Sino-Tibetan.

OpeningThe other big family.

Sino-Tibetan is the second-largest language family on earth by speakers — second only to Indo-European — and on present evidence the oldest continuously documented one.

The family includes Mandarin and the other Sinitic varieties (~1.3 billion speakers), Tibetan and Burmese (each over 30 million), and roughly four hundred mostly small languages spoken across the eastern Himalayas, Yunnan, Northeast India, Burma, Nepal, and Bhutan. Most of these languages have very few speakers; many are endangered.

This deck covers the family's structure, the question of its homeland and dispersal, the tonal systems that distinguish many of its members, the Chinese script and the 20th-century reforms, the Tibetan and Burmese literary traditions, and the current state of comparative reconstruction.

Chapter ISinitic and Tibeto-Burman.

The high-level division of Sino-Tibetan recognised in most modern classifications is between Sinitic — the Chinese topolects — and Tibeto-Burman, which contains everything else. Whether Sinitic is itself one branch among many in Tibeto-Burman, or stands as a coordinate primary branch, is contested and has shifted with each generation of researchers.

The major Tibeto-Burman subgroups: Bodish (Tibetan and Dzongkha and many smaller languages of Tibet, Bhutan, and Nepal), Lolo-Burmese (Burmese, Yi, Lisu, Naxi), Karenic (Karen languages of Burma and Thailand), Tani (the Apatani and related languages of Arunachal Pradesh), Naga, Kuki-Chin, Bodo-Garo, and a long tail of small subgroups in the eastern Himalayas.

The unity of the family was not obvious for a long time. The systematic comparative work began in the late 19th century with Ernst Kuhn, the Tibetologist Berthold Laufer, and Robert Shafer's Introduction to Sino-Tibetan (1966–1974). The family is now firmly established but the internal classification remains under active debate.

Chapter IIStandard Chinese.

Modern Standard Mandarin (普通话, Pǔtōnghuà, "common speech") is spoken natively by roughly 920 million people — more than any other language on earth — and as a second language by hundreds of millions more. It is the official language of the People's Republic of China, the Republic of China (Taiwan, where it is called Guóyǔ, 國語), and Singapore (where it is called Huáyǔ, 華語), and one of the six official languages of the United Nations.

The standard is based on the dialect of Beijing. Promoting it as a national medium has been a project of every Chinese government since the late Qing — the spread is rapid but incomplete. Hundreds of millions of Chinese citizens speak Mandarin only as a second language, with their first being one of the regional varieties.

Standard Mandarin has a relatively simple phonology — about twenty-one initial consonants, four tones (plus a neutral tone), and a limited set of syllable shapes. The language compensates for its narrow phonological inventory with extensive compounding: most modern words are two-syllable compounds rather than the monosyllables of classical Chinese.

Chapter IIIThe seven (or ten) Chineses.

The Chinese topolects — the term fāngyán (方言) is sometimes translated "dialect," but the divergence is at the scale of distinct languages — are conventionally grouped into seven or ten major divisions. Mandarin (Northern, by far the largest), Wu (Shanghainese, Suzhou), Cantonese / Yue (Hong Kong, Guangzhou), Min (Hokkien, Teochew, Taiwanese), Hakka, Xiang (Hunanese), Gan (Jiangxi), and the smaller divisions Jin, Hui, Pinghua.

The mutual intelligibility is low. A Mandarin speaker and a Cantonese speaker hearing each other speak rarely understand more than a few isolated words. The shared writing system masks the divergence — the same characters represent both varieties — but spoken Mandarin and spoken Cantonese are as distinct as Spanish and French.

The question of whether to call them dialects or languages is partly political. The PRC official line treats them as dialects of a unitary Chinese language; sociolinguists outside that frame more often call them separate languages. Either way, the situation is unique: a writing system shared across distinct spoken languages by a single literate culture for over two millennia.

Chapter IVThe pitch system.

