On a prominent foreign language teachers’ List, the question was raised:

Why are conversations with partners not CI? If one student asks “What is your favorite class and why?” and the other student answers intelligibly, then why is this less optimal CI than storytelling?

Well, of course it is. Sort of.  But McDonald’s is food, too — barely. Pairwork is the junk food of the language class. It provides poor-quality input. There are other legitimate reasons to use it at times, but providing optimal, maximized input isn’t one of them.

First, let’s think about what CI should ideally be:

  1. Comprehensible (totally understood by the student);
  2. Unpredictable (so that comprehension makes more impact);
  3. yet Structured to provide what’s needed (to guarantee that curriculum is covered)
  4. Compelling (or at least interesting) to get students to be engaged in thinking about what the language means rather than focusing on the form.
  5. Accurate (it should be the best available model of the language being taught, both in grammar and usage, as well as accent, intonation, and so on.)

Let’s think about a couple of possibilities.

1. Two students read each other questions and answer them. (Gets points for #1, 3, 5)

The two students are reading the questions from a list or book, so we know the grammar and usage are (should be) correct. So the issue of formal accuracy is taken care of. What’s missing is the optimization of the model input — students do not always have the best accents to emulate, and reading out loud is not a natural language task so the output will be a bit stilted at best anyway. Interesting? I suppose if you’re one of those straight-A students, or if you’re paired up with your best friend, you might find it moderately interesting to ask a whole list of “do you like” questions of your friend, one after the other. Unpredictable? Hardly. It’s fairly obvious to everyone what’s coming next; by the time they get to middle school, kids have a pretty good idea how textbooks work.

Input grade: 60% (barely passing)

2. Two students ask each other questions (no list or textbook used). (#1, 3, 5).

Same idea as (1) above, except points will almost always be lost for accuracy unless the students have already acquired the patterns being practiced — in which case, why are they practicing them in a language class? (I am assuming that these Q&A are intended for pattern practice, because I simply cannot think of another reason why students in a typical lower-level language class would ask a bunch of questions of each other.)

Overall grade: 40% (fail)

3. Teacher asks questions based on a list or textbook (#1, 3, 5)

Although the teacher asking the questions at least guarantees (we hope) a good accent and accurate delivery in a more animated way, this still lacks unpredictability. And that’s assuming the teacher isn’t going in alphabetical order calling on kids! And the most animated teacher can’t make “#4. Cuál clase le gusta más a María?” interesting.

Grade: 60% (barely passing)

4. Teacher asks questions based on a developing story created by the class. (#1, 2, 3, 4, 5)
This method hits on all the optimal characteristics of comprehensible input. It’s accurate, it hits the structures being emphasized, it’s interesting because it’s talking about content the class has decided on, not something imposed on them, it’s unpredictable because the story hasn’t been completed yet — what could be less predictable than that? It’s structured because the teacher knows what structure or items she is emphasizing, and takes care to repeat them using techniques that get in a lot of reps without getting repetitious.

Input grade: 98% (teacher got tired, it was seventh period, and there was a fire drill in the middle of class. Just kidding…)