I'd Rather Count Snowflakes

Diana Grieman and I were talking about order of acquisition. She said this:
And as you’ve been saying, probably the very most important thing that we can do is to stay as much as possible in the target language.  That, I think, is bigger than order of acquisition and complexity.  From lots of CI, the rest flows.
I agree with that. Staying as much as possible in the target language is more important than fretting about order of acquisition and complexity, and all of those things that we can analyze and get all into but ultimately can’t control anyway, in spite of our best efforts.
I don’t see what’s wrong with just trusting in the natural unfolding of the language in our classrooms. Are we really that fearful that a certain expression may not show up unless we consciously plan for its occurrence in our classes?
To me, languages are just so big, and there are so many possible combinations of words, that, if we but use the language over 90% of the time in our classes, the words we want our students to know will occur and they will learn them.
Yes, we need target expressions, and we need to establish meaning and hinge our lesson on that establishment of meaning, of course. But whatever expressions are in front of us from a story, or a song, or a text, or just happen to be in the air that day, in my view, will serve just fine.
All we have to really do is make the CI that we deliver crystal clear and interesting to our students. We can have the best list of frequent expressions ever designed in a nice TPRS scope and sequence (to me, a contradiction in terms), and it won’t make any difference unless we do good CI.
I, for one, refuse to take on anything as big as figuring out what words to teach my CI students when. I’d rather count snowflakes. O.K. – all you proponents of high frequency lists, deconstruct my argument.