Indus Valley: A million-dollar challenge to crack the script of early Indians

Each week, computer scientist Rajesh P.N. Rao receives emails from various individuals claiming they have cracked an ancient script that has puzzled scholars for generations. These self-proclaimed decipherers, including engineers, IT professionals, retirees, and tax officials, mostly from India or the Indian diaspora, are all convinced they have deciphered the Indus script, a writing system composed of symbols and signs.

Mr. Rao, the Huang Endowed Professor at the University of Washington and author of peer-reviewed research on the Indus script, says, "They are claiming that they have solved it, 'case closed.'" The race to decipher the script has arguably been intensified by a recent announcement from the chief minister of Tamil Nadu, M.K. Stalin, in southern India, who offered a $1 million reward to anyone who can crack the script.

The Indus civilization, also known as the Harappan civilization, was one of the world's earliest urban societies, emerging 5,300 years ago in what is now northwest India and Pakistan. The civilization's humble farmers and merchants lived in walled, brick-built cities and thrived for centuries. Since its discovery a century ago, around 2,000 sites have been found in the region. The reason for the society's sudden decline is unclear, with no apparent evidence of war, famine, or natural disasters. But its biggest mystery is its undeciphered script, which keeps its language, governance, and beliefs shrouded in mystery.

For over a century, experts such as linguists, scientists, and archaeologists have been trying to crack the Indus script. Various theories have linked it to early Brahmi script, Dravidian and Indo-European languages, Sumerian, and even some claims that it's just made up of political or religious symbols. Yet, its secrets remain locked away. "The Indus script is probably the most important undeciphered writing system," said the renowned Indologist Asko Parpola.

Today, more popular, astonishing theories equate the script to the content of Hindu scriptures, attributing spiritual and magical meanings to the inscriptions. Mr. Rao believes that most of these attempts overlook the fact that the script, composed of symbols and signs, appears primarily on stone seals used for trade and commerce, making religious or mythological content unlikely. There are many challenges to deciphering the Indus script.

Firstly, the corpus of text is relatively small, with around 4,000 texts, nearly all appearing on small objects such as seals, pottery, and tablets. Secondly, each text is short, averaging around five symbols or signs in length, with no long texts found on walls, tablets, or upright stone slabs. Take a common square seal, for example: a line of symbols runs along the top, with an animal motif (usually a unicorn) in the middle and an object of unclear meaning next to it.

Moreover, there is no bilingual artifact like the Rosetta Stone, which helped scholars decode Egyptian hieroglyphs. Such artifacts contain text in two languages, allowing for a direct comparison between the known and unknown scripts. Recently, progress in deciphering the Indus script has leveraged computer science to tackle this ancient enigma. Researchers are using machine learning techniques to analyze the script, trying to identify patterns and structures that may aid understanding.

Nisha Yadav, a researcher at the Tata Institute of Fundamental Research (TIFR) in Mumbai, is one such person. She is collaborating with scientists like Mr. Rao, focusing on applying statistical and computational methods to analyze the undeciphered script. Using digitized datasets of Indus symbols from the texts, they have found interesting patterns. “We still don’t know if the symbols are complete words, or part of words, or part of sentences,” Ms. Yadav cautioned.

Ms. Yadav and her colleagues have identified 67 symbols that account for 80% of the script. A symbol that looks like a jar with two handles was found to be the most commonly used. Also, the texts start with a high number of symbols and end with fewer symbols. Some patterns of symbols appear more frequently than expected. Machine learning models of the script have also been created to recover illegible and damaged texts, paving the way for further research. “Our understanding is that the script is structured and there is an underlying logic to the writing,” Ms. Yadav said.

To be sure, several ancient scripts remain undeciphered, facing similar challenges to the Indus script. Mr. Rao cites scripts such as Proto-Elamite (Iran), Linear A (Crete), and Etruscan (Italy), where their underlying languages are unknown. Other scripts, such as Rongorongo (Easter Island) and Zapotec (Mexico), have known languages, “but their symbols remain unclear.” The mysterious fired clay disk from the Minoan civilization, the Phaistos Disk, "is very similar to the challenges of the Indus script—its language is unknown and there is only one known example."

Back in India, it is not entirely clear why Tamil Nadu's Mr. Stalin announced the reward for deciphering the script. His announcement came after a new study linked Indus Valley symbols to graffiti found in his state. K. Rajan and R. Sivanandam analyzed more than 14,000 pottery shards with graffiti, including over 2,000 symbols, from 140 excavated sites in Tamil Nadu. Many of these symbols are very similar to those in the Indus script, with 60% of the symbols matching, and over 90% of the South Indian graffiti markings having “similarities” with those of the Indus civilization, the researchers claimed.

Rajan and Sivanandam stated that this “suggests some kind of cultural link between the Indus Valley and South India.” Many believe that Mr. Stalin's reward announcement positions him as a strong champion of Tamil culture and heritage, in opposition to Narendra Modi's Bharatiya Janata Party (BJP) ruling in Delhi. But researchers believe that Mr. Stalin's prize will not be claimed anytime soon. Scholars have compiled complete, updated databases of all known inscribed artifacts—which is crucial for decipherment. “But what did the Indus people write? I wish we knew,” Ms. Yadav said.