Résumé:
For more than a decade now, first year students at the Université de Formation Continue (UFC) of Djelfa have been taught grammar at the sentence level with much focus on grammar rules which they have to internalize so that they can use and write grammatically correct sentences. However, day-to-day observation of these students show that they do have an acceptable mastery of grammatical structures but are unable to communicate meaningfully and appropriately. In other words, these learners have sufficient grasp of the rules of grammar but use grammatical features in non-native-like ways that cannot be simply categorized as grammatical errors or inappropriate English at the level of the sentence. In the case of tenses, for example, when learners write a piece of discourse of their own , they are unable to control the use of tenses in a sequence of related sentences: they use tenses arbitrarily, overextend the use of a single tense all over the paragraph or jump from one tense to another without a discernible cause or tense markers to justify their choices. This may be due the fact that these learners use tenses without knowing " their contextual or pragmatic motivations and discourse functions" (Celce-Murcia and Olstain, 2000: 68). Furthermore, feedback obtained from the teachers' questionnaire and classroom observation showed that the UFC grammar teachers' main objective is to develop their learners' grammatical competence. It also showed that they give little importance to the discourse functions of the grammar items, neglecting the fact that these items are often determined by factors beyond the sentence.
The present study investigates the effect of raising UFC first year students' consciousness to how the past perfect is used and distributed in longer stretches of sentences in view of making them understand and produce it efficiently and appropriately. To this aim, we conducted an experiment research with two matching groups of 12 students each. The experimental group were taught the past perfect through discourse, taking into account the discourse functions of this tense and also its relation with other tenses such as the past simple and the past continuous whereas the control group were taught the same tense using de- contextualized, sentence-level activities. Both groups were pre-tested at the beginning of the experiment and post-tested at its end. The results of the post test indicated a positive effect of the discourse-based approach as the experimental group outperformed the control group.
