The Madras Indus scholar

Interview BY I Sundar Ganesan

What first propelled you to study the Indus script?
Early in the 1960s, I began working on the cave inscriptions of Tamil Nadu. They are the earliest records of not only Tamil but of any Dravidian language. So I spent several years visiting the caves, copying the inscriptions and published a number of papers. In between, I spent a dozen years in New Delhi, and became enchanted with the Indus script specimens I saw in the National Museum. Soon thereafter, I began working on it. In addition to the concordance* that I ultimately prepared in cooperation with computer scientists in Bombay, I have published a series of papers at three levels.

First, there are about half a dozen papers on the statistical analysis and such linguistic features as can be recognised without reading the language. Second, I began working on the meaning of some of the obvious ideograms. These are pictures of objects which can be recognised directly as representing a subject – like a man carrying a bow and arrow, who can be an archer. A human being with two horns may represent an important person or god, and so on. The other method is called ‘rebus’, that is, the transfer of sound from one picture which can be easily recognised to another word with the same sound but different meaning. The well-known example of this is the Dravidian min, which means fish, but also means star. So a fish can be drawn to indicate a star considered as a deity.

The concordance you created seems to have required a Herculean effort. Do you see any scope for further expansion?
The first concordance in the pre-computer age was made by Hunter, an Englishman in India who was in the Indian Educational Service. He aligned all the signs from their outward form and prepared the concordance. But subsequently more seals have been found at Mohenjodaro, Harappa and other new sites. [Finnish scholar] Asko Parpola and his colleagues have published a concordance; and in India, I, with the help of computer scientists at the Tata Institute of Fundamental Research, published our concordance. The first healthy sign is there is a lot of common ground between these three concordances. While more seals have been found, they only confirm what has been found earlier; the concordance shows that there is an underlying order. This order can come only from an underlying language.

I have gone further in my analysis, and I claim to have isolated two kinds of suffixes in the language – nominal suffixes at the end of names, and suffixes which indicate what are called ‘cases’. We also know that the adjective appears before the noun it qualifies. Then, we know the numerals. Progress has also been made in discovering the direction of writing, which is mostly from right to left, with some exceptions. We can also segment words and phrases. Well, that is good progress. In my view, the Indian tradition, mythology, religion, history, folklore, art, etcetera form the Rosetta Stone for decipherment. We can apply what we know of the Indian tradition to the pictorial figures in the Indus seals and try to work out what they could have represented.

There are periodic reports of Indus script being deciphered. Are there standard methods to test the validity of claimed decipherments?
The best summary and evaluation of the work done so far is Gregory Possehl’s book, The Indus Age: Its writing. I myself have reviewed five claims to decipherment – two based on Sanskrit, two on Tamil and one claiming that the script is merely a collection of numbers. My conclusion is negative – that none of the decipherments has been successful.

The first test is the direction of the Indus script. The one fact on which most scholars agree is that the Indus script reads generally from right to left. So this is the first test, which can eliminate non-serious attempts. The second test comes out of the progress achieved in segmentation of words. An Indus text can be segmented into separate words and phrases. Any decipherment will have to conform to these segments.

Another method is to match the frequency-distribution analysis of the script with similar analysis for the candidate language. The two frequency-distributions should match. To give an example, in English the letter ‘e’ has the highest frequency, of about 12 percent. If I say that the Indus script is written in English and there is one character which occurs with 10 percent of total frequency, then that must be ‘e’. There are other restrictions. In some languages, certain sounds do not occur in the beginning. There are other languages where certain combinations of consonants are not permitted, and so on. Applying these three tests, I can say that none of the decipherments so far have passed all the tests.

Is research on the Indus civilisation active at the moment?
There is very little interest in the Indus script in the West – there are very few people working on the Indus script around the world. The one exception is India, but research in India has gotten inextricably mixed up with politics: the Hindu nationalistic scholars claim the language is Sanskrit, while the Tamil nationalistic scholars claim it to be a form of Dravidian. Both claims have become suspect because of their political background. Any claim from an Indian scholar becomes suspect because one immediately asks what is the mother tongue or political affiliation of the scholar. A scholar from another country is happily free of this problem. I envy that freedom, but I too have an advantage: I am a son of the soil. The traditions of India, its mythology, its religions, its culture, its art, are in my blood, and therefore I may have insights which people who are not the inheritors of this culture may not have. This is a subjective reaction, but such resources as we have must be put to best use.

Does the 2006 discovery of the Neolithic stone axe at Sembiyan Kandiyur in Tamil Nadu extend the area of influence of Indus civilisation?
Let me first say that this is the greatest epigraphical and archaeological discovery made in Tamil Nadu in the recent past. Two stone axes were discovered accidentally by a school teacher who was digging in his backyard to plant banana saplings. One of the axes is incised with four graffiti-like marks. Fortunately he gave the axes to his friend, a trained archaeologist. The inscribed stone was brought to me, and I was immediately able to identify the four characters as being in the Indus script.

But one can have differences of opinion in interpreting the signs. As the axe was found in the lower Kaveri Valley, where there are no hills, it could not have been made locally. So it must have come by trade. The nearest Neolithic centres in Tamil Nadu are in Dharmapuri District, adjoining Karnataka, and it is known that Harappans were in contact with Karnataka because the gold in the ornaments of Mohenjodaro is supposed to have come from there. And we also know about the existence of Daimabad, a Harappan site in the Godavari Valley, in Andhra Pradesh. So it is not farfetched to think that late Harappan influence could have spread to Tamil Nadu also.

One thing I would like to emphasise is that it is only in Tamil Nadu, and nowhere else in India, that the particular sign which I have identified as muruku occurs continuously. With the exception of a single seal found at Vaishali in Bihar, nowhere in India has this particular sign recurred in the post-Harappan period. Therefore I do think it is a continuation of the earlier tradition, and it is likely that a religious symbol would have survived. It is quite possible that after the Indus script was forgotten and was no longer a system of connected writing, individual symbols, particularly those which were considered to be divine, have persisted – such as the swastika and the muruku symbols.

Will Pakistani experts who are working in the Mohenjodaro and Harappa regions be welcomed at the Indus Research Centre?
Why not? I think our colleagues in Pakistan should be invited to deliver talks on their latest discoveries and share their experiences with the people here. Similarly, there are people in Sri Lanka who are interested in the Indus script. There is also the question as to whether the Brahmi script, which is the parent script of all Southasian scripts, is itself derived from the Indus script. The idea is not far-fetched, and requires looking into. Scholars from countries like Sri Lanka, Tibet, Nepal, Thailand, Indonesia would all be interested to join in the investigations. What is required is a truly free academic atmosphere – free of bias, nationalistic or linguistic, and with a commitment to get at the truth wherever it may lead.