The Bernoulli numbers are among the oldest objects in mathematics -- they appear in power sum formulas, in zeta values at even integers, and in the Taylor series for tangent. In 1840, von Staudt and Clausen proved that denom(B_2k) = product of all primes p where (p-1)|2k. Nobody noticed what happens when you combine this with the Cunningham map c(n) = 2n+1.
Each axiom prime p enters the Bernoulli denominator with a specific period -- how often it appears as k increases. Since p = c(q) for the Cunningham chain, the condition (p-1)|2k becomes q|k. The period of each prime IS the previous axiom prime.
| Prime | Period | Meaning |
|---|---|---|
| D = 2 | 1 (always) | Ground pair, never absent |
| K = 3 | 1 (always) | Closure, never absent |
| E = 5 | D = 2 | Every 2nd Bernoulli |
| b = 7 | K = 3 | Every 3rd Bernoulli |
| L = 11 | E = 5 | Every 5th Bernoulli |
First simultaneous appearance of all five: B_60 at k = 30 = D*K*E = lcm{1,1,2,3,5}. Not coincidence -- 30 = product of the first three axiom primes.
Watch as k increases. Each axiom prime dances in and out on its own Cunningham-inherited schedule.
| B_2k | k | Denom | Name |
|---|---|---|---|
| B_2 | 1 | 6 = D*K | Ground pair |
| B_4 | 2 | 30 = D*K*E | Observer enters |
| B_6 | 3 | 42 = D*K*b | THE ANSWER |
| B_8 | 4 | 30 = D*K*E | E returns, b gone |
| B_10 | 5 | 66 = D*K*L | Transcendental! L enters |
| B_12 | 6 | 2730 = D*K*E*b*13 | GATE enters |
| B_20 | 10 | 330 = D*K*E*L | ANIMAL (b missing) |
| B_60 | 30 | all 5 + intruders | First pentagonal |
At k=3, depth b=7 enters for the first time. The denominator is 2*3*7 = 42. The answer to the ultimate question is the Bernoulli denominator where depth first appears. At k=5, transcendental L=11 enters -- but E=5 is gone (period D=2, and 5 is odd).
At k=6: E=5 present (2|6) and b=7 present (3|6). But 6|6 means 12|12 means (13-1)|12. The gate 13 is FORCED open whenever observation and depth coexist.
The Ramanujan denominator 2730 = THIN + lambda = 2310 + 420. The famous 691 in the numerator of B_12? That is the scar left by L=11's absence: period(L) = E = 5, and 5 does not divide 6. The transcendental was locked out.
The Cunningham chain D=2 -> E=5 -> L=11 -> 23 -> 47 extends past the axiom. The prime 23 = c(L) has Bernoulli period (23-1)/2 = 11 = L. The chain L -> c(L) = 23 -> period(23) = L is a fixed-point loop. The contaminator's Bernoulli period IS the prime that generated it.
How many Bernoulli numbers have 11-smooth denominators? Each intruder p > 11 contaminates at rate 1/s where s = (p-1)/2:
| Intruder p | Half-period s | Factored | Survival |
|---|---|---|---|
| 13 (GATE) | 6 | D*K | 5/6 |
| 17 (ESCAPE) | 8 | D^3 | 7/8 |
| 19 | 9 | K^2 | 8/9 |
| 23 = c(L) | 11 | L | 10/11 |
Four-contamination product: (5*7*8*10)/(6*8*9*11) = 175/297 = E^2*b / K^3*L. Exact axiom expression. Asymptotic density decays; at k = 770 = D*E*b*L, it hits exactly K/(D*E) = 3/10.
Eisenstein series E_k have coefficients related to divisor sums. Modular enrichment -- structure visible mod c(k-1) versus generic -- follows a precise pattern when c(k-1) is prime:
| Weight | c(k-1) | Enrichment |
|---|---|---|
| 4 | b = 7 | b/D = 3.5x |
| 6 | L = 11 | L/D = 5.5x |
| 12 | 23 | 23/D = 11.5x |
| 22 | 43 | 43/D = 21.5x |
| 24 | 47 | 47/D = 23.5x |
Enrichment at prime c(k-1) is exactly c(k-1)/D by Fermat's little theorem. When c(k-1) is composite -- like weight 8 where c(7) = 15 = K*E -- enrichment drops to 1.9x. Only primes carry the full signal.
This work is and will always be free.
No paywall. No copyright. No exceptions.
If it ever earns anything, every cent goes to the communities that need it most.
This sacred vow is permanent and irrevocable.
— Anton Alexandrovich Lebed
Source code · Public domain (CC0)
Contributions in equal measure: Anthropic's Claude, Anton A. Lebed, and the giants whose shoulders we stand on.
Rendered by .ax via WASM DOM imports. Zero HTML authored.