Delete one axiom, keep everything else
How do you tell a good theory from a merely correct one? Correctness is the price of entry — the theorems have to hold. Past that, mathematicians reach for a word that sounds soft but does real work: elegance. This chapter is about one particular, slightly contrarian idea of what elegance means — the idea that quietly governs every choice in this book.
There is a habit in how mathematics rewards its practitioners. You make your name by adding: a new axiom, a new structure, a new object that unlocks a whole landscape of theorems. Nobody has ever become famous for removing an axiom. Adding looks like discovery; subtracting looks like retreat.
The bias isn't baseless — a well-chosen new axiom really can open enormous territory, and mathematics is richer for the great ones. But it has trained a reflex: when something doesn't fit, add machinery to handle it. And that reflex runs up a bill nobody adds together.
The method behind this book runs the other way. Its name is a mouthful — Principled Emergence Metatheory, PEM for short — but its rule of taste is one sentence: a theory is more elegant the more of its structure emerges from fewer, simpler axioms stated plainly, rather than being installed by hand, one axiom per feature.1
Emergence is the operative word. In an emergent theory you do not write in the reciprocal of zero, or the imaginary unit, or the circle. You state a few plain demands, and those things appear on their own — forced into being by what you already committed to. The less you have to declare, and the more that falls out unbidden, the better the theory, by this measure. Elegance is fewness at the top paying for richness below.
Here is the line item the "add more" reflex skips. An axiom's cost is not just itself. It is every exception, caveat, and special case it forces downstream. Each time a theorem has to say "for all x, provided x isn't the troublesome one," that little provided is a tax — and the whole landscape of results pays it, permanently. Judge an object not by how simple it looks but by how many caveats it spawns, and some very "simple" objects turn out to be ruinously expensive.
By that accounting, the absorbing zero — the ordinary 0, carrying its rule that 0·a = 0 — is the most expensive object in all of mathematics. Look at what it charges:
In fairness, absorption does buy something: it keeps the ordinary number systems tidy, a clean field with one well-behaved zero. PEM's question is simply whether that tidiness is worth the sprawl of exceptions it demands everywhere else — and PEM's answer is no. An absorbing zero is not an elegant object. It is a landmine, and it has been claiming victims since the third grade, the first time "you can't divide by zero" arrived with no reason attached to it. It pays no rent and charges everyone else.
Faced with a landmine, the reflex is to build better warning signs: more axioms, sharper caveats, careful fences around the hole. PEM says something blunter. Don't fence the landmine — remove the thing that planted it. Find the single axiom responsible, delete it, and delete nothing else, so you can see exactly what it was costing you all along.
That axiom is absorption. Everything from here is what happens when you pull it out and add nothing back in its place: the reciprocal of zero, the imaginary unit, an exact circle — none of them posited, all of them emergent, each a piece of structure that was being held just out of reach by one line nobody thought to question.