What "Endangered Language" Actually Measures: Why the First Question Should Not Be "How Many Speakers?"

A language with two million speakers can be more endangered than one with two thousand. The difference lies in who is learning it at home. UNESCO's framework for assessing language vitality, first published in 2003 and elaborated in the 2009 Atlas of the World's Languages in Danger, treats intergenerational transmission as the decisive factor. Speaker counts matter, but they matter as one element among nine. The framework asks whether children are acquiring the language in ordinary household use, not merely how many people can speak it at all.

English · 1186 words

Transmission, Not Just Headcount

UNESCO's 2003 Language Vitality and Endangerment document identifies transmission between generations as the most commonly used factor in evaluation. This is not a technical preference. It reflects a practical reality: a language that moves from grandparents to parents to children can sustain itself even with small numbers. A language that skips the parent generation or loses the household domain cannot sustain itself even with large numbers.

The Atlas categories make this explicit. A "Safe" language, in UNESCO's terms, is spoken by all generations with uninterrupted transmission. "Vulnerable" means most children speak it, but perhaps only in restricted domains like the home. "Definitely endangered" means children no longer learn it as mother tongue in the home. "Severely endangered" means the parent generation understands it but does not speak it to children or among themselves.

Notice what these labels measure. They track the age structure of transmission and the domains of use. They do not track population size directly. A language with 100,000 speakers and no child acquisition is classified as more threatened than a language with 2,000 speakers and ordinary household transmission. UNESCO's 2016 Towards World Atlas of Languages notes that 100,000 persons is treated as a minimum population threshold for ensuring transmission, but the threshold is not a guarantee. The 2016 data showed 2,124 of 2,304 endangered languages with population data had fewer than 100,000 speakers. The number describes a risk factor, not a status.

The Published Scales

The UNESCO vulnerability scale binds status to age and home use. "Safe" requires use by all ages. "Vulnerable" allows restriction to certain domains. "Definitely endangered" marks the point where home transmission breaks. "Severely endangered" describes a language that has become the preserve of grandparents, with parents as passive understanders rather than active speakers.

These are not speaker-count bands. They are transmission categories. A researcher applying the framework must determine whether parents speak the language to children, not merely count how many people claim some ability to speak it. The distinction matters for practical assessment. A community with 50,000 speakers but no children learning the language in the home faces a different situation than a community with 5,000 speakers where three generations converse daily.

Why Domains Matter

UNESCO's expert recommendations list nine factors for determining endangerment. Intergenerational transmission comes first. Absolute number of speakers appears second. The list continues through shifts in domains of use, community members' attitudes toward their own language, and response to new domains and media. Statistics Canada's 2014 overview of language vitality factors, developed for domestic application, includes the same elements: transmission, speaker numbers, age range, domain shifts, community attitudes, and institutional support.

The domain factor is easily misunderstood. It asks not just whether people can speak the language, but where they speak it. A language restricted to ceremonial occasions differs from one used for shopping, media, and schooling. The UNESCO framework explicitly includes "response to new domains and media" as an assessment factor. A language with few speakers but vigorous use in film, texting, and workplace communication faces a different trajectory than a language with more speakers but shrinking domains.

Why Two Languages With Different Speaker Counts Can Be Ranked Differently

The corrective example is built into the framework itself. Consider two hypothetical languages. Language A has 150,000 speakers, but the parent generation has shifted to a majority language for home use. Children understand the heritage language but respond in the dominant language. Language B has 8,000 speakers, but children acquire it as mother tongue, parents speak it among themselves, and it is used across daily activities.

Language A is "definitely endangered" or worse. Language B may be "vulnerable" or even "safe." The speaker count is higher for Language A. The vitality is higher for Language B. This is not a paradox. It is the intended application of the UNESCO framework. The 2003 document emphasizes transmission as the most commonly used factor precisely because it captures this distinction.

UNESCO's 2016 discussion of the 100,000-person threshold reinforces the point. The threshold is cited as a minimum for transmission, not as a definition of safety. Languages below the threshold are endangered in population terms. Languages above the threshold may still be endangered in transmission terms. The categories are designed to capture this double structure.

Why Labels Vary Across Documents

The same language may carry different labels in different publications. This is not necessarily error. UNESCO's own Atlas uses the six-level scale: Safe, Vulnerable, Definitely Endangered, Severely Endangered, Critically Endangered, Extinct. Related frameworks, including the Gradual Intergenerational Disruption (GID) scale developed for the University of Kansas, adapt these categories with slightly different thresholds and terminology. A language listed as "vulnerable" in one document and "definitely endangered" in another may reflect different assessment dates, different frameworks, or different weighting of the nine factors.

The 2009 Atlas definitions and the 2003 Language Vitality and Endangerment document provide the authoritative source for UNESCO's terms. Other publications may apply them differently. A reader encountering conflicting labels should check which framework was used, which year the assessment reflects, and whether the label describes transmission status or some composite of factors.

What the Labels Do Not Tell You

The UNESCO framework has limits that careful readers should note. The labels do not capture community preference. A factor in the framework asks about community members' attitudes toward their own language, but the final label is an external assessment, not a community self-evaluation. The labels do not distinguish domain loss from speaker loss. A language may retain speakers while losing domains, or lose speakers while retaining domains; the framework combines these into a single status.

The labels also do not capture revitalization that has not restored home transmission. A language with growing numbers of adult learners through school programs, but no children acquiring it from parents, remains endangered in UNESCO's terms. The framework measures intergenerational transmission, not institutional learning. This is a feature of the method, not a flaw. It keeps the focus on sustainable, natural language use. But it means the labels may understate revival efforts that have not yet reached the household.

The nine-factor list includes "availability of materials for language education and literacy" and "amount and quality of documentation." These support vitality. They do not substitute for it. A well-documented language with enthusiastic learners but no child acquisition remains endangered. The label reports the condition of transmission, not the condition of the archive.

The same language can carry different labels across documents because the labels come from different frameworks and snapshots. The reader should treat any label as a report of method and moment, not as a timeless essence.