Why Languages Change, Except When They Don’t

Language is a complex dynamic system transmitted indirectly between individuals. It is inherently variable and subject to numerous competing constraints and biases. Such conditions make change over time the norm for all areas of language. Why is it, then, that some language families show a great deal of reconstructable changes involving the gain and loss of phonemic contrasts, while others (such as Pama-Nyungan in Australia) show few instances of such changes? In this talk, I review claims for how sound change (particularly contrast creation) occurs, from general evolutionary arguments to specific proposals for sound change. I argue that none of these claims alone is sufficient to explain the data in range of changes (and non-changes) we see in Pama-Nyungan, and instead sound changes arises through the interaction of primary and secondary (reinforcement) cues to contrast. Understanding Pama-Nyungan change thus gives us insight into how sound change occurs more generally, as well as elucidating how an understanding of diachrony informs synchrony (and vice versa).