R&R Rebuttal (jet-rebuttal)

When to trigger

• You received an R&R (or a reject-and-resubmit) from JET and must reply to referees
• A referee flagged a gap in a proof, an unstated assumption, or a missing case
• You need to position generalizations the editor and two referees asked for

How JET R&Rs work

• Decisions come from the editor synthesizing two-plus single-anonymized referee reports. Because review is single-blind, referees may know your prior work — engage substance, not identity.
• For theory, the highest-stakes comments are about correctness and scope: an alleged hole in a proof, an assumption doing more work than acknowledged, a claimed "general" result that secretly needs finiteness/continuity/convexity, or a counterexample to a stated proposition.

Response strategy

1. Triage by severity. Correctness issues first (a broken lemma sinks the paper), then generality and necessity-of-assumption questions, then exposition.
2. For a proof gap: either supply the corrected/complete argument or, if the result is false as stated, narrow the claim to what the proof supports and say so plainly. Never paper over a hole.
3. For "is assumption X necessary?": answer with a counterexample (X dropped → result fails) or a generalization (X weakened, result survives). Add the weaker statement as the new main result where credible.
4. For exposition: restate the theorem in the body, move heavy machinery to an appendix (thm-restate), and tighten notation.
5. Show the diff. Quote each comment, give a direct answer, and cite the exact section / equation / theorem number in the revised .tex that implements the change.

Severity ladder for theory reports

When the referee's counterexample is right

1. Verify first. Check it satisfies all your numbered assumptions — many alleged counterexamples violate a regularity condition, and showing which line fails ends the issue.
2. Locate the first false step. If the example is valid, find the exact line of the proof it breaks; everything before that line survives.
3. Repair or retreat. Add the missing condition loudly (own the generality loss in the letter and the introduction), or restate the result at the strength the proof actually delivers.
4. Credit it. If the referee's example now appears in the paper, acknowledge the referee.

Response-letter skeleton

Editor summary (≤1 page, theorem-level):
  Theorem 1 unchanged; Proposition 2 narrowed (now requires A4 — see R1.3);
  new Example 3 shows A2 cannot be dropped (R2.1); Appendix C rewritten for readability.
Per referee, per point:
  [R1.3] "<verbatim comment>"
    → Type: correctness
    → Response: corrected the Lemma 2 step; the missing case is handled by new eq. (12)
    → Location: §4.2 and Appendix B.1; flagged in the attached latexdiff
Attachments: revised .tex, latexdiff PDF, this letter.

Anti-patterns

• Defending a result you now know is false instead of narrowing it
• "We thank the referee" with no concrete change or location
• Adding assumptions silently to rescue a proof without flagging the reduced generality
• Re-disclosing or hiding generative-AI use inconsistently from the original submission

Output format

【Referee N · Comment k】<verbatim>
  → Type: correctness | generality/necessity | exposition
  → Response: <corrected proof / counterexample / narrowed claim>
  → Change at: §x / Prop y / eq (z) in revised .tex
【Net effect】result strengthened / scope clarified / claim narrowed