ABSTRACT
A series of studies were designed to examine the role of working memory and its subcomponents in written sentence comprehension. In particular, the studies addressed the question of whether working memory plays a role at the syntactic analysis stage and/or at the post-syntactic interpretive stage of sentence comprehension. In four experiments college students performed semantic acceptability judgements about four types of sentences which differed on two dimensions; syntactic complexity and number of propositions. The syntactic complexity variable was assumed to reflect processes occurring at the syntactic analysis stage of sentence comprehension and the number of proposition variable was assumed to reflect post-syntactic interpretive processes. In the first experiment subjects made the acceptability judgements without performing any concurrent task, while in the other three experiments subjects concurrently performed experimental tasks which stressed different components of the working memory system, as a way of determining what parts of the entire system are involved in these two different stages of sentence comprehension.
Results of the first experiment showed that processing time and errors increased as a function of syntactic complexity and of proposition density. The next two experiments showed that when subjects were required to retain a memory load and perform the judgement task concurrently, there was a greater decrement in their performance on syntactically complex than on simple sentences, and on two-proposition than on one-proposition sentences. This result was interpreted as evidence that both of these stages of sentence comprehension draw upon the limited capacity of the central executive component of working memory.
The results of the final experiment showed that a concurrent articulation 532task, which has been shown to interfere with the articulatory loop component of working memory, resulted in a decrement in subjects’ performance on the acceptability judgement task, while a control task (tapping) which was equated with the articulation task for general level of difficulty did not. However, concurrent articulation did not interfere any more with syntactically complex sentences than with simple sentences, although it did interfere more with sentences with two propositions than with sentences with one proposition. This result suggests that the articulatory loop is not involved in the syntactic analysis of a sentence but in post-syntactic interpretive processes involved in the judgement of the acceptability of a sentence.
