1. Sources:

[xiii] Pāli literature, in fact [the] ancient literature of India is a vast treasure-house of information with regard to the geographical condition and situation of the numerous cities, countries, villages and other localities as well as of rivers, lakes, parks, forests, caityas, vihāras, etc., of the vast continent of India. It is not unoften that such geographical information is supplemented by historical accounts of interest as well; and when they are collated together, we have before us a picture of the entire country of the times of which this literature may be said to have a faithful record.

Early Pāli literature is mainly canonical relating in most cases to rules and regulations of conduct of the monks of the Order as well as of the laity. Incidentally there are also Jātakas or birth stories of the Buddha as well as many other anecdotes and narratives having obviously an aim or purpose. Texts or narratives of purely historical or geographical nature are thus altogether absent in the literature of the early Buddhists; and whatever historical or geographical information can be gathered are mainly incidental and, therefore, more reliable.

From a time when Indian History emerges from confusion and uncertainties of semi-historical legends and traditions to a more sure and definite historical plane, that is from about the time of the Buddha to about the time of Asoka the Great, the canonical literature of the early Buddhists is certainly the main, if not the only, source of all historical and geographical information of ancient India supplemented, however, by Jaina and Brahmanical sources here and there.

Thus, for the history of the rise and vicissitudes as well as for the geographical situation and other details of the Soḷasa Mahājanapadas, the sixteen Great States, the most important chapter of Indian History and geography before and about the time of the Buddha, the Pāli Aṅguttara Nikāya is the main and important source of information which, however, is supplemented by that contained in the Jaina Bhagavatī Sūtra and in the Karṇaparva of the Mahābhārata. Cf. PHAI., p. 60

Even for later periods when epigraphical and archaeological sources are abundant, and literary sources are mainly brahmanical or are derived from foreign treatises such as those of the Greek geographers and Chinese travellers, the importance of geographical information as supplied by Pāli and Sanskrit Buddhist sources is considerable.

The commentaries of Buddhaghosa and the Ceylonese chronicles – Dīpavaṁsa and Mahāvaṁsa – for instance, contain information [xiv] with regard to the contemporary geography of India whose value can hardly be overestimated. The non-canonical Pāli and Sanskrit Buddhist literature belong no doubt to a later date, but being mostly commentaries on older texts, or treatises of a historical nature they speak of a time when Buddhism had just launched on its eventful career and was gradually gaining new converts and adherents. The information contained in them is, therefore, almost equally useful and trustworthy.

It has already been said that early Pāli literature is mainly canonical. The huge bulk of texts included in it contains in each of them incidental references to cities and places in connection with the gradual spread of Buddhism mainly within the borders of Majjhimadesa or the Middle Country and the localities bordering it. For such information, the Vinaya Piṭaka is a most important source and it is here perhaps for the first time that we find an accurate description of the four boundaries of the Madhyadeśa as understood by the Buddhists of the time.

No less important are the Dīgha, the Majjhima and the Aṅguttara Nikāyas of the Suttapiṭaka wherefrom can be gleaned a systematic survey of the entire geographical knowledge of the Middle Country, as well as of some other localities of Northern and Southern India.

The Jātakas also contain incidental references to places and localities which add to our geographical knowledge of Buddhist India. Such incidental references can also be found in almost each and every treatise, early or late, canonical or non-canonical. But of non-canonical literature which introduces us to important geographical notices, mention should be made of the Milindapañho or the questions of King Milinda, and the Mahāvastu, a Buddhist Sanskrit work of great importance. Of later texts, the most important from our point of view are the commentaries of Buddhaghosa and some of his colleagues. Mention must also be made of the two important Ceylonese Chronicles – the Dīpavaṁsa and Mahāvaṁsa as well as the huge Commentary literature of Ceylon and Burma.

Other sources from which we can gather chips of information as to the geographical knowledge of the early Buddhists may be mentioned the inscriptions of Asoka and those at the Khaṇḍagiri and Udayagiri hills of Orissa. Coins too, sometimes, enable us to locate a particular nation or tribe, as for example, the location of the kingdom of King Sivi of the Sivi Jātaka has been determined by the discovery of some copper coins at Nāgri, a small town 11 miles north of Chitor.

