Thinking about (comprehensible) input and ouput
Sunday February 6, 2011
"Comprehensible input" is a theory about language learning, which has this simple and appealingly (to me) mathematical explanation. If somebody knows "n", they learn "n+1" when they encounter input that includes "n+1". The example being, and I think I'm paraphrasing all of this from the wiki, that if a kid knows what "kick your brother" means, and the teacher says "kick your sister" and points to the sister, then the kid learns what that means.
This is tempting, but in the end I think it is not such a great theory, at least not for adult learners. I do appreciate the related application of the term comprehensible input that means that language classes should be made comprehensible to the students, i.e., don't talk way over their heads or whatever, and even that the appropriate level of discourse in the classroom is just slightly over the students' current level. That's all cool.
But I don't think that's how most of second language acquisition happens. As an example, I just don't trust stuff that I figure out or guess myself until I confirm it with a dictionary or by asking about it. It's too easy to make mistakes. Just yesterday I saw a book with a big picture of a shark on it and a big word that I didn't know, but that was kind of familiar. So I guessed that it was shark. But it wasn't even related to sharks. So that's one problem. People make mistakes. The other problem, I guess, is that people just won't figure out enough on their own, fast enough, to make good progress.
People need to make the target language comprehensible to themselves first. That happens by learning vocabulary and grammar, by studying explicitly, by being taught explicitly. Things must be made comprehensible.
I've just thought of a caveat to my acceptance of comprehensible input for adapting the language level to that of learners. I think that should be limited, because it is not the goal level of the language. You don't want to read children's books your whole life, and you certainly don't want to be stuck in the goofy stage of a partially formed and cumbersome, awkward grammar. It is necessary for learning some basic concepts, but as much and as soon as possible the language input should be native.
I've been appreciating the value of a lot of native language input recently as I watch a lot of episodes from 지붕뚫고 하이킥, a Korean sitcom. But what I find is not that I learn new things from it, usually, but rather that I practice (or am reminded of) things that I have learned, and moreover learn how things that I have learned are actually used. I can almost never make out words that I don't know, and I can't really make out foreign grammar (not that there is much of it) but the bigger issue is that after learning vocabulary and grammar in other places, there are infinitely many ways to combine them, only a handful of which are actually used by native speakers. Watching a show like this provides a lot of known-good expressions, which I can understand because I have already learned their components elsewhere. This reinforces things I've learned and guides me toward more natural expressions.
Somehow I feel like this happens more when I watch TV than when I read - maybe because I slow down and break apart everything I read so much. Or maybe because I go fast and don't spare thought about sentence structures etc. Huh.
I guess really the main thing that caught my eye about comprehensible input was a corollary on the wiki page: "speaking (output) is not practice." This kind of resonated for me, or maybe just gave me permission to think about things like this.
How can you learn from speaking, or writing? If you make a mistake, you won't necessarily know it. It's like waving around a flag; it won't make the flag any bigger or cleaner. And it isn't hard to think of people who speak a language plenty and yet still make tons of the same mistakes, with terrible accents, etc.
But then back at input, you can easily imagine someone spending a lot of time listening to (and/or watching) a foreign language, and never learning anything. It's not easy to bootstrap up from nothing to competency. I have to think that those folks who go and learn undiscovered languages or whatever are pretty incredible. That's not easy. Even if you could already read at some level, you could read and read and never improve your language skills if you weren't trying to. Consider Korea, where tons of foreigners come for a year or more, the writing system can be learned in a matter of hours, and while yes a lot of foreigners do avoid Korean language scenarios, they certainly aren't isolated from it, and yet they never learn anything.
For all that, I think all of listening, reading, writing, and speaking should be called practice. Importantly practice, as distinct from learning. They are all useful and even important for taking what we have begun to learn but not yet perfected and making those skills more secure, fluent and useful. The productive skills are especially important for developing new skills and knowledge as they are learned. Listening and reading become more valuable after a significant base has been built; otherwise too much will just whiz by.
Conversation... Conversation seems like more than just listening and speaking. It gives some opportunity for feedback, although that will not necessarily be perfect. I think it can still be taken as a kind of practice. All the language skills are practice, aren't they, the practices of native speakers too. I think they are unlikely to lead to acquisition of new forms, at least not quickly - I think that's my point here.
I guess that doesn't make it wrong to call a class that's just "free talking" a class... I just usually think of teachers teaching new material.
Summarizing: "Just" reading, or listening, or speaking, or writing, is not an efficient way to learn new material. It is, however, a valuable way to review and reinforce material that has already been initially learned. The same goes for conversation, with the qualification that because of interactivity it has more potential for learning than the others, but like the others the learning value will depend on the participants.
This post was originally hosted elsewhere.