Relationships between grammatical encoding and decoding: an experimental psycholinguistic study

Abstract

Although usually considered distinct processes, grammatical encoding and decoding have many theoretical and empirical commonalities. In two series of experiments relationships between the two processes are explored. The first series uses a dual task (edited reading aloud (ERA)) paradigm to test the implications of the standard dual-processor model that syntactic encoding and decoding should function independently and in parallel. Reaction times measured in both pluralising and paraphrasing editing tasks suggest that participants are unable to maintain separate syntactic trees for (input) decoding and (output) encoding, but, at least in this task focused mainly on the output structure. Whether this effect is due to a common grammatical coder or limited working memory resources is unclear. The second series explored the overlap between representations by means of comprehension-to-production syntactic priming. Both cued picture discription and sentence completion experiments failed to demonstrate a reaction time effect of word order priming, even if an efffect in terms of response tendencies did obtain. The results provide support for an architecture in which linearisation processes are either determined in parallel with hierarchical sentence construction, or can even precede it when strong word order clues are provided, even if the actual words aren__t available yet.