A History of Sanskrit Literature

For the post-Vedic period we have, in addition to the results of internal evidence, a few landmarks of general chronological importance in the visits of foreigners. The earliest date of this kind is that of the invasion of India by Alexander in 326 B.C. This was followed by the sojourn in India of various Greeks, of whom the most notable was Megasthenes. He resided for some years about 300 B.C. at the court of Pāṭaliputra (the modern Patna), and has left a valuable though fragmentary account of the contemporary state of Indian society. Many centuries later India was visited by three Chinese Buddhist pilgrims, Fa Hian (399 A.D.), Hiouen Thsang (630–645), and I Tsing (671–695). The records of their travels, which have been preserved, and are all now translated into English, shed much light on the social conditions, the religious thought, and the Buddhist antiquities of India in their day. Some general and specific facts about Indian literature also can be gathered from them. Hiouen Thsang especially supplies some important statements about contemporary Sanskrit poets. It is not till his time that we can say of any Sanskrit writer that he was alive in any particular year, excepting only the three Indian astronomers, whose exact dates in the fifth and sixth centuries have been recorded by themselves. It was only the information supplied by the two earlier Chinese writers that made possible the greatest archæological discovery of the present century in India, that of the site of Buddha’s birthplace, Kapila-vastu, identified in December 1896. At the close of our period we have the very valuable account of the country at the time of the Muhammadan conquest by the Arabic author Albērūnī, who wrote his India in 1030 A.D.

It is evident from what has been said, that before 500 A.D. literary chronology, even in the Sanskrit period, is almost entirely relative, priority or posteriority being determined by such criteria as development of style or thought, the mention of earlier authors by name, stray political references as to the Greeks or to some well-known dynasty, and allusions to astronomical facts which cannot have been known before a certain epoch. Recent research, owing to increased specialisation, has made considerable progress towards greater chronological definiteness. More light will doubtless in course of time come from the political history of early India, which is being reconstructed, with great industry and ability, by various distinguished scholars from the evidence of coins, copper-plate grants, and rock or pillar inscriptions. These have been or are being published in the Corpus Inscriptionum Indicarum, the Epigraphia Indica, and various journals devoted to the study of Indian antiquities. The rise in the study of epigraphy during the last twenty years has, indeed, already yielded some direct information of importance about the literary and religious history of India, by fixing the date of some of the later poets as well as by throwing light on religious systems and whole classes of literature. Thus some metrical inscriptions of considerable length have been deciphered, which prove the existence of court poetry in Sanskrit and vernacular dialects from the first century of our era onwards. No direct evidence of this fact had previously been known.

The older inscriptions are also important in connection with Sanskrit literature as illustrating both the early history of Indian writing and the state of the language at the time. The oldest of them are the rock and pillar inscriptions, dating from the middle of the third century B.C., of the great Buddhist king Açoka, who ruled over Northern India from 259 to 222 B.C., and during whose reign was held the third Buddhist council, at which the canon of the Buddhist scriptures was probably fixed. The importance of these inscriptions can hardly be over-rated for the value of the information to be derived from them about the political, religious, and linguistic conditions of the age. Found scattered all over India, from Girnar (Giri-nagara) in Kathiawar to Dhauli in Orissa, from Kapur-di-Giri, north of the Kabul river, to Khalsi, they have been reproduced, deciphered, and translated. One of them, engraved on a pillar erected by Açoka to commemorate the actual birthplace of Buddha, was discovered only at the close of 1896.

These Açoka inscriptions are the earliest records of Indian writing. The question of the origin and age of writing in India, long involved in doubt and controversy, has been greatly cleared up by the recent palæographical researches of Professor Bühler. That great scholar has shown, that of the two kinds of script known in ancient India, the one called Kharoshṭhī employed in the country of Gandhāra (Eastern Afghanistan and Northern Panjāb) from the fourth century B.C. to 200 A.D., was borrowed from the Aramaic type of Semitic writing in use during the fifth century B.C. It was always written from right to left, like its original. The other ancient Indian script, called Brāhmī, is, as Bühler shows, the true national writing of India, because all later Indian alphabets are descended from it, however dissimilar many of them may appear at the present day. It was regularly written from left to right; but that this was not its original direction is indicated by a coin of the fourth century B.C., the inscription on which runs from right to left. Dr. Bühler has shown that this writing is based on the oldest Northern Semitic or Phœnician type, represented on Assyrian weights and on the Moabite stone, which dates from about 890 B.C. He argues, with much probability, that it was introduced about 800 B.C. into India by traders coming by way of Mesopotamia.

