Theorem
For an admissible target candidate write
Status: proved elementary theorem
Date: 2026-08-15
Claim boundary: this corrects the local parameter domain used in some C1/C2/CN proof-mining notes. It does not prove universal Type A/B coverage or Erdős-Straus.
Setup
For an admissible target candidate write
Let
The exact Dirichlet condition needed for a prime progression is
The parameter itself need not be a unit modulo Q.
Every admissible CRT base satisfies
because the hard residue is coprime to 840 and every Type A/B target trap is a unit modulo its target modulus.
Theorem
Assume gcd(r,L)=1. For every prime p|Q:
- if
p|L, then
for every s;
- if
p\nmid L, thenLis invertible modulop, and
Therefore
Higher powers of p in Q do not change the gcd condition.
Define
Then
|\mathcal D_{r,L}(Q)| =Q\prod_{\substack{p\mid Q\\p\nmid L}} \left(1-\frac1p\right). }</div>
Important correction
The auxiliary set
is generally not the exact reduced parameter domain.
In particular, for the hard-class program
Hence any local coordinate supported only on 3,5,7 carries the full residue-ring parameter domain as far as Dirichlet reducedness is concerned.
For q=3 specifically,
Thus two singleton rows forbidding {1} and {2} do not form a reduced obstruction: s=0 mod3 remains prime-compatible.
Combined with Q3-SINGLETON-PULLBACK.md, a genuine q=3 local cover requires three distinct singleton rows covering
The constructive core in DSC-COUNTEREXAMPLE.md shows that such a three-row cover can in fact occur.
Research rule
Future local escape and covering arguments must distinguish:
- primes already dividing
L, where parameter residues are unrestricted by reducedness; - primes outside
L, where exactly one affine class modulopis excluded.
Do not replace this exact affine condition by gcd(s,Q)=1 unless an equivalence has been proved for the stated special case.