SUMARY The main aim of this doctorate research is to design an instructional model- a macro-grammar strategy- which combines grammar and strategy researchers contributions to help foreign language learners learn an L2 more autonomously. However, this research is specially motivated by our concern about the poor results which Spanish secondary school students achieve in Spanish schools in their written exams, since passing a written exam to get into university is these students main objective. The reasons for this failure may be due to an excessively communicative approach and too wide a variety of students sharing the same classroom. Therefore, though the purpose of this research is to help all students to learn grammar more autonomously thanks to their acquisition of metacognitive (selective attention, self-monitoring and self-evaluation) and cognitive strategies (elaboration, induction, translation, deduction, repetition and transfer), this study is specially concerned about unsuccessful learners who lack or do not apply the metacognitive and cognitive strategies necessary to learn grammar and transfer their acquired knowledge to the acquisition of new grammatical structures, since this model may enable students to learn on their own and improve their attitude towards foreign language learning. Our research is based on cognitive theory of learning. The theoretical background which supports our research attempts to demonnstrate how cognitive theory of learning manages to explain how learning, an L2, grammar and strategies, are learnt; and how the automatization of grammar and strategies helps students learn an L2. According to this approach the learner becomes an active organizer of incoming information and an L2 acquisition is seen as the acquisition of a complex, cognitive skill in which the learning of grammar plays an important role and LS facilitate its learning (McLaughlin, Rossman, and McLeod, 1983). This cognitive approach also states that grammar and learning strategies (LS) are complex skills; and explains how a skill is learnt so that the learning of an L2, grammar and LS can be understood. Thus, Andersons (1980) theory of knowledge acquisition, McLaughlins (1987), Schmidts (1990), Sharwood-Smiths (1993), Longs (1998), Dekeysers (1998), and Skehans (1999) position on grammar learning and Wendens (1987) and OMalley et al.s approach to LS manage to complement each other and offer an appropriate theoretical framework for designing an instructional model in which strategy training enables students to learn a grammatical structure and transfer this knowledge to the learning of other structures. The automatization of grammatical items is believed to take place when formally learned grammatical items have been explicitly noticed, have been structured into working hypotheses and have been proceduralized/automatized through practice in context (e.g. McLaughlin, 1987; Schmidt, 1990; Johnson, 1994; Batstone, 1994). And by using metacognitive strategies- planning, monitoring and evaluating strategies- students can come to understand that they have power over their own learning (Chamot and OMalley, 1996: 266); and the combination of these metacognitive strategies with cognitive strategies- mental steps which involve interacting with the material- will also facilitate learners to learn grammar on their own. The two main hypothesis of this research are that students, specially fair and poor learners, following the grammar strategy instruction will acquire the 2nd conditional structure better than the students who do not follow this instruction: that is to say, their results in the 1st post-test, 2nd post-test, and final post-test will be better. And that the experimental group students, specially poor and fair learners, will transfer their way of acquiring 2nd conditionals to another grammatical structure, in this case the 3rd conditionals, better than the students who do not follow this instruction. Next, the method to be followed in our empirical work is explained. This research in which 60 subjects- 30 experimental subjects and 30 control subjects- participated has been conducted in Luis Vives secondary school in Valencia during the school year 2000-2001. Their textbook, their teacher and the conditions under which they study English are the same and their standard is low-intermadiate. Due to the great range of results which students get in the pre-test and their last years results participants are ranked into three subgroups good, fair and poor (Izumi and Bigelow, 2000: 248). Students also fill in three metacognitive questionnaires, which analyse students characteristics, objectives, difficulties, and strategies used to learn grammar, before instruction takes place so that the homogeneity of both groups can be established. Our instructional model combines the explicit training of metacognitive and cognitive strategies when practising comprehension, consciousness-raising, controlled and spontaneous written grammar activities- filling in the blanks with the right tenses, correcting mistakes, rewriting, translating from Spanish into English and