Universals of Prenasalized Consonants: Phonemic or Derived, Single or Complex?
Received: Jul 10, 2011; Revised: Aug 31, 2011; Accepted: Sep 09, 2011
Published Online: Jan 01, 2017
Abstract
Most previous research on prenasalized consonants (Herbert 1975, 1977, 1986; Hayes 1991; Hyman 1992) focuses mainly on those derived from two underlying segments and functioning as a unitary segment. However, by examining further data from various languages, this paper proposes that in fact there are four types of prenasalized consonants: phonemic prenasalized consonants as in Fijian, derived prenasalized consonants functioning as a unitary segment as in Kikuyu, derived prenasalized consonants that behave as a complex segment as in Japanese, and phonetically enhanced prenasalized consonants as in Southern Barasano. Additionally, this paper compares phonologically driven prenasalized consonants to morphologically driven ones, arguing that in most cases phonologically driven consonants are developed into a single segment, while in a number of the languages the morphologically driven nasal-plus- consonant sequences turn out to be sequences of a syllabic nasal and a consonant (i.e., NC). Furthermore, based on the typology of prenasalized consonants, this paper argues that phonetic adjustments, such as vowel lengthening or duration of prenasalized consonants, cannot be a cue in deciding their phonemic or phonetic status since some Bantu languages, wherein prenasalized consonants are derived from underlying two segments, behave as a singleton and do not lengthen a preceding vowel. Finally, this paper proposes that irrespective of their diverse realizations, there is only one underlying representation NC, and their distributions can be accounted for in a uniform way based on the roles of universal constraints such as Align-Root-Left, Uniformity, Max-μ, *Complex, and *Coda within the framework of Optimality Theory (McCarthy & Prince 1986)