The Bhattiprolu Stupa in Andhra Pradesh Gave Alexander Rea Three Inscribed Relic Caskets in 1892, and Their Brahmi Letters, Placed by Georg Bühler at No Later Than 200 BCE, Write Five Consonants and the Vowel A Differently From Ashoka's Edicts

In the first months of 1892, Alexander Rea of the Archaeological Survey at Madras cut a shaft through the middle of a ruined brick stupa at Bhattiprolu, on the Krishna delta, and found three stone relic caskets at depths of about 14, 17 and 18 feet. All three carried inscriptions. The letters are a southern variety of Brahmi, the script of Ashoka’s edicts, and the epigraphist Richard Salomon describes it as “a unique and radically different southern variety of the script”.

Georg Bühler, who edited the texts in Epigraphia Indica, counted ten inscriptions from the stupa: nine incised on the three caskets and one scratched on a small piece of crystal. They name donors, a guild and a king, and Bühler reads at least two of them as dedications of caskets to hold relics of the Buddha.

Why Rea dug at the centre

South of the village stands the mound. Rea’s book places the village in the Repalle taluk of Krishna district, about four miles from the right bank of the main Krishna channel. Local people called it Lanja dibba. A report of 1870 had already listed it, and in the years after that its large bricks went into road-making and its marble slabs into a sluice on the Krishna canal. An assistant engineer named Norris estimated that the stupa had once stood about 66 feet high.

During that demolition a small stone casket had been found just above the summit, with a clay pot and a soapstone casket inside it and, in the innermost one, a crystal phial holding a pearl, bits of gold leaf and ashes. According to Rea, the stone casket was broken in the lifting and the crystal phial went to a Dr Burnell.

Rea drew a conclusion from how such stupas are laid out. The principal deposit, he wrote, is usually placed in the centre near the foundations, and the summit of the mound stood 15 feet above that level. He found the true centre well to the north of the middle of what remained, under a covering of loose bricks and earth. Around a small central well lay a ring of eight large bricks and outside it a ring of sixteen wedge-shaped ones, so that each course of brickwork, in his words, resembled a lotus flower.

Three caskets at three depths

The first casket lay 14 feet 6 inches down, where a three-sided slab of black stone, about 2 feet 11 inches across, proved to be a lid. Two lines of inscription ran around its stone receptacle. Inside stood a globular stone casket, and inside that a crystal phial containing a flat piece of bone half an inch broad. In and around it lay gold flowers, hollow gold beads, pierced pearls, an amethyst bead and 24 small silver coins laid out in a svastika pattern. Nearby lay a hexagonal crystal pierced along its axis, with letters scratched on its sides, which Rea took to be an amulet worn on a cord.

At 17 feet the second casket had a 19-line inscription on its lid and two more lines on the receptacle. Its crystal phial held no relic that Rea could see, but the cavity contained 164 gold flowers, pearls, a garnet, coral beads and a rolled sheet of thin silver with three lines of letters pricked into it. The third, at 18 feet, had an eight-line inscription under the lid and a miniature beryl casket inside its crystal phial. Three small pieces of bone lay in the beryl casket. Rea put the foundations at 23 feet.

By Rea’s account the finds went to the Madras Museum, and Live History India says the caskets are now in the Egmore museum in Chennai.

What the inscriptions say

Bühler’s translations are plain dedications. Inscription I, on the first casket, says that the father and mother of a man named Kura, Kura himself and a person called Siva ordered a casket and gave a box of crystal “in order to deposit some relics of Budha”. Inscription VI, on the second casket, records a gift by the sons of the Shagathi nigama (a guild or town), whose chief was King Khubiraka, or Kuberaka, son of a man whose name is partly lost. They gave another casket, a crystal box and a stone box.

Inscription III is a list of the members of a committee, called a gothi, with names such as Hiranyavyaghrapad and Sramanadasa. Bühler noted that no king, family or committee named at Bhattiprolu can be traced in any other historical source, so the inscriptions give no outside check on their own date.

Two cautions apply to these readings. Fleet revisited Inscription No. 1 A in the Journal of the Royal Asiatic Society in 1908, and Salomon, writing in 1998, calls the readings of H. Lüders “very different from and much superior to” Bühler’s. Individual names and words have been revised since Bühler’s edition, though the sense of the dedications is not in doubt.

The letters that differ from Ashoka’s

Two peculiarities mark the script, in Salomon’s account. Five consonants, gha, ja, ma, la and sa, are formed differently from standard Brahmi. Ma is upside-down compared with the usual shape, and gha looks like a development of the sign for ga. Second, uniquely among the early Indian scripts, the Bhattiprolu writers dropped the “inherent vowel”, the built-in short a that a Brahmi consonant carries unless told otherwise. A bare consonant stands for the consonant alone, and a short horizontal stroke at the upper right marks a following a. Encoding the script, the Unicode Standard records that the Bhattiprolu inscriptions added a separate sign for long a, made by adding a vertical stroke.