Tone — pitch contour distinguishing otherwise identical syllables — is a defining feature of most Sinitic and Lolo-Burmese languages. Mandarin has four tones plus a neutral; Cantonese has six (or nine, counting the entering-tone categories); Hokkien has seven or eight; Vietnamese (which is Mon-Khmer, not Sino-Tibetan, but has been heavily influenced by Chinese) has six.

Tones evolved historically. Old Chinese (~1000 BCE) had no tones; what became tones were originally final consonants whose loss conditioned pitch differences. The reconstructions of Bernhard Karlgren in the early 20th century and William Baxter's A Handbook of Old Chinese Phonology (1992) work backward to a non-tonal proto-language with much richer consonant clusters.

The same process has been observed in real time in some Tibeto-Burman languages and in non-Sino-Tibetan tonal languages of Southeast Asia. Tone is not innate to the family; it is a recurrent areal phenomenon.

Chapter VReconstructing Karlgren and after.

Bernhard Karlgren's Études sur la phonologie chinoise (1915–1926) was the first systematic reconstruction of Old Chinese phonology. Karlgren worked from the Middle Chinese of the Qieyun rhyme dictionary (601 CE), the rhyme practice of the Shijing (Book of Odes, ~1000–600 BCE), and the phonetic components of Chinese characters. His reconstruction stood for half a century.

The modern reconstructions of Li Fang-Kuei (1971), William Baxter (1992), and Baxter and Sagart (2014) have substantially revised Karlgren — proposing a richer system of consonant clusters, prefixes, and infixes that explain morphological alternations Karlgren could not. Old Chinese on the Baxter-Sagart system looks much less like modern Mandarin and much more like a typical Tibeto-Burman language.

The reconstruction matters because it brings Sinitic into typological alignment with the rest of the family. The features that make modern Mandarin unusual — the analytic syntax, the heavy use of tone, the compound-heavy vocabulary — turn out to be late innovations rather than ancient features. The Sinitic-Tibeto-Burman common ancestor was probably not very different from Tibetan or Burmese.

Chapter VIThe script.

The Chinese script in continuous use since the late Shang (~1200 BCE oracle bones) is the most-used logographic system on earth and the only one in everyday daily use. About 3,000–4,000 characters cover full literacy in modern Chinese; some 7,000 are in regular educated use; the comprehensive Hanyu Da Zidian (1986) lists 54,678.

The structural principle is the radical-phonetic compound. Most characters consist of a semantic radical (one of about 200 conventional shapes signalling category — water, hand, wood, mouth, heart) and a phonetic component (signalling pronunciation). 河 ("river") has the water radical 氵 plus the phonetic 可 (read kě; the character is read hé).

The system has costs and benefits. Costs: high learning load, especially for second-language learners; difficulty inputting in pre-Pinyin computer systems; visual complexity for the visually impaired. Benefits: extreme compactness; visual disambiguation of homophones (Chinese has many); cross-dialect intelligibility (the same character is read differently in Mandarin, Cantonese, Hokkien — but is the same character).

Chapter VIIThe script reform.

Twentieth-century debate over whether to abolish or simplify Chinese characters was ferocious. The May Fourth movement (1919) included serious proposals to romanise. Lu Xun, the great early-20th-century writer, wrote that Chinese characters "must die or China will die."

The 1956 PRC simplified character reform reduced the stroke count of about 2,000 characters by replacing them with simpler historically-attested forms or new abbreviations. 國 became 国, 學 became 学, 漢 became 汉. A second-round 1977 reform was rolled back; the 1956 reform stands.

Taiwan, Hong Kong, and most overseas Chinese communities continued with the older traditional characters. The split is a matter of political identity as well as orthographic preference. A literate user of either system can usually read the other with some effort. Pinyin (1958) — the romanisation system used for primary-school instruction and for input methods on computers — is now taught universally in PRC schools as a complement to characters, not a replacement.

Chapter VIIIThe roof of the world.