References to writing in ancient Indian literature are, it is true, very rare and late; in no case, perhaps, earlier than the fourth century B.C., or not very long before the date of the Açoka inscriptions. Little weight, however, can be attached to the argumentum ex silentio in this instance. For though writing has now been extensively in use for an immense period, the native learning of the modern Indian is still based on oral tradition. The sacred scriptures as well as the sciences can only be acquired from the lips of a teacher, not from a manuscript; and as only memorial knowledge is accounted of value, writing and MSS. are rarely mentioned. Even modern poets do not wish to be read, but cherish the hope that their works may be recited. This immemorial practice, indeed, shows that the beginnings of Indian poetry and science go back to a time when writing was unknown, and a system of oral tradition, such as is referred to in the Rigveda, was developed before writing was introduced. The latter could, therefore, have been in use long before it began to be mentioned. The palæographical evidence of the Açoka inscriptions, in any case, clearly shows that writing was no recent invention in the third century B.C., for most of the letters have several, often very divergent forms, sometimes as many as nine or ten. A considerable length of time was, moreover, needed to elaborate from the twenty-two borrowed Semitic symbols the full Brāhmī alphabet of forty-six letters. This complete alphabet, which was evidently worked out by learned Brahmans on phonetic principles, must have existed by 500 B.C., according to the strong arguments adduced by Professor Bühler. This is the alphabet which is recognised in Pāṇnini’s great Sanskrit grammar of about the fourth century B.C., and has remained unmodified ever since. It not only represents all the sounds of the Sanskrit language, but is arranged on a thoroughly scientific method, the simple vowels (short and long) coming first, then the diphthongs, and lastly the consonants in uniform groups according to the organs of speech with which they are pronounced. Thus the dental consonants appear together as t, th, d, dh, n, and the labials as p, ph, b, bh, m. We Europeans, on the other hand, 2500 years later, and in a scientific age, still employ an alphabet which is not only inadequate to represent all the sounds of our languages, but even preserves the random order in which vowels and consonants are jumbled up as they were in the Greek adaptation of the primitive Semitic arrangement of 3000 years ago.

In the inscriptions of the third century B.C. two types, the Northern and the Southern, may be distinguished in the Brāhmī writing. From the former is descended the group of Northern scripts which gradually prevailed in all the Aryan dialects of India. The most important of them is the Nāgarī (also called Devanāgarī), in which Sanskrit MSS. are usually written, and Sanskrit as well as Marāṭhī and Hindī books are regularly printed. It is recognisable by the characteristic horizontal line at the top of the letters. The oldest inscription engraved entirely in Nāgarī belongs to the eighth, and the oldest MS. written in it to the eleventh century. From the Southern variety of the Brāhmī writing are descended five types of script, all in use south of the Vindhya range. Among them are the characters employed in the Canarese and the Telugu country.

Owing to the perishability of the material on which they are written, Sanskrit MSS. older than the fourteenth century A.D. are rare. The two ancient materials used in India were strips of birch bark and palm leaves. The employment of the former, beginning in the North-West of India, where extensive birch forests clothe the slopes of the Himālaya, gradually spread to Central, Eastern, and Western India. The oldest known Sanskrit MS. written on birch bark dates from the fifth century A.D., and a Pāli MS. in Kharoshṭhī which became known in 1897, is still older, but the use of this material doubtless goes back to far earlier days. Thus we have the statement of Quintus Curtius that the Indians employed it for writing on at the time of Alexander. The testimony of classical Sanskrit authors, as well as of Albērūnī, shows that leaves of birch bark (bhūrja-pattra) were also regularly used for letter-writing in early mediæval India.

