Newton’s Mistake Wasn’t the Physics

An introduction to Designed Functional Maturity

Every creationist has made the move at some point. Adam was created as an adult, not as an infant. The wine at Cana was wine without grapes, without a harvest, without a fermentation anyone could have watched. The loaves multiplied without a field. So when someone says the universe looks old, the answer is that of course it does. It was made working. A world made working can contain states that, under ordinary inference, look like the results of a long prior development.

I think that’s correct. I also think that, as usually stated, it’s worthless.

Not wrong. Worthless, which is different and worse. Because as usually stated it explains everything, and a claim that explains everything has stopped doing any work at all.

Watch how it operates. Distant starlight is hard on a young timescale? Created in transit. Radiometric dates come back old? Created with the daughter isotopes already in place. Ice cores show a hundred thousand layers? The initial state was specified that way. Every time, the same move, and the move always works. That’s the tell. Nothing that always works is telling you anything.

I’ve watched critics call this dishonest. It isn’t. It’s something more ordinary and more fixable. There’s no rule saying when you’re allowed to reach for it.

The best illustration of the problem isn’t from a creationist at all.

Newton looked at the accumulating perturbations in the planetary orbits and concluded the solar system wasn’t stable on its own. God must adjust it periodically. That was a serious inference from a serious man, and Laplace dismantled it a century later. Everyone knows this story as the origin of the god-of-the-gaps charge, and it’s usually told as a lesson about religion retreating before advancing science.

Wrong lesson. Here’s what almost nobody mentions: modern dynamical astronomy has found the solar system is chaotic. Run the inner planets forward far enough and Mercury has a real, if small, chance of destabilizing. Newton’s physical intuition was closer to the truth than Laplace’s confidence was.

He was still wrong. He was wrong in a way that had nothing to do with the physics.

He moved from this is hard to reconstruct dynamically to therefore God acted here, and the first fact does not license the second. Being accidentally right about the instability doesn’t rescue the move. If it did, a lucky guess would count as method.

So here’s the rule I want to propose, and it’s aimed at my own side first.

Difficulty in historical reconstruction does not license a creative act. Not once.

✦

Which raises the obvious question. If “this is hard to explain otherwise” isn’t a license, what is?

The answer has to run forward instead of backward. You don’t start with an awkward observation and reverse-engineer a reason God would have made it that way. You start with what the text says was commissioned, derive what state those functions actually require, and then go look.

Genesis 1 assigns functions. The luminaries give light on the earth and serve for signs and seasons and days and years. The vegetation bears seed after its kind. The man is made in the image of God and given dominion, told to work the garden and name the creatures. These are ends. Each requires a state for the end to obtain at all.

Take the luminaries. A sign that cannot be seen is not a sign. If the commissioned function includes reckoning time from the earth’s surface, then observability from that surface is part of what was commissioned, and light in transit follows from the function rather than from anyone’s embarrassment about distance. That’s a derivation. It could have failed. It didn’t.

Take Adam. Dominion, cultivation, naming, marriage. Every one presupposes a mature agent with language and a working body. The maturity isn’t posited to explain how he looked. It falls out of what he was told to do.

There’s a second constraint, and someone is already forming the objection. Fine, light was initialized in transit, but why does it carry mutually consistent spectral, geometric, and temporal information? Because a created mature state can’t be a bag of individually convenient facts. It has to hang together. If light is present because the luminaries were commissioned to be seen, then its wavelength, its source geometry, the fields around it, and everything else correlated with it cannot contradict each other when you apply ordinary physics. Function licenses the state. Coherence constrains its contents. That is not permission to invent whatever historical record I find convenient. It means only that whatever is legitimately initialized has to belong to one coherent world.

Notice what this does to the constraint problem. The set of commissioned functions isn’t something I get to expand when I run into trouble, because Genesis 2 says the work was finished. The text closes the list. That’s a far stronger fence than any I could build myself, because I can always talk myself into moving a fence I built.

✦

Now the part that will annoy people on my own side.

A rule that never refuses anything isn’t a rule. So here’s what this one refuses.

The radiogenic inventory. Radiogenic heat contributes to the long-term thermal budget supporting mantle dynamics and geological cycling, and that looked like a clean route to licensing uranium, thorium, and potassium in the crust. I wanted it to work. It doesn’t. Over a few thousand years, leftover primordial heat is ample. A short-lived world doesn’t need a replenishing radiogenic source, so the function doesn’t entail the inventory. Unlicensed.

High-redshift metallicity. The luminary function entails that stars shine and are visible. It doesn’t entail that distant galaxies contain heavy elements. You can notice that Genesis puts water before stars, and that’s a genuinely interesting observation, but noticing is not deriving. Unlicensed, pending an actual chain.

Then there’s the hardest case, and it deserves stating at full strength rather than in the version that’s easy to answer.

