[amaybaum]

Independent review: ACCEPT WITH NITS at abd4b45a; its one nit is optional and not applied. Merge: authorized by the owner after independent review and green CI.

This PR corrects a citation error that #927 put on main. It adds no new scientific claim: every changed sentence restates its cited source.

• Reviewed head: abd4b45a638c530b0c64d0abf56267260b5fe298, based on main 83d1da1b.
• Up-to-date head: ebbb1a97b66a0a14f19b429602b70c9e0d9df771. This is an ordinary merge commit of main 61930e90 (Release gate: citation-roles, support ids reachable from the commit under check (#915) #916) into the reviewed head, required because the ruleset demands up-to-date branches.
  • It resolved no conflicts.
  • git diff main ebbb1a97 is byte-identical to the reviewed change set.
  • citation-roles and legacy-records (303 intact) pass.
  • The three required checks are green.

The error (F28, from the #918 novelty audit)

After #927, Main:332 credited Rivas–Huelga–Plenio 2010 [4] with "monotonicity of system-plus-ancilla distinguishability characterizes CP-divisibility". [4] does not state that. It gives an entanglement-based witness and a Choi–Jamiołkowski test.

The distinguishability characterization is Bylicka–Johansson–Acín [9], Theorem 1, and it holds for bijective maps. The sentence also omitted that scope. The misattribution came in through a review nit on #927.

Changes

• Main:332.
  • [4] keeps the entanglement witness, whose increase witnesses CP-indivisibility.
  • [9] gets its own theorem: for bijective (invertible) dynamical maps on ℂ^d, CP-divisibility is equivalent to monotonic non-increase of the trace distance between every pair of equiprobable states of the system and a (d+1)-dimensional ancilla.
• Main:606. The XOR example also cites [5], Eq. (15), "up to flipping the final bit". No priority is claimed.
• appendix-b:96 / FULL:5885. This site carried the same unscoped claim (§A.25). It is now restricted to bijective (invertible) maps, and its pointer is corrected to [Main §3.2], where the statement is.
• Generated forms. Main.tex/.pdf and the FULL book are regenerated.

Sources, read in open full text by the author and again independently by the reviewer

• [9] Bylicka, Johansson and Acín, PRL 118, 120501 (2017), arXiv:1603.04288, Theorem 1.
• [4] Rivas, Huelga and Plenio, PRL 105, 050403 (2010), arXiv:0911.4270. Its abstract calls the entanglement test "a sufficient condition for non-Markovianity", which is the same one-way direction as the new text.
• [5] Rivas, Huelga and Plenio, Rep. Prog. Phys. 77, 094001 (2014), arXiv:1405.0303, Eq. (15), p. 7. Its support is {001, 010, 100, 111}, so x3 = 1 ⊕ x1 ⊕ x2.

Review

The review confirmed:

• every hypothesis of BJA Theorem 1 is carried, and for §A.34 both directions rest on that single iff theorem;
• [4] is quoted in its own direction;
• "up to flipping the final bit" is exact;
• the §3.4 → §3.2 pointer correction is right;
• no remaining site misattributes the characterization or leaves it unscoped;
• the .tex and PDF text regenerate identically;
• the battery is green.

Recorded, not changed: Main:332's clause "ordinary system trace distance … can stay monotone on a CP-indivisible family" carries no citation.

Related: #927 (merged; source of the error), #918.

🤖 Generated with Claude Code

https://claude.ai/code/session_013T6pfKuv5mi3MXtKjH2tZb