Most journals asking for a revised manuscript want it submitted twice over: once as a clean version — the manuscript exactly as it would be typeset, with no visible edit marks — and once as a marked-up (tracked-changes) version that shows every addition, deletion, and edit made since the previous round. Confusing the two, or submitting only one when a journal requires both, is one of the most common — and most easily avoided — mistakes authors make at the revise-and-resubmit stage.
Why journals ask for two versions
A revised submission serves two different readers with two different needs, and no single file format serves both well.
- The editor and reviewers need to verify, efficiently, that the concerns raised in the previous round were actually addressed, and where. Re-reading the entire manuscript from scratch to spot what changed is slow and error-prone; a marked-up file lets them jump straight to the edits.
- Future readers — anyone who eventually reads the published article — need a document free of edit marks, strikethrough text, and colored insertions. A marked-up file is unreadable as a finished piece of writing and is never the version that goes to typesetting.
Rather than ask authors to choose, most journals ask for both, alongside the response to reviewers letter, which explains in prose what changed and why. The three documents work together: the letter tells the story, the marked-up file lets reviewers verify the story against the actual text, and the clean file is what moves forward if the manuscript is accepted.
This convention is standard practice across the scholarly publishing pipeline, though the exact wording varies by publisher — some call it a “clean copy” and “annotated” or “highlighted” copy, others “unmarked” and “marked” manuscript. Author instructions almost always specify it explicitly in the revision-submission checklist rather than leaving it to convention alone, so always check the specific journal’s author guidelines rather than assuming a generic pair of file names will satisfy the system.
What each version is actually for
The clean version
The clean version is the manuscript as it will exist going forward: no tracked-changes markup, no comment balloons, no highlighting. It is the file used for:
- Copyediting and typesetting if the manuscript is accepted
- A second round of peer review, if reviewers are shown the manuscript itself rather than only the change log
- Plagiarism/similarity screening in some editorial systems, where leftover markup or comments can distort a similarity score
Nothing about the clean version signals what changed — that is intentional. Its job is to read as a finished, coherent manuscript, exactly as if it had been written that way from the start.
The marked-up (tracked-changes) version
The marked-up version exists purely to make the editorial team’s verification job faster. It should show, unambiguously:
- Every sentence, phrase, or word that was added, deleted, or altered since the version reviewers last saw
- Ideally, changes attributable to the correct author/reviewer round, if the editorial system or word processor supports it
A marked-up file that shows changes going back several revision rounds, rather than only the most recent one, is a common and avoidable source of confusion — see the mistakes section below.
Tools and conventions for producing both versions
Microsoft Word: Track Changes
The most common path for word-processor manuscripts. Practical points specific to producing a clean and marked-up pair:
- Marked-up file: save/export the document with Track Changes left on and all insertions/deletions visible (Word’s “All Markup” view). Some editorial systems want the changes visible inline; others accept a Word document with tracked changes embedded that the reviewer can toggle themselves.
- Clean file: use Word’s “Accept All Changes” (Review tab) to produce a version with no visible markup, then save it as a separate file. Accepting changes in the working file and forgetting to also keep an unaccepted, tracked copy is the most common way authors end up unable to produce a proper marked-up version after the fact.
- Check for hidden markup before submitting the clean file. Word can retain tracked changes, comments, or hidden text that isn’t visible on screen depending on view settings. Before exporting the clean file, confirm under Review → Track Changes that tracking is switched off, and use Word’s “Check for Issues” / Document Inspector to catch anything still embedded in the file that isn’t showing in the current view.
- For teams collaborating across co-authors before the marked-up file is finalized, see CASRAI’s guide to track-changes etiquette for co-authored manuscripts and to managing manuscript versions across co-authors — both cover the collaborative editing stage that precedes producing the final submission pair described here.
Google Docs: Suggesting mode
Google Docs’ “Suggesting” mode is the rough equivalent of Track Changes. A marked-up export can be produced with suggestions still open; a clean export requires accepting all suggestions first (Tools → Review suggested edits → Accept all) before downloading as .docx or PDF. As with Word, always re-open the accepted, “clean” export to confirm no suggestion markers or open comments remain before submission.
