Constraints on contrast motivate nasal cluster dissimilation

Many languages exhibit nasal cluster dissimilation, in which an illicit sequence of nasal-stop clusters is modified in some way (e.g. NC1VNC2 → N1VNC2). This article discusses generalisations in the typology of nasal cluster dissimilation, and claims that nasal cluster dissimilation is driven by constraints on contrast distinctiveness: It occurs preferentially in those environments where the first NC is most confusable with a plain nasal consonant. I propose an analysis that appeals to auditory factors, and provide acoustic and perceptual evidence that is consistent with it.