Aspects of the morphology and phonology of phases
Heather Newell
December 2008

This thesis offers evidence that phases (Chomsky 1995) induce word-internal cycles of morphological and phonological interpretation. Phases proposed in the syntactic literature are shown to have effects word-internally, therefore supporting a representational theory of morpho-phonology (e.g. Distributed Morphology (Halle & Marantz 1994)). It is argued that phases exist at the nP, aP, vP, vP, DP, and CP syntactic levels. These phases are shown to have differing behaviour with regards to the domain which is sent to PF upon merger of the phase head. DP, CP, and vP are argued to be complement spellout phases following Nissenbaum (2000). nP, aP, and vP, however, offer evidence that the head of a phase is interpreted at PF with its complement. A possible motivation for this difference in interpretation domain is discussed. It is in derivations where syntactic material spans one (or more) of these boundaries that cyclic domains may be found within words at PF. Phonological and morpho-syntactic patterns induced by word-internal phases are investigated. Main stress patterns in Cupeño, Turkish, and Ojibwa are analysed. Turkish and Cupeño seemingly irregular main stress patterns are argued to be regular at the phase level. Main stress is assigned in these languages at the interpretation of the first phase. In other words, main stress is cyclic and immovable in these languages. Ojibwa main stress assignment is then shown to be insensitive to word-internal phase boundaries. Word internal phases are present in Ojibwa, as demonstrated by hiatus resolution strategies and footing patterns in the language (Piggott & Newell 2007). Main stress is assigned to the word, regardless of its internal cyclic domains – it is post-syntactic. These two patterns are argued to be the only possibilities for main stress assignment. Some morpho-syntactic paradoxes are then investigated. It is argued that word internal phases, in combination with late adjunction (Lebeaux 1988), are responsible for bracketing paradoxes, the dichotomous (phrase/word) nature of particle verbs, and semantically vacuous double affixation. Languages discussed in this section are English, German, Breton, and Yiddish. It is concluded that structural paradoxes arise only when an adjunct is late adjoined into a previously interpreted morpho-syntactic structure. None of the data presented here arise solely in the phonological, morphological, or syntactic component of language. The effects of syntactic phases on morpho-phonology argue for the necessity of an integrated approach to linguistic investigation.
Format: [ pdf ]
Reference: lingbuzz/000807
(please use that when you cite this article)
keywords: phases, stress, distributed morphology, bracketing paradoxes, phonology, morphology, syntax, ojibwa, turkish, cupe˜o, english, german, breton, yiddish
Downloaded:2034 times


[ edit this article | back to article list ]