Inscription X, on the crystal, follows Ashoka’s edicts closely, with two small exceptions that Bühler lists.

To Bühler the casket texts carried a bigger lesson. A fully worked-out local system in use by about 200 BCE, he argued, removed a favourite argument of scholars who held that writing reached India under the Maurya dynasty. In Salomon’s view Bühler leaned too hard on the odd consonants, and Bhattiprolu is more likely a provincial offshoot of early Brahmi in the south than a separate line from a Semitic ancestor.

How old the stupa and the script are

Bühler compared the letters with the edicts of Ashoka on one side and with the Nanaghat, Hathigumpha and Bharhut inscriptions, which he put around the middle of the 2nd century BCE, on the other. He found the Bhattiprolu letters in between and closer to Ashoka’s. His conclusion was that they “cannot be placed later than 200 B.C., but may be somewhat earlier”, and he added that this method “may mislead”. Elsewhere on the same page he called them probably a few decades younger than Ashoka’s edicts.

Later dates have their defenders. D. C. Sircar’s date, as Salomon reports it, is about the end of the 2nd century BCE, and the Unicode Standard says the inscriptions are from the 2nd century BCE. J. Halévy, cited by Salomon, argued that they are at least fifty years younger than Ashoka’s.

A popular account goes earlier. Live History India calls Bhattiprolu the earliest stupa in South India, older than Ashoka’s reign by about a hundred years, and gives the caskets’ inscriptions as 3rd to 1st century BCE. Bühler’s estimate puts the inscriptions after Ashoka’s edicts, not a century before them. Rea himself described the solid-domed stupas at Bhattiprolu and Gudivada as “undoubtedly the earliest” of the Krishna district group, and argued that inscriptions buried in the centre of a building date its foundation.

A Dravidian link, and its limits

The Tamil-Brahmi cave inscriptions of Tamil Nadu have features in common with Bhattiprolu. Salomon notes four entirely new characters in them, made for Dravidian sounds that northern Brahmi had no letters for. In the earlier form of that script, he adds, a consonant written without a mark is the bare consonant, as at Bhattiprolu. The letter shapes of the two scripts are not directly related, but Salomon sees a systemic relationship between them. He reads Bhattiprolu as the reapplication to a Prakrit language of a form of Brahmi that had already been modified for Dravidian, noting that Dravidian words commonly end in consonants and Prakrit words do not.

This reading depends on the dates of Tamil-Brahmi, which Salomon calls uncertain and controversial. Iravatham Mahadevan placed the earliest examples in the 2nd and 1st centuries BCE, while Dani, citing potsherds from Arikamedu, dated them to the 1st century CE. Mahadevan’s dating seemed preferable to Salomon. The Unicode Standard reads the relationship the other way, saying the Bhattiprolu spelling “seems to be derived from the Tamil Brahmi system”. For the Tamil poetry of the same centuries, see Sangam literature.

Live History India quotes a Dr Champakalakshmi calling the script a Rosetta Stone for reading Tamil-Brahmi, and says most historians regard it as the source of the Telugu script. Salomon and Bühler are more cautious. Bühler wrote that, apart from the position of the letter da, none of the Bhattiprolu peculiarities left a trace in later Indian alphabets, and Salomon adds that the southern ability to write bare consonants was not carried into later southern scripts.

The stupa itself, and what happened to the relics

In Rea’s survey the dome is 132 feet across and the base, with its surrounding path, at 148 feet, against about 138 feet for the base of the Great Stupa at Amaravati. It was solid brick throughout. Around it he found the marble bases of a railing and carved fragments in the same early style as the stupas at Jaggayyapeta and Garikapadu.

In 1969 to 1970 the Archaeological Survey returned under R. Subrahmanyam. Its review records bricks of 45 by 30 by 8 centimetres, platforms at the four cardinal points and a pathway 1.3 metres wide around the dome. The diameter came out at 47 metres against the roughly 44 metres the review attributes to Rea. Two wings of a vihara (monks’ residence), each of four rooms, were cleared beside the stupa, along with a square cell with curved walls on three sides, which the review calls possibly the earliest form of a chaitya-griha, a prayer hall, in Andhra Pradesh. A stucco head of the Buddha was among the finds. For another early stupa, see Sanchi and the stupa; for the emperor whose edicts are the benchmark for early Brahmi, see Ashoka the Great.

One relic from the stupa went to Calcutta. According to The Week, the British government agreed to give the Maha Bodhi Society a relic found at Bhattiprolu on condition that the society build a suitable vihara in Calcutta, and that the Bengal governor handed it over on 26 November 1920, the year the Dharmarajika Chetiya Vihara was completed. Live History India adds that bone fragments from the Bhattiprolu stupa were enshrined in the Global Vipassana Pagoda at Gorai, Mumbai, in 2006.