On a teachers’ site, this question recently surfaced:

Ok suppose a kid said, “Mr/Ms Smith, I want more grammar teaching/practice” and you’re like, “no way, there but for the grace of Blaine go I.” You know C.I. works but you need a fast, simple demo to show that grammar (or really any explicit teaching of language) doesn’t work. How do you do that?

The fastest way to show that memorized rules application breaks fluency is a way I stole from Susie Gross, who used it in workshops.

Just find a native speaker of English, like, for example, the kid who’s asking for more grammar work, and say, “Can you tell me what you did yesterday? Just three sentences.” Let them say the three sentences. Then tell them, “Hmmm…you’re pretty fluent in English, aren’t you? I wonder if you could tell me what you’re going to do tomorrow — but don’t use any words that contain the letter ‘r'”.

Immediately the eyes will go up as the student applies this new rule, and the fluency will break as output becomes hesitant.

“You’re applying one memorized rule to a language you already know very well. Can you imagine how ineffective it would be to apply five or six memorized rules at the same time to a language you aren’t fluent in?” Then offer a good book in place of grammar exercises.