Empires of the Word: A Language History of the World

by Nicholas Ostler

  One characteristic of Indian civilisation that they brought with them was a tendency to modify and customise the alphabet. Just as there are now at least ten major scripts* derived in India from the Brahmi characters (diffused all over the subcontinent in Aśoka’s time), there are another nine that developed in South-East Asia, Indonesia and the Philippines,† all derived from Indian scripts, many through the Pallava script of the south. The origin of this diversity lies in the variety of writing materials available in different places, but the different styles evidently came to be national icons. In the Cambodian pillars that carry rules for monasteries, Sanskrit in Khmer script on one side is paralleled by Sanskrit in a North Indian script on the other: perhaps there were North Indian devotees as well as Khmers resident here.31

  This is just one of many signs that there was heavy cultural traffic in both directions between India and Indo-China during this period. Another example is given by the life of Atīśa, a monk born in Bengal in 982, who went on to become one of the founders of Buddhism in Tibet in his sixties. He had spent his student days in Śri Vijaya, in Sumatra.

  In a way, the culture as the Indians brought it will always be a mystery to us. The splendours of Shwe Dagon in Burma, Borobodur in Java, Angkor

  Wat in Cambodia, as well as less well-known magnificences in Pagan, Champa, Laos, Bali and Sumatra, built over a millennium from about AD 500, all stemmed from the seminal ideas of the Indians, but at least in terms of architecture there is nothing now quite like them back in India. We can only speculate that styles executed in stone at Borobodur and Angkor Wat may echo the architecture of wooden buildings long vanished from southern India.

  Nevertheless, this roll-call of states and civilisations that took their beginnings from India reminds us how vast, how varied and how long lasting this influence was, all the more remarkable because no military force seems to have been applied anywhere to bring in the new, more organised, Indian society. This contrasts sharply with the record of incursions from the other developed civilisation to the north. Ever since the first century AD, China had been putting constant pressure on the Annamite kingdom of northern Vietnam, periodically invading it, and insisting on recognition of China’s emperor as its overlord.

  The earliest documented Indianised kingdom—the documentation is Chinese—was set on the lower Mekong, in modern Cambodia and southern Vietnam, probably in the first century AD. It is usually known as Funan, which is a Chinese version of its name. It was really called, in Khmer, Bnam, ‘the mountain’,* and its king as kurung bnam, a translation of Parvatabhū-pala or Śailarāja: bearing this title of ‘King of the Mountain’, he would have established a cult of the god Siva in a high place, so reconciling his legitimacy as an Indian king with the native spirits of the land.32

  Funan’s foundation myth, read from a Sanskrit inscription in Champa,33 confirms this. A Brahman named Kauinya (derived from Kuin, one of Siva’s titles) received a javelin from another Brahman, a hero from the Mahabharata named Aśvattāman, and threw it to find the right site for the city. He married a local princess named Soma, daughter of the king of the Nāgas, the many-headed water cobras worshipped as protectors of Khmer riches.

  Thereafter, major Sanskrit-speaking states were set up all over South-East Asia, Sumatra and Java.† Their names are themselves in Sanskrit, and show either a sentimental link with other Indian holy places far away, or an attempt to Indianise local names. It is often difficult now to locate them exactly. In Malaya, Lankasuka, controlling one much-used overland route from the Bay of Bengal to the Gulf of Siam, beside Tāmbralinga (Ligor), Takkola (Takuapa) and Kāaha (Kedah); in Cham, the south of modern Vietnam, Amarāvatī (Dong-duong), Vijaya (Binh-dinh), Kauhara (Nha-trang), Pāuranga (Phanrang); in Java, Tārumā (round Jakarta) and Kaarāja in the east; in Sumatra, Malāyu (Jambi), Śrī Vijaya (Palembang); in Burma, Sudhammavatī (Thaton), Śrīketra (Prome or Thayekhettaya), Hasavatī (Pegu), Śrī Deva (Si Thep); and in the region of modern Thailand Dvāravatī, north of Bangkok.

