| The words for One Health already exist in French. Keeping the concept intact — ecocentric, dynamic, credited — is the actual job. One Health was built to tear down walls. Between human medicine and veterinary science. Between plant health and the health of the rivers, soils and forests that hold everything else up. The whole point of the framework is that these were never separate in the first place. So there is a quiet irony in what happens when the concept crosses into French. The hardest part of the work is not finding the words. The words exist. The hardest part is refusing to let the walls go back up — one careful, defensible, invisible decision at a time. I have just finished translating a major institutional report on One Health and protected areas. I want to tell you what that job actually is, because it is not the job most people imagine when they picture translation. Nobody hands you a sentence and asks for its twin in another language. They hand you a way of seeing the world, and ask you to make it survive the crossing intact. |
The words Exist
The obvious question: how do you say One Health in French? This one is, mercifully, decided. Une seule santé is now the canonical rendering, consolidated across the WHO, the Convention on Biological Diversity, France’s Office français de la biodiversité, INRAE and the environment ministry. The older Un monde, une santé survives only in a few regional pockets; it leans on geography and connectivity, where the modern term leans on something more radical: the indivisibility of living things. The doctrine is clear, and a translator who reaches for anything else is being wrong. This discipline runs down into the operational vocabulary. Translating One Health material, I had to defend riposte, the WHO’s consecrated word for epidemic response, against a native reviewer’s instinct to soften it to réponse. That one word turned out to need an essay of its own. [→ Translating “response”]
Ecosystem Health
If the terminology were the whole game, the article would end here. It doesn’t, because underneath the settled word sits a fault line few name. The environmental pillar of One Health is routinely flattened into santé environnementale. And santé environnementale, in French institutional history, means something quietly different. It is anthropocentric. It studies how pollution, contaminants, degraded surroundings and more affect human health. It puts us back at the center of the frame. But One Health does not put us at the center. Its environmental component is ecocentric: the health of the ecosystem in itself, river and wetland and pathogen-buffering forest, whether or not a human is standing nearby to benefit. Translate ecosystem health as santé environnementale and you have done something subtly corrosive: you have smuggled the human back to the middle of the framework. Dictionaries and glossaries don’t flag this.
A River is Not A Number
Watch the same problem move into hydrology. Environmental flows has a precise French equivalent, débits environnementaux, used by the IUCN and the Ramsar Secretariat. The concept is dynamic: the quantity, quality and timing of water a living river needs, seasonal floods and natural low-water periods included. It is the rhythm of a watershed, not a number. French environmental law offers a tempting neighbor: débit écologique, or débit réservé — a fixed minimum, a regulatory floor, a single constant figure released downstream of a dam. Reach for that word because it sounds official, and you collapse a living rhythm into a flat line. Where planners have rendered environmental flows as a static minimum volume, they have legitimized artificial, constant-discharge regimes; silted riverbeds; blocked fish migrations; wetlands of international importance quietly drained of the variability that kept them alive. A mistranslation becomes an engineering specification, and the engineering specification kills the marsh.
The same discipline governs eaux intérieures (not eaux continentales, which carries a narrower hydrographic sense), bassin versant, régime hydrologique. Each choice is a small act of conceptual loyalty. None of them is a synonym swap.
When the wrong word becomes the law
Principle 15 of the 1992 Rio Declaration speaks, in English, of a precautionary approach; it’s deliberately soft. A flexible posture for managing scientific uncertainty. The French does not keep it quite so soft: des mesures de précaution, precautionary measures, things a state is expected to take. And in the French and European legal tradition built around the same idea, through the Maastricht Treaty, the loi Barnier of 1995, and ultimately the Charte de l’environnement, that posture acquired a far harder twin: le principe de précaution, a binding norm of constitutional rank. One precautionary idea; two very different degrees of obligation, sorted in part by which word a given text reaches for: approach, measures, principle.
Whatever you think of the outcome, notice the mechanism: the register is never neutral. At this altitude, the word a text chooses tracks — and helps settle — how tightly a state is actually bound. This is the stake that institutional clients feel in their stomachs even when they cannot name it. At the highest levels of environmental and health governance, the text is the policy. A concept-framework rendered loosely does not just read poorly. It reshapes what states are obligated to do.
Who answers for the words that land
There is one more integration the report demands, and it is the one most likely to be forgotten; not out of malice, but out of the ordinary friction of production.
When a translation of this weight is finished, it goes to layout. Text flows into InDesign; the file becomes an IDML; the designer threads it into columns, captions, credits. The ecologist is named. The hydrologist is named. The photographer is named, down to the caption. And somewhere in that handoff, with nobody intending harm, the one name that tends to evaporate is the translator’s.
Here French law is not ambiguous. A translation is an œuvre dérivée — a derivative work — and its author holds a moral right over it, including the droit de paternité, the right to be named. Under the Code de la propriété intellectuelle, that right is perpetual, inalienable and imprescriptible. The translator cannot sign it away even if they wanted to; an employment or commissioning contract does not extinguish it. ISO 17100 governs the process: the four-eyes revision, the qualified linguist. But it says nothing about public credit. The law says everything.
And the name does one thing the law does not require: it tells the reader what they are holding. An anonymous translation reaches them as if it were the original: neutral, authoritative, the binding text itself. A credited one admits what it is: a rendering by a named hand, one that made choices and can be asked to account for them. That admission quietly guards against the French being read as the source of the obligation rather than its faithful carrier.
In a digital workflow, before the final PDF is exported, the name must be carried into the document’s metadata: the dc:contributor field, Dublin Core’s element for agents other than the author, tagged with a translator role.
Conclusion
None of this is about finding words; the words exist. It is about refusing to let the walls go back up, one invisible decision at a time, so that what reaches the reader in French is the whole idea and not a quieter, lesser twin of it. That refusal is the work. It leaves no trace, which is exactly how you know it succeeded.