The Indus script has been called, with irony, the most deciphered script in the world. The first claim to a decipherment, based on the Sumerian language, was published as early as 1925. More than a hundred published claims have been made since then, including the controversial example published in 2000 in the Deciphered Indus Script: that the Indus language is Vedic Sanskrit. But although there has been some definite progress in understanding the script over the decades, none of the “decipherments” has persuaded anyone other than the proposer and a few others.
Even the number of texts is open to debate. Parpola counts about 5,000, Mahadevan 2,906 and Bryan Wells 3,835 inscribed objects. The total depends on how one assesses fragmentary and damaged inscriptions. The vast majority of the inscriptions – some 85%, according to Mahadevan – were found at Harappa and Mohenjo-daro. About 60% are on seals, but some 40% of these are duplicate inscriptions, so the useful total for the would-be decipherer is not as large as it seems. More inscriptions were found in the 1990s, but the Indus script corpus is not abundant. Then an inscription is tantalisingly brief: many consist of only a single character, the average has fewer than four characters in a line and five in a text, while the longest has only 26 characters divided among the three sides of a triangular terracotta prism.
In addition to the characters, many seal stones are of course engraved with an animal motif or an anthropomorphic figure (such as “proto-Shiva”). How, or even whether, the animal motif relates to its accompanying inscription is totally unclear. A particular motif, such as the unicorn, may appear on many seals, each with a different inscription. Conversely, there are four instances of one and the same inscription occurring on two seals with different motifs. There are even two seals with the same inscription on either side but two different motifs.
The brevity of the inscriptions, and the dearth of both monumental inscriptions and temporary inscriptions (such as scribbles on potsherds) of the kind common in Mesopotamia, Egypt and elsewhere, have led a trio of scholars – historian Steve Farmer, computational linguist Richard Sproat and Indologist Michael Witzel – to doubt whether the Indus inscriptions belong to a writing system. In their view, the characters do not represent the sounds of a specific language: they are merely non-linguistic symbols, probably of religious significance. For a number of reasons, this theory seems unlikely to be true. Perhaps the most important are: that groups of characters recur in the same sequence at different sites; and that the characters can mostly be shown to have been written in only one direction (as we shall see). Sequential ordering and an agreed direction of writing are universal features of writing systems, as opposed to symbolic systems or artistic decoration, where such rules are not crucial. Moreover, the Indus civilisation must have been keenly aware of the functioning of cuneiform writing in Mesopotamia through its trade links. On the other hand, the brevity might mean that the Indus characters belong to a partial writing system, capable of representing only limited aspects of the Indus language – rather than a full writing system such as Akkadian cuneiform and the English alphabet, that is, “a system of graphic symbols that can be used to convey any and all thought” (as defined in 1989 by John DeFrancis in Visible Speech: The Diverse Oneness of Writing Systems).
Assuming that the Indus script is a full writing system, several hopeful scholars have drawn a comparison with the recent Mayan decipherment. Like the Mayan script, the Indus script involves a large, complex and partly iconographic set of signs; and the names of neither places nor rulers are independently known (in contrast to the Egyptian hieroglyphs, where decipherers could turn to Greek and Roman historical sources). But there are clear-cut differences, as the Mayanist Michael Coe has pointed out. Mayan mathematics and calendrics were well understood long before the first linguistic breakthrough in 1952. Mayan inscriptions are “numerous, often lengthy, and encode complete sentences”, notes Coe. Modern Mayan languages are well known: “The cultural context is rich and detailed, and many aspects of it survived the Spanish conquest.” Lastly, and crucially, a bilingual is available (in the form of a Spanish-Mayan “alphabet”). None of these advantages applies to the Indus script. Leaving aside the brevity of the Indus inscriptions, which may well contain mostly names and titles, scholars know virtually nothing of the calendrical system, are uncertain about the numerical signs, can make only informed conjectures about the language and culture, and lack anything resembling a bilingual. Above all, the Indus civilisation disappeared more than 2,500 years before that of the Classic Maya, which is a long time, speaking either archaeologically or linguistically.
A vacuum of knowledge has been filled by serious scholarship as well as bizarre theories (some from reputable scholars) linking the Indus area with far-flung places. Both have been surveyed by Possehl and separately by Parpola, who notes:
Connections have been sought with the manuscripts of the Lolos living in southern China and in Southeast Asia, dating back to the sixteenth century ad; with proto-Elamite accounting tablets; with ideograms carved some two centuries ago on Easter Island in the southeastern Pacific Ocean; with Etruscan pot marks [‘More Seven League Boots!’ comments Possehl]; with the numerical system of Primitive Indonesian; with Egyptian, Minoan and Hittite hieroglyphs; with the auspicious symbols carved on a ‘footprint of the Buddha’ in the Maldivian archipelago; and with the [Mayan] glyphs of ancient Central America.
Let us examine quite briefly four of the more serious claims, each by a respected scholar. Although they have been almost universally rejected, they have something worthwhile to teach us about how to tackle this difficult problem and how not to. Note in what follows that all the inscriptions are read from right to left (we shall come to the evidence for this later).
The first of these “decipherments”, published in 1932, treated the Indus script as if it functioned like Egyptian hieroglyphs. Its author, the Egyptologist Flinders Petrie, did not postulate any connection between the Indus and Egyptian languages, but he did suggest that the pictographic quality of some Indus signs, their variants and their syntax might indicate their meanings on the Egyptian model, assuming that the seals belonged to officials and contained their titles. Thus Petrie read the Indus sign, which is by far the most common sign, frequently found at the end of inscriptions, as a title meaning “agent” (“wakil”, in Petrie’s terminology). The sign he took to be a tree with lopped branches, and the sign was said to be a writing tablet with a handle, a kind of hornbook. The sequence was therefore said to mean (reading from right to left) “wakil [agent] of the registrar of timber”.
On a similar basis, judging that the Indus sign looked like the Egyptian hieroglyphic symbol for “irrigated field”, Petrie translated the sequence as ‘wakil [agent] of irrigated land’. He also noted another very common sign and its apparent allographs (variants): These signs are often found doubled. Petrie decided that the first of the five allographs (on the far left) might indicate a title, inspector or intendant, while in combination the signs might indicate various grades of inspector, such as subinspector and deputy inspector – a “most imaginative explanation” for the doubling, notes an amused Possehl. To be fair to Petrie, he stressed that “these are only suggestions, or speculations, and for ‘is’ read ‘may possibly be’ in all instances” – a sound caution, not often heeded by subsequent Indus script decipherers.6 For even if Petrie were right, there would be no way to prove it, since his methodology was largely intuitive. But his suggestions did have one definite merit: they reminded everyone of the likely bureaucratic subject matter of the inscriptions. That is, unless one takes the view, as some scholars have, that the Indus civilisation is fundamentally different from Egypt, and indeed Mesopotamia, in its use of writing, and that the seals therefore probably contain esoteric ideas rather than economic matters.
Excerpted with permission from The Indus: Lost Civilizations, Andrew Robinson, Pan Macmillan India.
