Is Hakka an Endangered Language

Hakka is not dead. But the official figure of around 2.5 million Hakka people in Taiwan counts self-identified ethnicity, not functional fluency. The real question is who still speaks it at home. Fluent speakers under 40 are a much smaller fraction.

The gap is generational. Grandparents speak Hakka. Parents code-mix with Mandarin. Children respond in Mandarin. That is the chain of everyday use breaking, and it is the single most important signal in any language vitality classification.

The Expanded Graded Intergenerational Disruption Scale gives Hakka a specific place. EGIDS runs from 0 for international to 10 for extinct. Taiwan Hakka sits in the 6b to 7 range depending on dialect and region. That means it is threatened: still spoken by child-bearing generations, but no longer learned by most children as a first language.

The test is whether a three-year-old hears it in the kitchen. A dialect with Hakka TV programming, radio broadcasts, and certification exams but almost no toddler speakers is in a different condition from one with both. Sixian and Hailu have media presence but weak home use. Dabu and Raoping have almost no media footprint at all.

Treating Hakka as one language is the first mistake a visitor or heritage researcher makes. Taiwan has at least four distinct varieties with their own tone inventories and sandhi rules. Hakka language vitality Taiwan is not one number.

Sixian: The Standard-Bearer

Sixian, the default for Hakka TV and most education, has six citation tones and is the most standardised. It has the strongest rural base, particularly in Miaoli.

Hailu: More Tones, Fewer Voices

Hailu has seven tones and a stronger literary-layer retention, but fewer speakers. Home use is patchy, especially in urban areas like Taipei where the chain of everyday use broke two generations ago.

Dabu And Raoping: The Small-Pocket Varieties

Dabu and Raoping survive in small pockets, mostly among older speakers in Taichung and Hsinchu/Miaoli. If a variety has no home use, media presence does not keep it alive. It keeps it documented.

The Hakka Basic Act of 2018 made Hakka a national language. That is a legal status, not a demographic recovery. The Act created a policy architecture: the Hakka Affairs Council gained authority over language certification, broadcasting mandates, and designated Hakka cultural regions. Hakka TV and FORMOSA Hakka Radio operate under its terms.

A designated Hakka cultural region gets funding, signage, language requirements, and tourism infrastructure. What it does not automatically get is fluent Hakka speakers in shops and on the street. The gap between policy designation and lived linguistic reality is the core finding. You can walk through a designated region and hear almost no Hakka outside a heritage performance.

Hakka language revitalisation policy in Taiwan has three main levers: certification, broadcasting, and cultural regions. The Hakka language certification exam, run by the Hakka Affairs Council, tests across multiple dialects and shapes education and media hiring. That creates jobs. It does not create households where Hakka is the language of bedtime stories.

Broadcasting matters more than certification for passive retention. Hakka TV subtitles everything, which means a monolingual Mandarin speaker can watch. That is good for awareness. It is bad for acquisition. The failure case is the code-mixed Mandarin of Hakka social media, where young speakers write in romanisation or ad-hoc characters and default to Mandarin for anything abstract.

Ask any Hakka policy officer what keeps them awake and the answer is the broken chain from grandparent to parent to child. The EGIDS score collapses if that link fails. In most urban Hakka families in Taiwan, it failed two generations ago.

The intergenerational score is a blunt instrument. It counts whether children still learn the language at home as a first language. For Sixian in rural Miaoli, the score is higher. For Hailu in Taipei, the score is near zero. The ethnicity-vs-fluency gap is widest among people under 40, who may mark Hakka on a census form but cannot order a bowl of noodles in the dialect.

Hakka TV broadcasts in standardised Sixian and Hailu. That is a policy choice, not a reflection of speaker numbers. Dabu and Raoping have no dedicated broadcast slots. Dabu is kept alive mainly by community radio and elderly speakers in Taichung. Raoping survives in small pockets in Hsinchu and Miaoli.

What does media presence without home use mean for vitality classification? It means the dialect is documented, standardised, and performable, but not living. A language that exists only on air is a museum exhibit. A language that exists only in a grandmother's kitchen is endangered but alive. Most Taiwan Hakka dialects are between those two states.

Taiwan's six Hakka cultural regions have distinct administrative boundaries and funding streams. The Liudui region in the south, Miaoli, Hsinchu, Taichung, Hualien-Taitung, and the central region. Designation brings signage, festival funding, and heritage programming. What it does not bring is a generation of children who speak Hakka at home.

The confusion pair here is policy designation status versus language vitality. A visitor who assumes a designated Hakka cultural region has fluent Hakka speakers in shops and on the street is making the language vitality overestimation failure. The designation marks a historical settlement pattern, not a current speech community. Sixian in Miaoli still has some home use. Hailu in Hsinchu is patchier. The southern Liudui region has the strongest ritual calendar but uneven daily use.

The quickest test is not the certification exam. It is who speaks Hakka to a child in a shop. If the answer is no one, the dialect is in decline regardless of its EGIDS number. The second test is code-mixed Mandarin. When Hakka speakers under 40 switch to Mandarin for numbers, dates, and abstract nouns, the language has lost the domains that keep it necessary.

Do not plan fieldwork or a heritage visit on the assumption that a designated region is a Hakka-speaking environment. Official surveys report around 2.5 million self-identified Hakka people in Taiwan. The real fluent speaker count under 40 is a fraction of that. Check who is actually using it before you book.

The single thing that most often goes wrong here is mistaking policy for revival. The Hakka Basic Act of 2018 was a legal milestone. It did not restore the chain of everyday use. A national language designation on paper does not make Hakka the language of the playground or the dinner table.

Hakka in Taiwan has a future only if the home chain is rebuilt. That is not a policy problem. It is a household problem. No amount of broadcasting mandates or designated cultural regions can solve it.