A grammatical problem which has perplexed scholars since the 5th century BC has been solved by a Cambridge University student and could “revolutionise the study of Sanskrit”, a professor has said.