Proceedings of IATL 23

A large literature has been devoted to the distribution of reduced forms of English auxiliary verbs (as in Fred's tall, Fred's already eaten, I'll have the fish sandwich, etc.). On the face of it, these are simply weak forms of the corresponding full auxiliaries, and as such they would seem good candidates for the status of simple clitics: elements whose only idiosyncrasy is a prosodic property that requires them to attach to an adjacent full word. Limitations on their distribution, however, seem to involve syntactic conditions, such as the requirement that they not occur in positions immediately followed by the trace of a displaced element. Such syntactic conditions seem to put these elements in a class apart from purely phonological ("simple") clitics.

The reduced auxiliaries are idiosyncratic, lexically listed alternative forms of the corresponding full words. They are prosodically "deficient" in consisting of only a single consonant (/z/, /l/, /r/, /d/, etc.), and thus cannot constitute syllables of their own; phonologically, they attach to the end of a preceding full word. These phonological facts are apparently irreducible, and it would obviously be desirable to account for the full range of behavior of these elements on the basis of that alone. In fact, this seems possible. On the view of most phonologists, the interface between syntactic and phonological form is a prosodic structure that largely reflects syntactic organization, though the two are not isomorphic, and some independently motivated interface conditions govern their relation. Once the properties of this structure are recognized, however, it turns out that a single phonological condition suffices to constrain the distribution of the reduced auxiliaries: a prosodic structure containing a motivated P-Phrase with no phonological content is ill-formed (a condition which itself follows as a theorem from the properties of the prosodic hierarchy). The role of syntax in constraining the distribution of reduced auxiliaries is limited to its effect in inducing prosodic structure which will, when the prosodically deficient elements in question attach as required to preceding material, contain violations of this condition. Traces and other phenomena that have been adduced in this connection are simply some of the instances in which the independently required prosodic structure of a sentence results in such violations. From the point of view of their own grammar, English reduced auxiliaries have no special properties or conditions beyond what follows from their prosodically weak nature.

[full paper] [back to Table of Contents]