Uniqueness in element signatures

Representations of segments using elements or features face the challenge of whether cross-linguistic representations of the same sound should have the same representation. In line with a universal view of phonology, it is generally argued that the same phonetic output should have identical representations across systems. This paper discusses some cases where this position is challenged, particularly in cases involving language change and so-called crazy rules. It is argued that the phonological patterning of segments is what should be the crucial deciding factor in their representation, allowing different elemental combinations of the same sound from different languages to map onto identical patterns in the speech signal.