Sonority and initial consonant mutation in modern Celtic

Sonority can be thought of as the perceived loudness, or clarity, of a specific speech sound. This is correlated with the degree of obstruction within the supralaryngeal vocal tract. Thus, a low vowel (with the least amount of obstruction) has a very high degree of sonority, as compared to a plosive (the greatest amount of obstruction) which has a very low degree of sonority. Sonority has in recent years become again the focus of some attention within phonology. In this chapter we outline the main concepts behind the notion of sonority. In particular, we describe the debate around how many sonority classes there should be, and we outline the Sonority Sequencing Principle (that attempts to account for syllable shapes in natural language), and the Sonority Dispersal Principle (that deals with ideal consonant clusters). We then look at the phenomenon of initial consonant mutation within the Celtic languages, where word-initial consonants are subject to sets of phonological changes triggered by morpho-syntactic contexts. The different sets of changes are described for Welsh, Breton, and Irish. The effect of initial consonant mutation is then examined in terms of the changes in sonority profiles between the unmutated initial consonant and its mutated reflexes. First, this is considered in terms of syllable shape, thus testing the Sonority Sequencing Principle. Mutations in all three languages generally change syllable shapes such that they become less like the ideal syllable onset predicted by the Sonority Sequencing Principle. Next, the effect of mutations on initial consonant clusters in the three languages is examined, thus testing the Sonority Dispersal Principle. Widely differing results were found for different mutation types in the different languages, and we show how these results also depend on how many sonority classes one adopts in an analysis. We conclude that the initial consonant mutation systems of Celtic languages are not amenable to explanation via the main tenets of sonority theory, and that the data presented support the notion that sonority is an epiphenomenon rather than a constraining principle of phonological systems.