The Branching Hand

A polished stone surface stands in a field, taller than a man, its face covered in neat rows of small angular marks. For more than a thousand years people walked past it and could not read a word. The marks were letters once. Then the knowledge of how to sound them died, and the stone went silent.

This is the strange fact at the center of Indian writing. The oldest readable Indian script was, for centuries, unreadable. And from that one script, recovered only in the nineteenth century, descends almost every hand used today from the Himalaya to the islands of Indonesia.

One question runs through this chapter. How did a single ancient script branch into the hundreds of different writing systems of India and Southeast Asia, and how did scholars, long after the old letters had been forgotten, learn to read them again?

The earliest hand

We can fix a starting point with unusual confidence. The earliest securely dated and deciphered Indian writing is the set of edicts ordered by the emperor Ashoka, carved roughly between 268 and 232 BCE on rocks and on tall polished pillars across his territory.

Most of those edicts are in a script called Brahmi, the angular letterform that spread across the bulk of the subcontinent. In the northwest, in the region around modern Pakistan and Afghanistan, the same imperial messages appear in a different script called Kharosthi, a flowing right-to-left hand that grew up under Persian and later Greek contact.

The texts themselves are administrative and ethical, the orders of a ruler. The standard published collection is E. Hultzsch’s Inscriptions of Asoka of 1925, still the reference edition. The author identifies himself by a title, not a personal name. He calls himself Devanampiya Piyadasi, “Beloved of the Gods, He Who Looks On With Kindness.”

For the moment hold that name. It will matter later.

What matters here is the shape of the evidence. We are not guessing about an undated tradition. We have stone, cut on the order of a known emperor, in a known century. From this fixed point the whole family tree of Indian scripts can be traced forward.

The family tree

Brahmi is the ancestor of nearly all the scripts of South and Southeast Asia. Paleographers group them together as the “Brahmic” family. The standard modern survey is Richard Salomon’s Indian Epigraphy of 1998, and what follows is the broad consensus shape it describes. One caution belongs at the front: the exact branchings are debated, and specialists argue over which script gave rise to which. The tree below is the general picture, not a settled map of every twig.

From Brahmi the family splits, broadly, into a northern line and a southern line.

The northern line passes through the Gupta script of the fourth to sixth centuries CE and then fans out. From it come Siddham, the script in which much Buddhist material travelled to East Asia; Sharada, the old script of Kashmir; the Nagari hand that would later be standardized as Devanagari; and an eastern branch that yields the proto-Bengali-Assamese scripts and Odia.

The southern line runs differently. A key node is the Pallava and Grantha tradition of the south, from which descend the Tamil line of scripts, the closely related Telugu and Kannada (both from a shared ancestor), Malayalam, and the Sinhala script of Sri Lanka.

Then the family leaves India entirely. Carried by traders, monks, and the long maritime contact between South India and the lands across the Bay of Bengal, the southern Brahmic scripts seeded the writing systems of Southeast Asia. Khmer, Burmese, Thai, Balinese, Javanese, and the old Philippine Baybayin all sit on branches of this same tree.

Stand back and the scale is plain. A child writing Thai in Bangkok, a clerk filling a Devanagari form in Delhi, a monk copying Sinhala in Kandy: all of them are using descendants of the script Ashoka’s masons cut into rock in the third century BCE.

That is the inheritance. Now the question of how it actually works on the page.

How the letters work

A reader who knows the Latin alphabet often assumes every script does the same job in the same way. The Brahmic scripts do not. They belong to a different type, and the difference is structural.

The technical name is abugida, sometimes called an alphasyllabary. The word abugida was coined by the linguist Peter T. Daniels in 1990, borrowed from the names of the first letters of the Ethiopian script. It names a system that sits between an alphabet and a syllabary.

Here is the core of it. In a pure alphabet like Greek, a consonant letter is just a consonant; you must add a separate vowel letter to make a syllable. In an abugida, every consonant letter already contains a vowel, by default the short a. Write the letter for k, and unless you mark it otherwise, you have written ka.

To get a different vowel you do not swap in a new letter. You attach a mark to the consonant. These vowel-marks are called matras, small strokes and hooks placed above, below, before, or after the base letter to turn ka into ki, ku, ke, and so on.

Two more devices complete the system. When two or more consonants meet with no vowel between them, the letters fuse into a single blended sign called a conjunct, in Sanskrit a samyuktakshara. And when you need a bare consonant with no vowel at all, you add a small stroke called the virama or halant, whose only job is to switch off the inherent a.

So a single written unit in these scripts is usually a whole syllable, consonant plus vowel, built up from a base letter and its marks. That is why the type matters. It shapes how the script looks, how it is taught, and how it was eventually decoded.

And decoding it was a real problem. By the early nineteenth century, no one alive could read the oldest of these letters at all.

