Let me start by clearing up a point of confusion. Professor Moulin appears to believe that, in my article "How much does correctness matter?", I was claiming that grammatical correctness, as such, is unimportant for foreign learners (Swan, 1997). This is not the case, and I must take some of the blame for the misunderstanding: in a brief paper in which I addressed several different issues, I did not perhaps develop this part of my argument in sufficient detail.
Of course grammar matters, and in the limited time most of us have available we must teach enough of it, along with enough of everything else. However, what we call "grammar" is a number of different things, and some of these matter more than others. Since we don't have time to provide learners with a perfect command of all the structures in the language, we have to select. And if students don't learn everything we teach, we must know when to cut our losses: if we fill our intermediate and advanced lessons with remedial work on trivial points, there won't be enough time left for things that matter more.
The problem is to decide what points are really important to get right. How much does one mistake or another prejudice your chances of communicating effectively? Is it worse to mix up present tenses, leave out a definite article, get your modal perfect structures wrong, or drop third-person "s"? Early in his paper, Moulin promises to "examine to what extent disregarding current grammatical rules may jeopardize intelligibility and thus handicap or prevent communication." That would be nice. What Moulin actually does is to quote three mistakes which could in some cases (he does not consider the role of context in repairing error) lead to misunderstanding, and then say "I could adduce scores of similar examples ...." This leaves us no wiser than before. In the absence of hard information about the functional load of different grammatical structures, we are driven back on experience, common sense, and hunch. These tell us, surely, that dropping third-person "s" probably doesn't matter as much as mixing up present tenses, and that this probably doesn't matter as much as saying, for instance, "has should go" instead of "should have gone."
Misunderstandings aside, I believe that Moulin and I disagree on two central issues. The first of these is the notion, which I suspect Moulin subscribes to, that the grammar of a language is a single interconnected "system," and that accuracy is important because mistakes in one area somehow affect the working of the whole (in the way that an ignition or fuel supply fault can cause a car engine to stop running). I regard this view as profoundly mistaken, and responsible for a great deal of ill directed and ineffective teaching. Grammar is much more realistically seen as an agglomeration or heap of sub-systems; some fairly central, connected, and interdependent (like the English tense/aspect system, or the modal verbs), others relatively peripheral and separate, so that if they disappeared from the language, it would make little difference to the rest (like the presence or absence of "to" with infinitives, or our few remaining morphological person and case distinctions).
In support of his view, Moulin quotes Master as saying, in a 1994 paper, that "systematicity and completeness are essential." Quoting people who agree with you is a common academic strategy; it is not a very effective substitute for reasoned argument, and it is even less effective if, as here, the quotation is both inaccurate and inapposite. Master, in the paper referred to, is not talking about grammar teaching in general, but about a successful approach to the teaching of the English article system (which he describes on the same page as "an aspect of grammar that contributes little to communicative effectiveness"). And he does not say at all that systematicity and completeness are "essential." What he does say, rather cautiously, is that in this particular case "it is perhaps the systematic presentation of the article system that makes the difference" (some aspects of the article system tend to operate simultaneously, so that piecemeal teaching of article rules doesn't work well). Systematic presentation of "the whole picture" may possibly be valuable for the teaching of some other aspects of English grammar (such as tense contrasts); there is no reason to believe this is so for the language as a whole. What is certain is that mistakes in one area, whatever their local effect, do not cause the whole of communication to break down, because language is not that kind of "system."
My other central disagreement with Professor Moulin concerns the alleged absolute value of correctness. I entirely agree that "a concern for correctness and precision is part of any type of education"; but so is a sense of proportion. We require different degrees of accuracy for, different purposesﾑmore for building aero-engines than for building kitchen chairs. Precision without concern for its application often leads to aberration: at its most harmless extreme to the mindless pedantry of the researcher tabulating, for instance, all of Shakespeare's references to dragons, though more sinister illustrations are not difficult to think of. We will waste a lot of our own and our students' time if we pursue correctness for its own sake, on the illusory grounds that learners "need some grammatical discipline," that it is necessarily always a bad thing if mistakes fossilize, or that "absolute and immediate accuracy" is of any value at all to the average foreign-language user. Moulin says that "fluency based on inaccuracy and imprecision is simply a form of camouflage." If what he means by this is that people who make mistakes in foreign languages necessarily communicate badly, he is totally wrong. World languages such as English, Spanish, French, Chinese, and Russian are used effectively, for diplomatic, commercial, scientific, and other purposes, by many people who do not speak them correctly. Such people are often extremely skilled at communicating their meanings clearly and precisely, whether or not they get all their tenses and articles right. And where people are unsuccessful at communicating (like the undergraduates Moulin complains of who have trouble understanding complex arguments and constructing coherent discourse in their mother tongue), we have to ask whether the problem lies in their command of the language they are using or in some other area. We do not train people to think clearly by teaching them to use discourse markers.
Our task as teachers, or as researchers advising teachers, is not to make large hand-waving claims about the overall importance or unimportance of grammar and accuracy, but to encourage a cost-effective approach whereby those aspects of the language which really matter in the light of our students' aims are given the attention they deserve. If getting a particular structure right, or using it correctly, contributes significantly to comprehensibility, acceptability, exam marks, or career prospects, it is worth spending time on teaching and re-teaching the point. If not, we have better things to do.
References
Swan, M. (1997). How much does correctness matter? The Language Teacher, 21(9), 54-55, 61.