The modelling and recognition of the phonetic composition of speech using a number of independent levels or tiers is contrasted with conventional linear phone models both theoretically and experimentally.
Keywords: Phonetic recognition, speech recognition, non-linear phonology.