SN 1987A arrived in three independent channels. A neutrino burst, an optical light curve, and light echoes off a ring that was already there. All three agree under one event model. Critics point at that and say: succession is the only mechanism anyone has ever seen produce correlation like this, so this is a real event with a real history.

That inference is sound. I want to be precise here, because the tempting reply is to call it question-begging and the tempting reply is wrong. The critic is reasoning from the best available explanation, which is what anyone should do. What his argument depends on is the assumption that ordinary dynamics produced the state, and that assumption is exactly what’s in dispute. His reasoning is fine. It just can’t settle the question without helping itself to the answer.

Now, what can coherence actually do here, and what can’t it?

It handles the agreement between channels cleanly. Given an optical signature with particular parameters at a particular place, the neutrino front and the echo geometry consistent with those parameters have to be there too, because the physics relates them. A state with the light but not the neutrinos would be incoherent. So the alignment across channels isn’t a measure of historical depth. It’s a measure of how tightly the laws bind those channels together.

What coherence does not do, by itself, is explain why there’s a supernova signature there at all instead of nothing. That’s the harder half, and it’s what the critic is really pressing on.

The answer, if there is one, sits at the level of populations rather than events. A lawful cosmos of this scale and composition has stars distributed across evolutionary stages. A universe containing only main-sequence stars, with no remnants and no ejecta anywhere, would violate the mass function and the distributions the laws themselves govern. It would be an incoherent state. So event content of that general kind is entailed at the population level even though no particular event was ever specified.

Is that enough? I’m not sure yet, and I’d rather say so than claim the case is closed. It’s the hardest test the rule faces, and it’s the one I’d want a critic to push on first.

All of these are things I’d have preferred to keep. That’s rather the point.

✦

Before I go further, a word about the other side of the table, because a rule that only binds one party isn’t a rule either.

The move I’ve just described as the central defect of mature creation is not a creationist invention. It is what any comprehensive framework does when an observation doesn’t fit, and the naturalist account does it constantly.

JWST finds galaxies at high redshift that are too massive, too structured, and too numerous for the models? Raise the star formation efficiency. Adjust the early mass function. Posit feedback-free starbursts. Soft tissue and sequenceable collagen turn up in a fossil dated at sixty-eight million years? Propose iron-mediated crosslinking as a preservation mechanism, after the discovery rather than before it. Origin of life? Given enough time and chemistry we don’t yet understand.

Each of those may turn out to be correct. That isn’t the point. The point is that every one of them was produced in response to the anomaly it explains, and unbounded time plus emergence will absorb almost anything you hand it. “It evolved” and “given enough time” have the same logical shape as “it was created that way,” and the same immunity.

I want to be careful about the scope here, because overstating it would forfeit the argument. This deficit is real in specific places and it is not general. Nobody has a route from undirected chemistry to homochirality; percent-level enantiomeric excess from polarized light or crystal surfaces isn’t close to what life requires. Nobody has a route to the first functionally specified sequences. But stellar evolution and nucleosynthesis are among the best-modeled processes in physics, and pretending otherwise would be exactly the kind of convenient move I’m objecting to.

There’s also an asymmetry worth noting where it actually exists. In January 2022, before JWST returned any data, a creationist astrophysicist published the prediction that the telescope would find fully formed galaxies at unprecedented distances, carrying elements heavier than lithium. Dated, public, and risky. It held. Whether it follows from the framework I’m proposing here is a separate question, and I don’t think it does yet, for reasons I’ll get to elsewhere. But it was made in advance, and the accommodations on the other side were not.

So I’m not offering this rule as a concession. I’m offering it as a demand, and the first person it binds is me. If difficulty in historical reconstruction doesn’t license a creative act, then difficulty in mechanistic reconstruction doesn’t license an appeal to unknown chemistry and unlimited time. Neither side currently has a stopping rule. I’d like both to.

✦

I want to be careful about what I’m claiming, because the temptation to overclaim here is enormous.

This doesn’t prove the earth is young. It is not a chronometric argument. It’s a rule governing what mature creation is allowed to explain.

It also doesn’t touch what God can do. He acts in history constantly and the text says so: the Flood, Cana, the healings, the empty tomb. The restriction falls on me, not on him. I may invoke a divine act where the text reports one or where a commissioned function entails the state. Past that I’m speculating, and I should say so out loud rather than dress it up.

And the Flood cuts both ways. That it happened is reported. How it happened is not. So flood geology gets no protection at all. It carries the full empirical weight, and it can fail.

✦

Here’s what I think is at stake.

Mature creation has spent a hundred and sixty years in a defensive crouch, absorbing anomalies and calling it an answer. It absorbs because it can. Nothing stops it.

Put a stopping rule on it and the character of the thing changes. It starts generating claims that can be checked instead of accommodations that can’t. It starts refusing things. It becomes the sort of position that could turn out to be wrong, which is the only sort worth defending.

Whether it survives its own rule, I don’t know yet. Several cases I expected it to preserve have already been refused, but that may be early results, not final ones.

I’d rather hold a position that can lose than one engineered never