The first example of a palm leaf Sanskrit MS. belongs to the sixth century A.D. It is preserved in Japan, but there is a facsimile of it in the Bodleian Library. According to the Chinese pilgrim Hiouen Thsang, the use of the palm leaf was common all over India in the seventh century; but that it was known many centuries earlier is proved by the fact that an inscribed copper-plate, dating from the first century A.D. at the latest, imitates a palm leaf in shape.

Paper was introduced by the Muhammadan conquest, and has been very extensively used since that time for the writing of MSS. The oldest known example of a paper Sanskrit MS. written in India is one from Gujarat, belonging to the early part of the thirteenth century. In Northern India, where ink was employed for writing, palm leaves went out of use after the introduction of paper. But in the South, where a stilus has always been employed for scratching in the character, palm leaves are still common for writing both MSS. and letters. The birch bark and palm leaf MSS. are held together by a cord drawn through a single hole in the middle, or through two placed some distance apart. This explains how the Sanskrit word for “knot,” grantha, came to acquire the sense of “book.”

Leather or parchment has never been utilised in India for MSS., owing to the ritual impurity of animal materials. For inscriptions copper-plates were early and frequently employed. They regularly imitate the shape of either palm leaves or strips of birch bark.

The actual use of ink (the oldest Indian name of which is mashi) is proved for the second century B.C. by an inscription from a Buddhist relic mound, and is rendered very probable for the fourth century B.C. by the statements of Nearchos and Quintus Curtius.

All the old palm leaf, birch bark, and paper Sanskrit MSS. have been written with ink and a reed pen, usually called kalama (a term borrowed from the Greek kalamos). In Southern India, on the other hand, it has always been the practice to scratch the writing on palm leaves with a stilus, the characters being subsequently blackened by soot or charcoal being rubbed into them.

Sanskrit MSS. of every kind are usually kept between thin strips of wood with cords wound round them, and wrapped up in coloured, sometimes embroidered, cloths. They have been, and still are, preserved in the libraries of temples, monasteries, colleges, the courts of princes, as well as private houses. A famous library was owned by King Bhoja of Dhār in the eleventh century. That considerable private libraries existed in fairly early times is shown by the fact that the Sanskrit author Bāna (about 620 A.D.) had in his employment a reader of manuscripts. Even at the present day there are many excellent libraries of Sanskrit MSS. in the possession of Brahmans all over India.

The ancient Indian language, like the literature composed in it, falls into the two main divisions of Vedic and Sanskrit. The former differs from the latter on the whole about as much as Homeric from classical Greek, or the Latin of the Salic hymns from that of Varro. Within the Vedic language, in which the sacred literature of India is written, several stages can be distinguished. In its transitions from one to the other it gradually grows more modern till it is ultimately merged in Sanskrit. Even in its earliest phase Vedic cannot be regarded as a popular tongue, but is rather an artificially archaic dialect, handed down from one generation to the other within the class of priestly singers. Of this the language itself supplies several indications. One of them is the employment side by side of forms belonging to different linguistic periods, a practice in which, however, the Vedic does not go so far as the Homeric dialect. The spoken language of the Vedic priests probably differed from this dialect of the hymns only in the absence of poetical constructions and archaisms. There was, in fact, even in the earlier Vedic age, a caste language, such as is to be found more or less wherever a literature has grown up; but in India it has been more strongly marked than in any other country.

If, however, Vedic was no longer a natural tongue, but was already the scholastic dialect of a class, how much truer is this of the language of the later literature! Sanskrit differs from Vedic, but not in conformity with the natural development which appears in living languages. The phonetic condition of Sanskrit remains almost exactly the same as that of the earliest Vedic. In the matter of grammatical forms, too, the language shows itself to be almost stationary; for hardly any new formations or inflexions have made their appearance. Yet even from a grammatical point of view the later language has become very different from the earlier. This change was therefore brought about, not by new creations, but by successive losses. The most notable of these were the disappearance of the subjunctive mood and the reduction of a dozen infinitives to a single one. In declension the change consisted chiefly in the dropping of a number of synonymous by-forms. It is probable that the spoken Vedic, more modern and less complex than that of the hymns, to some extent affected the later literary language in the direction of simplification. But the changes in the language were mainly due to the regulating efforts of the grammarians, which were more powerful in India than anywhere else, owing to the early and exceptional development of grammatical studies in that country. Their influence alone can explain the elaborate nature of the phonetic combinations (called Sandhi) between the finals and initials of words in the Sanskrit sentence.