Chinese Buddhist texts, especially the itineraries of travellers, though later in date, are of inestimable value as sources of the geography of Buddhist India. Of the various Chinese accounts, those of Song-yun and Hwiseng are short and describe only a few places of North-Western India. I-tsing [xv] who landed at Tāmralipti (or modern Tamluk in Midnapur) in A.D. 673, gives us a more detailed account. He visited Nālandā, Gijjhakūṭa, Buddhagayā, Vesālī, Kuśīnagara, Kapilavastu, Sāvatthī, Isipatana Migadāva and the Kukkuṭapabbata.

But more important are the accounts of Fa-Hien and Yuan Chwang. Fa-Hien entered India from the north-west (399–414 A.D.), toured all over northern India and left it at the port of Tāmralipti. Yuan Chwang also covered the same tract (629–645 A.D.), but his account is fuller and more exhaustive. The geographical notices of both the pilgrims are precise and definite, and for one who wants to get a correct and exhaustive idea of the geography of Northern India during the fourth and seventh centuries of the Christian era, they are, in fact the most important sources of information. But as we are here concerned with the geography primarily of the early Buddhists, We shall turn to them only when they would enable us and help us to explain earlier notices and information.

It will be noticed that in the earlier canons and texts as well as in those later texts and canons that speak of earlier times, Majjhimadesa is the country par excellence that is elaborately noticed. Its towns and cities, parks and gardens, lakes and rivers have been mentioned time and again. Its villages have not even been neglected. Repetitions of the same information are often irritating and it seems that the Middle Country was almost exclusively the world in which the early Buddhists confined themselves. That was, in fact, what happened.

It was in an eastern district of the Madhyadeśa that Gotama became the Buddha, and the drama of his whole life was staged on the plains of the Middle Country. He travelled independently or with his disciples from city to city, and village to village moving as if it were within a circumscribed area. The demand near home was so great and insistent that he had no occasion during his lifetime to stir outside the limits of the Middle Country. And as early Buddhism is mainly concerned with his life and propagation of his teaching, early Buddhist literature, therefore, abounds with geographical information mainly of the Majjhimadesa within the limits of which the first converts to the religion confined themselves.

The border countries and kingdoms were undoubtedly known and were oftentimes visited by Buddhist monks, but those of the distant south or north or north-west seem to have been known only by names handed down to them by traditions. Thus the Mahājanapadas of Gandhāra and Kamboja were known, but they hardly had any direct and detailed knowledge about them. Of the south, they hardly knew any country beyond Assaka, Māhissati (Avanti Dakṣiṇāpatha), Kaliṅga and Vidarbha. But with the progress of time as Buddhism spread itself beyond the boundaries of the Middle Country, and its priests and preachers were out for making new converts, their geographical [xvi] knowledge naturally expanded itself, and by the time Asoka became Emperor of almost the whole of India, it had come to embrace not only Gandhāra and Kamboja on one side, and Puṇḍra and Kaliṅga on the other, but also the countries that later on came to be occupied by the Cheras, Choḷas and Pāṇḍyas.

The position of the early Buddhists as regards their geographical knowledge may thus be summarised:– they were primarily concerned with the Middle Country, the cradle of the Buddha and Buddhism, but even as early as the Buddha’s time they knew the entire tract of country from Gandhāra-Kamboja to Vaṅga, Puṇḍra and Kaliṅga on one side, and from Kāśmīr to Assaka, Vidarbha and Māhissati on the other. But knowledge of these outlying tracts of country were not as intimate, and they come to find mention in the earlier texts only when their incidental relations with the Middle Country are related or recalled.

2. Buddhist Conception of India.

The Brahmanical conception of the world has been vaguely preserved in the Epics and the Purāṇas wherein the world is said to have consisted of seven concentric islands – Jambu, Sāka, Kusa, Sālmala, Kraunca, Gomeda, and Pushkara – encircled by seven samudras, the order, however, varying in different sources. Of these seven islands, the Jambudīpa is the most alluded to in various sources and is the one which is generally identified with Bhāratavarṣa, or the Indian Peninsula.

Jambudīpa is one of the four Mahādīpas or the four great continents including India. When opposed to Sīhaladīpa, Jambudīpa means the continent of India as Childers points out (Pāli Dictionary, p. 165). The ancient name of India according to the Chinese was shin-tuh or sindhu (Legge’s Fa-Hian, p. 26). Jambudīpa is called a vana or forest. Papañcasūdanī, II, p.423 (P.T.S).

