Every week, Rajesh PN Rao, a computer scientist, gets emails from people who claim to have cracked an ancient script that has baffled scientists for generations.
These self-proclaimed codebreakers – ranging from engineers and IT workers to retirees and tax officials – are mostly from India or of Indian origin living abroad. All of them are convinced that they have deciphered the script of the Indus Valley Civilization, a mixture of signs and symbols.
“They claim that they have solved it and that it is ‘case closed,'” says Mr. Rao, Hwang Endowed Professor at the University of Washington and author of peer-reviewed studies of Indus script.
Adding fuel to the race, MK Stalin, chief minister of the southern Indian state of Tamil Nadu, recently raised the stakes, announcing a $1 million reward for anyone who could crack the code.
The Indus or Harappan civilization – one of the earliest urban societies in the world – emerged 5,300 years ago in present-day northwestern India and Pakistan. Its austere farmers and merchants, who lived in walled cities of baked brick, thrived for centuries. Since its discovery a century ago, around 2,000 sites have been discovered across the region.
The reasons behind the society’s sudden decline remain unclear, with no obvious evidence of war, famine or natural disaster. But its greatest mystery is the undeciphered script, which keeps its language, governance and beliefs shrouded in mystery.
For more than a century, experts – linguists, scientists and archaeologists – have been trying to crack the Indus script. Theories linked him to early Brahmi scriptDravidian and Indo-Aryan languages, Sumerianhe even claimed that it consisted only of political or religious symbols.
However, its secrets remain locked away. “The Indus script is perhaps the most important writing system that has not been deciphered,” says Asko Parpola, a leading Indologist.
These days, the more popular spectacular theories identify the script with content from Hindu scriptures and attribute spiritual and magical meanings to the inscriptions.
Most of these attempts ignore that the script, composed of signs and symbols, appears mostly on stone seals used for trade and commerce, making it unlikely that they contain religious or mythological content, according to Mr. Rao.
There are many challenges in deciphering the Indus script.
First, the relatively small number of letters – about 4,000 of them, almost all on small objects such as seals, ceramics and tablets.
There is also the brevity of each script – an average length of about five characters or symbols – with no long texts on walls, plaques or upright stone slabs.
Consider the common square seals: rows of characters run along their top, with a central motif of an animal – often a unicorn – and an object beside it, the meaning of which remains unknown.
There is also no bilingual artifact like Rosetta Stonewho helped scientists decipher Egyptian hieroglyphs. Such artifacts contain text in two languages, offering a direct comparison between a known and an unknown script.
Recent advances in deciphering the Indus script have used computer science to solve this ancient enigma. The researchers used machine learning techniques to analyze the script, trying to identify patterns and structures that could lead to understanding it.
Nisha Yadav, a researcher at the Mumbai-based Tata Institute of Fundamental Research (TIFR), is one of them. In collaboration with scientists like Mr. Rao, her work focused on applying statistical and computational methods to analyze the undeciphered script.
Using a digitized dataset of Indus script characters, they found interesting patterns. A caveat: “We still don’t know if the characters are whole words, parts of words or sentences,” Ms Yadav says.
Ms. Yadav and co-researchers found 67 characters that make up 80% of the text on the script. A sign that looks like a jar with two handles turned out to be the most used sign. Also, scripts started with a large number of characters and ended with fewer. Some character patterns appear more often than expected.
Also, a model of machine learning of the script was created in order to restore illegible and damaged texts, opening the way for further research.
“Our understanding is that the script is structured and there is an underlying logic to the writing,” says Ms Yadav.
Certainly, several ancient scripts remain undeciphered, facing challenges similar to the Indus script.
Mr. Rao cites scenarios like Proto-Elamite (Iran), Linear A (Crete), i Etruscan (Italy), whose basic language is unknown.
Others, like Rongorongo (Easter Island) i Zapotec (Mexico), they knew the languages, “but their symbols remain obscure”. The Phaistos Disc of Crete – a mysterious baked clay disc from the Minoan civilization – “closely reflects the challenges of the Indus script – its language is unknown, and there is only one known example”.
In India, it is not entirely clear why Mr. Stalin of Tamil Nadu offered a reward for deciphering the letter. His announcement follows a new study linking Indus Valley signs to graffiti found in his state.
K Rajan and R Sivananthan analyzed more than 14,000 graffiti pottery fragments from 140 excavated sites in Tamil Nadu, which included more than 2,000 characters. Many of these characters are very similar to those of the Indus script, with 60% of the characters matching, and more than 90% of the characters of South Indian graffiti having “parallels” to those of the Indus civilization, according to the researchers.
This “suggests some kind of cultural contact” between the Indus Valley and southern India, Mr Rajan and Mr Sivananthan say.
Many believe that Stalin’s move to announce the award positions him as a staunch champion of Tamil heritage and culture, opposing Narendra Modi’s Bharatiya Janata Party (BJP), which rules Delhi.
But the researchers are convinced that soon there will be no more candidates for the Stalin Prize. Scientists have compiled complete, up-to-date databases of all known printed artifacts – key to decipherment. “But what did the Indus people write? I wish we knew,” says Ms Yadav.
