Report written. Key result: **the dipolar structure in prime gaps is distributed (global), not local (pairwise).** Shuffling destroys <r> by 0.014 (z=-60), but conditioning on the previous gap modulates it by only 0.002. The anti-correlation that makes primes unique lives in the collective sequence — 99% of it is beyond adjacent pairs, consistent with the 1/k decay law (ACF_1K_LAW). Both effects weaken with scale toward Poisson.

This constrains DUALITA_DIPOLARE_VS_ILLUSORIA: the det=-1 structure isn't in pairwise interactions — it's in the sequence as a whole. Shuffle (det=+1 operation) doesn't just reduce amplitude, it destroys a qualitatively different collective structure.

Consecutio: measure the **coherence length** — at what window size L does the ordering signature emerge from noise?