On the existence of moraic onset geminates

Abstract

Moraic theory standardly syllabifies geminates in a coda-onset configuration whereby the coda bears a mora. Initial geminates pose a serious problem for the theory since word-initially no coda exists to host the first half of the geminate. Previous proposals have addressed this issue but have not resolved it satisfactorily, because they have created new difficulties pertaining to prosodification, syllabification or generation of insufficient or incorrect patterns. I propose that treating the geminate as a moraic onset simultaneously resolves all the issues above, provided we dispense with the stipulation that onsets are never moraic. An important prediction emerges from this proposal: onset geminates could also occur word-medially. I claim that such prediction is empirically confirmed in languages like Marshallese and Trique. I also argue that moraic theory is right in claiming that geminates are underlyingly moraic consonants – rather than simply long – and demonstrate how in the current model the contrast between geminates and singletons is preserved in all positions in the word.