Steve Kaufmann just put out a podcast on his blog entitled "Why I find grammar study largely a waste of time". Unlike my last contrarian title, which I intended to be more provocative than anything else, I actually stand fully behind this one.
As background, my approach to grammar can be found here, and a comparison of my approach to Steve's can be found here.
Why I largely disagree with what Steve's saying in his podcast, after the jump.
Let me start by offering a simple, utilitarian definition of grammar: grammar is the set of rules in a language that tells you (1) when and how words differ from their dictionary forms, (2) what order words go in when used together, and (3) when you need to use certain words with others. In his definition of grammar, Steve refers to this as "what happens in a language", but I find the reality of what grammar rules do to be much more concrete than such an amorphous definition.
Steve's approach to grammar is to review it as necessary. He finds that this doesn't do much good for him in turns of making things stick. And I can't say I'm surprised. Passive exposure to any piece of knowledge is a very weak form of exposure, and unlikely to lead to retention.
I also start off learning the grammar with a weak form of passive exposure by reading the rules. Because of how weak this exposure is, it's not a place to spend a lot of time. Sure, you'll get an initial repetition of all the rules through this step, but the big goal is to get an idea of the broad strokes of the language's grammar. You won't retain much, so reading (and rereading) is not the best way to spend your time.
To get something into your head quickly, you need to manipulate it, to interact with it. To give an example of this, I can still recall topics for which I wrote papers about in college to some level of detail, but ask me what I read in the books for those same classes and you'll be left with a blank stare. That's because the active output led to a much higher level of retention than the passive input.
So clearly active output can get something burned into your brain much more effectively than passive input. The question for me then becomes what kind of active output is best. And that's where I turn to an outline to start to get grammar rules burned into my brain.
In his podcast, Steve says that he thinks spending a lot of time on grammar is a waste of time. And I would agree, but I suspect our definitions of "a lot" are not aligned here. I typically spend my first few weeks of learning a language focusing on getting the grammar into an outline that's easy for me to understand. As I know I'm going to be learning a language for years, I don't see that level of time commitment to be a lot. After that, I refer back to the outline as necessary, fix it if I realize that something I put in there is screwed up, and add to it when I come across something new
Steve criticizes the use of terminology in his podcast. I'm with him here. Indeed, that's one of the roles of the outline; for you to create grammar explanations that you understand. I typically do away with all but the most basic terminology in these cases. So instead of "the first person singular present indicative conjugation of estar is estoy", I'd be much more likely to write something like "estar used with yo in the present becomes estoy". That is easy to understand and uses no jargon whatsoever.
The grammar book that Steve reads from in his podcast was quite a doozy with all its terminology. I'd say that if your book is that dense, you either need to just figure it out and put it into terms that you understand in your outline and never look at it again, or ditch the book completely for one that will require less effort. When I've encountered rules that, as written in the grammar I was using, just didn't make any sense, I usually asked a native speaker to write a few sentences for me using the rule. By looking at those sentences and discussing it with them, I can usually parse out the rule myself, which then goes into my outline.
I'm completely with Steve that things will start to feel natural because of exposure to the language. But how many exposures do you need to a given rule before you finally figure it out if you do nothing else? If you need 15-20 exposures just to learn a single word, and grammar rules aren't often as clear or as simple as learning the meaning of a word, I'd imagine the number of exposures you'd need would be a great deal more. Sure, it'll eventually get clearer, but this just doesn't seem like the most efficient way to approach figuring out grammar rules.
On the other hand, if you get the rule in your head right off the bat, those 15-20+ exposures will all be a repetition of something you already know. Instead of needing to puzzle out the grammar, you'll be reviewing it. Using a more input-focused method, it'll be a longer time before those exposures you're getting to the grammar become meaningful exposures. And meaningful exposures are the ones that are going to get burned into your brain.
Steve notes some examples of how long-time language learners continue to screw up their use of the language. I would say that this is less a problem of whether or not they focused on the grammar than it is a problem of how much output they've produced and whether or not they've been corrected when producing that output. I would further venture that Steve's own method is more likely to lead to such output mistakes than a method that pushes output early and often and actively seeks to get such output corrected. But that, of course, is a wholly separate issue from grammar.
The final point I'll make is that there is no dichotomy between grammar and vocabulary / input. Anyone who thinks you don't need to focus on expanding your vocabulary is just flat-out wrong, regardless of how they feel about grammar. The real question here is whether to spend some time focusing on materials that aim to tell you the grammar rules up front, or whether you simply try to absorb grammar rules from exposure to the language.