A joint paper needs writing, and the obvious way to share the load is to split it by section: you take methods, I take results, someone else takes the introduction.
This reliably produces a first draft that reads like four separate papers, with inconsistent terminology, repeated content, and a discussion that does not follow from the results as presented. Someone then spends more time reconciling it than they would have spent writing the whole thing.
Why the obvious split fails
Three reasons:
Sections are not independent. What the introduction promises determines what the discussion must deliver; how the methods are described determines what the results can claim. Writing them in parallel means each author is guessing what the others will say.
Terminology drifts. Four people writing about the same construct will use four slightly different terms, and the differences are hard to see once you are inside the draft.
Nobody owns the argument. A paper makes one argument. Divided authorship without a single person holding that argument produces a document where each part is defensible and the whole does not land.
A sequence that works
Five steps, in order. The first two are where the time is actually saved.
One: agree the claim, in one sentence. What is this paper's single main finding? Write it down and circulate it. Disagreements at this stage are cheap; the same disagreements discovered at draft three are expensive.
Two: agree a detailed outline — section headings plus, under each, the specific points in order. Two pages. This is the document to argue about.
Three: one person writes the first full draft — rough, fast, complete. Not polished. The point is a whole document with one voice and one argument.
Four: the others write into it, each contributing the parts where they hold the expertise or the data — but into an existing structure rather than into a blank section.
Five: one person — usually the same one — does a full pass for consistency.
This concentrates the work rather than sharing it evenly, and that is deliberate. It should be reflected in authorship position, agreed in advance.
Who should hold the pen
Not automatically the most senior person, and not automatically the native English speaker. Three better criteria:
Who understands the whole study, not just their part.
Who has the time in the next few weeks. A first draft that stalls for two months because the designated writer is teaching is worse than a slightly weaker draft delivered on time.
Who benefits most from the position that this work justifies — often an early-career member, and giving them the draft with real support is one of the more useful things a group can do.
Language differences, handled fairly
On a paper written in English with non-native speakers among the authors, two failure modes are common and both are avoidable.
The first: the native speaker rewrites everything, and in doing so changes meaning in places where the original was precise and the rewriting was not. The fix is a simple rule: language edits are proposed, not applied silently, wherever they touch a claim.
The second: non-native authors are assigned only data and figures because "the writing will be faster that way". Over several papers this removes the one thing that develops writing ability, and it becomes self-fulfilling.
A better arrangement: everyone writes their part in whatever state they can; language polishing is a separate, named task done at the end, by whoever is best placed. Separating drafting from polishing removes the pressure that produces both failure modes.
Keeping the revision loop short
Four practices:
One current version, one place. Circulating attachments produces versions that fork silently.
A deadline for comments, with a default. "Comments by Friday; after that I will assume the section is fine and move on." Without the default, one slow co-author stalls everything, and nobody wants to be the one chasing.
Comment, do not rewrite — except in your own section. A draft where four people have silently rewritten each other's sentences is unmergeable.
Resolve disagreements in a call, not in comment threads. Two rounds of written argument is the point to pick up the phone; across a language difference, that point comes sooner.
One thing to settle at the outline stage
Who decides when the paper is finished.
There is always a co-author who wants one more analysis and one who wants it submitted last month. Naming, in advance, the person who makes that call saves a specific and very common argument — the one that happens when the draft is essentially done and everyone is tired of it.
Why does splitting a paper by section fail?
Sections are not independent, terminology drifts between writers, and nobody ends up owning the single argument the paper makes.
What sequence works better?
Agree the main claim in one sentence, agree a detailed outline, have one person write a rough complete draft, let others write into it, then one consistency pass.
Who should write the first draft?
Whoever understands the whole study, has time in the next few weeks, and benefits most from the authorship position — not automatically the most senior or the native speaker.
How should language editing be handled?
As a separate named task at the end, with the rule that edits touching a claim are proposed rather than applied silently.
