The Indus script, also called the Harappan script and the Indus Valley script, is a collection of symbols created by the Indus Valley Civilization. Most writings with these symbols are very short, which makes it hard to know if they formed a writing system used to record a language spoken by the Harappan people. No language from this civilization has been identified yet. Many attempts have been made to understand the script, but it has not been deciphered. There are no known writings that include two languages to help translate the script, and the symbols have not changed much over time. However, the way the symbols are arranged sometimes differs depending on where they were found.
The first recorded drawing of a seal with Harappan symbols was made in 1875 by Alexander Cunningham. By 1992, about 4,000 objects with inscriptions had been discovered, some as far away as Mesopotamia because of trade between the Indus Valley and Mesopotamia. These inscriptions include more than 400 different symbols.
Some scholars, including G. R. Hunter, S. R. Rao, John Newberry, and Krishna Rao, have suggested that the Brahmi script might be connected to the Indus script. Raymond Allchin has cautiously agreed that the Brahmi script could have been influenced by the Indus script, but this idea has not been proven. Another possibility is that the Indus tradition continued in the megalithic graffiti symbols found in southern and central India and Sri Lanka. These symbols likely are not a written language but may share some similarities with the Indus symbols. Linguists such as Iravatham Mahadevan, Kamil Zvelebil, and Asko Parpola have suggested that the Indus script may be related to a Dravidian language.
Corpus
By 1977, at least 2,906 objects with readable inscriptions had been found, and by 1992, about 4,000 inscribed objects had been discovered. In 2025, it was reported that around 5,000 inscriptions had been found since 1924.
Indus script symbols have mostly been found on stamp seals, pottery, bronze and copper plates, tools, and weapons. Most of the written material consists of seals, impressions of these seals, and markings on pottery. Seals and their impressions were usually small and easy to carry, with most measuring about 2–3 centimeters on each side. No examples of the Indus script have been found on materials that can rot or decay, such as papyrus, paper, textiles, leaves, wood, or bark.
Early examples of the Indus script were found on pottery and clay impressions of Harappan seals from around 2800–2600 BCE during the Early Harappan period. These symbols appeared alongside objects used for administration, such as seals and standard weights, during the Kot Diji phase. Excavations at Harappa showed that some symbols developed from marks made by potters and graffiti from the earlier Ravi phase, which dates to about 3500–2800 BCE.
During the Mature Harappan period (about 2600–1900 BCE), strings of Indus symbols were often found on flat, rectangular stamp seals and written on many other objects, including pottery, tools, tablets, and ornaments. Symbols were created using methods like carving, chiseling, embossing, and painting on materials such as terracotta, sandstone, soapstone, bone, shell, copper, silver, and gold. By 1977, Iravatham Mahadevan noted that about 90% of the Indus script seals and objects found were from sites in Pakistan along the Indus River, such as Mohenjo-daro and Harappa, while the remaining 10% came from other locations. Many seals also included images of animals like bulls, water buffaloes, elephants, rhinoceroses, and the mythical "unicorn," possibly to help people who could not read identify the seal’s origin.
The Late Harappan period (about 1900–1300 BCE) followed the more developed Mature Harappan period and was a time of decline and local changes before the early Iron Age. Inscriptions from this period have been found at sites linked to its local phases. At Harappa, the use of the script stopped around 1900 BCE as inscribed seals were no longer used, but in other areas like Rangpur and Gujarat, the script may have continued, especially in graffiti on pottery. Seals from the Jhukar phase of the Late Harappan period, located in present-day Sindh, Pakistan, do not include the Indus script, but some pottery fragments from this time have inscriptions. Both seals and pottery fragments with Indus script, dated to around 2200–1600 BCE, have been found at sites connected to the Daimabad culture in present-day Maharashtra.
Many artifacts, especially pottery fragments and tools with markings, have been found in Central India, South India, and Sri Lanka from the Megalithic Iron Age, which came after the Late Harappan period. These markings include inscriptions in the Brahmi and Tamil-Brahmi scripts, as well as non-Brahmi symbols that existed at the same time as the Tamil-Brahmi script. Scholars have not reached a clear agreement on the meaning of these non-Brahmi symbols. Some researchers, like Gregory Possehl, suggest that these symbols may be a continuation of the Indus script into the 1st millennium BCE. In 1960, archaeologist B. B. Lal found that many megalithic symbols he studied were similar to those in the Indus script, showing a possible connection between the Indus Valley Civilization and later Megalithic cultures. Similarly, Iravatham Mahadevan has argued that sequences of Megalithic graffiti symbols match those on Harappan inscriptions, suggesting that the language of Iron Age people in South India may have been related to or the same as that of the late Harappans.
Characteristics
The symbols in the Indus script are mostly picture-like, showing objects from the ancient world, items found in the Harappan culture, or things from nature. However, many symbols are abstract. Some symbols are made by combining simpler picture-like symbols, while others only appear as parts of more complex symbols. Some symbols look like tally marks and may be early forms of numbers.
There are more than 400 main symbols, which is too many to be used for sounds alone. Because of this, most experts believe the script is a mix of symbols and syllables. The exact number of symbols is unclear, as some symbols may be different versions of the same symbol. In the 1970s, an expert named Iravatham Mahadevan listed 419 unique symbols. However, in 2015, another expert, Bryan Wells, estimated there were about 694 unique symbols.
Of Mahadevan’s symbols, 113 appear only once, 47 appear twice, and 59 appear fewer than five times. Only 67 symbols make up 80% of all symbols used. The most common symbol is the "jar" symbol, labeled as "sign 311" by scholar Asko Parpola.
