In defense of an onset-rime syllable structure for English.

Many linguists and psycholinguists have suggested that English syllables have an onset-rime structure. Pierrehumbert and Nair (1995) have recently argued against this view and in favor of the idea that syllables have a moraic structure. We show that the results of word game experiments reported by Pierrehumbert and Nair are consistent with the onset-rime theory and that there are problems with their idea of output templates. Although people may learn about the phonological structure of a word game's output when they have a chance to do so, they tend to divide syllables at the boundary between the onset and the rime even when they do not have the opportunity to memorize a model. Moreover, the results of our recent statistical study of the distributions of phonemes in English syllables suggest that the rime is a domain for particularly close dependencies among phonemes. We conclude that Pierrehumbert and Nair's rejection of the onset-rime model was overly hasty and was based on a limited set of data. When a broad range of data is considered, the evidence supports the idea that English syllables have an onset-rime structure.