It is recorded in the Visuddhimagga that a single world-system is 1,203,450 yojanas in length and breadth, and 3,610,350 yojanas in circumference. Within this world-system lies this earth (Vasundharā) which is 24 nahutas Nahuta = ten thousand. in thickness. The wind girt water flows 48 nahutas in thickness; the wind climbs for ninety-six myriad yojanas unto the lower ether.

The highest of the mountain peaks is the Sineru which sinks 84,000 yojanas in the great deep and ascends to the same height. The Sineru is compassed by seven celestial ranges named Yugandhara, Isadhara, Kāravika, Sudassana, Nemindhara, Vinataka and Assakanna. The Himavā is 500 yojanas in height and 3,000 yojanas in length and breadth. It is crowned with 84,000 peaks. The Jambudīpa has been named after the Jambu tree which others name Naga (Vis. M., I, pp. 205–206; cf. VT., I, p. 127 and Asl., p. 298).

Buddhaghosa points out that [xvii] Jambudīpa is 10,000 yojanas in extent and it is called mahā or great (Smv., II, p. 429). Of these 10,000 yojanas, 4,000 are, according to Spence Hardy, covered by the ocean, 3,000 by the forest of the range of the Himalayan mountains and 3,000 are inhabited by men (Manual of Buddhism, p. 4). He further points out that the five great rivers, Gaṅga, Yamunā, Aciravatī, Sarabhū and Mahī, after watering Jambudīpa, fall into the sea (Ibid., p. 17).

Jambudīpa has 500 islands (Ibid., p. 449). In the earlier ages, there were 199,000 kingdoms in Jambudīpa, in the middle ages, at one time, 84,000 and at another, 63,000; and in more recent ages about a hundred. In the time of Gotama Buddha this continent contained 9,600,000 towns, 9,900,000 seaports, and 56 treasure cities (Ibid., p. 4).

The Dīgha Nikāya of the Suttapiṭaka narrates that the Exalted One, while relating the Cakkavattsīhanāda Suttanta, predicted thus: ‘Jambudīpa will be mighty and prosperous, the villages, towns and royal cities will be so close that a cock could fly from each one to the next.’

This Jambudīpa – one might think it a ‘Waveless Deep’ – will be pervaded by mankind even as a jungle is by reeds and rushes. In this Continent of India there will be 84,000 towns with Ketumatī (Benares), the royal city, at their head (DN., III, p. 75).

We learn from the Aṅguttara Nikāya that in Jambudīpa trifling in number are the pleasant parks, the pleasant groves, the pleasant grounds and lakes, while more numerous are the steep precipitous places, unfordable rivers, dense thickets of stakes and thorns and inaccessible mountains (Vol. I, p. 35).

We are informed by the Papañcasūdanī that gold is collected from the whole of Jambudīpa (II, p. 123). The Dīpavaṁsa records that Asoka built 84,000 monasteries in 84,000 towns of Jambudīpa (p. 49). This is supported by the Visuddhimagga which states that Asoka, the Great King, put up 84,000 monasteries in the whole of Jambudīpa (Vol. I, p. 201).

The Milinda Pañho (p. 3) informs us that in Jambudīpa many arts and sciences were taught, e.g. the Sāṅkhya, Yoga, Nyāya and Vaiseśika systems of philosophy; arithmetic, music, medicine, the four Vedas, the Purāṇas and the Itihasas; astronomy, magic, causation, and spells, the art of war; poetry and conveyancing. We learn from the commentary on the Therīgāthā that there were disputants here well versed in arts and sciences (P.T.S., p. 87).

It is interesting to note that merchants made sea-voyages for trade from Jambudīpa. Law, A study of the Mahāvastu, p.128. Once a dreadful famine visited it (Dh.C., III, pp. 368, 370 and 374). There were heretics and bhikkhus here and the unruliness of the heretics was so very great that the bhikkhus stopped holding uposatha ceremony in [xviii] Jambudīpa for seven years (Mv., p. 51).

The importance of Jambudīpa is very great as it was often visited by Gautama Buddha besides Mahinda who paid a visit to it with an assembly of bhikkhus (Dv., p. 65). The whole of Jambudīpa was stirred up by Sānu, the only son of a female lay disciple, who mastered the Tripiṭaka and lived one hundred and twenty years (Dh.C., IV, p. 25). The Kathāvatthu informs us that the people of Jambudīpa led a virtuous life (p. 99). There is a reference to the great Bo-tree at Jambudīpa (Cv., Vol. I, p. 36).