It is, however, the vocabulary of the language that has undergone the greatest modifications, as is indeed the case in all literary dialects; for it is beyond the power of grammarians to control change in this direction. Thus we find that the vocabulary has been greatly extended by derivation and composition according to recognised types. At the same time there are numerous words which, though old, seem to be new only because they happen by accident not to occur in the Vedic literature. Many really new words have, however, come in through continual borrowings from a lower stratum of language, while already existing words have undergone great changes of meaning.

This later phase of the ancient language of India was stereotyped by the great grammarian Pāṇini towards the end of the fourth century B.C. It came to be called Sanskrit, the “refined” or “elaborate” (saṃ-skṛi-ta, literally “put together”), a term not found in the older grammarians, but occurring in the earliest epic, the Rāmāyaṇa. The name is meant to be opposed to that of the popular dialects called Prākṛita, and is so opposed, for instance, in the Kāvyādarça, or Mirror of Poetry, a work of the sixth century A.D. The older grammarians themselves, from Yāska (fifth century B.C.) onwards, speak of this classical dialect as Bhāshā, “the speech,” in distinction from Vedic. The remarks they make about it point to a spoken language. Thus one of them, Patanjali, refers to it as used “in the world,” and designates the words of his Sanskrit as “current in the world.” Pāṇini himself gives many rules which have no significance except in connection with living speech; as when he describes the accent or the lengthening of vowels in calling from a distance, in salutation, or in question and answer. Again, Sanskrit cannot have been a mere literary and school language, because there are early traces of its having had dialectic variations. Thus Yāska and Pāṇini mention the peculiarities of the “Easterns” and “Northerners,” Kātyāyana refers to local divergences, and Patanjali specifies words occurring in single districts only. There is, indeed, no doubt that in the second century B.C. Sanskrit was actually spoken in the whole country called by Sanskrit writers Āryāvarta, or “Land of the Aryans,” which lies between the Himālaya and the Vindhya range. But who spoke it there? Brahmans certainly did; for Patanjali speaks of them as the “instructed” (çishṭa), the employers of correct speech. Its use, however, extended beyond the Brahmans; for we read in Patanjali about a head-groom disputing with a grammarian as to the etymology of the Sanskrit word for “charioteer” (sūta). This agrees with the distribution of the dialects in the Indian drama, a distribution doubtless based on a tradition much older than the plays themselves. Here the king and those of superior rank speak Sanskrit, while the various forms of the popular dialect are assigned to women and to men of the people. The dramas also show that whoever did not speak Sanskrit at any rate understood it, for Sanskrit is there employed in conversation with speakers of Prākrit. The theatrical public, and that before which, as we know from frequent references in the literature, the epics were recited, must also have understood Sanskrit. Thus, though classical Sanskrit was from the beginning a literary and, in a sense, an artificial dialect, it would be erroneous to deny to it altogether the character of a colloquial language. It is indeed, as has already been mentioned, even now actually spoken in India by learned Brahmans, as well as written by them, for every-day purposes. The position of Sanskrit, in short, has all along been, and still is, much like that of Hebrew among the Jews or of Latin in the Middle Ages.

Whoever was familiar with Sanskrit at the same time spoke one popular language or more. The question as to what these popular languages were brings us to the relation of Sanskrit to the vernaculars of India. The linguistic importance of the ancient literary speech for the India of to-day will become apparent when it is pointed out that all the modern dialects—excepting those of a few isolated aboriginal hill tribes—spoken over the whole vast territory between the mouths of the Indus and those of the Ganges, between the Himālaya and the Vindhya range, besides the Bombay Presidency as far south as the Portuguese settlement of Goa, are descended from the oldest form of Sanskrit. Starting from their ancient source in the north-west, they have overflowed in more and more diverging streams the whole peninsula except the extreme south-east. The beginnings of these popular dialects go back to a period of great antiquity. Even at the time when the Vedic hymns were composed, there must have existed a popular language which already differed widely in its phonetic aspect from the literary dialect. For the Vedic hymns contain several words of a phonetic type which can only be explained by borrowings on the part of their composers from popular speech.