  Names of rulers too are typically Sanskritic. Good examples are the more than thirty Cambodian kings whose names end in -varman, ‘bastion’, from Jayavarman, who died in AD 514, to Śrīndrajayavarman, 1307-27, and the Majapahit kings of Indonesia from Rājasa in 1222-7 to Suhitā, 1429—47.*

  These led to many more Sanskrit place names, since it was customary to name a city after the king that founded it. To give one example among many dozens, Śrehapura (literally ‘best of cities’), capital of Cambodia, was named after its founder, King Śrehavarman (’best bastion’). It is likely also that Śri Vijaya, the dominant kingdom in southern Sumatra, was named after a king named Vijaya, ‘Victorious’.

  This is just a sample of some of the better known; as could be expected, the history of the relations of all these cities and kings over a thousand years is a vast and labyrinthine subject, and not one to be broached here.

  It is easy to overlook what a major change the introduction of Sanskrit must have been for the local peoples. Sanskrit, as a type of language, was fearsomely different from the local languages, now classified as Burman, Austro-Asiatic and Austronesian. Sanskrit is polysyllabic, and highly inflected, with a complicated consonant system that is not averse to long clusters. Word order is free. This language was being taken up by speakers of other languages where words were short, often distinguished by tone, and made up of simple syllables with single consonants at beginning and end. Inflections were simple or absent, but word order was rigid. It was at least as radical a change as it would be to bring Japanese in as an elite language where previously everyone had known only English or Dutch. What a wrench it was can be seen in the mangled remains of some of the Sanskrit names: Śriketra came out as Thayekhettaya, Śrī Deva as Si Thep.

  Nevertheless, the quality of written Sanskrit that the natives acquired in this part of the world deviated hardly at all from that of India. We do not see strong ‘substrate influence’ in the texts written here. Talking of Cambodia, R. C. Majumdar remarks that its inscriptions, known from AD 475 to 1327, are generally ‘composed in beautiful and almost flawless kāvya—i.e. poetic—style, and some of them run to great lengths… Almost all the Sanskrit metres have been successfully used in these verses, and they exhibit a thorough acquaintance with the most developed rules and conventions of Sanskrit rhetoric and prosody.’34 The inscriptions are also full of learned, even witty, allusions to the Vedas and all the different branches of Indian learning, especially grammar.

  Particularly accomplished was Queen Indradevi, consort of Jayavarman VII (who ruled in Cambodia 1181—c.1218): she was a pious Buddhist and taught the Buddhist nuns of three convents. She has left an inscription, in praise of her younger sister, another scholar, who had sadly died young: it runs to 102 verses in several different metres.35

  Some of the literature written in Indo-China joined the canon of Sanskrit classics. Vararuci’s Sārasamuccaya (’collection of essences’) could be hard hitting: to show how views can differ, he evokes a woman’s breast—seen by her child, and by her husband; and then her dead body, seen first by an ascetic, then by her lover, and then by a dog. Later on, he prefigures Pascal’s wager in his advice to the atheist (nāstika—literally the ‘isn’t-ist’): if there is no world after death, there is nothing to fear either way; but if there is, it will be the atheists who stand to suffer.36

  Sanskrit texts apparently played an important role in the foundation of new Hindu cults, which might be founded to buttress newly independent states: so when Jayavarman freed Cambodia from Javanese control in the twelfth century, he invited a Brahman named Hirayadāma (’Golden Cord’) to perform Tantric rites to guarantee this freedom, under its own ruler. The resulting cult of Devarāja (’god-king’) lasted for 250 years, explicitly based on four named śāstra texts. It could not have been done without Sanskrit, and the access to ancient wisdom that it implied.

  The sense of numinous power infusing Sanskrit led on occasions to a sort of spiritual nostalgia. One king of Champa, Gangara
ja, is said to have abdicated his throne so as to have the chance to give up the ghost on the banks of the Ganges. And, more public-spiritedly, there is evidence from an inscription put up at Vat Luong Kau in Laos that a king called Śrī Devanīka planned to set up a new Kuruketra at home as a substitute for the sheer holiness of the real Kuruketra north of Delhi. As the site of the Mahabharata’s great battle, it was peerless among shrines, but sadly inaccessible. He quotes the epic:

  Pthivyā Naimiam puyam antarīke tu Pukaram Trāyānām api lokānām Kuruketram viśiyate.