The Buddhist system includes Jambudīpa as one of the islands that comprise the world, but counts eight dvīpas (in stead of seven) and has different names for some of the samud ras. See Pullee’s Studi Italini di Filologia Indo-Iranica, Vol. IV, pp. 15–16. Also see J.R.A.S., 1902, p. 142; 1907, p. 42 and CAGI., Intro., p. XXXVI, and footnote. The Jaina tradition has, however, new names for the several dvīpas as well as for the samudras. The Bhuvanakoṣa section of the Mārkaṇḍeya, Matsya and Purāṇas as well as Bhāṣkarācārya and the Mahābhārata allude to nine divisions of India. Of these nine dvīpas eight have been shown to be divisions not of India proper, i.e. they are not so many provinces of India, but of Greater India, CAGI., App. I, pp. 749–754. and are islands and coun tries that encircle the Indian Peninsula. This Indian Peninsula is the ninth dvīpa which is girt by sea (sāgara-samvritaḥ) and is called Kumārīdvīpa. This description of India is, however, unknown to Buddhist tradition.

Early Buddhist sources are, however, silent about the size and shape of India, though the ancient Indians had a very accurate knowledge of the true shape and size of their country. Alexander’s informants gathered their knowledge from the people of the country, and described India as a rhomboid or unequal quadrilateral in shape, with the Indus on the West, the mountains on the north and the sea on the east and south CAGI., p. 2. . . . . At a somewhat later date the shape of India is described in the Mahābhārata as an equilateral triangle which was divided into four smaller equal triangles Ibid., p. 5. . . . . Another description of India is that of the Navakhaṇḍa or nine divisions which was first described by the astronomers, Parāsara and Varāhamihira, and was afterwards adopted by the authors of several of the Purāṇas. Ibid., pp. 6–7. According to this description, India of the times had the shape of an eight-petalled lotus encircling a round central division. ‘In the geography of Ptolemy, however, the true shape of India is completely distorted, and its most striking feature, the acute angle formed by the meeting of the two coasts of the Peninsula at Cape Comorin is changed to a single [xix] coast line running almost straight from the mouth of the Indus to the mouth of the Ganges.’ CAGI., p.9.

For a Buddhist conception of the shape of India, we have to turn to the Mahāgovinda Suttanta (DN., Il, p. 235), and to the itinerary of Yuan Chwang, the celebrated Chinese traveller. The former authority states that the great earth (i.e. India) is broad on the north whereas in the south it is ‘Sakaṭamukham,’ i.e. has the form of the front portion of a cart, and is divided into seven equal parts. The description of the shape of India as given in the Mahāgovinda Suttanta thus corresponds to a great extent to the actual shape of the country which is broad on the north having the Himalayas extending from east to west and ‘Sakaṭamukham’, i.e. triangular towards the south.

The description of the shape as we read in the Mahāgovinda Suttanta agrees wonderfully with that given by the Chinese author Fah-Kai-lih-to. According to him, the country in shape is broad towards the north and narrow towards the south, a description to which he humorously adds the ‘people’s faces are of the same shape as the country’. Fa-Hien’s travels, trans. by S. Beal, p. 36, note.

The next important information in this connection is derived from Yuan Chwang’s itinerary; and it is interesting to compare his description with those just noted. He describes the shape of the country as a half-moon with the diameter or broadside to the north, and the narrow end to the south. This description, however, is just like what Yuan Chwang’s conception could possibly be; for he did not visit the south; in fact, he hardly crossed the Vindhyas. His travels were thus mainly confined to the north of India which may be said to resemble a half-moon with the Vindhyās as its base and the Himalayas spreading its two arms on two sides as the diameter.

3. Division of India.

Indian literature, whether Buddhist or Brahmanical, divides India into five traditional divisions. These five divisions are clearly stated in the Kāvya Mīmāṁsā (p. 93) :–

‘Tatra Bārāṇasyā parataḥ purvadeśaḥ
Māhiṣmatyā parataḥ Dakṣināpathaḥ
Devasabhāyā parataḥ paschātdesaḥ
Prithudakāt parataḥ Uttarāpathaḥ
Vinasanaprayāgayoḥ Gaṅgā-Yamunāyosca antaraṁ Antaravedī’

To the east of Bārāṇasī is the eastern country; to the south of Māhiṣmatī is the Dakṣiṇāpatha or the Deccan; to the West of Devasabhā (not yet identified) is the Western country; to the north of Prithudaka (modern Pehoa, about 14 miles West of Thaneswar) is the Uttarāpatha or the northern country; and the tract lying between Vinasana and Prayāga, [xx] i.e. the confluence of the Yamunā and the Ganges, is called the Antaravedī.