Tibetan in the broad sense covers a continuum of related languages spoken from Ladakh through the Tibetan plateau to Bhutan and parts of Nepal. The literary language — Classical Tibetan — was developed in the 7th century under King Songtsen Gampo, who sent his minister Thonmi Sambhota to India to design a script (modelled on the Brahmi-derived scripts of north India) and a grammar.

The Tibetan script has been frozen since the 9th century. Modern spoken Tibetan has changed substantially — many of the consonant clusters are simplified, tone has emerged in central Tibetan dialects — but the spelling preserves the older pronunciation. The result: written Tibetan is often called the most conservative attestation of any modern Sino-Tibetan language, like written English compared to spoken English in scale of mismatch.

The literary tradition is overwhelmingly Buddhist: the Kangyur (the translated word of the Buddha, ~108 volumes) and Tengyur (commentaries, ~225 volumes) are among the largest scriptural collections of any religion. Tibetan also preserves substantial pre-Buddhist Bön traditions, indigenous historiography, medicine (the Gyüshi corpus), and a rich poetic tradition.

Chapter IXThe Lolo-Burmese branch.

Burmese is the largest member of the Lolo-Burmese subgroup of Tibeto-Burman, with about 33 million native speakers. The language is the official language of Myanmar; the Burmese script is a Brahmic abugida derived ultimately from a 5th–6th-century Pyu and Mon predecessor.

The Lolo-Burmese subgroup includes Yi (also called Lolo, ~9 million speakers in Yunnan), Lisu, Naxi (with the unique Dongba pictographic script), Akha, and dozens of smaller languages. The subgroup is well-defined and shares characteristic developments — for instance, the loss of certain Old Chinese-style consonant clusters and the development of complex tonal systems.

The Burmese literary tradition begins with the 12th-century Pagan-period inscriptions. The classical literature is heavily Buddhist (Theravāda, in contrast to Tibet's Vajrayāna), influenced by Pali and Sanskrit through scriptural translation. Modern Burmese diglossia is significant: the literary language differs substantially from the colloquial spoken one, in vocabulary, particle inventory, and to some extent grammar.

Chapter XThe Yue language.

Cantonese — Yue Chinese — is spoken by perhaps 80 million people, mostly in Guangdong, Guangxi, Hong Kong, Macau, and the global diaspora. It is the most-used non-Mandarin Chinese variety in international contexts and the only Chinese variety with substantial standardised written colloquial use.

Cantonese is more conservative than Mandarin in several respects. It retains final consonants (-p, -t, -k, -m) lost in Mandarin; it has six (or nine) tones to Mandarin's four; its vocabulary preserves Old Chinese forms displaced in Mandarin. Tang-dynasty poetry — written in a Middle Chinese close to Cantonese in some features — often rhymes better in Cantonese than in Mandarin reading.

The language has a written colloquial tradition, especially strong in Hong Kong, where popular publications, advertising, and informal text use distinctly Cantonese characters and constructions (the character 嘅 for the possessive particle, 咗 for the perfective marker, 喺 for "to be at" — all unattested in Mandarin writing). The political pressure on Cantonese under PRC media policy has been substantial; the language's future in mainland Guangdong is uncertain.

Chapter XIThe southeastern coast.

The Min languages of Fujian and the diaspora — Hokkien (the largest, also called Southern Min or Taiwanese), Teochew, Eastern Min (Fuzhounese), Hainanese — are the most divergent of the Sinitic varieties, conventionally treated as having split off earliest from the rest. They preserve features of pre-Middle Chinese phonology that other Chinese languages have lost.

Hokkien (Southern Min) is the dominant language of Taiwan's native population, the lingua franca of much of the southern Chinese diaspora (Singapore, Malaysia, Philippines), and the household language of perhaps 50 million speakers. Its written form has historically been irregular — it is rarely taught as a written language in schools — though the Taiwanese ROC government has supported standardised orthography.

The Min languages have eight tones in their conservative varieties and complex tone-sandhi rules — automatic tone changes when syllables combine. The result is that nearly every two-syllable word is pronounced with different surface tones than its citation forms predict. Min phonology is a permanent challenge for learners.

