| Summary: | A questionnaire study and a self paced reading experiment investigate the off-line and on-line comprehension of adjunct and argument prepositional phrases (PP) in Brazilian Portuguese. The first study compares NP-attached adjunct and argument PPs and VP-attached adjunct and argument PPs and show a general preference for argument PPs in both cases. The second study presents an experimental design crossing the argument status of PPs (argument or adjunct) and the segmentation type of the sentences that contain the PPs (short or long). Results do not indicate differences in reading times (RT) between adjunct and argument PPs in the first pass of the parser, in contrast with models of sentence processing that predict a rapid initial access to lexical information. Nonetheless, RT differences between adjunct and argument PPs are found in a garden-path configuration, something which is attributed to the reanalysis stage. Segmentation effects are also found and discussed with relation to the Implicit Prosody Hypothesis (Fodor, 1998; 2002).
|
|---|