Search

Ambiguous relative clauses with two possible antecedent nouns have been widely studied in Spanish since they cast doubt on the universality of syntactic parsing principles. Current research suggests that there is an early contribution of prosodic information in the disambiguation of this kind or sentences in oral comprehension and production tasks. In this paper we present two comprehension experiments of locally ambiguous relative clauses with two possible attachment sites in Spanish, in which different prosodic parameters were manipulated to yield either congruous or incongruous pairings of prosody and syntax, with the aim of testing their influence on listeners' attachment preferences of the ambiguous clause and on their response times. Additionally, fundamental frequency (F0) and temporal (pause distribution) prosodic parameters were manipulated in different subexperiments so as to assess their separate contribution to participants' performance. Our results confirm an effective and relatively early use of prosodic information by listeners during language processing. In particular, congruous prosody with syntax generally facilitates participants' responses, though this effect was found to vary across the two attachment options. Finally, F0 and temporal prosodic features were found to make a similar contribution to participants' responses, with few occasional differences between them. [ABSTRACT FROM AUTHOR]

Two kinds of adjectives are commonly distinguished in linguistic theory: 'individual-level' (e.g. tall) and 'stage-level' (e.g. dirty) adjectives. According to Chierchia (1995), the former possess an eventive argument that is saturated in the lexicon and understood as generic, whereas in the latter, the eventive argument comes unsaturated. On the contrary, Kratzer (1995) considers that only s-level adjectives have an event argument in their argument structure. The purpose of this work is to provide evidence in support of one of these two theories. After reviewing previous theoretical and empirical evidence, a behavioral experiment (with a reading-time task) will be presented. The results of this experiment seem to support Chierchia's proposal, suggesting that individual-level adjectives carry more lexical weight (i.e. show lexical coercion) than stage-level adjectives. [ABSTRACT FROM AUTHOR]