Option 2: Focus on meaning

A typical response to frustration with Option 1 has been a radical pendulum swing: a shift of allegiance to Option 2, and an equally single-minded focus on meaning. This position is implicit in much of the writing of Corder, Felix, Wode, Allwright, and others, in Prabhu's procedural syllabus, in part of the rationale for French immersion programs in Canada, in Newmark and Reibel's Minimal Language Teaching Program, and more recently in Krashen's ideas about sheltered subject-matter teaching, and Krashen and Terrell's Natural Approach.

Unlike Option 1, the starting point in Option 2 is not the language, but the learner and learning processes. While the rationales and terminology have differed greatly, advocates of Option 2 typically invoke one or more of the following in support of their proposals: (i) the alleged failures or irrelevance of Option 1; (ii) (more positively) the repeated observations of putatively universal "natural" processes in L2 learning referred to above, reflected, among other ways, in relatively common error types and developmental sequences across learner age groups, L1 backgrounds and (naturalistic, instructed and mixed) learning contexts; (iii) the futility of trying to impose an external linguistic syllabus on learners; and (iv) the belief that much first and second language learning is not intentional, but incidental (i.e., while doing something else), and implicit (i.e., without awareness). L2A, in other words, is thought to be essentially similar to L1A, so that recreation of something approaching the conditions for L1A, which is widely successful, should be necessary and sufficient for L2A. Accordingly, Option 2 lessons with a focus on meaning are purely communicative (in theory, at least). Learners are presented with gestalt, comprehensible samples of communicative L2 use, e.g., in the form of content-based lessons in sheltered subject-matter or immersion classrooms, lessons that are often interesting, relevant, and relatively successful. It is the learner, not the teacher or textbook writer, who must analyze the L2, albeit at a subconscious level, inducing grammar rules simply from exposure to the input, i.e., from positive evidence alone. Grammar is considered to be best learned incidentally and implicitly, and in the case of complex grammatical constructions and some aspects of pragmatic competence, only to be learnable that way.

Although arguably a great improvement on Option 1, a focus on meaning suffers from at least five problems:

1. While not inevitable, in practice there is usually no learner needs or means analysis guiding curriculum content and delivery, respectively.
   
   
2. In the view of many (but not all) researchers, there is increasing evidence for the operation of maturational constraints, including sensitive periods, in (S)LA (for review, see, e.g., Curtiss, 1988; Long, 1990, 1993; Newport, 1990). The jury is still out on this, but a number of studies suggest that older children, adolescents and adults regularly fail to achieve native-like levels in an L2 not because of lack of opportunity, motivation or ability, important though all these clearly are in many cases, but because they have lost access to whatever innate abilities they used to learn language(s) in early childhood. If so, it will be insufficient for later L2 learning simply to recreate the conditions for L1A in the classroom.
   
   
3. Although considerable progress in an L2 is clearly achieved in Option 2 classrooms, as evidenced, e.g., by the ability of some graduates of Canadian French immersion programs to comprehend the L2 at levels statistically indistinguishable from those of native-speaker age peers, evaluations of those programs have also found that even after as much as12 years of classroom immersion, students' productive skills remain "far from native-like, particularly with respect to grammatical competence" (Swain, 1991), exhibiting, e.g., a failure to mark articles for gender. Such items have been in the input all the time, but perhaps not with sufficient salience, and with inadequate sanction (e.g., negative feedback) on their accurate suppliance. Similar findings of premature stabilization have been reported in studies of adult learners with prolonged natural exposure by Pavesi (1986), Schmidt (1983), and others.
   
   
4. White (1991 and elsewhere) has pointed out that some L1-L2 contrasts, such as the grammaticality of adverb-placement between verb and direct object in (L1) French , but its ungrammaticality in (L2) English (*He closed quickly the door), appear to be unlearnable from positive evidence alone, i.e., simply from exposure to the input. English speakers should have no trouble learning that in addition to 'Je bois du cafe tous les jours' (I drink coffee every day), it is possible to say 'Je bois toujours du cafe' (*I drink every day coffee), which is ungrammatical in English. It should be easy because the learners will hear plenty of examples of each structure in the French L2 input, i.e., positive evidence. The reverse is not true, however. French speakers trying to learning English in an Option 2 classroom will be faced with the task of noticing the absence of the alternative French construction in the input. Worse, the deviant structure (*He opened carefully the door) causes no communication breakdown, making it likely that learners will remain unaware of their error. Positive evidence alone may suffice to show the learner what is grammatical, but not what is ungrammatical.
   
   
5. A pure focus on meaning is inefficient. Studies show rate advantages for learners who receive instruction with attention to code features (for review, see Ellis, 1994; Long, 1983, 1988). As I have argued for many years, comprehensible L2 input is necessary, but not sufficient.