My MA thesis investigated optional sibilant harmony in Navajo using the first person possessive morpheme, which contains an underlyingly palatal sibilant that may harmonize to alveolar when affixed to noun stems that contain [+anterior] sibilants. The literature commonly describes sibilant harmony as being mandatory in Navajo when sibilants are in adjacent syllables, and optional when there is more distance between sibilants. In other words, sibilant disharmony is ungrammatical, but gradiently rather than categorically; in some instances disharmony is ungrammatical enough that it must be repaired through assimilation, while in other instances it is less ungrammatical and may be tolerated. The statistical nature of the variation in these optional harmony settings is not fully understood, however, and the three studies contained within this thesis were designed to investigate how often assimilation occurs in nonmandatory environments and to identify factors that contribute to the variability observed. In the first study, a Google search was used to evaluate sibilant harmony in online Navajo language use in the Spring of 2008 and again in the Spring of 2010. The findings present a picture of optional sibilant harmony that differs somewhat from the traditional view; harmony seems to be optional even in the environment that has traditionally been described as mandatory, and it occurs far less frequently than anticipated. These results led to the creation of an online survey wherein fluent speakers of Navajo provided grammaticality judgments of both assimilated and unassimilated forms. Almost universally, respondents preferred the unassimilated shi- even in those environments where assimilation would previously have been considered mandatory. The third study involved the recording of data from three speakers of Navajo, none of whom use the assimilated si- either in writing or in speech–at least, not to a degree that is discernible to the naked ear. Acoustic analysis was performed to determine whether the prefix-initial palatal sibilant is acoustically consistent across the board–duration, spectral mean, onset of frication energy, and the second formant of the following vowel were measured to investigate whether the prefixal esh differs acoustically when it appears before words that contain potential triggers than when it does not. Analysis reveals some differences in the spectral mean and duration of the fricative portion of the first person possessive morpheme when it occurs before stems that contain [+anterior] sibilants. Taken together, the findings presented herein suggest that the mandatory sibilant harmony environment no longer exists in Navajo, at least with regards to the first person possessive morpheme. Harmony is far less prevalent than expected overall, and is wholly absent for some speakers. The factors of continuancy and adjacency were found to contribute significantly to the gradience observed in all three studies, however, even for those speakers who do not overtly use assimilation.
A PDF copy of the thesis can be downloaded from the open-access KU Scholarworks site found here.