Chapter XIIShanghai and the Yangtze delta.

The Wu languages of the Yangtze River delta — Shanghainese is the most prominent, with Suzhounese, Hangzhounese, Ningbonese, Wenzhounese alongside — count perhaps 80 million speakers. Shanghainese in particular has been dramatically reshaped by the city's rapid urbanisation and immigration.

Wu is distinguished by retaining a three-way voicing contrast in initial consonants — voiceless aspirated, voiceless unaspirated, voiced — that other Chinese languages have largely lost. This is the kind of conservative feature that lets historical linguists work backward to Old Chinese.

The Wu literary tradition — centred on Suzhou — was one of the great vernacular literatures of late imperial China. The 17th–19th-century Suzhou ballads (tanci), the Wu-language sections of the great novel Haishang Hua (1894), the Shanghainese-language opera tradition (Huju). Modern Shanghainese is under heavy pressure from Mandarin in primary education and media.

Chapter XIIIWhere ST began.

The Sino-Tibetan homeland question is less heated than the Indo-European one but no less open. Two main proposals.

The Yellow River hypothesis. Proto-Sino-Tibetan was spoken in the upper Yellow River basin (modern Gansu, Shaanxi) by Neolithic millet-farmers, perhaps 7,000–6,000 BP. The expansion was westward and southward into the Himalayas and Southeast Asia, and the Sinitic branch developed in place. This is the older view.

The Yunnan-Sichuan hypothesis. Proto-Sino-Tibetan was spoken in the upland regions of southwest China and northeast India — the area of present-day Tibeto-Burman maximum diversity. Sinitic moved north; the rest stayed roughly in place. This view is favoured by some recent phylogenetic work.

The 2019 paper by Zhang, Yan, and others (Nature) used Bayesian methods on a 109-language dataset and dated the family at ~7,200 years before present, located in northern China — supporting the older Yellow River view. The 2020 paper by Sagart and others reached similar conclusions. The current consensus, such as it is, leans toward the northern homeland.

Chapter XIVThe eastern Himalayan diversity.

The greatest concentration of Tibeto-Burman linguistic diversity is in northeast India and adjacent Burma. The Naga languages alone — perhaps thirty distinct languages spoken in Nagaland, Manipur, parts of Arunachal Pradesh — have a combined speaker population under three million but exhibit phonological and grammatical diversity comparable to the entirety of Romance.

The Kuki-Chin (or Kukish) subgroup, spoken in Mizoram, Manipur, and the Chin Hills of Burma, has been a particular focus of Sino-Tibetan reconstruction because its languages preserve features of the proto-language lost elsewhere. The Tani subgroup of central Arunachal Pradesh is similarly conservative and similarly under-documented.

Most of these languages have fewer than 50,000 speakers and many fewer than 10,000. The combination of small populations, mountainous terrain, postcolonial state pressure (English in India, Burmese in Myanmar), and limited literary tradition makes them among the most endangered language groups in Asia.

Chapter XVThe independent branch.

The Karen languages — Sgaw Karen, Pwo Karen, and several smaller varieties spoken in eastern Myanmar and northern Thailand — are unusual within Tibeto-Burman. Their basic word order is SVO (subject-verb-object), like Chinese; most Tibeto-Burman is SOV, like Tibetan and Burmese. Their morphology is unusually analytic.

The position of Karenic within Tibeto-Burman has been debated for over a century. Some classifications make Karen a primary branch coordinate with all the rest of Tibeto-Burman; others place it within a more central position. The distinctive features may be areal influence from the Tai-Kadai and Mon-Khmer languages of mainland Southeast Asia.

The Karen are a politically distinct ethnic group in Myanmar and have been at intermittent civil war with the central Burmese state for most of the post-1948 period. The Karen National Union is one of the longest-running insurgent movements in Asia. Karen-language Christianity (Sgaw Karen-language Bible from American Baptist missionaries, 19th century) is an important factor in modern Karen identity.

