Phi-formation in Italian: a new proposal

In early generative theory (Chomsky and Halle (1968)) the interaction of phonology with the rest of the grammar was strongly limited to an interface with syntax such that the output of the syntactic component was the input to the phonological component, even though some readjustment rules were necessary in cases where the lack of isomorphy between morphosyntactic and phonological constituents was blatant. In more recent work, the different subcomponents of phonology have their own principles applying within phonological units not necessarily isomorphic to any morpho-syntactic constituent. With particular reference to phrasal phonology, it has been shown in much work (Selkirk (1984), Nespor and Vogel (1986)) that, while the principles defming the various prosodic consituents make reference to nonphonological notions, it is of crucial importance that the resulting prosodic constituents are not necessarily isomorphic to any constituents found elsewhere in the grammar. In the approach of Nespor and Vogel (1986) (henceforth N&V), for example, primitive phonological phrases in Italian are formed by making reference to domains delimited by the heads of maximal projection (X head ). These phonological phrases can then undergo a rule of restructuring if they happen to be the first nonbranching complements on the recursive side (the right side in Italian) of X. Italian phonological phrases are then isomorphic to no other constituents in the grammar, even though their right edges always have to coincide with the right edges of either X head or of their immediate nonbranching complements. N&V's analysis of Italian Φ-constituency is the most detailed in the literature. Nonetheless, it does not appear to be empirically adequate. New data that will be presented here will lead us to a completely different picture of how Φ-constituency is assigned in Italian. More specifically, we will show that the core of Φ-constituent structure assignment in Italian is a set of prosodic principles that make reference to notions of phonological weight, balance, and symmetry that contribute to the eurhythmicity of the utterance.