Sentences like (i) the boy seems to the girl to be happy are unusual because the subject the boy is understood as being associated with the adjective happy despite the two being distant from each other. Such sentences are referred to as 'raising' sentences because this association is thought to arise from movement of the subject (raising) from a location closer to the adjective. Compare this to (ii) it seems to the girl that the boy is happy, where the subject occurs in its original (unraised) position. Previous research has shown that children have difficulty understanding raising sentences like (i), but not an unraised sentence like (ii). This dissertation research investigates whether the source of difficulty might be reduced to a processing factor referred to as 'intervention' -- the presence of a noun phrase (the girl) intervening between the subject and the location from which it moved. Intervention effects have been observed in many structures that involve such moved elements, including object relative clauses, object wh-questions, etc.
The goal of this dissertation is twofold. First, it corrects previously unnoticed flaws in current methodologies, and thus provides a better picture of children's abilities with respect to raising. Second, it offers an explanation for children's difficulty with raising that draws on a phenomenon (intervention) that generalizes across multiple constructions, thereby reducing the observed difficulties with raising to a more general facet of human language processing. The results of this work will allow us to evaluate and compare three different theories that have been proposed to explain intervention effects in the fields of language acquisition and adult sentence processing.
Five experiments will be conducted to investigate how children?s comprehension is affected by (i) the presence of an intervening noun in raising structures, (ii) the type of intervening noun (lexical/pronominal), and (iii) the position of the subject (raised and unraised).
This dissertation is interdisciplinary in bringing together three different subfields: acquisition, syntax and processing. Furthermore, the approach adopted in this dissertation has the potential to be extended to a variety of additional populations such as L2 learners of English and speakers of other languages that permit raising across another noun phrase.
Our research project investigates English-speaking children's comprehension of subject-to-subject raising sentences (e.g., 'the boy seems to the girl to be happy'), in which the subject 'the boy' is understood as being associated with the adjective 'happy' despite the two being distant from each other. Previous research has shown that children have difficulty understanding such raising sentences. In this project, children's difficulty with raising is explained through a processing factor referred to as 'intervention' - the presence of a noun ('the girl') between the subject and the location (closer to the adjective) from which it moved. Such processing limitations are understood to interfere with the proper functioning of grammar that is essentially adult-like. Previously unnoticed methodological flaws of studies on the acquisition of raising are corrected, thus providing a more reliable instrument to assess children's comprehension of raising sentences. One corpus study and five experiments are reported. The corpus study (the entire USA CHILDES database) reveals that children very rarely hear or produce raising sentences with an experiencer (an intervening noun). Moreover, not a single instance of 'seem' (in either child or child-directed speech) with an experiencer intervening between 'seem' and an infinitival complement was found. Second, children have difficulty comprehending raising sentences that contain an intervening experiencer ('Donald seems to Mickey to be short'), but are able to comprehend with ease sentences in which the experiencer is fronted ('To Mickey, Donald seems to be short'), and thus no longer interrupts the relationship between the two subject positions. Third, children's comprehension of raising sentences improves when the intervening experiencer is pronominal ('Donald seems to him to be short'), rather than a lexial noun phrase - a signature property of classic intervention effects reported from previous research. Crucially, no improvement is observed in the reverse situation, with a raised pronominal and a lexical noun phrase experiencer ('He seems to Mickey to be short). Finally, children exhibit difficulty comprehending nonstandard 'copy-raising' patterns ('Donald seems to Mickey like he is short') when no gender cue occurs to identify the referent of the pronoun. These results suggest that what has previously been considered a deficit in child grammar can be reduced to a performance-based limitation that is manifested in raising patterns with an intervening experiencer. Not only to the findings of this project contribute to research on the acquisition of raising, they also add to the growing body of literature demonstrating the presence of intervention effects in structures involving a broader range of dependencies, suggesting that this phenomenon is not limited to a single structure.
