Syntactic properties of spontaneous speech in the Language into Act Theory
In comparison with writing the performance of spontaneous speech is characterized by a strong reduction of the syntax , anyway a clear and general representation of speech syntax does not emerge with evidence from literature. Our framework , the theory of Language into Act (L-AcT), which is an extension of the Speech act theory (Austin, 1962), it is derived from corpus-based research carried out on large spontaneous speech corpora. Utterance is assumed as the reference unit; it is pragmatically defined such as corresponding to a speech act and is identified through prosodic devices. In the L-AcT perspective, the syntactic reduction can start finding its explication in favor of pragmatic information functionality, because the final linguistic nature of the utterance corresponds to the combination of semantic/syntactic islands, going hand by hand with the information units (IU), taking part to its Information Pattern (IP), and being characterized each for a pragmatic function. Therefore also the study of syntax is carried out considering the information performance of the utterance. First of all the most common syntactic structures, in reality occurring in speech, are expounded in a general survey, with data resulting from corpus-based research. But the central topic of the paper is the explication of the two different types of subordination (linearized and patternized): actually both types of completive clauses (objective, subjective, indirect interrogative but also reported speech, spoken thought and exemplification) and types of relative clauses (restrictive and non-restrictive) are analyzed according to the feature to be performed in a linearized way within a same IU, or in a patternized way across two IUs. The deriving syntactic and semantic implications are commented. Finally, on the base of the IPIC-DB, general quantitative data regarding all kinds of completive and relative clauses are presented , making emerge their low frequency.