In this paper two types of explicit instructional modes are of particular interest: structured input instruction and output-based instruction. In structured –input instruction, students pay attention to the form of the target structure and process input for meaning through tasks that do not require them to produce the target structure. The structured-input group receives explicit instruction on the key grammatical item and practices this feature through input-based activities. In meaning-oriented output- based instruction, students are intended to focus only on meaningful activities, in which students attend to the meaning of both the stimulus and the response, and are given opportunities to produce language. This area of research has found a fertile ground in Second Language Acquisition and there are now a number of studies that have contrasted structured-input and output-based instruction on tests of comprehension and production. These provide evidence that there are equivalent gains for both treatments on comprehension and interpretation tasks whereas output-based instruction may be superior on tests that require a pressured response. The understanding of how input and output affect comprehension and production of target forms and structures in one’s second language (L2) is a key issue in SLA research and has been the subject of several studies trying to examine the relative effects of input-based as compared to output-based instructional conditions (Allen, 2000; Collentine, 1998; DeKeyser & Sokalski, 1996; Erlam, 2003; Nagata, 1998; Salaberry, 1997). On the one hand, it is axiomatic to say that SLA can only take place when learners have access to L2 input. There are researchers who claim that language acquisition environments that entail naturalistic, comprehensible input can result in linguistic competence. Competence consists of the rules the learner has internalized, as opposed to performance which consists of the comprehension and production of the language. Thus, for example, Krashen’s (1985) Input Hypothesis proposed that adults acquire languages subconsciously by receiving and processing “comprehensible” input and learn about language consciously, however conscious learning cannot be converted into acquisition and can be used only as a Monitor of what they have already acquired via the consciously learnt rules. VanPatten’s processing instruction (PI) is an input-based instructional technique which posits that it affects the acquisition of target forms by actively engaging learners in processing structured input, that is, input that has been manipulated to contain many instances of the same grammatical meaning-form relationship. In this way learners are pushed to alter their existing processing strategies by changing the ways they attend to input data that result in better intake. On the other hand, the role of output practice in SLA remains to be clarified. From the teacher’s point of view, the key question is this: to what extent should instruction be directed at developing form-meaning associations through comprehension practice only as opposed to providing opportunities for learners to practice in production tasks. The purpose of the study is to compare the linguistic development of learners who have received structured-input instruction (comprehension practice) for reflexive pronouns in English to the linguistic development of learners who have received output-based instruction (production practice).

The experiment had a between subjects design. The independent variable is an explicit instruction type with two experimental conditions (a) structured-input instruction and (b) output-based instruction, and one control condition. What we seek to investigate is the outcome on four dependent variables (a) listening comprehension, (b) reading comprehension, (c) written production and (d) oral production. Automaticity (speed) and accuracy are incorporated in all language measures except the reading comprehension tasks. This investigation included the use of a pretest and two posttests. Improvement from pretest to posttest in the target structure was compared among subjects who received no instruction in reflexive pronouns in L2 English and subjects who were provided with an instructional treatment aimed at improving their ability to use reflexive pronouns. The experiment followed a pretest-posttest design. One-way repeated measures ANOVAs were performed on raw scores from both comprehension and production tests across all testing sessions to determine whether explicit instruction as manipulated by (a) information and practice regarding the morphological aspects of reflexive pronouns one at time or all at once and (b) activities eliciting nonverbal or verbal answers have a differential effect on intake of English reflexive pronouns by L2 learners. Pairwise post-hoc comparisons followed of all conditions. The alpha level was set at p < .05 for all statistical tests.

Output –based learners are superior to learners in other conditions on both comprehension and production tests and structured-input learners are superior to learners in the control group in their ability to comprehend the target structure more effectively as well as producing it. Because the instructed groups made the greatest gains in reflexivity, we draw the conclusion that form- focused instruction is effective in teaching a grammatical structure such as reflexive pronouns in L2 English. In R. Ellis’ review (2002), he concludes that FFI can have a significant effect on the accuracy of use of grammatical features if it is directed at simple morphological features (because it increases salience and processing) than at more complex syntactic structures (which require attention and maintenance). Further empirical research is necessary to examine whether these results are generalizable to more complex structures, the issue of the development of automaticity, as well as whether actual output favours form or emphasizes fluency at the expense of form