Working at the Australian National University was getting to be as exciting as I'd hoped it might when we'd taken the considerable decision to leave London. During 1972, I taught a full-year course on “Australian linguistics” – three hours a week for twenty-six weeks. The first term we spent in detailed study of Dyirbal – going through texts and generally getting to understand the structure of its nouns, pronouns, verbs and sentences – as a vantage point for the main endeavour, a comparative overview of the two hundred different Aboriginal languages of Australia. The aim was to see what sorts of features recurred across the continent and, eventually, to discover whether all the Aboriginal languages of Australia could be proved to be genetically related as one language family, descending from a unique ancestor language that must have been spoken some tens of millennia in the past.

I'd done some reading around what had been written on Australian languages from all parts of the continent – although not much work had been done, and a lot of it was of pretty mixed quality. But until now I hadn't really tried systematic comparison of languages. In that 1972 course we made all sorts of serendipitous discoveries. These were improved and refined as I taught the Australian course again, in 1974, 1975 and 1977, until they were eventually tied together in a long book called The Languages of Australia, published by Cambridge University Press in 1980.