Corridor
Let p be a Mordell-hard prime:
Status: proved structural theorem
Date: 2026-08-15
Depends on: FAB-COPRIME-DIVISOR-CRITERION.md
Claim boundary: this gives a necessary condition for coprime fab certificates on the Mordell-hard prime classes. It does not prove existence of such a certificate for every prime and therefore does not prove Erdős-Straus.
1. Setup
Let p be a Mordell-hard prime:
In particular
and p is a quadratic residue modulo 3, 5, and 7.
Suppose a,b are coprime positive integers below p, and k is a coprime fab certificate as in FAB-COPRIME-DIVISOR-CRITERION.md:
Write
with c>0, and put
2. Auxiliary factorization
From
and
we get
Hence
The coprime criterion gives gcd(a,k)=1, so
Write
Since the right side is 1 mod4 and k=3 mod4, necessarily
Also
because a common divisor would divide 1.
3. Nonresidue theorem
Theorem
For a Mordell-hard prime p, the leftover factor c satisfies
where the symbol is the Legendre/Jacobi symbol after removing square factors in the numerator.
Proof for odd c
Modulo k,
Since 4b^2 is a square modulo k,
because k≡3 mod4.
On the other hand
For odd c, Jacobi reciprocity and k≡3 mod4 give
Therefore
Because p≡1 mod4, reciprocity introduces no sign when reversing p and the odd part of c, hence
Even c
Write c=2^e c_0 with c_0 odd.
Since the hard classes satisfy p≡1 mod8,
If e>0, the identity p+k=4abc forces p+k≡0 mod8, so
and therefore
Removing the factor 2^e from the reciprocity calculation leaves the same sign as in the odd case. Thus again
QED.
4. External-prime corollary
For every hard class,
The first equality follows from p=1 mod8; the others follow by quadratic reciprocity from the fact that the six hard residues are quadratic residues modulo 3, 5, and 7.
If every prime factor of c belonged to {2,3,5,7} or occurred only through square contributions, then (c/p)=1, contradicting the theorem.
Therefore:
Corollary
Every coprime fab certificate for a Mordell-hard prime contains an odd prime factor
with odd valuation contribution and
Since p≡1 mod4, equivalently
5. Why this matters
The Type A/B / shadow program reached residual small-prime coordinates beginning at 11 and 13 after the 3,5,7 hard shield was imposed.
The divisor-parametrization framework reaches the same boundary from the opposite direction: a coprime certificate cannot live entirely inside the 2,3,5,7 squareclass support. It must import a genuine external nonresidue prime.
Thus a promising all-prime strategy is to coordinate:
- the first external quadratic nonresidue of the hard prime;
- the factorization of the linear forms
a+bp; - the divisor class
k=-p mod 4abfrom the coprime criterion.
This is a structural bridge between the two research languages, not a claim that the remaining existence theorem is closed.