The forgotten letters

The script on Ashoka’s pillars had drifted so far from any living hand that learned Indians and European scholars alike stood before it unable to read it. The pillars were famous. The writing on them was a wall.

The man who broke through was James Prinsep, an assay master of the East India Company’s mint in Calcutta and secretary of the Asiatic Society. Working with copies of inscriptions sent to him from across the country, he attacked Brahmi systematically and, in 1837, cracked it.

A breakthrough came from a small, repeated detail. Among the inscriptions Prinsep studied were short donative records at the great stupa of Sanchi, each ending in the same two-character sequence. He reasoned that these were gifts, and that the recurring ending was likely the word danam, “gift, donation.” That gave him secure values for two signs and a foothold to test the rest. Letter by letter the system opened.

Kharosthi, the northwestern script, fell next, in 1837-38. Here the key was bilingual money. The Indo-Greek kings who ruled the northwest struck coins with a Greek legend on one face and a Kharosthi legend on the other, each naming the same ruler. The Greek side was already readable. By matching the king-names across the two scripts, Prinsep could pin down Kharosthi letter values the way a cryptographer reads a cipher against its plain text.

Then came the moment that joined script to history. With Brahmi readable, the edicts everywhere announced their author as Devanampiya Piyadasi. That title was matched, by way of the Ceylonese chronicles that preserved the island’s dynastic memory, with the emperor Ashoka of the Maurya line. The silent stones now had a named voice and a place in time.

The recovery had a price built into it. The script had been lost so thoroughly that it took a mint official, a stack of copied inscriptions, and a few bilingual coins to bring it back. Writing, the thing meant to defeat forgetting, had itself been forgotten.

Sanskrit has no one script

There is a belief, widely held, that Sanskrit “is written in Devanagari,” as though the language and the script were born together. The record does not support it.

Sanskrit has no single native script. It is a language, and a language can be carried by any writing system that can hold its sounds. Across India it was written in whatever hand was local. A Sanskrit manuscript copied in the Tamil country might be in Grantha. The same text in Kashmir might be in Sharada. In Bengal it took the Bengali script, in Nepal the Newari hand, in much of the north the Nagari letters. The words were identical. The shapes on the page were entirely different.

Devanagari itself, the deva-nagari or “script of the city of the gods,” is roughly tenth century CE in origin. For most of its history it was one regional script among many. Its rise to the status of the standard script for Sanskrit is recent. It came with the printing press. In the nineteenth century, as Sanskrit texts began to be printed in volume, Devanagari was adopted as the convenient common type, and the printed page slowly made it look like the natural and only home of the language.

This cuts against a strand of tradition that treats certain scripts, and Devanagari above all, as eternal or divinely given, with the very sound-set of the language, the varnamala, fixed for all time by the Maheshvara-sutras, the fourteen sound-aphorisms traditionally ascribed to Shiva. The careful ordering of those sounds is real and ancient, and we traced its preservation in an earlier volume. But the letters that write them are another matter. Every Indian script in use descends from third-century-BCE Brahmi, and the dominance of Devanagari for printed Sanskrit is a product of the print era, not of eternity.

Hold the distinction clearly. The sounds were guarded for millennia by the spoken tradition that produced the Vedas. The shapes were always local, always changing.

Where Brahmi came from

One question Prinsep’s success could not answer is where Brahmi itself came from. It is still open, and it should be stated honestly because it has drawn more heat than almost any other problem in Indian paleography.

Two broad positions have stood for more than a century. The first is the Semitic hypothesis, argued most fully by Georg Bühler in his On the Origin of the Indian Brahma Alphabet of 1895, which derives Brahmi from a Semitic source, ultimately a North Semitic or Aramaic script carried into India by trade. The second is the indigenous hypothesis, which holds that Brahmi was developed within India, whether freshly invented or grown from earlier local marks.

Salomon’s reading in 1998 is balanced and worth following. A broadly Semitic, most plausibly Aramaic, source for the consonantal core of Brahmi is widely accepted among specialists. But Bühler’s specific derivation, which traced the letters to a particular Phoenician model, is weak in its details and has not held up. Salomon also notes the uncomfortable history of the debate: both camps have at times been driven by bias, the imperial impulse to see India as borrowing its letters, and the nationalist impulse to see them as wholly home-grown.

So the honest answer has a shape. The likeliest source for the bones of the script is Semitic, the exact pathway is unproven, and anyone who declares the matter settled in either direction is going beyond the evidence. We do not know how Brahmi was born. We know, with great precision, what it became.

That second knowledge is enough to carry us forward. We can read the letters now. The next question is what the letters were written on, and why so little of it survived.

The stone in the field is no longer silent. We taught ourselves its alphabet, and in doing so learned how nearly we had lost it. In the next chapter, Leaf and Bark, we turn from the letters to the page, and to the perishable materials that carried India’s words and so easily let them rot away.