The paper presents a new analysis of the moraic structure of English monomorphemic and dimorphemic monosyllables. This analysis is based on experimental data which shows that, in monomorphemic monosyllables, vowels are longer when following by one coda consonant than when followed by two (Munhall et al. 1992), and that, in dimorphemic monosyllables, vowels are longer than they are in monomorphemic words composed of the same segments (as demonstrated by one perception and two production experiments reported in this paper), such that the word 'passed' has a longer vowel than the word 'past'. In order to account for the consistent vowel length differences, the bimoraic structures for these monosyllables make use of mora-sharing (Broselow et al. 1997). The distinction between mono- and dimorphemic words is explained by high-ranking output-output constraints. The vowel in 'passed' is longer than the vowel in 'past' because the word is being faithful to the moraic structure of its base 'pass'.
 



This analysis shows that moraic structure is predictable in English monosyllables, i.e. it is governed by markedness and OO faithfulness constraints. The moraic structure of a given monosyllable is determined by the number of segments in the rime and whether or not there is a morpheme boundary in the rime. This paper thus motivates the need for OO constraints that reference moraic structure.