On the interaction of velars and labials

The paper treats data from a wide range of languages to show, on the one hand, that there is pervasive direct interaction between labials and velars to the exclusion of coronals, and, on the other hand, that the drive behind these phenomena is simply the presence of labiality in labials and the lack of any place specifications in velars. These data then further support the view that velars lack place specifications altogether, a view presented in Huber 2004b contra Paradis and Prunet 1991. Most importantly, the paper clearly shows that all these phenomena are in fact prosodic phonologically conditioned and absolutely regular, rather than random or unprincipled changes. On this basis, the paper sets up a new typology of the phenomena, which better captures the phonological conditions underlying them.