Published online by Cambridge University Press: 02 May 2003
This paper argues in favour of the analysis of Tohono O'odham stress originally presented in Fitzgerald (1997). This analysis relies on a single grammar in which metrical and rhythmic constraints are interleaved with the MORPHEME-TO-STRESS PRINCIPLE (MSP), a constraint that requires morphemes to be stressed. The single-grammar approach is contrasted with the co-phonology account offered for the same data in Yu (2000). The two analyses use the same basic constraints, with the exception of the MSP. Instead of this constraint, Yu uses three separate co-phonologies. The comparison between the two theories is focused on this difference, the MSP vs. the multiple grammars of the co-phonology model. The point made here is that Tohono O'odham does not need co-phonologies.