Darwin Skepticism Deserves Better Than Handwaving

What if the most important debate about Darwinism is not whether evolution happens, but whether the mechanism we can actually observe is powerful enough to do the work routinely assigned to it?

That distinction matters. A lot.

Because there is a lazy way to defend Darwinism, and it goes something like this: point to antibiotic resistance, point to Galápagos finches, point to dog breeding, say “evolution is a fact,” and then act as though the entire case has been closed. Checkmate, creationists. Science has spoken. Everyone may now go home.

But that is not an argument. It is a fog machine.

No serious critic of Darwinism should deny microevolution. Bacteria develop antibiotic resistance. Finch beaks shift in response to drought. Breeders produce dramatic variation in dogs across generations. These things are real, measurable, repeatable, and scientifically unimpeachable. They show heritable variation acted on by natural selection. Darwin was right about that.

This is not a small concession. It is a genuine one.

But it is not the end of the conversation. It is where the real conversation begins.

The disputed question is not whether organisms adapt. They do. The disputed question is whether the same mechanism that explains beak shape, fur density, and bacterial resistance can also explain the origin of eyes, limbs, nervous systems, new protein folds, new animal body plans, and the large-scale biological architecture required to build the living world from the bottom up.

That leap—from the observable small to the inferred large—is not a finding. It is an extrapolation. And that extrapolation has problems. Serious ones.

Yale computer scientist David Gelernter, writing in his 2019 Claremont Review of Books essay “Giving Up Darwin,” put the distinction bluntly:

“There’s no reason to doubt that Darwin successfully explained the small adjustments by which an organism adapts to local circumstances: changes to fur density or wing style or beak shape. Yet there are many reasons to doubt whether he can answer the hard questions and explain the big picture—not the fine-tuning of existing species but the emergence of new ones. The origin of species is exactly what Darwin cannot explain.”

Gelernter is not a creationist. He is not a theologian. He is not arguing from Genesis. He is a secular Jewish intellectual who calls Darwin’s theory “brilliant and beautiful,” and who nevertheless finds the extrapolation from microevolution to macroevolution scientifically unconvincing.

That makes him difficult to dismiss in the usual way. Which is probably why the usual way is still attempted.

But dismissal is not a rebuttal. Handwaving is not science. And if Darwinism is as strong as its defenders claim, it should survive contact with its strongest objections.

So let us be very clear about what is being argued here.

This article does not deny microevolution. It does not deny common ancestry as a broad proposal. It does not prove Intelligent Design. It does not smuggle Christianity into a lab coat and call it biology.

The claim is narrower and more defensible: the evidence for small-scale evolutionary change does not automatically establish the sufficiency of Darwinian mechanisms to produce large-scale biological innovation. That sufficiency must be demonstrated, not assumed.

And on three major fronts—protein rarity, the Cambrian explosion, and gene regulatory networks—the assumption is under enormous pressure.

Microevolution Is Real. The Question Is Whether It Can Be Infinitely Scaled.

The first move in any honest discussion of evolution is to separate what we observe from what we infer.

Microevolution is directly observed. It refers to small-scale changes within populations: shifts in allele frequencies, adaptive variation, changes in traits already present in some form. Antibiotic resistance is microevolution. Finch beak variation is microevolution. Dog breeding is microevolution.

No problem so far.

Macroevolution, however, is a much larger claim. It refers to the origin of fundamentally new biological structures and body plans: new protein folds, new cell types, new developmental architectures, new phyla-level forms of life. It is not merely a Chihuahua becoming a Great Dane. It is the emergence of the biological machinery needed to build eyes, wings, hearts, nervous systems, and vertebrate body plans.

The standard Darwinian answer is simple: macroevolution is just microevolution plus time. Accumulate enough small changes over enough generations, and eventually the small becomes large. Given enough mutation, selection, and geological patience, bacteria-to-biologist evolution becomes plausible.

