Pronunciation GB: ˌɔːtəʊseɡˈmentl, GA: ˌɑːt̬oʊseɡˈmenl
A theory of phonology in which representations are viewed as consisting of independent tiers of units. Typically, a central or skeletal tier consists of place-holders or timing slots. Other tiers are associated to the central tier. These tiers contain information about such things as vowel quality, features of articulation for consonants, nasalisation, lexical tone. The main function of this multi-tiered representation is to capture the fact that phonological processes may operate upon certain aspects of the phonological form (delete a lexical tone, for example) while leaving others unaltered.
Reference: Goldsmith J. A.(1990) Autosegmental and metrical phonology. Oxford: Blackwell.