ABSTRACT
Sonorants are only underspecified for [voice] in languages in which it is a redundant value. In languages that have voiced and voiceless sonorants, [voice] is not redundant in sonorants. Mester and Ito propose that "voiceless" sonorants are actually underlyingly aspirated. This chapter reviews the cross-linguistic evidence from consonant systems of languages with voiceless sonorants. It discusses some additional phonological evidence. Maddieson's table of phonemes for Hopi does not include aspirated obstruents. One of his references (Whorf 1946) says that Hopi has "preaspirated" obstruent phonemes and voiceless sonorants. Icelandic has a process of sonorant devoicing that also gives evidence that voiceless sonorants are marked with the same feature as aspirated stops. Turning to the facts of neutralization for sonorants and the Laryngeal Constraint, then, these are the possibilities: If the language has only voiced sonorants, the Constraint will never apply to sonorants, since the language has no sonorants with a Laryngeal node.
