gabbard@ling.ohio-state.edu
Amendments to feature theory phonology have been proposed to account for the guttural consonants behaving as a class. Data from Somali reinforce the notion of a guttural class, but the nature of the rules in Somali suggest a different feature organization than has been previously argued. The Somali data comes from my field work with Somali speakers in Columbus, Ohio. There are two facts of Somali that are of particular interest. First, stop consonants become fricative consonants between vowels. Second, the class of gutturals: uvulars [q], pharyngeals [ħ] and [ʕ], and glottals [ʔ] and [h] do not block this rule with respect to the stop [t]. This is shown in the suffix [-ta]; between vowels and across gutturals the suffix [-ta] becomes [-ða]:
[æɣæb] "thing [æɣæb-ta] "the thing"
[ǰɪlɪf]" "tree bark" [ǰɪlɪf-ta "the tree bark"
[ukun] "egg" [ukun-ta] "the egg"
[baka:ɾ] "grain pit" [baka:ɾ-ta] "the grain pit"
[ho:yo] "mother" [ho:ya-ða] "the mother"
[ǰaʕ] "scandal" [ǰaʕ-ða] "the scandal"
[laħ] "ewe" [ laħ-ða] "the ewe"
Steriade's 19871 theory of feature structure cannot account for the gutturals being both specified for place and simultaneously transparent to a rule. McCarthy's 19892 theory of feature structure captures the idea that gutturals act as a class, but his theory would not allow for guttural transparency. I propose that the pharyngeal node be on the same tier as the place node and not dominated by it as McCarthy proposes.
