Statistical Analysis of the Tables in Mahadevan’s Concordance of the Indus Valley Script

Abstract

The Indus Script originates from the culture known as the Indus Valley Civilization, which flourished from approximately 2600 to 1900 bc. Several thousand objects bearing these signs have been found over a wide area of Northern India and Pakistan. In 1977, Iravatham Mahadevan published a concordance of all of the scripts that had been discovered so far. Accompanying the concordance are a set of nine tables showing the distribution of individual signs by position, archaeological site, object type, field symbol (accompanying image), and direction of writing. Analysis of the frequencies of the signs found so far using Large Numbers of Rare Events (LNRE) models estimated the total vocabulary of the language, including signs not yet found, to be about 857. All the tables were analysed using Pearson’s residuals, and it was found that the signs were not randomly distributed, but some showed statistically significant associations with position, object, field symbol or direction of writing. A more detailed analysis of the relation between signs and field symbols was made using correspondence analysis, which showed that certain signs were associated with the unicorn symbol, while others were associated with the gharial and dotted circle symbols.