Chapter XVIRomanisation.

Modern Hanyu Pinyin (汉语拼音), adopted by the PRC in 1958 and the international ISO standard for Chinese romanisation since 1982, is a Latin-script system designed by Zhou Youguang and a committee of Chinese linguists. It uses standard Latin letters with diacritics for the four tones and conventional spellings for sounds (q, x, c, z) that may surprise English-speaking learners.

Pinyin replaced the older Wade-Giles system (developed by Thomas Wade in the 1860s and refined by Herbert Giles in the 1890s) — Wade-Giles spellings still survive in some pre-1980 transliterations: Peking (Beijing), Mao Tse-tung (Mao Zedong), Lao Tzu (Laozi), Chiang Kai-shek (Jiang Jieshi). Wade-Giles uses apostrophes to mark aspiration (t' for the aspirated stop, t for the unaspirated); the apostrophes were often omitted in casual use, producing systematic ambiguity that Pinyin removed.

Other systems exist for other purposes. Bopomofo (注音符號, the Mandarin Phonetic Symbols) is a non-Latin phonetic system used in Taiwan since the 1910s, particularly for primary education and computer input. Yale romanisation is used in some American academic contexts. Pe̍h-ōe-jī is the romanisation of Hokkien used by Taiwanese Christians since the 19th century.

Chapter XVIIWenyan.

Classical Chinese (文言, wényán) — the literary language of imperial China from roughly the late Zhou through the early 20th century — is a written register descended from the spoken Chinese of approximately 500 BCE, frozen as the medium of literary, administrative, and scholarly writing.

The grammar is heavily monosyllabic and analytic. Particles do most of the work that inflection does in Indo-European languages. Reading Classical Chinese is closer to reading mathematics than to reading prose — each character is a packed semantic unit, and the relations among them must be inferred rather than explicit.

The classical canon was the basis of the imperial examination system (suspended 1905) for over a millennium. Every educated person across the Sinosphere — China, Korea, Japan, Vietnam — was literate in Classical Chinese, regardless of their spoken language. The result was an East Asian intellectual continuum mediated by a single written language until the 20th-century vernacular revolutions.

Modern Mandarin literary writing (the baihua reform, championed by Hu Shi from 1917 onward) is closer to spoken language but still highly influenced by Classical conventions, idioms, and four-character expressions (chengyu).

Chapter XVIIIBeyond the family.

The cultural reach of Sinitic extends well beyond its native speakers through the historical Sinosphere — the East Asian polities that adopted Chinese characters, Classical Chinese as a literary medium, and Confucian-Buddhist intellectual culture.

Korean is genetically unrelated to Chinese — it is a language isolate, possibly part of a small Koreanic family with the extinct Old Koguryo. Until the 15th-century invention of Hangul, Korean was written almost exclusively in Chinese characters. Modern Korean retains a substantial Sino-Korean vocabulary stratum, and characters (hanja) appear in some formal contexts.

Japanese, a member of the small Japonic family, also borrowed the Chinese script and a substantial Sino-Japanese vocabulary stratum. The kanji system is the residue.

Vietnamese, a member of the Mon-Khmer family of Austroasiatic, was written in Chinese characters (Chữ Hán) and adapted Chinese characters (Chữ Nôm) for its native vocabulary until the early 20th-century romanisation. The Chinese-derived component of Vietnamese vocabulary is enormous.

The cultural unity of the Sinosphere is a non-genetic kinship — based on shared writing, shared classical literature, and shared institutional traditions — rather than on language family.

Chapter XIXThe 20th-century vernacular revolution.

The May Fourth Movement of 1919 produced a wholesale rejection of Classical Chinese in favour of the spoken vernacular as the medium of writing. Hu Shi's 1917 essay A Preliminary Discussion of Literary Reform set the agenda. Lu Xun's stories, written in colloquial Mandarin and published from 1918 onward, were the proof of concept.

