I have come fairly late to this thread, so there are a few things that I would like to comment on. Let me first say that I refuse to accept Krashens opposition between acquisition and learning (where the latter includes explicit grammar studies). I know from my own activities that I learn more from input when I'm prepared for it than when I'm totally blank - and grammar studies is one way to get prepared for the acquisition task. Which leaves the question of
how to study grammar.
On page 1 ...
BeaP wrote:I find explicit grammar instruction necessary. The question is rather: when and how much grammar should one learn. I think this area has improved a lot in the past years. We used to have B2-C1 grammars only, that wanted to teach everything about the articles, then everything about the pronouns, and so on. Now everything is nicely graded, when you're a beginner, you can read through a grammar for beginners to learn the most important things. You won't be overwhelmed.
Another thing that seems to be common in coursebooks now is that first you have to meet something 'in the wild', by examining authentic texts, then you can be made aware of the logic that's behind the usage, then you can try to practice the same thing yourself, preferably in a concrete situation. Learning solely from input results in a lot of grammar mistakes (in my experience), that the speaker doesn't realise, but they can be corrected.
I don't agree with statement 1 (although I understand why BeaP supports it). For me a grammar book should always be systematic - that is: the articles (and ALL the articles) should be explained in one chapter, then the inflection of nouns, then that of the adjectives etc. etc The order of the topics can be changed, but I like the global overview and compartmentalized logical structure of an 'oldfashioned' grammar book, and I hate been served the knowledge in small bits in a
random pedagogical order.
The reason that I can survive such a book is precisely that I don't read everything sequentially. For instance many of those old grammar book spent the first fifty pages on phonology, and if I can't hear the language in action while I read that chapter then it is crushingly boring to read - like watching cooking on TV without having the smells and sizzling sounds. When I do my green sheets I actually disregard the order in the books. For instancce I always try to get the morphology of both articles and adjectives and nouns squeezed into one table, but I also know that you can't do that in a full grammar which should be logical, reasonably complete and nevertheless readable for the average learner. In other words MY personal resumé has to be structured differently from the grammar book (or books) I use - but that doesn't mean that I want grammar books to switch to 'green sheet' format (except maybe in an appendix).
The same thing applies to grammar teaching: I want the full monty in the form of a classical grammar book, but if I discover a curious fact in real life or get a correction then it's logical that I get the relevant bit served on a platter without the ten pages of information which it may be surrounded by in the grammar book. And such tidbits will of course have to be served rather randomly.
As for using standardized names: I prefer using the Latin ones wherever they are relevant. But in some Latin grammars there are long lists of the different uses of the subjunctive, and in the old days you were supposed to learn the Latin name for each of those uses. I do acknowledge that having a list of typical uses can be handy, maybe even with a name tag on each kind of use - but ultimately you have to learn from examples, and in spite of the misgivings of some theoricians,
that essentially means that you have to convert the grammatical explanations into automatized hunches based on semantics. The conjunctivus in Latin may be optative, but that information is irrelevant for you when you speak or write the language - here you just have to
feel that you can get that semantic nuance by using a subjunctive. OK, there are cases where you don't have the choice: for instance "si" in French
must be combined with an indicative, whereas you can play somewhat more expressively with the modi in Italian after "se" - but this just means that the semantics have lost the battle in one isolated case in French, not that semantics never is relevant.
Chomsky fought hard against semantical descriptions in grammar for a reason: they can't be expressed in formulae, and they are mostly rather vague. But
all grammar is fundamentally based on semantics,
except where something has been automatized to the degree of French "si". A grammatical rule is just the formalization of a habit - and the surprise is that random habits can form systematic structures... but that's apparently the way things work ..