Although the Chinese languages definitely belong to the same family as Tibetan, Burmese and their relatives, they differ radically. Proto-Sino-Tibetan must be dated many thousands of years ago. Possibly it was spoken in what is now southern China. Proto-Chinese, most likely, developed along the coast and the coastal valleys south of the Yángzĕ delta – the area of Chinese speech where dialect diversity is still at its peak.

The oracle bones of about 1400 BC, discovered at archaeological sites near the lower Yangze, are the oldest written documents of Chinese, which thus has a recorded history as long as that of Greek. True Chinese literature begins later: with collections of poetry, and with the works of Confucius and Mencius, around 500 BC. By that time Chinese was the language not only of the lower Yangze but of the wide lands to the north, in which the historic Chinese capitals, Beijing not the least of them, grew to eminence.

The classical literature of China is extremely rich, including short and long poetry, history, memoirs and many other prose genres. Chinese can claim the oldest printed literature in the world, dating back to a Buddhist sutra printed from wood blocks in AD 868.

Prose fiction has been cultivated in Chinese for much longer than in any other language. The Water Margin and The Story of the Stone (Dream of the Red Chamber) are, in their different styles, landmarks of world literature.

It is not surprising that there are several major languages extending across the vast and heavily populated region that is modern China. It is more surprising that one language alone – Ptōnghuà, the ‘common language’ or Mandarin – has so many hundreds of millions of speakers across northern China. Elsewhere in the world, single languages have not managed to extend themselves so far, except for rather brief periods. Examples are the Latin of the Roman Empire, the English of the British Dominions and the United States, the Russian of the Soviet Union: but not one of these languages has yet achieved such an enormous number of first-language speakers. Here Putonghua is unique, and its uniqueness must come from three factors: North China has been under a single government, most of the time, for well over a thousand years; administration and culture have shown impressive stability and uniformity over all this time; and, probably, northern peasants have found it more necessary than those in the south to travel and to resettle in order to escape famine and to find work.

In the modern context Putonghua has every reason to spread further and faster. It is propagated by the administration of a centralised state, by its media, by its education and by ever increasing nationwide travel and migration. In the past, very few people in southern China knew the language of the capital: nowadays, very many do, especially the young.

The ‘internal’ history of Chinese – the history of its words and its sounds – is very difficult to trace, since the script does not directly reflect the sounds of the language. A ‘rhyming dictionary’ compiled in AD 601 contributed to the researches of Bernhard Karlgren, who reconstructed the sounds of Ancient Chinese in Grammata Serica (1940) and Grammata Serica recensa. Karlgren's work has revolutionised many aspects of Chinese studies. Thanks to him we know far more about the patterns and forms of ancient poetry, the origins and histories of the other Chinese languages, and, much further back, the phonology and grammar of proto-Sino-Tibetan, the ancestral language out of which, many thousands of years ago, Chinese, Karen and the Tibeto-Burman languages all developed.

The best known of all monosyllabic languages, Chinese has about 1,600 possible syllables in its sound pattern. How can all the concepts of the modern world be specified in only 1,600 words? They cannot: although each syllable can be regarded as an independent word, Chinese strings them together, as would any other language, to specify ideas.

Together, the Chinese languages have more mother-tongue speakers than any other language on earth. Even on its own, Putonghua probably achieves this first place.

The Chinese languages are not mutually intelligible, but they share the Chinese writing system – and thus Chinese literature belongs to all of them equally, though in informal writing plenty of local variations may occur. That is why all the languages are treated here under the single heading ‘Chinese’.

Putonghua or Mandarin first spread as the language of the capital, Beijing, and thus of administrators and scholars throughout the vast empire. China has for many centuries had a highly centralised system of higher education and centrally organised recruitment to the higher echelons of the civil service. The speech of Beijing has for all this time been the natural language for communication among Chinese from different regions, and the obvious first language to learn for non-Chinese living or working in China.

Putonghua is distinguished among the Chinese languages in having only four tones (high; high rising; falling-rising; falling), no voiced stops and no syllable-final consonants.

Wú is named after the old state of Wu, whose capital, two thousand years ago, was Soochow. Shanghai is nowadays the centre of Wu speech. The language has three ranges of stops – voiced, voiceless unaspirated, voiceless aspirated – as did middle Chinese. Like Putonghua it is one of the few languages of the world in which z, or a sound very like it, functions as a vowel.

Gàn is named after the river Gan, flowing through the province of Jiangxi.

Xiāng is the language of Hunan province: but while ‘Old Xiang’ of the mountains and valleys of the south and east is a quite distinct language, ‘New Xiang’ of the north-west and the cities is becoming a mixed language, not too different from neighbouring Putonghua dialects.

