Abstract
The Indus script is one of the major undeciphered scripts ofthe ancient world. I will discuss the difficulties associated withdeciphering this script and present a theoretical framework, based onMarkov chains, within which the syntax of the script can be exploredsystematically. I will present key results from the Markov chain analysiswhich provide evidence in favour the hypothesis that the script was usedto represent natural language. The use of other probabilistic methods foranalysis corpora of linear sequences will also be briefly outlined.