It is tidy. It is intuitive. It is also doing a lot of work it has not earned.

Because the crucial question is not whether small changes occur. The question is whether the known mechanism can generate the kind of novelty required for large-scale innovation. Can random mutation and natural selection produce new functional proteins? Can they build new body plans? Can they rewrite developmental operating systems without killing the organism in the process?

Those are not rhetorical flourishes. They are empirical questions.

Richard Lenski’s long-term evolution experiment at Michigan State University is one of the most important experimental projects in evolutionary biology. Since 1988, Lenski and his team have observed twelve populations of E. coli across more than 70,000 generations. The bacteria have adapted. They have changed. They have demonstrated evolution under controlled conditions.

And after decades of experimental evolution, they are still E. coli.

That does not make the experiment a failure. Quite the opposite. It is a triumph of experimental rigor. But what it demonstrates is refined microevolutionary adaptation, not the origin of new animal body plans, not the emergence of novel protein folds, and not the transformation of one fundamental biological architecture into another.

This is the epistemological chasm at the center of the debate: microevolution is observed science; macroevolution by the same mechanism is an inference. The inference may be true. But it does not inherit the evidentiary status of the observation simply because the same word—evolution—is used for both.

That is the bait-and-switch.

Pointing to antibiotic resistance to prove the origin of animal body plans is like pointing to a child’s Lego tower and declaring the Hoover Dam explained. Yes, both involve building. No, the first does not automatically account for the second.

New Proteins Are Not Easy to Find in the Dark.

The first serious challenge comes from molecular biology.

Proteins are the workhorses of life. They catalyze chemical reactions, form cellular structures, transmit signals, regulate processes, and perform the countless tasks that make living systems possible. If macroevolution is going to build new biological structures, it needs access to new functional proteins—or at least new protein folds and domains capable of doing new work.

So the question becomes unavoidable: how hard is it to generate a new functional protein by random mutation?

Douglas Axe, a molecular biologist with a PhD from Caltech and postdoctoral work at Cambridge University’s Center for Protein Engineering, attempted to measure this. His 2004 paper in the Journal of Molecular Biology, “Estimating the Prevalence of Protein Sequences Adopting Functional Enzyme Folds,” examined how rare functional protein folds are within the vast space of possible amino acid sequences.

His estimate was astonishing: among the possible 150-amino-acid sequences, roughly 1 in 10^77 would fold into a stable, functional protein domain.

To put that in perspective, there are approximately 10^80 atoms in the observable universe. A probability of 1 in 10^77 is not merely “unlikely” in the ordinary sense. It is not like finding a needle in a haystack. It is like finding a specific atom in a cosmic mountain range of haystacks while blindfolded, underwater, during an earthquake.

Critics sometimes dismiss Axe’s work as a “back-of-the-napkin” calculation. That is false. Axe’s estimate came from site-directed mutagenesis experiments on a real enzyme. He physically altered amino acid clusters in a beta-lactamase protein domain and measured which variants retained function. This was bench science, performed in a serious research context and published in a leading peer-reviewed journal.

Now, a fair critic can say that Axe studied one protein fold, not every possible protein fold. That is true. And it matters. We should not overstate the result.

But the problem does not disappear. Because Axe’s work sits within a broader pattern of independent research suggesting that functional protein sequences are extraordinarily rare.

Reidhaar-Olson and Sauer, writing in Proteins: Structure, Function, and Genetics in 1990, studied the lambda repressor protein and found functional sequences as rare as roughly 1 in 10^63 for a 92-amino-acid sequence.

Taylor and colleagues, writing in Proceedings of the National Academy of Sciences in 2001, studied chorismate mutase and found functional sequences at a frequency of about 1 in 10^24 for a 93-amino-acid enzyme. That is far more generous than Axe’s figure, but still staggeringly rare.