LaTeX: latexdiff
latexdiff is the standard tool for producing a marked-up manuscript from two .tex source files (the previous submitted version and the revised version). Run against the two files, it generates a new .tex file with additions/deletions marked (by default, underlined additions and struck-through deletions), which compiles to a PDF showing the diff directly — this is the marked-up file. The clean file is simply the revised .tex source compiled normally, with no diff markup involved, since LaTeX has no equivalent of embedded “tracked changes” living inside the working document the way Word does. Because latexdiff operates on two complete files rather than an in-document edit history, authors need to keep an accurately labeled copy of the exact previously-submitted version to diff against — a stray edit to what should be the “before” file produces a diff that misrepresents what actually changed.
Manual highlighting
Where a journal’s system doesn’t accept native tracked-changes files (for example, some require a single marked-up PDF regardless of source format), authors sometimes manually mark changes using colored text, highlighting, or margin annotations, then flatten to PDF. This is more error-prone than software-generated tracking, since it depends on the author remembering to mark every change by hand, but it’s occasionally the only option a submission system allows — check the specific journal’s technical requirements before assuming a native Word or LaTeX diff will be accepted directly.
Common mistakes to avoid
- Submitting the wrong file as the wrong version — uploading the marked-up file into the “clean manuscript” slot of the submission system, or vice versa. Editorial systems generally don’t validate file contents against the label, so a mislabeled upload can go unnoticed until an editor opens it.
- Leftover tracked changes in the “clean” file — accepting changes on screen without checking that Track Changes is actually turned off, then submitting a file that still silently records subsequent edits, or that still contains hidden comments. Always run a document-inspection check (Word’s Document Inspector, or simply reopening the file and checking the Review tab) before calling a file clean.
- A marked-up file that mixes multiple revision rounds — showing changes accumulated across two or more rounds of revision rather than isolating only what changed since the version reviewers most recently saw. This makes it harder, not easier, for reviewers to verify the current round’s edits, defeating the point of the marked-up file.
- Mismatched pagination or numbering between the response letter and the marked-up file — if the response to reviewers letter references specific page or line numbers (“see p. 4, line 12”), those references need to match the marked-up file’s actual layout, which can shift if line numbering, headers, or formatting differ between the two documents.
- Assuming the diff file compiles correctly (LaTeX) — a
latexdiffoutput occasionally fails to compile cleanly, particularly with complex tables, equations, or custom macros, and needs a manual check before submission rather than being uploaded unverified. - Forgetting to keep an unmodified copy of the previously submitted version — needed as the “before” file for both Word’s compare-documents feature and
latexdiff. Once the working file has been edited in place, the exact prior version can be hard to reconstruct without it.
How this fits into the revision-submission package
A complete revision submission typically bundles three things: the response to reviewers letter (the point-by-point explanation of what changed and why), the marked-up manuscript (letting the editor and reviewers verify those claims against the actual text), and the clean manuscript (the version that moves forward toward publication if accepted). For the difference between the decision types that trigger this process in the first place, see CASRAI’s comparison of major revision vs. minor revision and its guide to what a revise-and-resubmit decision actually involves.
Frequently asked questions
Do all journals require both a clean and a marked-up version?
Most do, but not universally, and the acceptable format for the marked-up version varies (native Word tracked-changes file, a flattened PDF with visible markup, a latexdiff-generated PDF, or a system-generated diff for manuscripts submitted through certain editorial platforms). Always check the specific journal’s revision-submission instructions rather than assuming a generic pair of files will be accepted as-is.
Which version should I submit if a journal’s instructions don’t specify?
Submit both, clearly labeled in the filenames (for example, “Manuscript_clean.docx” and “Manuscript_trackedchanges.docx”), unless the submission system only provides a single upload slot — in that case, contact the editorial office rather than guessing, since submitting only a marked-up file as the sole manuscript can cause confusion if the paper proceeds toward typesetting.
Does the marked-up version need to show changes from every previous round, or just the most recent one?
Just the most recent round, unless a journal specifically asks for a full history. The marked-up file’s purpose is to let the current round’s reviewers verify the current round’s changes; older, already-resolved edits add noise rather than clarity.
Can I use LaTeX’s latexdiff output as the manuscript I submit if accepted?
No. The latexdiff output is markup-only, meant for reviewer verification; it should never be the file that proceeds to copyediting or typesetting. Always compile and submit the plain revised .tex source as the clean version.
