Declarative vs. procedural knowledge

Before we go any further, click on the link for a short post explaining the basic differences between declarative knowledge and procedural knowledge.

This is an e-mail I’ve just received:

Dear Luiz,
When teaching grammar, it seems we can never control what students use… and sometimes I’m not really sure what to do. I’m teaching a certain structure and, when the production phase starts, they use something different (but something that makes sense…)… do you understand what I mean? What do you think is better to do? I want them to use the new structure, but they can get what they want by using something different.
Laís
This is my reply, which addressed the issue of declarative knowledge vs. procedural knowledge:

Dear Laís,

Your question reminds me of an old anecdote that has become something of an urban legend in contemporary ELT and that connects interestingly with the concepts of declarative knowledge and procedural knowledge:

Teacher: Ok, guys, so now, let’s review the present perfect.
Student: Oh, no, not again. We already do the present perfect four times this month. Why teaching grammar? Give us a song!

This (seeming?) lack of interface between what is taught and practiced (declarative knowledge) and what is actually deployed in communication (procedural knowledge) has been haunting me for as long as I can remember. It nagged at me to the point of despair in the mid 90s, when I still believed that a task-based model was THE key to teaching grammar, as long as teachers could make it work in tandem with a pre-determined grammar syllabus. Little did I know.

But I digress. A few years later, still baffled by students’ apparent imperviousness to the way I was teaching grammar (i.e., their inability to turn declarative knowledge into procedural knowledge), I set out to write a whole M.A. dissertation investigating the relative merits of different options in teaching grammar. I wanted to discover which activity types were more likely to help students proceduralize the new grammar – in other words, turn declarative knowledge into procedural knowledge. Guess what I found?

That maybe I had the wrong research question.

So, again, Laís, I don’t have an answer to your question. I do have, however, a few observations regarding declarative knowledge vs. procedural knowledge that I’d like to share.

1. I think we ought to begin by asking ourselves why we want students to produce a given piece of language (“the structure of the day” in the words of Scott Thornbury) freely in the first place. This decision seems so intuitive that it’s hard to examine it critically, but, I insist, why have we traditionally always tried to engage students in “freer” activities focusing on a certain structure? To give our lessons a sense of closure and achievement? Maybe. And this, in itself, is an entirely valid reason, of course. But if we’re looking for solid evidence of student learning (assuming that learning = turning declarative knowledge into procedural knowledge), we might be barking up the wrong tree. Accuracy in the “production phrase” could be evidence of conformity(to use Dave Willis’ words) rather than activated declarative knowledge. So where does that leave us?

2. Certain structures will take a long time to cross over into spontaneous communication (i.e., to go from declarative knowledge to procedural knowledge). Period. The teaching of grammar (with a view to proceduralization), in that sense, is more like a time bomb rather than fireworks – in other words, it doesn’t explode as soon as it’s set off. We may try to speed up the process, but students will only use the “new stuff” (i.e., turn declarative knowledge into procedural knowledge) when they’re developmentally ready to. Now, if you work for a school that operates with a set syllabus, a coursebook (to be covered in X hours) and a series of achievement tests (on the basis of which pass/fail decisions are made), this whole discussion on declarative knowledge may sound really far-fetched and impractical. And you know what? I don’t blame you.
3. During controlled practice, certain activity types tend to ease students’ way into the “production” phase more effectively than others, as I argued in a recent post. This explains why, when teaching grammar, moving from a gap-fill to a role-play is usually a recipe for disaster. To help students convert declarative knowledge into procedural knowledge, there has to be some sort of “semi-controlled” practice before the “production” phase. For example, if the freer activity is story-telling (narrative tenses), during controlled practice, showing students pictures and asking them to describe what happened (with relatively few cues) might prove to be more effective than simply using a cloze exercise with 20 blanks and the base form of the verbs in parentheses (a coursebook favorite). In other words, not every kind of controlled practice is equally effective in terms of automaticity(turning declarative knowledge into procedural knowledge.)

4. There seems to be an inherent tension between getting things done and using the right language (activating declarative knowledge). Attentional resources are always competing against one other and the more attention students need to pay to what they’re going to say (content), the less attention they will pay to how they’re going to say it (form). So, when teaching grammar, it’s up to us to help students shift back and forth between meaning and form in the best possible way, so we can help students go beyond declarative knowledge. This can be done basically through task rehearsal (encouraging students to plan what to say before the actual activity) and task repetition (carrying out the same task with a different partner), both of which do a good job of freeing up “brain space” to help students perform better (i.e., turn declarative knowledge into procedural knowledge). Click here for an interesting article on the effects of task rehearsal, task repetition and task familiarity on students’ accuracy, fluency and precision.

5. And then there are lexical phrases to complicate matters further. What if it’s actually true that formulaic chunks(expressions that largely bypass grammar analysis) are stored, learned and used a little differently from grammar? Could it be that sentences such as “How long does it take to get from A to B?” or “No matter how hard I try, it’s still difficult for me to ___” can – and indeed should – be drilled over and over until students can use them more or less automatically (i.e, move from declarative knowledge to procedural knowledge)? There is, after all, a memory component to chunk learning, which means that maybe – and I say maybe – teaching grammar ( teaching, here, meaning helping students turn declarative knowledge into procedural knowledge) bears less resemblance to teaching lexical chunks than we’d like to believe.

Laís, I’m not sure this is the answer you were expecting to get, but it’s really the best I can give you right now. If you have entirely too much time on your hands, consider reading my critique of task-based learning, which I wrote back in my Lancaster days.

Thank you for writing.

Luiz