Most experts believe the Indus script was read from right to left, though some examples show it was written left to right or in a zigzag pattern. Even though the script has not been fully translated, the writing direction was determined by evidence, such as symbols being compressed on the left side, suggesting the writer ran out of space. On seals, which leave a mirror image on clay or ceramic, the symbols are read from right to left, similar to other inscriptions.
Some researchers have compared the Indus script to the Brahmi script, suggesting they might be related. Others have compared it to scripts from Mesopotamia and the Iranian plateau, like Sumerian proto-cuneiform and Elamite scripts. However, experts now believe the Indus script is not closely related to any other writing systems from 2000 to 3000 BCE, though some similarities with Proto-Elamite may exist. A clear connection to another script has not been proven.
Researchers have also compared the Indus script to Brahmi and Tamil-Brahmi scripts, noting possible similarities. Early scholars, such as John Marshall and Stephen Langdon, suggested these similarities, with some, like G. R. Hunter, proposing Brahmi may have developed from the Indus script.
The Indus script has also been compared to the Proto-Elamite script from Elam, an ancient civilization that existed at the same time as the Indus Valley civilization. Both scripts were mostly picture-like. Around 35 Proto-Elamite symbols may be similar to Indus symbols. In 1932, G. R. Hunter argued that the similarities between the two scripts were too close to be a coincidence.
Theories and attempts at decipherment
The following factors are often considered the greatest challenges in successfully decoding the Indus Valley Script:
Many attempts to decode the script have been made over the years, but experts do not agree on a single solution. Some areas where experts do agree include the direction of most inscriptions (written from right to left), the numerical nature of certain symbols, the similar function of some ending symbols, and methods used to divide inscriptions into groups. More than 100 different decoding attempts have been published since the 1920s, and the topic interests many non-experts.
In 2025, the Chief Minister of Tamil Nadu, M. K. Stalin, announced a $1 million prize for decoding the Indus Valley Script. He said, "Archaeologists, computer experts from Tamil Nadu, and other experts worldwide have tried to decode the script, but it remains a mystery even after 100 years."
Although no clear agreement exists, some experts believe the script recorded an early form of the Dravidian languages (Proto-Dravidian). Early supporters included archaeologist Henry Heras, who proposed interpretations of symbols based on Dravidian language assumptions.
Using computer analysis, Russian scholar Yuri Knorozov suggested the Dravidian language was likely the base for the script. In the 1960s–80s, Finnish scholar Asko Parpola led a team that used computer analysis to study the script, similar to Knorozov’s work. Parpola concluded the script likely belonged to the Dravidian language family. His research is detailed in his book Deciphering the Indus Script. Archaeologist Walter Fairservis argued that script on seals might show names, titles, or jobs, and that animals on seals could represent family groups. Computational linguist Rajesh P. N. Rao and colleagues used computer analysis to show the script has a structure similar to written language, like Old Tamil.
Some scholars have proposed meanings for symbols. For example, the Dravidian words for "fish" and "star" (mīn) were linked to drawings on Harappan seals. In 2011, Rajesh P. N. Rao noted that Iravatham Mahadevan and Asko Parpola had made progress but said their proposed meanings were not yet proven.
In 2014, epigraphist Iravatham Mahadevan identified a repeated group of four symbols he interpreted as "Merchant of the City" in an early Dravidian phrase. He stated he had not fully decoded the script but believed his work showed the script was Dravidian.
Indian archaeologist S. R. Rao argued the script recorded an early Indo-Aryan language. He compared Harappan symbols to Phoenician letters and claimed the Phoenician script evolved from Harappan symbols, not from the Proto-Sinaitic script as previously thought. He read the script left to right, like the Brahmi script, and said the script included numbers and was "Sanskritic." Rao’s ideas supported views that Indo-Aryan people were early inhabitants of the Indian subcontinent, but experts like Parpola pointed out differences between Indus culture and Indo-European cultures, such as the importance of horses in later cultures.
Another theory suggests the script belongs to the Munda language family, spoken in central and eastern India. However, Munda vocabulary does not match Harappan culture, making this theory less likely.
Some scholars argue the symbols are not language but represent families, clans, gods, or religious ideas, like parts of coats of arms. In 2004, Steve Farmer, Richard Sproat, and Michael Witzel claimed the script was non-linguistic, citing short inscriptions, too many rare symbols, and lack of repeated symbols. Parpola disagreed, noting similar rare symbols exist in Chinese writing and that short texts need not repeat symbols.
A 2009 study by Rajesh P. N. Rao, Iravatham Mahadevan, and others in Science compared the complexity of Indus symbols to linguistic systems like Sumerian or Sanskrit. They said this similarity does not prove the script is linguistic. Later studies examined longer symbol patterns, but Sproat argued the model used could produce similar results for non-linguistic systems like Mesopotamian symbols. Rao and others defended their work, and Sproat’s responses were published in Computational Linguistics. A 2014 paper in Language provided further evidence about the script’s nature.
Unicode
The Indus symbols have been given the ISO 15924 code "Inds." In 1999, Michael Everson proposed adding the Indus script to Unicode's Supplementary Multilingual Plane, but the proposal was not accepted by the Unicode Technical Committee. As of February 2022, the Script Encoding Initiative still lists the Indus script among those not officially included in the Unicode Standard (and ISO/IEC 10646).
The Indus Script Font is a Private Use Areas (PUA) font that represents the Indus script. It was created using data collected by Indologist Asko Parpola in his book Deciphering the Indus Script. Amar Fayaz Buriro, a language specialist, and Shabir Kumbhar, a font designer, were asked by the National Fund for Mohenjo-daro to develop the font. They presented it at an international meeting about Mohenjo-daro and the Indus Valley Civilization on February 8, 2017.