Keefe and Szostak, in a 2001 Nature paper, searched more than a trillion random sequences before finding functioning ATP-binding proteins. Again: not a napkin. A laboratory experiment.

Tian and Best, in PLOS Computational Biology in 2017, used a statistical model and reported functional probabilities for ten protein domains ranging from 1 in 10^24 to 1 in 10^126. This study is sometimes cited against Axe, but it does not exactly rescue Darwinian optimism. Even under more favorable modeling assumptions, the rarity problem remains severe.

More recently, a 2025 PLOS ONE paper using an independent mathematical method reportedly estimated the rarity of Axe’s beta-lactamase domain at approximately 1 in 10^72—close to Axe’s 1 in 10^77 figure.

The exact numbers vary. That is expected. Biology is messy. But the direction of the evidence is not hard to see: functional proteins appear to occupy tiny islands in a vast ocean of nonfunctional sequence space.

And here is the Darwinian difficulty: natural selection cannot select a protein that does not yet function.

Selection is powerful once a functional advantage exists. It can preserve, refine, and amplify. It can optimize. But it cannot guide a blind search toward a function before that function appears. Selection does not look into the future and say, “Only seven mutations to go—keep climbing.”

It is not prophetic. It is reactive.

So the hardest step is not the refinement of an already-working protein. The hardest step is getting to minimal function in the first place. That is precisely where the probability problem bites.

Stephen Meyer, in Darwin’s Doubt, calls this the combinatorial search problem. The number of possible amino acid sequences is astronomically large; the functional sequences appear to be exceedingly rare; and undirected mutation must somehow locate these functional islands before selection can do anything useful with them.

If that is the mechanism, then the mechanism has a problem.

The Cambrian Fossil Record Looks Backward for Darwinism.

The protein problem presses Darwinism from below, at the molecular level. The Cambrian explosion is evident in the fossil record.

Darwin knew this problem existed. In On the Origin of Species, he called the sudden appearance of major animal groups in the fossil record a “serious” difficulty for his theory. His hope was that future fossil discoveries would fill in the missing precursors.

They have not.

The Cambrian period, beginning about 541 million years ago, witnessed the geologically sudden appearance of many major animal body plans. Arthropods, chordates, echinoderms, and other phyla-level forms appear in a relatively narrow window of geological time. This is what is commonly called the Cambrian explosion.

Even defenders of evolutionary theory do not deny that something dramatic happened. The debate is over what explains it.

The Darwinian expectation is bottom-up. Small variations accumulate. Populations diverge. Species split. Over vast time, larger taxonomic categories emerge from lower-level variation. Body plans should appear as the end result of a long process of gradual differentiation.

But the Cambrian record has often been described as top-down. Major body plans appear early, suddenly, and with remarkable disparity. Lower-level diversification comes afterward.

Stephen Jay Gould, no friend of Intelligent Design, emphasized this pattern in Wonderful Life. He later observed that the explosion of multicellular life seemed “as abrupt as ever,” and even more so as Precambrian evidence improved. That last point matters. The problem is not merely that we lack fossils. The problem is that better fossil evidence has not produced the expected Darwinian staircase.

James Valentine, Douglas Erwin, and David Jablonski wrote in Paleobiology in 1991 that the Cambrian pattern “creates the impression that metazoan evolution has by and large proceeded from the ‘top down.’” Jeffrey Levinton, writing in Scientific American in 1992, called the Cambrian explosion “evolutionary biology’s deepest paradox.” Malcolm Gordon later suggested that the traditional version of common descent may not apply to kingdoms and may not apply to many, if not all, phyla.

These are not Sunday school pamphlets. These are mainstream evolutionary voices noticing that the record is strange.

The standard reply is preservation. Perhaps the ancestors existed before the Cambrian but were too small or soft-bodied to fossilize. That is possible in principle. But the escape hatch has empirical problems.

