The Proof We Settled For

Everyone treats "scientifically proven" as the strongest thing you can say about a matter. Actually, it's the weakest kind of proof there is — and the men who fashioned it knew exactly what they were giving up. But the generations that came after are the ones who forgot.

---

There's a move people reach for when they want to end an argument. They say it's been proven. Scientifically proven, the studies are in, the matter is settled. The line is meant to close the conversation, and most of the time it works — because almost everyone treats scientific proof as the strongest thing you can say. The ceiling. The last word. The place past which doubt is just stubbornness. Scientifically proven is as certain as it gets.

However, that ranking is backward. And the reason it's backward is sitting in a swap the British empiricists made across the seventeenth and eighteenth centuries. It's a swap they were honest about, one that worked better than anyone had a right to expect, and whose terms — what we got, and what we lost — have been completely forgotten.

Start with what proof used to mean, before the empiricists got to it. For about two thousand years, to prove something meant to demonstrate it. And "demonstrate" had a hard, specific sense: to show that a conclusion couldn't possibly be false, as long as you accepted what came before it.

The model for this was geometry. Think about how you learn that the angles inside a triangle add up to a straight line. You don't go out and measure a thousand triangles and notice they all come close. You prove it — once — from the definitions of what a triangle and an angle are, and the proof holds whether or not a single triangle exists anywhere in the world. That's demonstration. It doesn't need the world in order to check it.

The engine was the syllogism, the basic form of classical logic: all men are mortal; Socrates is a man; so Socrates is mortal — and the so there is iron. Grant the first two and the third is not likely, not well-supported, but locked. It cannot come out otherwise. That kind of certainty is what proof used to mean.

And here's the strange thing about that certainty, the thing nobody thought to call a weakness for a very long time. A syllogism never tells you anything you didn't already know. The fact that Socrates is mortal was hiding inside "all men are mortal" the whole time — the argument just walked it out into the open. Deduction works like a sealed pipe: whatever certainty comes out the bottom, you had to pour in at the top.

For geometry, that's no problem at all, because the things geometry talks about are made out of definitions, and you're the one who gets to set the definitions. But for the actual world, it's a disaster. You can't reason your way to what the world happens to contain. You have to go out and look. And looking was the one thing this older, stronger kind of proof had no room for.

This is exactly where Francis Bacon comes in, and he's the one who said the quiet part out loud. In 1620 he published a book called the Novum Organum — Latin for "new instrument" — and the whole thing is an attack on the syllogism as a tool for learning anything about nature. His complaint was sharp: a syllogism can win your agreement to a sentence without ever getting hold of the thing the sentence is supposed to be about. You can build a chain of flawless logic on top of words that were never properly fastened to the world, and all that flawlessness buys you exactly nothing.

Bacon wanted the world. He wanted knowledge that did something — that produced results, built things, worked. So he proposed a completely different method, called induction — reasoning the other direction, from many specific observations up to a general rule. Instead of starting with a grand certainty and unpacking it, you start with what you actually observe, case after case, and you build your way up toward the general claim.

But that method came with a bill, and Bacon knew precisely what it was. Every step of induction claims more than the evidence underneath it. You watch the sun come up a few thousand mornings and you conclude it always will — and that conclusion reaches far past the mornings you actually saw. Which means it can be wrong, in a way a syllogism never can. Bacon took that deal on purpose. He gave up the can't-possibly-be-false of the old proof in exchange for a method that could finally reach out and touch nature. He bet that enough care and discipline would keep it reliable, even though it could no longer be certain.

And that's the whole move, right there. The bar for what counts as proof quietly shifted — from "this conclusion cannot fail" to "this conclusion is well-supported." Everything that's happened since is a refinement of that one swap, but the swap itself has never changed. To make the world provable at all, Bacon changed what proof means. He bought reach, and he paid for it with certainty. It was the right trade. And it was still a trade.

About seventy years later, John Locke wrote the whole thing down in black and white. His Essay Concerning Human Understanding sorts out what human beings can actually know, and it sorts that into ranks. At the very top sits real knowledge — the certain kind, the kind the ancient Greeks would have recognized. We get that, Locke says, in mathematics and in a small handful of other places where we can inspect the ideas directly.

And everything else? Everything about the real, physical world — every "matter of fact," to borrow his phrase, beyond what happens to be sitting in front of your eyes this second — Locke drops into a lower rank. He calls it judgment, or probability. Not knowledge. Just probability.

He doesn't soften this and he doesn't bury it. He says it flat out: natural philosophy — the thing most people now simply call "science" — can never be made into a true science, by which he means a body of certain, demonstrated truth like geometry. The most careful experimenter alive, Locke says, earns better opinion — not knowledge; sharper judgment — not certainty. Read that as a man handing something over in full view of everyone. The strong, old kind of proof simply does not stretch to cover the world. And rather than pretend that it does, Locke draws the line clearly and tells you which side of it everything you care about is standing on.

