Languages as infrastructure

Knowing a language is rarely a profession in itself, but it often becomes the infrastructure of a profession.

A language gives access to:

• education;
• documentation;
• the labour market;
• clients;
• a professional community;
• another country;
• sources and professional settings where machine translation loses important shades of meaning.

Machine translation has sharply cut the price of first access to a foreign text. You can now grasp the general sense of a document, a letter or a conversation without the command of the language you once needed.

But translation doesn’t replace the language where ambiguity, trust, cultural norms, professional terms or legal consequences are involved. AI can get you through the door; it doesn’t always notice that it carried a different object out of the room.

English matters especially in many technical, scientific and international fields. But how necessary it is depends on the profession and the setting. For someone working in a local context, another language may matter more.

The number of languages isn’t a universal standard either. One new language at a high level is more use than three left in the state of “I can order coffee if the menu cooperates”.

The choice of language is best tied to a real function:

• moving abroad;
• work;
• family;
• the region;
• access to knowledge;
• cultural interest.

Professional value comes not from the language on its own but from the combination:

language plus a speciality.

An engineer, doctor, lawyer or salesperson who can work in another language environment gains an extra market. A translator, in turn, also needs subject knowledge once the text is harder than a tourist brochure.

A language widens the geography of your choices. But moving also depends on documents, recognition of your diploma, money, family and demand. It opens a door. Housing behind that door is not included.

The power of a language shows up in the negotiating room, not in a level test. While the sides are exchanging positions, translation is enough. But the difference between “we will consider it” and “we will probably consider it” decides whether there is a deal, and that is the first thing translation smooths over — out of politeness, or because it doesn’t fit into one word. Same in documentation: the description comes through correctly, the modality is lost, and a requirement reads as a recommendation.

The machine took understanding as far as the general sense and stopped exactly before the shades of meaning that obligations rest on. A language is worth learning where the price of one such shade is higher than the price of learning it, and not worth learning where the description is enough.