First, many Cambrian fossils are themselves soft-bodied. The Burgess Shale and similar deposits beautifully preserve delicate organisms. So the idea that soft-bodied precursors could not have been preserved is too simple. If exceptional preservation captured soft Cambrian animals, why did it not capture the necessary Precambrian ancestors?

Second, Stephen Jay Gould and Simon Conway Morris both noted the force of this problem in different ways: the conditions that could preserve soft-bodied Cambrian organisms should not be waved away as irrelevant, since the missing organisms are precisely the soft-bodied forms Darwinism needs.

Third, research published in PNAS in 2018 found that similar preservation conditions existed in Cambrian and Precambrian sediments. Arthropod traces appear in Cambrian rocks. In Precambrian rocks with comparable preservation potential, those traces are “strikingly absent.” That is a wonderfully understated scientific phrase. Translation: if they were there, we had a decent chance of seeing them. We do not.

Another reply is time. The Cambrian explosion, critics say, took millions of years. It was not an instantaneous event.

Fine. But this does not solve the problem. Even twenty or thirty million years is geologically brief for the origin of numerous major animal body plans. More importantly, time is not a mechanism. It is the arena in which mechanisms operate. Saying “millions of years” does not explain how biological information arose any more than saying “a large warehouse” explains how a dictionary wrote itself.

Time can give a mechanism room to work. It cannot substitute for the mechanism.

That is the real Cambrian question: where did the information required to build these body plans come from?

Meyer’s Darwin’s Doubt is controversial precisely because it presses that question with uncomfortable persistence. He examines the Ediacaran organisms that predate the Cambrian, the molecular clock hypotheses used to push animal divergence earlier, and the informational demands of new body plans. His central claim is not that the Cambrian explosion proves God by itself. It is that known Darwinian mechanisms have not demonstrated the creative capacity required to generate the biological information the Cambrian explosion demands.

That is a more modest claim than many critics pretend. It is also harder to answer.

Body Plans Are Built by Developmental Operating Systems, Not Loose Genes.

The third challenge may be the most technically devastating because it comes from developmental biology, not from religious polemic.

To build an animal body plan, genes do not simply fire in isolation like scattered light bulbs. They operate in coordinated networks. These are called gene regulatory networks, or GRNs. They control embryonic development by turning genes on and off in precise sequences, in precise places, at precise times.

Think of a GRN as part of the operating system of development. Not the paint color. Not the furniture. The operating system.

In 2006, Eric Davidson and Douglas Erwin published a landmark paper in Science titled “Gene Regulatory Networks and the Evolution of Animal Body Plans.” They identified deeply conserved subcircuits within these networks called “kernels.” Their description is worth quoting:

“Kernels are sub-circuits composed of recursively wired regulatory genes that operate during the initial phase of regional pattern formation for a particular body part. If any of the genes in the sub-circuit are prevented from functioning, the body part fails to develop.”

Let that land.

If any of the genes in the sub-circuit are prevented from functioning, the body part fails to develop.

Not: the organism becomes slightly less fit.

Not: natural selection has some rough material to work with.

Fails.

The technical phrase often used here is developmental catastrophe. Change the wrong part of the developmental architecture, and you do not get a hopeful monster. You get no viable organism at all.

This creates an obvious problem for Darwinian macroevolution. If major body plan changes require alterations to the developmental control systems that build the body plan, but those systems are resistant to change because alterations are catastrophic, then the mechanism is blocked at precisely the point where it needs access.

Davidson and Erwin noted that certain kernels appear essentially unchanged across enormous evolutionary distances. The kernel involved in heart specification is conserved in animals as different as flies and mice, whose last common ancestor lived hundreds of millions of years ago. The endomesoderm specification network is retained between sea urchins and starfish, lineages separated since the end of the Cambrian.

That is not merely slow evolution. That is a deep constraint.

The Darwinian reply, offered by paleontologist Charles Marshall and others, is that new body plans may not require entirely new genes or kernels. Perhaps they require rewiring: changing how existing genes are deployed, when they are expressed, and where their outputs are used.