  On the earth the blessed Naimisha, in the ether Pushkara, But in the three worlds, Kurukshetra holds the crown.37

  The long years of Indian influence came to an end only after a full millennium. A major jolt had already come in the thirteenth century, when the Mongols sacked Pagan and other Burmese kingdoms in the north. But it has been suggested by one of the leading scholars, not without nostalgia, that Indian civilisation was the victim of its own increasing popularity: ‘The underlying causes of this decline were the adoption of Indian civilization by an increasingly large number of natives who incorporated into it more and more of their original customs, and the gradual disappearance of a refined aristocracy, the guardian of Sanskrit culture.’38

  In any event, in the fifteenth century Vietnam expanded its influence into Champa, annexing permanently the south of Indo-China; and about the same time groups of mountain peoples, the Shan in Burma, and the Thai in Siam, established new kingdoms that thrust aside the old powers of Pagan and Angkor. Nonetheless, when founding their new capital, the Thai could not help calling it Ayutthaya, in direct tribute to the Hindu hero Rama’s residence, Ayodhya.

  Sanskrit carried by Buddhism: Central and eastern Asia

  So far we have largely spoken of Sanskrit as a vehicle of Hinduism. And it seems that for the most part this is what it conveyed at first in South-East Asia. Fa-Xian, returning to China via Ye-po-ti (Yava-dvīpa) in the East Indies in the early fifth century AD, remarked: ‘in this country heretics and Brahmans flourish, but the law of Buddha is not much known’.39

  To this day, Hinduism survives on the island of Bali, east of Java. However, elsewhere in South-East Asia the picture is now very different, Hinduism long ago replaced by Buddhism. This is the result of a long and complex, though not especially bloody, history of doctrinal contests between the two faiths. Hindu cults’ close associations with ruling dynasties ultimately worked against them, when those dynasties fell. But there was also competition among strains of Buddhism, Tantra, originally ‘the loom’ or ‘the framework’, Mahāyāna, ‘the great vehicle’ and Theravāda, ‘the docrine of the elders’. Theravada, buttressed by links with the Sinhala in Śri Lanka, ultimately triumphed in South-East Asia. Nevertheless, all these struggles took place against an unchallenged background of Indian learning.

  Buddhist missionaries actually came very soon after the first Indian buccaneers and traders, if not along with them. Ceylonese chronicles tell of Aśoka sending two monks, Soa and Uttara, to Suvaabhūmi in the third century BC,40 although the first archaeological records of Buddhist activity in South-East Asia (in the areas of modern Burma and Thailand) are from the fifth century AD. Hinduism was always a religion likely to appeal to kings and a ruling elite, but not voluntarily to the lower orders, the Śudras and outcastes, who are singularly downtrodden in the Hindu caste system; by contrast, Buddhism, with its egalitarian emphasis on personal quest for enlightenment, could in principle appeal much more widely. It seems likely that in the early days of Indian advance into the region both religions were represented; their complementary charms may even have served to back each other up, while promoting Indian culture among outsiders.

  The religious distinction always had some linguistic implications, the Hindus favouring classical Sanskrit, while the Buddhists preferred the closely related but somewhat simpler Pali. As time wore on, there was also a tendency for Pali to be reclothed in archaic Sanskrit forms, giving rise to the particular style of Buddhist Hybrid Sanskrit. Real learning, and creativity, in classical Sanskrit tended to be at its best in the Hindu areas, such as Champa, Cambodia, Java and Bali.

  Despite the Buddha’s original urgings to his disciples to leave behind strict linguistic codes and work in any vernacular (sakayā niruttiyā) in order to get the message across, the Buddhist scriptures remained in Pali in South-East Asia, where—in contrast with China and Tibet—there was no major effort to translate them into local languages. Pali became an esoteric liturgical language, unknown to the general population, but apparently without adverse effects on the spread of Buddhism.

