Globalization and language death: a planet of 7,000 tongues, half of them going

There are roughly 7,000 languages spoken on Earth in 2025. Linguists estimate, based on documented intergenerational transmission patterns, that between 50% and 90% of those languages will have no remaining native speakers by 2100 (Krauss, 1992; Simons & Lewis, 2013). The trajectory is steeper than past centuries because the structural pressures producing it — mass schooling in official languages, urban migration, monolingual media — have intensified.

The headline story tends to focus on Indigenous languages with small remaining speaker bases. The more demographically consequential story is about immigrant languages in second-generation contexts. The mechanism is similar; the scale is much larger.

1. How a language dies

Languages don't usually die by mass forgetting. They die by not being transmitted to the next generation. Parents speak Language A; children grow up speaking Language B at school, with peers, in media; children become functional bilinguals in childhood, monolingual adults; their children grow up in Language B only.

This pattern — intergenerational language shift — is the dominant mechanism. It typically completes in three generations. Joshua Fishman's Graded Intergenerational Disruption Scale (1991) provides the standard framework: full vitality at level 1, ceremonial-only use at level 6, no living speakers at level 8.

2. The immigrant case

In immigrant families, the same process runs faster because of selection pressure. Parents arrive speaking Language A. Children acquire Language B from the local school and become socially dominant in it. Grandchildren typically have only passive comprehension of Language A; great-grandchildren, none.

Studies of Spanish in the United States, Turkish in Germany, and Punjabi in Canada all show the same trajectory at roughly the same speed: full fluency at generation 1, partial at generation 2, vestigial at generation 3 (Portes & Rumbaut, 2006).

3. What's lost

The framing "what's lost when a language dies" is contested. Several losses are real:

Specific knowledge. Indigenous languages often encode environmental, medical, and ecological knowledge that translates imperfectly. Botanical and astronomical detail is sometimes lost.

Cognitive variety. Different languages structure aspects of thought differently — counting systems, color categories, evidentiality marking. These don't determine thought but they shape habits of attention (Lucy, 1997).

Cultural continuity. Songs, ceremonies, jokes, and oral history bound to a specific language often don't survive translation into a dominant language.

The argument that "language death is just like species evolution, natural and inevitable" misreads both processes. Most language deaths are not driven by natural shift but by specific policy choices — school language, media access, official-language laws. Different choices yield different outcomes.

4. The reverse cases

A small number of languages are actively reviving from near-death:

• Hebrew: revived from near-zero native speakers in the 19th century to several million today
• Welsh: declined for 200 years, stabilized since 1990s policy interventions
• Māori: similar trajectory — substantial revival since the 1980s through education policy and media support
• Hawaiian: lost most native speakers in the 20th century, revived through immersion schools

These cases show that language shift is reversible but expensive. Reversal requires substantial institutional investment over decades, and it has to be sustained politically across multiple election cycles.

5. The practical implication for individual families

For an immigrant family wanting to maintain a heritage language across generations, the literature is reasonably clear:

• Children need substantial daily exposure to the heritage language to maintain it (roughly 30%+ of waking hours through early childhood)
• The exposure needs to be embedded in social and emotional contexts that the child cares about (relatives, peers in the language, media the child actually wants)
• Reading and writing instruction in the heritage language matters; oral-only transmission decays faster
• Mixed-language households need explicit norms; "we speak Spanish at home" works better than implicit defaults that drift toward the dominant language

The work of maintaining a heritage language is substantial and largely invisible. The cost of not doing it is also substantial and largely invisible. Most families do not realize the loss is happening until it's complete.