But rewiring is not magic. It requires coordinated changes in regulatory DNA. It requires multiple parts of an integrated system to shift in the right way without destroying development. It still requires information. It still faces the problem of getting several necessary changes before selection has a selectable advantage to preserve.

Meyer’s response to Marshall is essentially this: the rewiring objection does not remove the information problem; it relocates it. If new body plans require new regulatory arrangements, where do those arrangements come from? And if partial arrangements are not useful until enough of the system is in place, what exactly is natural selection selecting along the way?

This is where Douglas Erwin’s own words become especially important. Erwin is not an Intelligent Design advocate. He is a mainstream evolutionary biologist. Yet he wrote:

“The crucial difference between the developmental events of the Cambrian and subsequent events is that the former involved the establishment of these developmental patterns, not their modification.”

And elsewhere:

“There is every indication that the range of morphological innovation possible in the early Cambrian is simply not possible today.”

That sentence should stop us.

The range of morphological innovation possible in the early Cambrian is simply not possible today.

If the mechanism available now cannot generate the kind of innovation that apparently occurred then, then we do not have a small explanatory gap. We have a deep mechanistic puzzle.

The Usual Dismissals Answer a Caricature, Not the Argument.

At this point the handwaving usually begins.

Not always. There are serious evolutionary biologists who understand these problems and try to answer them carefully. But popular-level defenses of Darwinism often do something much easier: they pretend the skeptic has denied finch beaks, fossil evidence, or science itself.

That is convenient. It is also false.

There are four common evasions.

First, credentialism. Gelernter is not a biologist. Axe works with the Discovery Institute. Meyer is an Intelligent Design advocate. Therefore, ignore the argument.

This is lazy. Credentials matter insofar as they help us weigh expertise, but they do not replace argument. Axe’s 2004 paper was published in the Journal of Molecular Biology. Davidson and Erwin published in Science. Keefe and Szostak published in Nature. Gould, Levinton, Valentine, Erwin, and Davidson were not writing for Answers in Genesis. If the argument is wrong, show where it is wrong. Do not merely point at the author’s tribe and declare the matter settled.

That may work on Twitter. It should not work in the pursuit of truth.

Second, conflation. Critics answer challenges to macroevolution by citing evidence for microevolution. Antibiotic resistance. Finch beaks. Peppered moths. Lenski’s bacteria.

Again: all granted. Gladly.

But this does not answer the question. Antibiotic resistance does not demonstrate the origin of new protein folds. Finch beak variation does not explain the Cambrian explosion. Dog breeding does not show how gene regulatory network kernels can be rewritten without a developmental catastrophe.

The move is rhetorically effective because it borrows the certainty of observed adaptation and spends it on unobserved innovation. But borrowing is not earning.

Third, time-as-magic. Given enough time, anything can happen.

No, it cannot.

Given enough time, a man can walk from Los Angeles to New York. Given enough time, he cannot flap his arms and fly to Jupiter. Time helps only when the mechanism is capable of the task. If the probabilistic resources are inadequate, if selection cannot act before function exists, if developmental systems collapse under major perturbation, then adding time does not solve the problem. It merely gives the unsolved problem a longer runway.

Bravo. The plane still does not fly.

Fourth, promissory science. We do not have the explanation yet, but science will eventually find one.

Maybe. I am not opposed to future discoveries. I like future discoveries. They are very useful things to have.

But “science may explain it later” is not a rebuttal to an argument now. It is a promissory note. Sometimes, promissory notes are paid. Sometimes they bounce.

Darrel Falk, a geneticist associated with BioLogos and a critic of Meyer’s conclusions, nevertheless conceded in reviewing Darwin’s Doubt that natural selection is not the “driving mechanism” of macroevolutionary change and that the mystery of the Cambrian explosion still awaits a solution.

