Translation, and What Gets Lost

Multi-country organisations translate their policies, which is right: a document nobody can read in their own language is not a policy, it is a filing requirement.

The workflow in “Translation, and What Gets Lost” becomes more dependable when scheduled work, actual time and later corrections can be distinguished. For teams exploring how to detect mouse jigglers, learn more on this page can provide operational context, while policy ownership, employee explanation and consequential decisions remain with accountable people.

The difficulty is that working time documents are full of terms that look ordinary and are technical. A translator rendering them will reach for the nearest local equivalent, and in employment law the nearest equivalent is frequently a term with its own statutory definition.

For an independent reference relevant to “Translation, and What Gets Lost”, consult the SHRM workplace-investigation resources. Use it to challenge assumptions about working time, privacy, recordkeeping and exception handling against the organisation’s real operating model.

The terms that cause trouble

Working time itself, which has a specific meaning in several regimes and a general one in ordinary speech. Rest period. Night worker. Overtime, which may or may not be a defined category locally. Standby and on-call, which are distinguished in some languages and not in others.

Also: reference period, averaging, opt-out, compensatory rest. Each of these is either a term of art locally or a phrase that will be read as one.

Where the organisation's document uses a term loosely and the translation renders it as the local term of art, the translated policy says something more specific than the original intended.

The two-pass approach

First, a normal translation. Second, a review by somebody in the country who understands both the language and the employment context — usually local HR or the local adviser, not a linguist.

The second pass is looking for one thing: places where the translation has landed on a defined local term, and whether that is what the organisation meant. It takes an hour per document and it is the step that gets skipped.

Brief the reviewer on what to look for, because asked to review a translation they will check the language.

Marking terms that must not be localised

Where the organisation uses a term in its own internal sense — its own standard, its own category of hours — mark it in the source document as an internal term and have it carried through with a definition rather than translated to a local equivalent.

A short glossary at the end of the global policy, listing the terms and their intended meanings, makes this possible and is useful in the source language too.

It also surfaces, during drafting, the places where the organisation has been using a word loosely. That is worth finding before the document is translated into six languages.

Which version governs

State it. Usually the source language version governs and the translations are provided for convenience, and that sentence belongs in every translation.

Where local law requires a document to be provided in the local language, which it does in several places, the position may be different and is worth checking rather than assuming. A policy that is only available in a language the workforce does not read may be ineffective locally whatever the governing-version clause says.

This is one of the few questions in this collection where the answer genuinely differs and matters: ask it per country.

Keeping translations in step

The failure mode is not a bad translation; it is a good translation of an old version. The source document is updated, the translations are not, and for a year or two the organisation has two policies.

Attach the translations to the version number. A change to the source triggers a retranslation task with the same deadline, and a translation whose version number does not match the current source is marked as out of date wherever it is published.

That is a small administrative discipline and it prevents the state where nobody can say which version a given country is operating under.

The annexes, which are usually worse

Annexes are often drafted locally in the local language and summarised into the central set in the source language.

The summary is then what the centre reads and relies on, and it loses detail. Where an annex is drafted locally, keep the local version as the governing one and treat the central summary explicitly as a summary.

Labelling it that way is honest and it stops the centre answering questions from a document that was never intended to be relied on.

The practical minimum

A glossary in the source document. A second-pass review by somebody local with an employment background. A governing-version sentence. And translations tied to version numbers.

Four things, none of which costs much, and together they prevent most of the ways a multilingual policy set quietly becomes several different policies.

Who the reader is in each language

The global policy is read by managers; the annex is read by managers and sometimes by employees and representatives. Those are different audiences with different needs from a translation.

Where a document will be read by a works council or an inspector, the translation has to be good enough to be relied on formally, which is a higher bar than internal comprehension.

Decide which documents are in that category per country and treat them differently: a certified or locally reviewed translation for those, a working translation for the rest.