writing- on the learning of 2nd conditional sentences in the foreign language classroom. Since it is acknowledged that neither a forms-based instruction nor meaning-based instruction alone can lead to SLA (Doughty and Williams, 1998: 11), our approach is eclectic and combined. Our strategy training is provided in four sessions in three weeks running in January. The declarative knowledge about the 2nd conditional, their induction, practice in controlled and spontaneous written activities, and immediate feedback is provided in each session, while metacognitive and cognitive strategies are practised. Students answers in the metacognitive questionnaires and their results in the pre-test and four post-tests are analysed following statistical procedures so that we can be more certain of the reliability of the results. A previous pilot study (Viñes, 2000) was also carried out so that possible deficiencies could be avoided, and results could be compared. The results obtained in the two short-term and long-term post-tests confirm our first hypotheses; our instructional model helped learners, specially unsuccessful learners, to learn the 2nd conditional. And the results of the Discovering the 3rd conditional test confirms our 2nd hypothesis; our instructional model helped experimental, specially unsuccessful learners, to transfer the knowledge acquired in learning the 2nd conditional to the learning of the 3rd conditionals. However, the results which students get are not 100% correct in all the different activities, that it is why, it is considered that they are still in the associative stage of learning in relation to their learning of 2nd and 3rd conditional sentences, and metacognitive and cognitive strategies. More spontaneous written activities and more strategy training sessions should have been provided in order to help them get to the autonomous stage of learning. Thus, the results of the four post-tests offer relevant empirical evidence on the effectiveness of our instructional model to help learners, specially unsuccessful learners to improve their inductive ability, acquire declarative knowledge, and transfer the declarative knowledge acquired in learning the 2nd conditional to the learning of the 3rd conditional. Finally, some pedagogical implications, which can be specially useful to help secondary school students learn an L2, are suggested. First of all, encouraging students to work on their own in silence may aid poor learners in performing activities. Secondly, providing strategy training which combines metacognitive and cognitive strategies is essential. Thirdly, adapting the material to students level, and allowing the L1 to be used in the classroom when needed may encourage students to learn. Fourthly, providing smaller amounts of new information in meaningful contexts, but allowing students to discover it, make it explicit, practise it immediately and get feedback, may allow students to perceive learning grammar as a rewarding game because they can be aware of their success. Moreover, repeating this material applied in different contexts within long intervals of time may help all learners to acquire the new structures since different opportunities for their learning are provided. This doctorate research presents the findings of an empirical study carried out in a Spanish secondary school to test the effectiveness of an instructional model designed to help these secondary school students, specially the unsuccessful ones, learn grammar more autonomously and improve their attitude towards foreign language learning. First, the theoretical background which supports our empirical work is described. Anderson (1980, 1981, 1982, 1983), Anderson and Fincham (1994) and Anderson et al.s (1997) theory of knowledge acquisition, McLaughlin (1987), Schmidt (1990), Sharwood-Smith (1993), Long (1998), Dekeyser (1998), and Skehans (1999) position on grammar learning and Wenden (1987) and OMalley and Chamots (1990) approach to learning strategies complement each other, and offer an adequate framework for designing an instructional model which combines the explicit training of metacognitive and cognitive strategies when practising controlled and spontaneous written grammar activities on the learning of 2nd conditional sentences in the foreign language classroom. Second, the description of our empirical work follows. The experimental and control subjects characteristics, the description of our instructional model, its instruction, the short-term and long term post-tests taken, and the statistical tests applied on the students answers make up this description. Third, relevant empirical evidence on the effectiveness of our instructional model to help learners, specially unsuccessful learners to improve their inductive ability, acquire declarative knowledge, and transfer the declarative knowledge acquired in learning the 2nd conditional to the learning of the 3rd conditional are presented and discussed. Finally, some pedagogical implications, which can be specially useful for secondary school teachers when helping secondary school students learn an L2, are suggested.