Translating intelligence: bilingual analysis as a default

Half of institutional legal work is bilingual. A note on what that means for the analysis layer.

In Geneva, Zurich, Brussels, and most places where cross-border capital is documented, the working language of a contract is rarely the working language of everyone reading it. A German-law agreement is reviewed by a French-speaking principal. An English master agreement governs a relationship negotiated in three tongues. Translation is not an occasional need; it is the ambient condition.

The usual workflow treats language as a separate step: send the document to a translation agency, wait, then analyse the result. Two problems follow. The analysis is only ever as good as the translation it inherited, and the round trip adds a hand-off precisely where legal nuance is most fragile — in idiom, in clause architecture, in the words a jurisdiction gives specific weight.

There is a cleaner default. Perform the analysis in the source language, where the legal precision lives, and deliver both the translation and the reasoning in the target language the reader actually needs. One process instead of two hand-offs. The analysis never inherits a translator's guess about what a clause was trying to do.

This matters most where a single word carries legal freight — where a term of art in one system has no clean equivalent in another, and a fluent translation can quietly erase the distinction. Keeping the analysis attached to the original text is how the distinction survives the crossing.

Bilingual by default is not a feature bolted onto a monolingual tool. It is an assumption about how institutional work is actually done, built in from the start.