engleski

Derived Environment effects in a surface-oriented theory

There seems to exist a systematic distinction between processes which apply across-the-board and those which are limited to the derived environment, yet OT did not work out a systematic approach to derived environments effects. The solution in terms of the universal ranking of Stem-Faithfulness over general Faithfulness proposed by Pater (1999) cannot apply to Polish data because it does not explain why some stem consonants (stem-final) do alternate whereas some others (stem-internal) do not. The solution offered by Łubowicz (1998) in terms of a conjunction of constraints with ANCHOR, basically relying on the relationship between the palatalization and morpheme and syllable edges does not account for two sets of data. First, it does not cover the palatalization without the surface trigger. Second, it makes wrong predictions for morpheme-internal palatalization, which in earlier approaches has been regarded as the second case of derived environment (cf. the definition in (1)), with the trigger derived not from the concatenation of morphemes but from the prior application of another rule. Finally, McCarthy’ s (2003) Comparative Markedness is not a surface-oriented solution, as it requires a reference to the input. The new proposal in terms of the alternating environment does not refer to either derivation or to the underlying representation, but instead refers to simple surface regularities. It is argued that the derived environment effects apply actually in an alternating environment, defined as a relation of the segment to the adjacent segments, which does not surface as uniform throughout the paradigm. The prediction is that the two cases of the derived environment, i.e., those which involve a morphological boundary, and those which are triggered by segments violating faithfulness, can and should be treated in the same way. This seems to be a desired effect with respect to the Polish (and Finnish) data. The notion of the alternating environment does not need to rely on any OT theory internal concept and, thus, can be adopted in any output-oriented phonological theory.

phonology, Derived Environment, Optimality Theory

nije evidentirano