The description of initial consonant mutations in the Celtic languages has frequently been attempted. Theoretical treatments have tended to focus on either the phonological aspects of the alternations or the syntactic aspects of distribution. Both of these perspectives, however, leave the topic incompletely covered. On the one hand, there is no reliable synchronic phonetic conditioning generally to be found in the modern Celtic languages. On the other, the syntactic conditions are not unified and frequently make reference to strictly local, rather than hierarchical, relations between “triggers”, which seem to condition the mutations, and “targets”, the word or words which actually instantiate the particular mutations. Attempts to bridge the theoretical gap directly by means of a so-called “syntax-phonology interface” consistently miss the functions of the mutations as part of word formation, i.e. the morphological function of mutations.
This dissertation treats consonant mutation in Scottish Gaelic (SG) as a set of morphological processes, operative in relating one lexeme to another, a lexeme to its various inflected word-forms, and word-forms to particular shapes of those word-forms required by particular syntactic constructions or collocations. In this way, mutations are shown to be deeply integrated in the realizational and demarcative morphological systems of SG. Mutations are used in constellations of functions that are characterized by partial formal generalizations, and so they are unified only abstractly.