Subpower upper bounds for the actual two-character residue #
The target character's original L-value is retained. Its induction correction, and the other two L-values at the common modulus, cost only logarithms. No coprimality or primitivity is needed for these upper bounds.
theorem
DirichletCharacter.norm_inductionEulerProduct_one_le_one_add_log
{q : ℕ}
(χ : DirichletCharacter ℂ q)
(Q : ℕ)
[NeZero Q]
:
Inducing any character to a nonzero common modulus has logarithmic Euler loss at one, including overlapping levels.
theorem
DirichletCharacter.twoCharacterResidue_re_le_original_value_mul_subpower
{q Q : ℕ}
[NeZero q]
[NeZero Q]
(hq : q ∣ Q)
(χ : DirichletCharacter ℂ q)
(ψ : DirichletCharacter ℂ Q)
(hχ : χ ≠ 1)
(hquad : χ ^ 2 = 1)
(hψ : ψ ≠ 1)
(hprod : (changeLevel hq) χ * ψ ≠ 1)
{a : ℝ}
(ha : 0 < a)
:
Only the target's L-value is pulled back to its original modulus.
The correction remains explicit; the other two factors are bounded at Q.