Darwinism needs mutations to build. The measured ledger of mutation runs the other way.
Every version of the Darwinian story depends on one raw material. Not selection — selection is a filter, and a filter cannot invent anything. The raw material is mutation. Copying errors in DNA, arriving at random, occasionally producing something useful, accumulating over deep time into eyes and wings and brains.
Everything rests there. So it is worth asking a straightforward accounting question, and it is a question geneticists have been asking out loud for sixty years: over generations, does the mutational ledger of a population run up or down?
The Numbers Nobody Disputes
Start with what is not controversial. Each human being inherits somewhere in the range of a hundred new mutations that neither parent had — the figure varies by study and by paternal age, but the order of magnitude is agreed on across the field.
Also not controversial: the overwhelming majority of mutations that have any effect at all have a bad one. This is not a creationist talking point. It is the reason radiation is dangerous, the reason paternal age correlates with genetic disease, and the assumption behind every genetic-counseling appointment in the country.
Beneficial mutations exist. Nobody serious denies it. Lactase persistence is real. Sickle-cell heterozygosity confers malaria resistance — at a brutal price. Bacterial antibiotic resistance happens in a petri dish in an afternoon. The question was never whether the beneficial column has entries in it. The question is the ratio, and the ratio is not close.
Haldane’s Bill
In 1957 J. B. S. Haldane — a founder of the modern evolutionary synthesis, a Marxist, an atheist, and no friend of anything resembling this argument — worked out the cost of substituting one beneficial gene for another in a population.
Selection is not free. For a new beneficial trait to become universal, the individuals without it have to fail to reproduce. In a species with a long generation time and few offspring, that reproductive cost is severe, and it limits how fast fixed changes can accumulate. Haldane’s own estimate landed at roughly one substitution per three hundred generations in a mammal population.
Run that against the timeline usually given for the split between humans and chimpanzees and the arithmetic gets uncomfortable fast. The number of generations available is not remotely enough to fix the number of coordinated differences observed — not the differences in bone and muscle alone, but in language capacity, cognition, gestation, and development.
This is called Haldane’s Dilemma, and it is a live problem inside the literature, not a rumor from outside it. Various answers have been proposed — soft selection, truncation selection, the argument that most change was neutral drift rather than selection. Those are real proposals and deserve a fair hearing. But notice what the neutral-drift answer concedes: if most of the difference arrived by drift rather than selection, then most of the difference was not built by the mechanism Darwin proposed. That is not a rescue. That is a change of subject.
The Deeper Problem: Selection Cannot See Small Damage
Here is the part that does the real work in the argument, and it comes straight out of population genetics rather than theology.
Natural selection operates on the whole organism. It cannot reach into a genome and remove one bad letter. It can only remove — or fail to remove — an entire individual. So a mutation with a very small negative effect is effectively invisible to selection. It reduces fitness by an amount far smaller than the noise of ordinary life: weather, accident, predation, chance.
Those slightly-deleterious mutations are not filtered out. They are inherited. And they arrive every generation, in the dozens, in every individual.
Meanwhile a rare beneficial mutation, if it is also small in effect, faces the same invisibility problem — but it arrives at a rate orders of magnitude lower. So both columns of the ledger fill with entries too small for selection to sort, and one column is filling vastly faster than the other.
Michael Lynch, a mainstream evolutionary geneticist with no theological axe to grind, has written plainly that the human mutation load is a real and worsening problem, and that the deterioration in fitness per generation is measurable rather than hypothetical. Geneticist John Sanford has argued the stronger case that this decline is terminal for the Darwinian account itself — and his argument is contested, sharply, by researchers who dispute his fitness modeling. That dispute is legitimate and I will not pretend it is settled in my favor.
But note where the disagreement actually lies. It is not over whether the load exists. It is over how fast it accumulates and whether compensating mechanisms can absorb it. The direction of the arrow is not in dispute.
The Photocopier
Take a page of text and photocopy it. Then copy the copy. Then copy that. Do it ten thousand times.
Somewhere in that stack a stray toner speck will land in exactly the right spot to turn a comma into a period, and the sentence will read slightly better. That will happen. It is not impossible; over ten thousand generations of copies it is nearly certain.
And it will be swamped. By generation ten thousand you have a gray page. The occasional lucky improvement is real and completely irrelevant to the outcome, because the process degrades information far faster than it improves it, and no amount of additional time reverses that. Time is what makes it worse.
Now demand something further: that the photocopier, running long enough, produce a chapter that was not in the original. That is what the Darwinian account requires — not tweaks to existing text, but new text, new function, new coordinated systems, authored by the copying process itself.
The Honest Objections
A fair argument answers the best response, not the worst one, so here are the serious ones.
- Recombination and sexual reproduction shuffle the deck. True, and it genuinely helps — sex allows offspring to inherit fewer bad alleles than the average parent carried. It does not remove the mutations from the population; it changes their distribution.
- Most of the genome does not matter, so most mutations are harmless. This was the standard reply for decades, resting on the “junk DNA” assumption. That assumption has been eroding steadily as functional roles for non-coding regions keep turning up. The more of the genome that turns out to matter, the more of the mutation load counts.
- Synergistic epistasis lets selection remove bad mutations in batches. A real proposal, mathematically elegant, and requiring an enormous and largely unobserved rate of reproductive elimination to work.
- Populations are obviously not collapsing. Correct, and I am not predicting imminent extinction. The claim is narrower and harder to answer: a process running net-negative on a measurable timescale is not the process that wrote the genome in the first place.
What the Ledger Points At
None of this proves Genesis. An argument from mutation load is a negative argument — it shows a proposed mechanism is inadequate to the job assigned to it. That is worth something, and it is not worth everything. Christians who overclaim here do the same thing they accuse the other side of doing.
But it does fit a shape. A creation that was made complete and is now running down is exactly what the Bible describes, and the Bible describes it in language that predates population genetics by three thousand years.
For the creature was made subject to vanity, not willingly, but by reason of him who hath subjected the same in hope, Because the creature itself also shall be delivered from the bondage of corruption into the glorious liberty of the children of God. For we know that the whole creation groaneth and travaileth in pain together until now. (KJV)
Romans 8:20–22
The word the King James translators reached for is corruption — decay, degradation, the running-down of a thing that was once whole. Paul was not writing about nucleotides. He was writing about the direction the whole system moves, and the direction he named is the direction the ledger runs.
The old earth-science phrase for it was the arrow of time. Scripture had a better word for it first, and a better ending: subjected the same in hope. The decay is real and it is not final.
The Question Under the Question
If you are a skeptic reading this, I am not asking you to accept a Bible verse as data. I am asking you to notice that you already believe the thing this argument is about. You believe your genome is degrading — you take folic acid, you avoid radiation, you accept that paternal age matters. The whole medical apparatus of your life assumes it.
The question is only whether a process you already know is running downhill is the same process that ran the whole thing uphill, unaided, from chemistry to consciousness.
Something wrote the text. Copying errors are not a plausible author, and every year the ledger makes that harder rather than easier to argue. In the beginning was the Word — and the genome reads like something that was said, and has been echoing ever since, a little fainter each generation, waiting for the voice that spoke it to speak again.
Teaching the Word. Watching the Times.
— SmithForChrist