Abstract:
The purpose of this study is to examine the effectiveness of connected speech instruction in improving learners’ decoding of spoken English and their reactions to it. It follows the trend of a balanced listening session that addresses bottom-up processes and decoding skills without compromising the need to develop top-down processes. Thus, it attempts to examine the effectiveness of a teaching framework which allows for integrating connected speech instruction into the listening comprehension lessons following the principles of the diagnostic approach (suggested by Field, 2008a). The study included a sample of thirty-eight second year students of English from the University of Mohamed Boudiaf -M’sila, divided into two equal groups; an experimental group and a control group. After a pre-test on connected speech decoding, the experimental group received listening comprehension lessons with an extended post-listening phase to address the connected speech features which were diagnosed as problematic during the listening phase. The control group received similar listening lessons without a specific focus on connected speech. After the treatment, a posttest was administered to both groups and the results were compared. The pre-test results revealed that the connected speech aspects of assimilation, elision, weak forms and liaison pose serious obstacles in the processes of lexical segmentation and word recognition. The analysis of the mishearings showed that the participants may resort rather unsuccessfully to top-down processing in a bid to compensate for their inability to decode connected speech features. The mishearings also revealed a poor application and coordination of some topdown strategies, namely predicting and monitoring. Poorly decoded connected speech at the beginning of utterances was found to constrain the way in which the unfolding text is processed by triggering inferences about syntactic and lexical forms (cognitive effects). Even if they are wrong, the participants tended to stick to these inferences despite the presence of sufficient contradictory evidence in the unfolding text. Quantitative data analysis confirmed the hypothesis that connected speech-integrated listening instruction is effective in improving learners’ decoding of speech. A post-instruction interview with the treatment group revealed positive reactions to the lessons.