We further know that in the sixth century B.C., Buddha preached his gospel in the language of the people, as opposed to that of the learned, in order that all might understand him. Thus all the oldest Buddhist literature dating from the fourth or fifth century B.C. was composed in the vernacular, originally doubtless in the dialect of Magadha (the modern Behar), the birthplace of Buddhism. Like Italian, as compared with Latin, this early popular speech is characterised by the avoidance of conjunct consonants and by fondness for final vowels. Thus the Sanskrit sūtra, “thread,” and dharma, “duty,” become sutta and dhamma respectively, while vidyut, “lightning,” is transformed into vijju. The particular form of the popular language which became the sacred idiom of Southern Buddhism is known by the name of Pāli. Its original home is still uncertain, but its existence as early as the third century B.C. is proved beyond the range of doubt by the numerous rock and pillar inscriptions of Açoka. This dialect was in the third century B.C. introduced into Ceylon, and became the basis of Singhalese, the modern language of the island. It was through the influence of Buddhism that, from Açoka’s time onwards, the official decrees and documents preserved in inscriptions were for centuries composed exclusively in Middle Indian (Prākrit) dialects. Sanskrit was not familiar to the chanceries during these centuries, though the introduction of Sanskrit verses in Prākrit inscriptions shows that Sanskrit was alive during this period, and proves its continuity for literary purposes. The older tradition of both the Buddhist and the Jain religion, in fact, ignored Sanskrit entirely, using only the popular dialects for all purposes.

But in course of time both the Buddhists and the Jains endeavoured to acquire a knowledge of Sanskrit. This led to the formation of an idiom which, being in the main Prākrit, was made to resemble the old language by receiving Sanskrit endings and undergoing other adaptations. It is therefore decidedly wrong to consider this artificial dialect an intermediate stage between Sanskrit and Pāli. This peculiar type of language is most pronounced in the poetical pieces called gāthā or “song,” which occur in the canonical works of the Northern Buddhists, especially in the Lalita-vistara, a life of Buddha. Hence it was formerly called the Gāthā dialect. The term is, however, inaccurate, as Buddhist prose works have also been written in this mixed language.

The testimony of the inscriptions is instructive in showing the gradual encroachment of Sanskrit on the popular dialects used by the two non-Brahmanical religions. Thus in the Jain inscriptions of Mathurā (now Muttra), an almost pure Prākrit prevails down to the first century A.D. After that Sanskritisms become more and more frequent, till at last simple Sanskrit is written. Similarly in Buddhist inscriptions pure Prākrit is relieved by the mixed dialect, the latter by Sanskrit. Thus in the inscriptions of Nāsik, in Western India, the mixed dialect extends into the third, while Sanskrit first begins in the second century A.D. From the sixth century onwards Sanskrit prevails exclusively (except among the Jains) in inscriptions, though Prākritisms often occur in them. Even in the literature of Buddhism the mixed dialect was gradually supplanted by Sanskrit. Hence most of the Northern Buddhist texts have come down to us in Sanskrit, which, however, diverges widely in vocabulary from that of the sacred texts of the Brahmans, as well as from that of the classical literature, since they are full of Prākrit words. It is expressly attested by the Chinese pilgrim, Hiouen Thsang, that in the seventh century the Buddhists used Sanskrit even in oral theological discussions. The Jains finally did the same, though without entirely giving up Prākrit. Thus by the time of the Muhammadan conquest Sanskrit was almost the only written language of India. But while Sanskrit was recovering its ancient supremacy, the Prākrits had exercised a lasting influence upon it in two respects. They had supplied its vocabulary with a number of new words, and had transformed into a stress accent the old musical accent which still prevailed after the days of Pāṇini.