The reform was successful within a generation. By 1949 modern Chinese journalism, popular fiction, school textbooks, and government documents were written in Mandarin-based vernacular. Classical Chinese persisted only in scholarly study, classical poetry composition, and ceremonial contexts.

The reform had costs. Pre-modern Chinese literature, fully accessible to literate adults in 1900, requires specialist study by educated Chinese in 2026. The cultural continuity of two and a half millennia was deliberately broken — the pragmatic gain was the democratisation of literacy across a population whose first spoken language was already vernacular Mandarin or its relatives.

Chapter XXThe Buddhist canon.

The Tibetan literary tradition is overwhelmingly Buddhist and overwhelmingly translated. From the 8th century onward, Tibetan scholars produced what is probably the most complete and most accurate translation of Sanskrit Buddhist literature ever achieved. The Kangyur (translated word of the Buddha) and Tengyur (Indian commentaries) total over 4,500 distinct texts.

Beyond the imported canon: the great Tibetan vernacular literature includes the Gesar epic — perhaps the longest epic poem in the world, with versions running to over a million verses, performed by hereditary singer-bards of eastern Tibet — and the lyrical poetry of figures like Milarepa (1052–1135), the Sixth Dalai Lama Tsangyang Gyatso (1683–1706), and Patrul Rinpoche (1808–1887).

Modern Tibetan literature, both inside Tibet and in the diaspora, is in difficult circumstances. The Chinese state's pressure on the language is substantial; the diaspora literary community in Dharamsala and elsewhere is small. Writers like Jangbu (the pen name of Dorje Tsering Chenaktsang) and Tsering Woeser have produced significant late-20th and early-21st-century work.

Chapter XXIThe small languages.

Of the approximately 450 Sino-Tibetan languages, the great majority are spoken by fewer than 100,000 people, and many by fewer than 10,000. UNESCO and Ethnologue list dozens of Sino-Tibetan languages as critically endangered.

The pressures are familiar: state languages displace local languages in education and media; out-migration to cities removes young speakers from village contexts; intermarriage between speakers of different small languages defaults to a regional lingua franca; the prestige economy favours the language of national markets.

Documentation projects — the long-running Tibeto-Burman Linguistics group at Berkeley (under James Matisoff and successors), the Endangered Languages Project, the Linguistic Survey of Nepal (Bandhu and others), the Northeast India linguistic projects (Burling, Post, DeLancey) — have made important recordings. The race is fast. Several Sino-Tibetan languages have lost their last fluent speakers in the last twenty years.

Chapter XXIIThe pictographic survival.

Among the small Tibeto-Burman languages of northwest Yunnan, Naxi (Nakhi) preserves something extraordinary: the Dongba script, perhaps the only living pictographic writing system in the world. The script, used by Naxi religious specialists for ritual texts, has been in continuous use for at least eight centuries (the earliest dated manuscripts are 13th–14th-century).

The Dongba pictographs do not encode every word of the spoken language; they function as memory aids for chanted ritual texts that the Dongba priest knows by heart. The system is intermediate between picture-writing and full script.

About 2,000 distinct Dongba glyphs are catalogued. The Naxi community has perhaps 300,000 ethnic members, of whom a few thousand still actively use the language and a much smaller number — some hundreds — read the Dongba script. The script is one of the great ongoing curiosities of linguistic anthropology.

Chapter XXIIITwenty essentials.