Yuè or Cantonese, named after the old southern state of Yue, is the major language of southern China, centred on the great trading city of Canton. Yue preserves all the middle Chinese syllable-final consonants, p t k m n ng. It sounds markedly different from Putonghua.

Kejia or Hakka is the language of northerners who moved south in medieval times. Hakka is in fact a Yue term meaning ‘guests’, and Kèjiā is the Putonghua reading of the same word. Kejia is close to Gan but with phonological and lexical differences.

Mn, Hokkien or Teochew is a group of nine mutually unintelligible dialects of Fújiàn and Taiwan. Min is the typical language of the Chinese fishing community of Singapore. Abroad it is sometimes known as Swatownese, Swatow being the major emigration port for the speakers of this language. Swatow itself is in Guangdong province, but its cultural links are with Fujian to its north. The whole region is geographically isolated from the rest of China – except by coastal trade – and this has maintained its linguistic diversity.

Ptōnghuà or Mandarin (800,000,000 speakers) is in origin the language of the neighbourhood of Beijing. It is spoken in the north of China generally; by the 7,000,000 Chinese Muslims, an official nationality of Chinese Mongolia; and by rapidly increasing numbers throughout China. Overlying the regional languages, there are regional dialects of Putonghua. In its own homeland the major dialect divisions are northern (within which there is a distinct north-western dialect), southern or eastern – spoken around Nanking and the lower Yangze, and south-western, the language of Sichuan. Putonghua has also been called Kuanhua, Pekingese, Kuo-yü and by other names.

A medieval offshoot of what is now called Putonghua, the Chinese that is the language of traditional learning in Vietnam is pronounced on a system established in the 10th century: that was when Vietnam became independent and thus ceased to draw on the administrators and educators of the empire. More recently, Vietnam's Chinese community has used Yue (Cantonese) as its lingua franca.

Gàn (25,000,000 speakers) is the language of Jiăngxì.

Kèjiā (40,000,000 speakers), once well known as Hakka, is spoken throughout south-eastern China in agricultural communities in Yue and Min-speaking areas. It is the language of the New Territories of Hong Kong.

Mn (50,000,000 speakers) is the language of Fújiàn, also spoken across the water in Taiwan. It is the language of many Chinese speakers abroad, including some very important communities in south-east Asia, notably Singapore and Bangkok. It has had many names: Hokkien, Fukien, Teochew, Teochiu, Chiuchow, Cháozhōu, Taeciw, Swatownese. The coastal dialects of Swatow, Amoy and Foochow are the best known abroad: inland dialects are said to be quite distinct. Mn Nán (‘Southern Min’) extends along the coast of Gungng and is also spoken on Hinán island.

Wú (90,000,000 speakers) is the language of Shanghai, the heavily populated coastal districts nearby, and Zhejiang. With the recent vast growth of Shanghai, the dialect of the old Wu capital of Soochow is now ‘old-fashioned’ rather than standard.

Xing or Hunanese (55,000,000 speakers) is the language of Hunan.

Yuè or Cantonese (65,000,000 speakers) is spoken in a large area of southern China. A form of Yue, influenced by English and Malay, is the lingua franca of Hong Kong, and is the majority language of the Chinese communities in the United States and in Britain.

Dungan or Tung-an has about 50,000 speakers in Kyrgyzstan and Kazakhstan and was a national language of the Soviet Union. Unlike all other varieties of Chinese, Dungan is written in Cyrillic script.

The oldest known passages of written Chinese are ‘oracle bones’ and inscriptions of 3,500 years ago. The brief texts scratched on bones were clearly used for fortune-telling.

In those 3,500 years Chinese script has developed massively: yet some characters are still recognisable. As shown in the box, they name concrete and familiar objects.

A thousand years later, when Chinese literature had begun to be written down, all words – however abstract or concrete their meaning – could be recorded. The stock of characters had grown in several ways, but the most productive was by the doubling of characters: often the right-hand element was borrowed to give a sound or a ‘sounds-like’, the left-hand element to suggest a meaning.

Single or doubled, each character makes up a square block of text and each represents a single syllable of the spoken language.

In communist China a simplification of the individual characters and a reduction in the total number of characters have both been promulgated. Here Chinese is now normally printed from left to right across the page, like the Latin alphabet. The older pattern, still widespread in Taiwan, Hong Kong, Singapore and elsewhere, is to lay out the page with lines read vertically and following from right to left.

Most Chinese dictionaries arrange characters not by sound but according to the number of strokes (from 1 to 17) used to write the character, or used to write the ‘radical’ element of a doubled character.