The role of connectives in the comprehension of spontaneous spoken discourse has been investigated by testing the effect of the connective ‘but’ in the realization of causal inferences and the integration of adjacent statements. The role of this connective in the realization of causal inferences has been tested through a judgment task. The role of ‘but’ in the integration of the adjacent statements has been tested through a word monitoring task. The presence of the connective resulted in shorter reaction times for the realization of causal inferences in the judgment task, but it did not result in shorter reaction times for the integration of adjacent statements, as measured by the word monitoring task. These results suggest that listeners are able to make use of connectives to help them create and decide on the existence of causal connections, but not to process and recognize the surface form of the second statement of the pair.

This study investigated the role of working memory capacity on the making of reinstatement and causal elaborative inferences during the reading of natural texts. In order to determine participants' working memory capacity, they were asked to take the reading span task before they took part in the study. Those participants that were identified as high or low working memory capacity readers were asked to perform a lexical decision task in two conditions: pre-inference and inference. In the pre-inference condition, target words representing reinstatement or causal elaborative inferences were presented immediately before the sentences that were predicted to prompt them. In the inference condition, the target words were presented immediately after the sentences that were predicted to prompt the inferences. Results indicated that, for the high working memory capacity readers, lexical decision times were faster at the inference compared to the pre-inference locations for both types of inferences. In the case of low working capacity readers, lexical decision times were faster at the inference compared to the pre-inference locations only for reinstatement inferences. These findings suggest that working memory capacity plays a role in the making of causal inferences during the comprehension of natural texts.

No abstract