure 2: The three basic types of organisms branching off from the last universal common ancestor.

The whole previous figure you just saw is contained on the little twig of the right-hand branch labeled “animals.” I love the modesty of that label, “animals.” The whole of the animal kingdom is contained in that little twig. And notice the twig to the right of it: “plants.” “Plants!” The whole of the plant kingdom is contained in that twig. And these are just two twigs on the branch of Eucaryotes. There are still the two other domains of the Bacteria and Archaea to be accounted for. The extrapolation of the Darwinian mechanisms from peppered moths and fruit flies and finch beaks to the production and evolution of every living thing is a breathtaking extrapolation of gargantuan, brobdingnagian proportions. We know that, in science, such extrapolations often fail. So, again, I think we are compelled to ask, what is the evidence for this extraordinary extrapolation?

Unfortunately, Dr. Rosenberg does not answer that question. What he argues is that if naturalism is true, then Darwinism is the only game in town. I think a good many evolutionary biologists would be skeptical of Rosenberg’s claim. But if he’s right, then that, at best, only gives us good reason to think that naturalism is not true.

Finally, Dr. Rosenberg argues that theism and Darwinism are incompatible. He argues,

Darwin’s theory tells us that we, and every other creature that roams the earth or the deep, are at best improbable outcomes of natural selection. It is, therefore, a highly unreliable means of making us or anything that looks like us. . . . Any omnipotent deity or even a very powerful agency who decided to employ natural selection to produce us would be disappointed many more than 99 times out of 100 attempts. An omnipotent agency that employed natural selection to make Homo sapiens would certainly not be an omniscient one. He would have had to be rather dim.

This argument is, I think, confused. Far from being the problem, God’s omniscience may be the best hope for saving Darwinism. For an omniscient God would know the true counterfactuals concerning evolutionary development such as, “If the laws of nature were applied to a certain set of initial conditions, then biologically complex life forms of such and such a kind would evolve.” God need not even intervene miraculously in order for the evolutionary process to have an intended goal.

Of course, if God did intervene miraculously at certain critical junctures in the evolutionary process to bring about certain mutations, biologists would never know it. How do we know that God did not cause the critical mutations that facilitated the reptile-to-bird transition?

Dr. Rosenberg replies that in such a case, Darwinism is just false. He says,

Darwinian theory tells us that evolutionary outcomes are the result of probabilities—chance reproduction, chance variation, chance environmental events, operating on low population numbers. But the theist thinks that the process God chose was fiercely complicated and absolutely certain to have produced us.

I think this claim is based on a common misunderstanding of evolutionary theory. When biologists say that the mutations responsible for evolutionary change occur randomly, they do not mean “by chance” or “purposelessly.” Rather, as Francisco Ayala explains, they mean “irrespective of their usefulness to the organism.”

Now, this is hugely significant! The biologist is not making the presumptuous philosophical claim that biological mutations occur by chance, and hence, that the evolutionary process is undirected or purposeless. Rather, he means that mutations do not occur for the benefit of the host organism. If we take “random” to mean “irrespective of usefulness to the organism,” then randomness is not incompatible with divine direction or purpose.

For example, suppose that God in His providence causes a mutation to occur in an organism, not for the benefit of the organism, but for some other reason (say, because it will produce easy prey for other organisms that He wants to flourish). In such a case, the mutation is both purposeful and random.

And as for probabilities, Dr. Rosenberg seems to have forgotten that these are relative to background information. It is perfectly consistent to say that some mutation is improbable, relative to certain sets of physical data, but not improbable, when God’s action is taken into account. This is the case even on the purely physical level: when we say that some event is highly improbable, we do not mean that it “just happens” without causes relative to which its occurrence is highly probable.

So, I’d say that the theist can be a Darwinist if he wants to. But I have yet to be convinced that the neo-Darwinian account of biological evolution really gives us the full story.

Let the hounds run on! We’ve followed them long enough.

Sean Carroll’s Response

I think that Alex’s views provide a great opportunity for us to step back a little bit and draw some of the connections between cosmology, entropy, Darwinism, and the origin of life. It’s often said, and again not in this particular forum, but there is an argument against natural selection and Darwinian evolution that creationists sometimes bring up, which is clearly false, but it’s false in an interesting way. It goes, the second law of thermodynamics says that disorder increases in the universe. Things run down, and eventually, you will reach a heat death of the universe. So the argument says, “How in the world can a universe governed by the second law develop such wonderful, organized, complex structures like we see here in the biosphere on earth?”

There is a simple, one-line refutation of that argument, which is that the second law says that entropy increases in closed systems, and the biosphere is not a closed system; therefore, the argument simply does not apply. There is no law of physics that says the entropy of our biosphere has to go up. And that’s true, but it still leaves unanswered the question of, “Okay, but why did all of this complexity, and order, and interestingness arise? Even if it’s compatible with the second law, the second law doesn’t say that it has to become more complex over time.” That’s a very interesting process that’s not been completely understood. It’s very understandable—I think that it’s an open area of science that we’re having a lot of fun with right now, so it’s fun to think about exactly what’s going on.

Our earth is very far from thermal equilibrium. Thermal equilibrium is the state you would be in when you max out your entropy—everything is as dull as possible. The air in this room is pretty close to thermal equilibrium; it’s all uniform, and nothing is happening. The earth is very far from that, and the reason why it is so is because the Sun is a hot spot in a cold sky. If you ask people sometimes why the Sun is important for biology, for life on earth, they might tell you, “Well, the radiation—the rays from the Sun—gives us energy.” That’s actually not the true reason.