But when the Kāvyamīmāṁsā says that the Western boundary of the eastern country (Purvadeśa) is Benares, it seems to extend the eastern boundary of Manu’s Madhyadeśa up to Benares. This is exactly what it should be. For, by the time when the Kāvyamīmāṁsā came to be written the Aryans had already outstripped the older limits of the Madhyadeśa and Aryandom had extended up to Benares.

In the Dharmasūtras and Dharmaśāstras, Aryandom, i.e. Āryāvarta, is described to have extended from the region where the river Saraswatī disappears (i.e. the Vinasana of Manu and Kāvyamīmāṁsā) in the west, to the Kālakavana or Black Forest (identified with a locality near Prayāga by S. N. Majumdar; see CAGI., Intro., p. xli, footnote) in the east; and from the Himalayas in the north to the Pāripātra in the south.

The Dharmaśāstra of Manu calls the Āryāvarta of the Sūtras to be the Madhyadeśa or the Middle Country and his boundaries of Aryandom are almost identical. Almost all Brahmanical sources give a description of Madhyadeśa or Āryāvarta, the most important division of India, but very few except the Kāvyamīmāṁsā, as stated above, and the Bhuvanakoṣa section of the Purāṇas give any detail about the four remaining divisions of the country. And this is exactly the case with Buddhist sources as well. A detailed description of the Middle Country is as old as the Vinaya Piṭaka as well as references to Majjhimadesa all over early Pāli texts; but an accurate description of the other divisions of India is not found earlier than Yuan Chwang. The reason is not very far to seek. As with the Brahmanical Aryans, so with the Buddhists, Middle Country was the cradle on which they staged the entire drama of their career, and it is to the description and information of this tract of land (by whatever name they called it) that they bestowed all their care and attention.

Outside the pale of Madhyadeśa there were countries that were always looked down upon by the inhabitants of the favoured region. The five divisions as indicated in the Bhuvanakoṣa section of the Purāṇas are identical with those given in the Kāvyamīmāṁsā.

They are: (a) Madhyadeśa (Central India), (b) Udīcya (Northern India), (c) Prācya (Eastern India), (d) Dakṣiṇāpatha (Deccan), and (e) Aparānta (Western India). The same division of the country into five provinces was adopted by the Chinese as well. ‘In the official records of the Thang dynasty in the seventh century, India is described as consisting of “Five Divisions” called the East, West, North, South and Central, which are usually styled the Five Indies.’ CAGI., p. II.

Yuan Chwang also adopts the same divisions which Cunningham describes as follows: Ibid., pp. 13–14.

[xxi] 1. Northern India comprised the Punjab proper, including Kāśmīr and the adjoining hill States, with the whole of eastern Afghanisthan beyond the Indus and the present Cis-Satlej States to the West of the Saraswatī river.

2. Western India comprised Sindh and Western Rajputana with Cutch and Gujrat, and a portion of the adjoining coast on the lower course of the Narbadā river.

3. Central India comprised the whole of the Gangetic provinces from Thaneswar to the head of the Delta, and from the Himalaya mountains to the hanks of the Narbadā.

4. Eastern India comprised Assam and Bengal proper, including the whole of the delta of the Ganges together with Sambalpur, Orissa, and Ganjam.

5. Southern India comprised the whole of the Peninsula from Nasik on the West and Ganjam on the east to Cape Kumārī (Comorin) on the south, including the modern dis tricts of Berar and Telingana, Mahārāshtra and the Konkan, with the separate States of Hyderabad, Mysore, and Travancore, or very nearly the whole of the Peninsula to the south of the Narbadā and the Mahānadī rivers.

It is thus obvious that the Chinese system of five divisions was directly borrowed, as Cunningham rightly points out, from the Hindu Brahmanical system as described in the Purāṇas and the Kavyāmīmāṁsa. The only difference is that the Antaravedī of the Kavyāmīmāṁsa was replaced by the ‘Middle Country’ (i.e. the Majjhimadesa of early Pāli texts or Mid-India of the Chinese) which included the Western portion of the Prācya country or Eastern India.