• 1915Études sur la phonologie chinoiseKarlgren
• 1948Grammata Serica RecensaKarlgren
• 1966Introduction to Sino-TibetanShafer
• 1973The Loloish Tonal Split RevisitedMatisoff
• 1986The Languages of ChinaRamsey
• 1988The Chinese Language: Fact and FantasyDeFrancis
• 1992A Handbook of Old Chinese PhonologyBaxter
• 2003Handbook of Proto-Tibeto-BurmanMatisoff
• 2003The Sino-Tibetan Languages (Routledge)Thurgood & LaPolla (eds.)
• 2008The Origin of Chinese CharactersWieger / Baxter
• 2010Chinese Lexical SemanticsCai & Liu
• 2014Old Chinese: A New ReconstructionBaxter & Sagart
• 2015The Tibetan Empire in Central AsiaBeckwith
• 2015A Reference Grammar of TibetanTournadre & Dorje
• 2017The Languages and Linguistics of South AsiaHock & Bashir (eds.)
• 2019Phylogenetic evidence for ST origin in N. China (Nature)Zhang et al.
• 2020Dated language phylogenies (ST)Sagart et al.
• 2021Sino-Tibetan Etymological Dictionary (online)Matisoff (continuing)
• 2022Visible Speech: Tibeto-Burman TonesMazaudon
• 2023The Cambridge Handbook of Chinese LinguisticsHuang & Shi (eds.)

Chapter XXIVWatch & read.

↑ Chinese — The Sinitic Languages — overview

More on YouTube

Watch · Master Chinese Tones — Pronunciation Training
Watch · Unseen Tibet — language and the high plateau

Chapter XXVThe contemporary picture.

Sino-Tibetan in 2026 is dominated by Mandarin. Standard Mandarin is the medium of education, government, media, and increasingly daily speech across the PRC, displacing the regional Sinitic varieties at all socioeconomic levels. The non-Sinitic Tibeto-Burman languages within China — Tibetan, Yi, Naxi, the Bodish languages of the high plateau, the dozens of small Yunnan languages — face similar pressure, more acute because their speaker populations are smaller.

Outside the PRC: Burmese is secure as a national language; Tibetan is secure in Bhutan (where Dzongkha, a Tibetan variety, is the national language) and in some Indian Buddhist communities; the Tibeto-Burman languages of Northeast India face Hindi and English pressure but have several active literary projects.

The diaspora: Hokkien remains vital in Malaysia, Singapore, and Taiwan. Cantonese continues in Hong Kong, Vancouver, San Francisco, and elsewhere — though under pressure even in Hong Kong from PRC educational policy. Mandarin's global reach as a second language is growing, slowly, in line with the PRC's economic weight.

Chapter XXVIWhere the field goes.

Three open frontiers in Sino-Tibetan studies.

Documentation of small Tibeto-Burman languages remains the most urgent. Each year, a few more pass the last-speaker threshold. The work on Apatani (Tani), Mizo (Kuki-Chin), and the small Bodish languages of Bhutan and northeast India is being done; the funding does not match the need.

Reconstruction of Proto-Sino-Tibetan at the level Baxter-Sagart achieved for Old Chinese is still in progress. Matisoff's Handbook of Proto-Tibeto-Burman (2003) is the major synthesis; the Sinitic-Tibeto-Burman correspondence is still being refined.

Phylogenetics, as in Indo-European, has been transformative. The 2019 Zhang et al. paper and the Sagart 2020 paper give convergent dates and homelands. Whether the model can be extended down to subgroup-internal phylogenies, with statistically meaningful divergence dates, is the next frontier.

Chapter XXVIIWhat we know.

What a century and a half of comparative work has established:

1. Sino-Tibetan is a real family with about 450 living languages, the second-largest by speakers in the world.

2. The internal structure has at least two major divisions — Sinitic and Tibeto-Burman — though the position of some intermediate groups (Karenic, Bai) remains contested.

3. The proto-language was probably spoken in northern China around 7,000 years before present, and the family expanded southward and westward into the Himalayas and Southeast Asia.

4. Tonal systems are a recurrent areal innovation rather than an inherited family feature; Old Chinese was non-tonal.

5. The Chinese script is a continuous tradition over 3,000 years and serves as the unifying cultural artefact across the Sinitic varieties despite their spoken-language divergence.

What we don't know: the timing details, the position of several intermediate subgroups, the relationships between Sino-Tibetan and other proposed macro-families (the Sino-Caucasian and Dene-Caucasian hypotheses are not respectable in mainstream linguistics).