Writing for publication

Writing for international journals: what “improve the English” usually means

The most common feedback given to non-Anglophone authors is also the least useful, because the underlying problem is usually not the sentences. It is what the sentences are being asked to do.

Published 15 August 2026Reading time 12 min
Portrait of Dr. Ngo Cong-Lem
Dr. Ngo Cong-Lem
Founder and editor, ELT Pub · Dalat University · Monash University

A manuscript comes back in eleven days without review. The editor's note says the work is not a good fit for the journal at this time, and suggests the English would benefit from professional editing before resubmission elsewhere.

This is the most common experience in international publishing, and it is also the least informative. The note is polite, generic, and — in a large share of cases — misdiagnoses the problem. Papers written in fluent, grammatical English are desk-rejected every day. Papers with visible second-language features are published every day. Something other than sentence-level correctness is doing most of the work.

This article is about what that something is.

What “improve the English” usually means

Editors are busy and their rejection notes are short. “The English needs work” is a phrase that gets used for at least four distinct problems, only one of which is actually grammar:

The practical consequence matters: a great deal of money is spent on language editing services that fix the first problem and leave the fourth untouched. The paper comes back cleaner and is rejected again.

A diagnostic. Ask a colleague outside your immediate subfield to read only your abstract and introduction, then tell you in one sentence what you are claiming and why it matters. If they cannot, the problem is not your grammar, and no proofreader will find it.

The contribution has to be visible early

International journals in most fields expect the introduction to perform a specific rhetorical job, and to perform it quickly. John Swales' Create a Research Space model — the standard description of this in English for academic purposes, and still the most useful — sets it out as three moves:

  1. Establish a territory. Show the area matters and summarise what is known.
  2. Establish a niche. Identify what is missing, unresolved, or contradicted in that body of work.
  3. Occupy the niche. State what this study does about it, and outline what follows.

Move 2 is where most rejected manuscripts fail. Many submissions establish a territory at length — sometimes for two thousand words — and then move directly to method, never having said what was wrong with the existing literature. The reader reaches the method section not knowing what problem is being solved.

This is frequently a transfer effect from local publishing conventions rather than a failure of thought. In several national and regional traditions, extensive deference to prior work is expected, and pointedly identifying a gap reads as presumptuous. In the international convention it is mandatory. You are not permitted to be modest about the gap: if nothing is missing, there is no reason to publish.

Practically, Move 2 should be explicit enough to be quotable. Something with this shape, in your own words:

Although X has been extensively documented in [context], studies have largely relied on [approach], and little is known about [specific gap]. This is a problem because [consequence].

Then Move 3, immediately, with no delay:

The present study addresses this by [what you did]. It contributes [specific claim] to [specific conversation].

If a reader cannot extract your contribution from the first page, the review is already going badly, regardless of what appears on page fourteen.

The literature review is an argument, not a catalogue

The second structural pattern that draws rejection is the literature review written as an annotated inventory: Nguyen (2019) found this, Smith (2020) found that, Lee (2021) found something else. Each summary is accurate. Nothing accumulates.

The international convention expects the review to be organised by idea, not by study, and to be building toward the gap you named. Every paragraph should be doing one of a small number of jobs: establishing consensus, identifying disagreement, exposing a limitation shared across studies, or narrowing toward your specific question.

A useful revision test: cover the citations and read only your topic sentences. If they form a coherent argument that arrives at your research question, the review is working. If they read as a list of names, it is a catalogue and reviewers will say it is descriptive — which is the word they use for this problem.

Synthesis also means grouping. Three studies using self-report measures found an association, whereas the two using observational data did not is synthesis. Listing all five in sequence is not.

Results and Discussion do different jobs

Conflating these two is so common that many journals' author guidelines address it directly, and it is a reliable trigger for major revisions.

SectionIts jobWhat does not belong
Results What you found. Findings reported neutrally, in a deliberate order, with statistics or excerpts as evidence. Explanations of why. Comparisons to other studies. Implications. Speculation.
Discussion What it means. Interpretation, relation to prior work, mechanism, limitations, implications. New results appearing for the first time. Restating every finding in the same order.

The Discussion also has a conventional shape worth following: restate the key finding, relate it to the literature you set up in the introduction, explain what might account for it, state limitations honestly, then draw out implications. Note that this closes the loop opened by Move 2 — a discussion that never returns to the gap it promised to fill leaves the paper structurally incomplete, even when every individual paragraph is sound.

The most frequent Discussion failure is a section that simply repeats the results in prose. If a sentence could be moved to Results without loss, it belongs there.

Claim strength

English-language academic writing hedges heavily, and the calibration is more specific than it first appears. The convention is not that you should be vague — it is that the strength of the claim must match the strength of the evidence, and the language marks that match precisely.

EvidenceAppropriate language
Correlational, single studywas associated with, suggests, appears to
Experimental, well-powered, replicatedshows, demonstrates, caused
Qualitative, small sampleparticipants described, in this context, these accounts indicate
Plausible but untested mechanismmay reflect, one possible explanation is

Two failure directions, both common. Overclaiming — writing proves where the design supports suggests, or generalising from one institution to a nation — is the faster route to rejection, because it signals that the author has not understood their own design's limits. Overhedging is subtler: stacked hedges (it may possibly be somewhat suggested that) drain the paper of any assertion, and reviewers report that they cannot identify the claim.

Aim for one hedge per claim, chosen deliberately. And state limitations as genuine constraints on interpretation rather than as ritual apology — a limitations paragraph that reads the sample was small and further research is needed tells a reviewer nothing they had not already noticed.

Fit, and the cover letter

A substantial share of desk rejections really are about fit, exactly as the note says. Before submitting:

The cover letter should be short and should do three things: name the contribution in one or two sentences, say why this journal specifically, and confirm the usual declarations. Editors read it to decide whether to send the paper out. It is not the place for a summary of the whole study.

Responding to reviewers

A revise-and-resubmit is good news that is routinely mishandled. The response document is a genre in its own right, and it has conventions:

Thank the reviewers once, briefly, at the top. Extended gratitude in every response reads as anxious rather than gracious.

On the unfairness of it

Everything above is advice for operating inside a system that is not neutral, and it would be dishonest to present it as though it were.

English-language international publishing imposes a real and unevenly distributed cost. Authors writing in an additional language spend time and money that Anglophone colleagues do not, on work of equal quality. Reviewers do sometimes conflate unfamiliar phrasing with unclear thinking. Regional scholarship of genuine value stays invisible because it is published in languages the indexes do not track. Naming this is not a complaint; it is an accurate description of the terrain, and it is the reason so much of the advice above is about convention rather than quality.

Two things follow. The first is practical: because the conventions are conventions, they are learnable, and learning them is the fastest available return on effort. Almost none of what this article describes is about how well you write English. It is about what the introduction is for, what the discussion is for, and how strongly to state a claim — and those can be learned deliberately, from the published record of your own target journals, which is free.

The second is that the distinction is worth keeping hold of. Meeting a convention competently is not the same as doing good research, and a paper that satisfies every structural expectation while contributing nothing has succeeded at the wrong thing. The conventions exist to make a contribution legible. They are worth learning for that reason, and worth keeping in their place for the same one.

Related: SSCI, SCIE, AHCI, ESCI: what your promotion committee actually counts covers choosing where to submit, and declaring AI use in a methods section covers disclosure if you use language tools during drafting.

Found an error in this article? Corrections are made openly and are welcome — see the editorial policy or write to ngoconglem@gmail.com.