That is not Intelligent Design triumphalism. It is a sober admission from someone who rejects Meyer’s broader conclusion.

Which is exactly the point: the honest position is not that every biological question has been answered and only religious cranks remain unconvinced. The honest position is that real problems remain.

The Honest Position Is Not Anti-Science. It Is Anti-Pretense.

None of this proves Christianity.

It does not prove Genesis. It does not prove the resurrection. It does not prove Intelligent Design as a full theory. It does not allow Christians to stop thinking and start gloating.

Good. We should not want that anyway.

What it does show is that Darwinian confidence often exceeds Darwinian evidence. The mechanism we observe producing small-scale adaptation has not been demonstrated to produce the large-scale biological innovations routinely attributed to it. The protein data raises a search problem. The Cambrian fossil record raises an informational and historical problem. Gene regulatory networks raise a developmental constraint problem.

Taken together, they do not form a knockdown proof. They form a cumulative case that the extrapolation from microevolution to macroevolution is not the simple, settled, obvious fact it is often claimed to be.

And that matters for Christian apologetics, not because Christianity needs biology to have gaps, but because naturalism often pretends it has none.

There is a kind of intellectual prison that forms when a worldview forbids certain conclusions before the evidence is allowed to speak. If every possible explanation must be material by definition, then materialism can never lose. It can only be “not yet fully explained.” The bars are invisible, which makes the prison feel like freedom.

But truth is not served by pretending open questions are closed. It is not served by calling every skeptic ignorant. It is not served by using microevolution as a shield against every challenge to macroevolutionary creativity.

If Darwinian mechanisms can generate new proteins, new body plans, and new developmental architectures, then show it. Not with slogans. Not with finch beaks. Not with handwaving about time. Show the mechanism doing the work required.

Until then, legitimate Darwin skepticism deserves better than dismissal.

It deserves an answer.

And if the answer is not yet known, then perhaps the most scientific words available are also the most humbling:

We do not know.

For some, that sentence feels like defeat.

For others, it is where the climb toward truth begins.

References

• Axe, D.D. (2004). “Estimating the Prevalence of Protein Sequences Adopting Functional Enzyme Folds.” Journal of Molecular Biology, 341(5), 1295–1315.
• Davidson, E.H. & Erwin, D.H. (2006). “Gene Regulatory Networks and the Evolution of Animal Body Plans.” Science, 311, 796–800.
• Erwin, D.H. & Davidson, E.H. (2009). “The Evolution of Hierarchical Gene Regulatory Networks.” Nature Reviews Genetics, 10, 141–148.
• Gelernter, D. (2019). “Giving Up Darwin.” Claremont Review of Books, Spring 2019.
• Gould, S.J. (1989). Wonderful Life: The Burgess Shale and the Nature of History. W.W. Norton.
• Keefe, A.D. & Szostak, J.W. (2001). “Functional Proteins from a Random-Sequence Library.” Nature, 410, 715–718.
• Levinton, J. (1992). “The Big Bang of Animal Evolution.” Scientific American, 267(5), 84–91.
• Meyer, S.C. (2013). Darwin’s Doubt: The Explosive Origin of Animal Life and the Case for Intelligent Design. HarperOne.
• Reidhaar-Olson, J.F. & Sauer, R.T. (1990). “Functionally Acceptable Substitutions in Two Alpha-Helical Regions of Lambda Repressor.” Proteins: Structure, Function, and Genetics, 7, 306–316.
• Taylor et al. (2001). “Searching Sequence Space for Protein Catalysts.” Proceedings of the National Academy of Sciences.
• Tian, P. & Best, R.B. (2017). “How Many Protein Sequences Fold to a Given Structure?” PLOS Computational Biology.
• Valentine, J.W., Erwin, D.H., & Jablonski, D. (1991). “Morphological Complexity Increase in Metazoans.” Paleobiology, 17(4), 1–12.