Explaining science in English: adapting scientific content for an international audience

Translating is not adapting

A translation restores the meaning; an adaptation rebuilds the text for a reader who has neither the same context nor the same expectations. The English-speaking reader of corporate scientific content expects a claim, evidence, then detail. The French text often starts with context, develops the method and concludes. Reversing the order is the first operation, before any question of vocabulary.

Second difference: length. For the same content, a well-written English text has 15 to 25% fewer words than its French version. If your translation is as long as the original, it probably has not been adapted.

Plain English rules applied to science

Plain English is not impoverished English, it is direct English. It applies very well to scientific content, provided you keep the exact technical terms and simplify everything else.

• One idea per sentence, twenty words on average, and an action verb rather than a nominalization (“we measured” rather than “the measurement was performed”).
• The active voice by default; the passive voice stays reserved for protocols where the agent does not matter.
• The exact technical term, defined once, then reused without synonyms: in scientific English, lexical variation reads as imprecision.
• French connectors (“indeed”, “moreover”, “it should be noted that”) disappear; the logical flow comes from the order of the sentences.
• Intensity adjectives (“innovative”, “revolutionary”) are replaced by the data that justifies them.

Figures, units and conventions

This is the most frequent source of errors, and the most visible to a scientific reader. The decimal separator becomes a point, the thousands separator a comma or a thin space depending on the style guide. The French “milliard” becomes “billion”, and the French “billion” becomes “trillion”. Dates are written out in full to avoid any ambiguity between American and British formats.

SI units are kept, but you check the conventions of the field: temperatures, volumes and some concentrations are expressed differently depending on the country and the journal. Finally, always distinguish percentages from percentage points, and state the calculation basis as soon as a relative figure appears.

Regulatory and medical vocabulary

Healthcare or medtech content must speak the vocabulary of the reader's authority. A clinical trial application is called a CTA in the European Union and an IND in the United States; a medical device is CE marked in Europe and goes through a 510(k) or a PMA with the FDA. Using the wrong term immediately signals that the text was not written for that market.

Trial phases, endpoints and indications are expressed with standardized terms; this is a case where the literal translation of a French term can be wrong. Review by someone who knows the regulation of the target market is essential.

Tone: evidence before adjectives

English-speaking readers, especially North Americans, are used to assertive corporate texts. But in scientific sectors, a claim without evidence backfires on the author. The rule we apply: every strong claim is followed, in the same sentence or the next, by the data, the publication or the result that supports it. The text stays confident, but every sentence is defensible.

Conversely, French rhetorical hedges (“it would seem that”, “could potentially”) weaken an English text beyond what the science requires. Keep the real uncertainty, expressed once and precisely, and remove the uncertainty of politeness.

The review that counts

Review by a native speaker is necessary, but insufficient if that person does not know the field. The right set-up combines two reviewers: a scientist in the field, who checks that the simplification betrays nothing, and an English-speaking writer, who checks the flow. When one person combines both skills, it is a rare asset worth keeping.

Finally, test the text on a real target reader before publishing it: an investor, a clinician or a partner from the target country. A question asked while reading is worth more than a correction after publication.

Pre-publication checklist

• The claim, evidence, detail order is respected in every section.
• The English text is at least 15% shorter than the French.
• Every technical term is defined once and reused identically.
• Figures, units, dates and percentages have been checked against the conventions of the target market.
• The regulatory vocabulary matches the reader's authority.
• Every strong claim is followed by its evidence.
• The text has been reviewed by a scientist in the field and by an English-speaking writer.

Going further