morpheme-structure rules/conditions

Terms used in GENERATIVE PHONOLOGY to refer to the processes which have attempted to cope with REDUNDANCY in carrying out an analysis. When segments CO-OCCUR, the presence of a FEATURE characterizing one segment may make it unnecessary to specify a certain feature in another segment: the CONSTRAINTS involved are handled by MORPHEME structure (or ‘LEXICAL-redundancy’) rules. For example, given an English morpheme which has an AFFRICATE in INITIAL position, it is predictable that the following segment will be a VOWEL. It would then be possible to leave the features for vowels (e.g. [−consonantal], [+sonorant]) blank in the UNDERLYING FORM of the morpheme, the appropriate values being filled in automatically by the application of the relevant morpheme-structure rule at some subsequent point in the DERIVATION. Several problems with this view led to a subsequent proposal to handle these redundancies in terms of morpheme-structure conditions, which state more explicitly the processes constraining the correspondences between segments, without recourse to the blank-filling procedure.