A Generation Earlier

A substantial share of the world’s roughly seven thousand languages have few enough speakers, and few enough young ones, that they are unlikely to be spoken in a century. That is a linguistic fact with a travel dimension: in a large number of destinations, the language you hear in the market is not the language on the signs, and the relationship between the two is often the most interesting political story in the region.

How a language actually dies

Not by decree, usually, and not by the last speaker dying — that is the end of a process, not the process. It happens through intergenerational transmission stopping: a generation that speaks the language to each other but not to its children, because the children’s prospects depend on the dominant language of school and work and the parents make a rational decision on their behalf.

Once that break occurs, the language has perhaps two generations left regardless of how many speakers it currently has. A language with fifty thousand speakers all over sixty is in a worse position than one with five thousand speakers of whom a thousand are children. Speaker counts, which are what get reported, are a poor indicator; the age profile is the whole story.

The pressures are consistent across continents: state-language schooling, internal migration to cities where the dominant language is the medium of everything, media and now algorithmic platforms in the dominant language, and — historically and in some places within living memory — deliberate suppression, including the punishment of children for speaking their own language at school. That last is why revitalisation is so often framed as recovery rather than preservation, and why the politics around it are rarely calm.

A language does not die when the last speaker does. It dies a generation earlier, on the day parents stop using it with children.

What revival looks like when it works

The successes share a structure. Immersion schooling, starting with the youngest children — nursery and early primary conducted entirely in the language — which is the intervention that most reliably produces new young speakers, because it puts the language back into the age group where it was lost. School speakers are not yet home transmission, and the honest reading of the successes is that the home remains the fragile link even where the classrooms are full.

Official status that puts the language in courts, forms, health services and signage, so that speaking it is a public act rather than exclusively a domestic one. Media: radio and broadcasting historically, and now streaming, subtitling and social platforms, which have been unexpectedly effective for languages whose young speakers are geographically dispersed. And a domain in which the language is genuinely necessary rather than symbolic — a workplace, an institution, a service — because a language that is never required becomes a hobby.

Hebrew is the outlier case usually cited and the least generalisable, since it involved a state building an entire education system, bureaucracy and army around it. The more instructive examples are the regional and indigenous languages in Europe, the Pacific and the Americas that have moved from steep decline to real growth in young speakers through schooling and status without a state of their own — within a single generation in the best cases, and with home transmission still the unfinished part everywhere.

What a traveller will actually encounter

Diglossia, usually: one language for official business and strangers, another for home and the market, with speakers moving between them mid-sentence without noticing. A visitor who only ever hears the first will conclude the second is not spoken, which is why so many travel accounts describe monolingual places that are nothing of the kind.

You will also encounter the politics. In a great many regions the language question is the sharpest live dispute there is — about schooling, about signage, about broadcasting money, about whether a variety is a language or a dialect, which is a political classification wearing a linguistic coat. A visitor asking about it will get a real answer and should be ready for a real one.

The traveller’s part

Small and not nothing. Learning the greeting in the local language rather than the state language is almost always noticed, and usually warmly — it signals that you know there is a distinction, which is most of what is being asked. In the tensest bilingual regions, read the room first; an outsider’s pointed language choice can also be read as taking a side. Buying the books, the music and the media in the language puts money into the only market it has. Attending the festival, the service or the performance where it is used.

And asking which language you are hearing, which is a better question than most travellers ask and is almost always welcome. What to avoid is treating it as an exhibit: a language with a hundred speakers is not a curiosity for a list, it is the working speech of a community that has usually had to fight to keep it, and the fight is generally still going on.