If the whole sky were the temperature of the Sun, we would be getting a lot more energy than we do now; however, what would happen is that the earth would attain the temperature of the Sun and life would cease. If the whole sky were dark, we would get no energy, the whole earth would attain the temperature of the night sky, and life would cease. The only reason we have life is because we get some photons from the Sun, but particularly, in a low-entropy form. We then radiate out that energy right back into the universe. Overall, the earth does not gain energy from the universe. Recently, it has been gaining energy because of global worming, but that’s minor, a little effect, and you’ll be debating that next year. Usually, the amount of the energy we give back to the universe is the same amount we get from the Sun, but for every one photon—for every one particle of light we get from the Sun—we radiate twenty photons back into the universe, so we have increased the entropy of that sunlight by a factor of twenty.

The earth is not only compatible with the second law of thermodynamics, it is an engine for increasing entropy, and that’s related to the origin of complexity. Complexity and entropy are very different things. If you think of stirring milk into a cup of coffee, before you stir the milk, the entropy is low. You have milk and you have coffee. After you stir it, the entropy is high. They’re all mixed up together. But before you stir the milk, it’s a simple system: milk, coffee. After you stir the milk, it’s also a simple system. Everything is mixed up together. It’s only in between, when you start mixing the milk into the coffee and the tendrils of milk begin to swirl in complicated ways, that you have a high degree of complexity.

The universe is the same way. The early universe had very low entropy, and the late universe will be very high entropy. But the early universe was a very simple place, and the late universe will also be a very simple place. It’s the middle phase of the history of the universe that is complex. The point is not that complexity requires low entropy. Complexity requires entropy to be increasing. The maintenance of complexity, as in a living being, requires that you are very far from equilibrium, living in an environment where entropy is going up to beat the band.

So, not only is the existence and evolution of life on earth compatible with the second law of thermodynamics, but in some very real sense, it is a spin-off; it is a consequence of the second law of thermodynamics. And the second law, in turn, happens because the early universe has a low entropy, which is a fundamentally cosmological fact. In that sense, life on earth is able to exist because of the cosmological features of our universe 14 billion years ago.

Alex Rosenberg’s Response

I’m not exactly sure why, in a colloquium among philosophers and physicists, some subjects should be ruled out just because they weren’t labeled in the title that brought us together. I was invited to offer my perspective on an issue of importance between science and theism, and I have done so. I hope it was of interest, and I hope perhaps it was edifying. I am not going to be too worried whether or not it was on the topic that Bill was prepared to argue here.

I want to make only two relatively brief points. Of course, Darwin used the word random, and by that he meant, “blind.” That is to say that variations are random in the sense that they are not correlated with, they are not brought about or caused by, the needs that they meet, the benefits they may confer, or the advantages that they may provide the organisms that they find themselves in. And that’s a reason why it’s far better to substitute the word “blind variation” for “random variation” in understanding the process that Darwin identified.

Similarly, Darwin used the word natural selection, and that’s an unfortunate term, too, because it keeps us thinking about an intentional process, an active process, that nature uses in order to cull among variants, those who are worse and dispose of them, allowing the better ones to proceed. Far better to think about that part of the process as environmental filtration, a passive process, a process that acts like a sieve, preventing some things from passing through, allowing everything that is below a minimum size, to pursue this metaphor a little further, to pass through. And the essence of Darwin’s discovery is that adaptations, things that look like they are the result of design, in fact, are the result of a purely causal process that involves no design at all, no selection at all, no foresight at all, but are merely the result of blind variations and passive environmental filtration.

If you are going to demand of yourself that you give a causal account of the nature of adaptations that honors the principles that I have identified in physics, you have no alternative but to give Darwin’s account. And, it’s only been one hundred years—one hundred fifty years now—since Darwin’s account began to have wide circulation, and, as I said, biology is much harder than quantum mechanics.

Why is it harder than quantum mechanics? Because, for one thing, it involves the operation of quantum principles, the operation of classical physical principles, the operation of chemical thermodynamic laws, on conjuries of initial conditions so complicated, so fine, so different across various environments that we now know relatively less about life on our planet than we know about the cosmology of the universe.

The standard model enables us to make predictions to twelve decimal places. Richard Feynman once said that it’s like predicting the distance from Los Angeles to New York to within one hair’s breadth. We’re not in that position in biology. We’re lucky that we have that wonderful graph that Bill showed us of the three-fold phylogenetic kingdom of the archaea, the eukaryotes, and the prokaryotes. But if we are going to be serious about applying what we know from physics and chemistry to explain the nature of living processes, we are not going to have any alternative, but to seek instances of this mechanism of blind variation and natural selection.

Now, let me get on to one other point. Bill began by saying that the one obvious piece of the story of how adaptations emerged that I neglected was the fine-tuning of this universe. I am in agreement with Sean and with other physicists that fine-tuning is immaterial to our characteristics. Bill introduced the concept of the red herring and I’ll use it here, too. But even if the universe were fine-tuned for us, think about the additional interventions that God would have had to have introduced in order to get from the fine-tuning to us. The fine-tuning of the universe, the particular combination of the constants that are so identified, according to the best of our science, need not have led to us. It most probably wouldn’t have led to us.

↑ Back to top