And then David Hume came along and drew that same line so precisely that nobody has managed to rub it out in the nearly three centuries since. In 1748, in a book called the Enquiry, Hume splits everything we can think about into two boxes. And those two boxes are this entire essay, squeezed down small.

In the first box go what Hume calls "relations of ideas" — things like geometry and arithmetic, truths you can work out by pure thinking without ever getting up from your chair. They are perfectly certain, every bit as certain as the ancients ever wanted. But — and this is the catch — they are completely empty of the world. Proving something about a triangle tells you nothing about whether there's a single triangle out there anywhere. The certainty is real, but it's sealed off from reality.

In the second box go "matters of fact" — meaning everything that is actually, concretely the case. This box does touch the real world. And the price of that contact is that nothing in it is ever certain, because the opposite of any fact is always possible. "The sun will rise tomorrow" is a matter of fact, and there's nothing the least bit contradictory about imagining that it doesn't. So it can never be proven the way a theorem is proven. Notice what's happened here: the certainty is in one box, the world is in the other, and there is nothing at all that sits in both.

Then Hume does something even more unsettling. He pulls the floor out. Everything we know about matters of fact, he points out, runs through cause and effect — we make sense of the world by learning that this causes that. But where does our knowledge of cause and effect actually come from? Only from experience — from having watched one thing follow another, over and over and over. And here's the trap. To get from "this has always followed that" to "this will follow that next time," you have to assume one thing: that the future will go on behaving like the past. That single assumption is quietly doing all the work. And you cannot prove it, because every argument you might offer for it already takes it for granted. You can't say "the future will resemble the past because it always has so far" — that's just helping yourself to the very thing you were trying to prove.

Induction — the method Bacon traded certainty to get — has no solid ground underneath it whatsoever. It runs on habit. We expect the sun to come up because we've gotten used to it coming up, not because we've reasoned our way there. Even the feeling that one thing must cause another, Hume says, isn't something we ever actually observe out in the world — it's a kind of mental reflex, a custom worn into us by repetition. Hume isn't trying to destroy scientific proof here. He does something quieter, and much harder to shake off: he measures, with total precision, exactly how much weaker this kind of proof is than the kind it replaced — and he shows that the weakness goes all the way to the bottom, and can't be patched.

Now step back and look at what these three men actually did — and, just as importantly, at what they didn't do. They demoted proof. They moved the standard from "a conclusion that cannot fail" down to "a conclusion that's well-supported." But not one of them ever once mixed up the new, weaker thing with the old, stronger one. Bacon sold the trade as a feature and named its cost in the same breath. Locke wrote that cost right into his book and put the entire physical world on the cheaper side of the page. Hume drew the boundary in permanent ink and flatly refused to pretend it could be crossed.

They reached for the weaker kind of proof on purpose, and for one very simple reason: the stronger kind had nothing to say about the world, and the world was the whole point. The certainty was never on offer for matters of fact in the first place. They knew that. They chose reach anyway — and, crucially, they told us plainly that they were choosing.

So where did it go wrong? The forgetting came later, and it came from the rest of us. Somewhere along the way — after this watered-down, well-supported, never-quite-certain kind of proof turned out to work spectacularly well, after it built the modern world brick by brick, every engine and vaccine and bridge in it — we quietly slid it back onto the throne the old proof had left empty. "Scientifically proven" stopped meaning the best support a claim can possibly have, given that this kind of claim can never be certain and started meaning, simply, as sure as anything ever gets. We took the runner-up and lost the paperwork that said it was only ever standing in for a champion. We put the consolation prize on the throne. The exact thing the empiricists accepted because it was weaker than real proof, we now wave around as though it were the strongest proof there is. That isn't a small slip. It's the precise opposite of what the thing was built to be.

There's one more thing, and it's the one that matters. That older, stronger proof didn't die when the empiricists made their swap. It's still here, untouched, exactly as certain as it ever was. It simply doesn't apply to anything that actually happens. Mathematics still demonstrates — still hands you conclusions that cannot fail — and mathematics still can't tell you whether the bridge will hold, because a bridge is a matter of fact, and matters of fact got sorted into the other box three hundred years ago and never came back out. Everything that touches the real world is stuck with the lesser proof — the one Bacon settled for so that the world could be studied at all.

And there's nothing wrong with using it. It's the finest instrument we have ever built for understanding the world, and the evidence for that is the world it built. The mistake isn't using it. The mistake is forgetting what it is — forgetting that it's the lesser proof, accepted deliberately, because the greater one couldn't reach this far. The moment you treat "proven" as though it meant "certain," you've thrown away the one thing the empiricists were so careful to hold onto: the memory that they gave certainty up, on purpose, and took something else in trade.

They settled with their eyes wide open, and they said so plainly. So the dishonesty isn't theirs. It's ours — and we repeat it every single time we say proven and mean sure.