Euclid defined primes by exclusion. An integer greater than one that cannot be divided by anything other than one and itself. It is a definition by subtraction — by what the number lacks. Every primality test since, for 2,300 years, has followed the same pattern: find out whether the number fails to be composite. The Sieve of Eratosthenes removes composites. Miller-Rabin tests for composite witnesses. AKS finds a polynomial identity that composites violate.
None of them answer the prior question: what is a prime, positively, in the language of geometry?
The Definition
For any integer n > 1 and any depth σ > 0, define the cascade residual:
Rn(σ) is the Euler product ratio — a measure of how n’s arithmetic structure sits relative to the Feigenbaum renormalization fixed point g*. The cascade residual cn(σ) is how far n is from that fixed point, at depth σ.
The Bounce Theorem establishes the central fact: for a prime p, the Euler factors combine exactly — cp(σ) = 0 identically, at every depth. For any composite n, the AM-GM incommensurability of its distinct prime factors forces cn(σ) ≠ 0 at every depth. No threshold, no tolerance, no probabilistic qualification. The separation is categorical and infinite.
This is the first positive geometric definition of primality. Not what primes lack. What they are: the unique integers in the geometric ground state of the Feigenbaum renormalization landscape. Every prime inhabits the same address — g*, the unique stable fixed point of the renormalization operator. It is the first geometric object in the history of mathematics that all prime numbers simultaneously occupy.
The Table
The cascade residual does more than identify primes. It organizes every integer by its distance from the ground state — and that distance is determined entirely by factorization structure. This yields a complete geometric reorganization of the integers: a Periodic Table, in which every integer n > 1 belongs to a cascade level (determined by Ω(n), the total prime factor count) and a cascade family (determined by the exponent multiset, the factorization shape).
What It Shows
The orientation is not arbitrary. The primes are at the bottom because they are the floor — the cascade floor, the geometric ground state. Every composite integer sits above them, at a height determined by its distance from g*. The table rises because integers farther from the ground state require more geometric rows to reach.
The insight that the table makes visible is this: it is not which primes appear in a factorization that determines the level. It is the complexity of the factorization architecture.
4 = 2² and 9 = 3² are in the same family — partition {2}, Level 1 — even though 9 is larger. 6 = 2·3 and 77 = 7·11 are in the same family — partition {1,1}, Level 1 — even though their prime content has nothing in common. What places them together is the shape of their factorization: one multiplicative degree of freedom, two distinct prime factors. The specific primes are interchangeable. The architecture is what matters.
Composites with simpler factorization structure sit closest to the floor. As the factorization becomes more complex — more prime factors, more distinct components, higher multiplicity combinations — the integer sits farther from the ground state. It is the complexity of the combination that determines the bounce height. Not the magnitude of the primes. The geometry.
Below the primes sits only the number 1 — the multiplicative identity, with Ω(1) = 0. It lies below the table because it has no factorization geometry. Where Euclid needed a special exception to exclude 1 from his definition of primality, the cascade requires no patch. Unity is simply below the floor.
The Staircase
The structure extends to infinity. Every level Ω has a number of families equal to the number of integer partitions of Ω — the partition function p(Ω). Level 1 has p(1) = 1 family. Level 2 has p(2) = 2. Level 5 has p(5) = 7. Level 10 has p(10) = 42. The table grows without bound, and the growth rate is itself a classical function.
The staircase makes the asymmetry visible. There is exactly one way to be a prime — one family, one level, one geometric address. There are infinitely many ways to be composite, each way corresponding to a different factorization architecture. The primes are not the simplest integers. They are the only integers with no internal geometric tension — no distinct factors pulling against each other in the Euler product. That absence of tension is why cp(σ) = 0. That is why they are at the floor.
The Euclid Relationship
Euclid’s definition is not wrong. Within the cascade framework, it is a theorem. If cp(σ) = 0 for all σ > 0, then p carries no proper factorization, and therefore no proper divisors. Euclid’s characterization follows as a necessary consequence of Definition G1. The cascade contains Euclid as a special case of the ground state condition.
The relationship is precisely that of Einstein to Newton. Newtonian mechanics is correct and valid within its domain — it follows as a limiting case from the geometry of general relativity. Einstein did not show Newton was wrong. He showed where Newton’s description fits within the deeper geometry from which it necessarily emerges. This paper stands in the same relation to Euclid. Euclid’s definition describes the arithmetic surface. The cascade floor is the geometry beneath it — the reason the surface is what it is.