Baroda's Oriental Institute Compared Ramayana Manuscripts Dated Back to 1020 CE and Published the Valmiki Ramayana in Seven Volumes Between 1960 and 1975

Between 1960 and 1975 the Oriental Institute of the Maharaja Sayajirao University of Baroda (now Vadodara) published the Valmiki Ramayana, which tradition honours as the first poem (what the Ramayana is), in seven volumes, one for each kanda, or book. Every volume’s title page says the epic was “critically edited for the first time”, and Sheldon Pollock, who translated two of its books into English, also calls it the first critical edition of the poem. Volume one alone rests on 37 manuscripts, chosen from 86 that the editors had collated, meaning compared line by line.

The oldest of them is a palm-leaf copy kept at the Bir Library in Kathmandu. Its colophon, the scribe’s closing note, mentions the year Samvat 1076, which Pandit Baburam Acharya, the library’s former librarian, read as about 1020 CE. Pollock names a palm-leaf manuscript from Nepal, dated about 1020 CE, as the earliest exemplar of the poem.

A department opened on 12 March 1951

Bhatt, the editor of the first volume, tells the start of the project in his preface, dated 11 January 1958. A critical edition of the Ramayana had long been wanted. Hansaben Mehta, the first Vice-Chancellor of the Maharaja Sayajirao University, and the Syndicate (the university’s governing body) accepted a suggestion from the president of the Gujarat Research Society of Bombay, and opened the Ramayana Department in the Oriental Institute on 12 March 1951.

Manuscripts came first. Bhatt writes that the Institute’s director collected them from public and private libraries by appeal and by personal visits. The Collation Section, where manuscripts are compared line by line, opened on 4 April 1954. Bhatt’s preface gives the day as Chaitra Shukla Pratipada, the Hindu New Year, and records that Hansaben Mehta wrote the first verse of the Ramayana on the collation sheets. A Critical Apparatus Section, which records every variant reading under the finished text, was added on 12 July 1955.

Funding came in parts. The University Grants Commission gave Rs 20,000 a year starting in 1955 to 1956, and the preface lists twelve donors, among them the Century Spinning and Manufacturing Company, the Dorabji Tata Trust and Sir Purshottamdas Thakordas, who together gave Rs 1,22,753. Lenders included the Government of Nepal and the Maharajas of Bikaner, Jodhpur and Darbhanga, who allowed manuscripts to be microfilmed or lent. Fourteen libraries and institutes sent manuscripts, from the Adyar Library in Madras to the Bhandarkar Oriental Research Institute in Poona and Visva-Bharati at Santiniketan.

Eighty-six manuscripts in ten scripts

Bhatt’s card index showed more than 2,000 manuscripts of the whole Ramayana, or of parts of it, in libraries in India and abroad. Pollock gives the same figure. For the Bala Kanda, the first book, the department chose 86 on the grounds of age, purity and completeness. They were written in ten scripts: Devanagari (45 manuscripts), Grantha (10), Malayalam (9), Telugu (7), Maithili (4), Bengali (4), Nandinagari (3), Newari (2), Sharada (1) and Kannada (1).

Some of the 86 were dropped as plain copies of older manuscripts in the set, and the Kannada and Nandinagari ones, which agreed with the Telugu, were dropped too. That left 37: 21 from the northern family of manuscripts and 16 from the southern. Twenty-three of the 37 carry dates. Bhatt wrote that the Nepal manuscript is probably the oldest Ramayana manuscript in the world, though damaged in several places. Vaidya, who edited the second book, wrote in 1962 that it was now almost impossible to find a Ramayana manuscript older than 1020 CE. Pollock adds that the earliest evidence he cites is a Cambodian temple inscription of about 600 CE recording the gift of a Ramayana codex.

Two recensions and a family tree

The manuscripts fall into two recensions, or families of text: a northern and a southern. Bhatt grouped the northern one into a north-eastern version (Nepali, Maithili and Bengali manuscripts), a north-western version (Sharada and some Devanagari manuscripts) and a western one. Southern manuscripts are the Telugu, Grantha and Malayalam manuscripts, along with some Devanagari ones.

Bhatt notes that most of the old commentators on the Ramayana came from South India, which he thought helped keep the text there steady. He used five commentaries as supporting evidence, those of Ramanuja, Maheshvaratirtha, Govindaraja, Kataka and Tilaka. Bhatt also looked at where the story is quoted in the Mahabharata and the Puranas. Retellings by Kamban, Krttivasa and Tulsidasa he called interesting and inspiring, and did not use them. Tulsidasa’s retelling has its own place in Tulsidas and the Ramcharitmanas.

