No, that's true for traditional teaching, which teaches rules and then encourages output as a means toward mastery of the language. The students have to learn rules to use the language. CI teaching does not teach rules (overtly), it presents comprehensible input. There is some degree of overt explanation, but it's generally under 5% these days (after considerable action research). Learning to do something is completely different from acquiring the language. There are no societies in which everyone -- every single member -- can play basketball, weave baskets, or milk a goat. But every normal member of every society acquires language without instruction.
This again is traditional teaching -- rules and output. The idea of this type of teaching is that students must be reminded of the rule if they do not use it correctly. In CI teaching, we rarely "explain" why you can't say "there is two dogs", we simply say "Oh, there are two dogs? There are two dogs. Where are the two dogs?" and go on.
Yes. This is called input. Not comprehensible input. I call it "linguistic waterboarding".
Traditional teaching. Students will someone magically generalize from a dialogue to fluency. It doesn't happen. In CI teaching (TPRS style at least) the idea is to present the new structures and vocabulary items frequently and in unexpected contexts, so that the brain gets the breadth of information it requires to generalize.
Let me translate this into CI: "Since I hadn't acquired the structure being used, I had no idea how to respond." You do not need to learn grammar rules or have them explained to you to have an idea of how to respond -- you need to have acquired the structures in question.
Yes, this is traditional rules-and-output teaching. You learn the rules, then you plug in some vocabulary and make sentences, with a teacher telling you "right" or "wrong". Then you've "mastered" that grammar point and that set of vocabulary and the class goes on to the next thing on the list.
And they all teach traditionally. There is no substantial presence of CI instruction in today's methods classes. I think that my teaching this summer at the STARTALK session in Hawaii might be the first time someone has put TPRS or CI into a STARTALK for Chinese -- and they do dozens of these things and have for some time now. All traditional teaching.
I feel accountability, of course. I'm teaching CI -- it's the best teaching strategy I know. But the fact is -- there is no assessment. I'm not around when the teacher has to make it work (or not), and the teacher can't exactly come after me and say, "Look here! I tried that CI stuff, and it didn't work!" Likewise, I don't have the opportunity to watch her in the classroom and say, "Well, you're going too fast, and see that student in the second row? See the glazed look? She doesn't understand what you're saying. Focus on her and...." (which is why we're developing coaching as a separate service, to help get people from the "aha!" moment of a workshop to being able to do this reliably day after day in a real classroom.)
It sounds to me -- and I'm not being snarky here -- that you have never experienced or probably seen anything close to CI teaching. You've experienced (the worst of) immersion, with lots of input that no one could understand, and that didn't work. You've experienced rules-and-output teaching. But nothing you describe in your own learning experiences is even vaguely CI.