A Phonetic Basis for the Sonority of [X]

Sonority is a property of segments that relates to phonotactics and syllable shape but whose nature remains in dispute. Clements (1990) uses this property to define the order that segments may occur in a syllable in what he calls the sonority sequencing principle (SSP), which says that “between any member of a syllable and the syllable peak, only sounds of higher sonority rank are permitted.” His principle relies on a sonority scale, which ranks segments by their natural class, with obstruents registering low on the scale, followed by nasals, liquids, glides, and finally vowels. The classes are defined by the features vocoid, approximant, sonorant, and what Clements calls “syllabic” (only vowels are [+syllabic]). The sum of the positive values of features for a given class yields the sonority value. Syllables are usually arranged with the sonority peak at the nucleus with falling sonority towards either syllable edge, so that low-sonority segments tend to precede higher-sonority segments in onsets but the opposite is usually true in codas. Some languages allow sonority plateaux, or multiple consecutive segments of the same sonority, such as a sequence of two stops or a stop plus a fricative. The implication is that languages that allow plateaux, which violate the SSP, also allow sequences that do not violate the SSP, but the reverse does not hold: there should not be languages that only allow SSP violations. Most accounts agree on the segment ranking above (or a close variant), though the basis for the hierarchy varies. Some definitions are based in phonological features, as in Clements, but others make more reference to phonetics. Parker (2002) creates a hierarchy that is mainly based on acoustics and aerodynamics and defines sonority as a function of intensity. This is later extended to create a universal phonological scale. Wright (2004) maintains that a sonority scale is not a phonological entity but rather a natural organization of sound patterns that falls out from how well place, manner, and voicing cues can be recovered by the listener from the acoustic signal. In contrast to the other definitions mentioned here, Wright does not relate sonority to phonology at all. French rhotics present a problem for the sonority hierarchy. The rhotic is usually realized as a voiced uvular approximant or occasionally a trill, both of which are considered liquids, but after a voiceless obstruent, the rhotic surfaces as a voiceless uvular fricative. The rhotic patterns like the other liquid in French [l]; both are permitted to be the second member of a complex onset. In theory, French does not allow sonority plateaux in onsets, yet the fricative realization of the rhotic, to the exclusion of all other fricatives, occurs as the second member of a complex onset, such as in /kKem/ [kXEm] ‘cream’. The sonority scale for French is thus internally inconsistent: other fricatives are not allowed in French complex onsets, but there seems to be an exception for this realization of the rhotic. UC Berkeley Phonology Lab Annual Report (2014)