Seven rules for choosing a reading

Bhatt wrote that no mechanical rule can build a text, and that a reading cannot be chosen simply because most manuscripts have it. He still set down seven working principles and said he followed in general those of Dr Sukthankar, the scholar behind the Mahabharata critical edition at Poona. His phrase for the two epics is that they “sail in the same boat”. In summary, the principles run like this:

• When the northern and southern recensions agree, the text is accepted.
• When they differ, the southern reading is generally preferred as the older form, unless it is absurd or does not fit the context. The northern reading is taken in that case.
• A passage missing from both recensions, or from either, is suspect and is rejected.
• A passage missing from any version of a recension is dropped when the story does not need it.
• A passage missing from several manuscripts is dropped as secondary, as long as the manuscripts are not closely related.
• A wavy line marks a reading where the manuscripts split evenly.
• Editors should avoid emending the text, as far as possible, in favour of interpretation.

Bhatt kept to the manuscripts. He wrote that the text had to be rebuilt solely on their evidence, leaving the question of how the epic grew to a later stage. He gave a reason for trusting the southern text. In his view, North India, where Sanskrit learning was strong, revised the Ramayana toward the formal standard of classical poetry, while the south kept the older form. His example is the opening verse, where Valmiki asks the sage Narada about the ideal man. Southern manuscripts give Narada three adjectives and Valmiki one, and the northern ones strip Narada of his adjectives and put Valmiki first. The finished Bala Kanda runs to 76 sargas (chapters) and 1,941 stanzas, against 77 sargas and 2,292 stanzas in the printed Vulgate, which Goldman’s team describes as widely known to traditional audiences.

Seven kandas, seven editors

Each volume was edited by a different scholar, under a board whose chairman was the Vice-Chancellor of the university. Title pages give the order and the years:

• Bala Kanda, Bhatt, 1960.
• Ayodhya Kanda, Vaidya, 1962. Vaidya was also general editor of the Mahabharata edition. It is based on 29 manuscripts chosen from 43 collated, dated between 1020 and 1860.
• Aranya Kanda, Divanji, 1963. Its title page calls him the late Divanji.
• Kishkindha Kanda, Mankad, 1965.
• Sundara Kanda, Jhala, 1966, which carries the story of Hanuman crossing to Lanka.
• Yuddha Kanda, Vaidya, 1971.
• Uttara Kanda, Umakant Premanand Shah, 1975.

The Ayodhya volume’s preface, from 1962, says the University Grants Commission had sent an expert committee to Baroda on 28 and 29 June 1960 and was satisfied with the work. On its advice the Commission agreed to pay two-thirds of the cost from 1960 to 1961 through 1966 to 1967, and the Government of Gujarat later sanctioned the remaining third. That preface hoped to finish the whole work in 1967. The last volume appeared in 1975. By 1966 the Sundara volume already lists Bhatt as the late general editor and Shah as secretary and general editor.

The last volume and the Kerala manuscripts

Shah’s introduction to the Uttara Kanda shows the same collecting habit at the end of the project. It is based on 41 manuscripts. The team first collated three Malayalam manuscripts and, after editing the first ten to fifteen sargas, found its southern evidence too thin. Shah and the assistant editor Nambiyar then searched the Kerala University manuscripts library on the spot, and Nambiyar acquired valuable codices for the Institute. In the end it used ten Malayalam manuscripts. Shah, who joined only in 1965 after collation for the other kandas was finished, hopes that a second edition will bring in further Grantha and Malayalam evidence for kandas one to six.

Alongside the seven volumes, the Institute published a Pada-Index in three volumes and a Bare Text of the constituted text in one book. The Institute’s own page says that more than 25 assistants worked on it for 25 years. In the 1960 volume, five illustrations by the university’s Faculty of Fine Arts show Valmiki’s request to Narada, his curse on the hunter who killed a krauncha bird, Dasharatha letting Rama go with Vishvamitra, Rama breaking the bow and the wedding of the brothers. Facsimiles of the Nepali manuscript are printed in the front.

How the text has been used and questioned

Robert Goldman’s team at Berkeley chose the Baroda text for its English translation, published by Princeton University Press, and in 2013 described it as the first and only translation of the Valmiki Ramayana based on that text. Pollock writes that the translators checked the edition’s sources and did not follow it where they thought it wrong. He adds that the Bala Kanda applied the principle of removing doubtful passages less strictly, so the translators kept some probable additions and flagged them in their notes. He also treats the older view that the north polished a southern original as only partly right. In his reading the northern text often rephrases the southern one like an early commentary.

John Brockington, according to Pradip Bhattacharya’s review of his essays, argues that the northern and southern divide does not capture the full chain of transmission, and that the edition’s status as a closed question should be reopened. Shah’s 1975 introduction, for its part, looks ahead to a second edition.