  Nor was there any converse tendency to have Pali, or some form of Sanskrit, taken up as a language of general communication outside Buddhist liturgy and debate. There is no secular literature in Pali, even if the Jātaka tales, which nominally recount the past lives of the Buddha, are rather like such other story books as Aesop’s Fables, or its Indian equivalent, the Pañcatantra. And in South-East Asia, where Pali survives as a liturgical language, the local vernacular has nothing to do with it: Burmese, Thai, Khmer, Acehnese, Malay and Javanese are all unrelated to Pali, heavy as they are with loans from the Indian languages.

  Buddhism has proved a faith of remarkable attractiveness from India outward to the north and east, and so Pali and Sanskrit are extremely well known in these vast areas. But they have remained no more than liturgical languages. As a result, Buddhism’s linguistic effects have been far weaker than those of Christianity or Islam. After all, Latin, the language of Western Christianity, provided the foundation for the growth of a common language in the monasteries and then the universities of Europe in this same period (AD 500-1500). Islam propagated Arabic all round North Africa, Arabia, Palestine and Mesopotamia, persisting up to the present day, both in unchanged form as an international lingua franca for the educated and, with local variations, as the basis of many vernaculars. There is no comparable linguistic union of Buddhists, in their daily languages.

  As for how the language was used in this part of Sanskrit’s story, there is little to say. In Hinduism, the virtue implicit in the very sound of the Vedas had long since been separated from any need to understand their meaning. Now once again for the Buddhists, with the language no longer widely understood, but still widely heard in chants and ceremonial, its substance and sound began to be given a mystic value of their own. Sanskrit became for many a language of mantra, ‘incantation’ and maala, ‘circle, sacred diagram’. In medieval Japan, repeating namu amida butsu, a version of nama Amitabha Buddha, ‘Bowing to you, O Resplendent Enlightened One’, was the infallible means of reaching the Pure Land after death. And to this day millions of Tibetans chant om mai padme hum, ‘Hail the jewel in the lotus’, a mystical phrase from Tantric Buddhism, its original sexual imagery now quite forgotten.

  More pragmatically, the technology and systems associated with writing and analysing Sanskrit provided the basis for literacy in other languages. In this way, sacred languages, unavailable for direct communication among people, could still go on inspiring developments in the local vernaculars.

  The advent of Sanskrit, known as fànwén, ‘Brahman writing’, in China, bongo, ‘Brahman talk’, in Japan, had only a small effect on the character-based system of writing in use in East Asia, since this had already been well established in China for over a millennium: rather, Chinese characters are often used (though only phonetically) to represent Sanskrit itself in the Buddhist practice of these countries.

  One effect it did have was on Chinese phonetics. Chinese scholars of the Tang period (seventh to eighth centuries), knowing the Sanskrit alphabetic tradition could identify the initial consonants of characters, called them zìmŭ, ‘word mothers’, apparently after the Sanskrit term mātkā, ‘maternal’, which is also a letter of the alphabet. These were used to systematise the traditional practice for indicating pronunciation in dictionaries: Chinese dictionaries have always done this by what is called fànqiě, linking a character with two others, one w
ith the same initial consonant, and the other with the same tone and rhyme. Putting this into a systematic chart was a very modest step in linguistic understanding, since no further analysis of the rhyme part (for example, into vowels and consonants) was undertaken.41

  There is also an interesting curiosity in one of the other writing systems used in this vast area of Asia.* Japan owes the order of symbols in its syllabary, the so-called kana, or go-jū-on, ‘fifty sounds’, to the order of letters in Indian alphabets. The order of Sanskrit letters is conventionally

  This is not an arbitrary order like our ABCD… † Rather it appeals to various purely phonetic properties of the sounds represented. So, for example, all the consonants are placed in an order where tongue contact gradually advances from the back to the front of the mouth cavity. And the nasal consonants (m, n, etc.) always come immediately after the other consonants formed at the same place of articulation. The strange order of the vowels is partly conditioned by the fact that most instances of e and o in Sanskrit actually derive from the diphthongs ai and au, and so are well classified next to their long equivalents āi and āu.