Question From Anne

Here is a question from Anne. I guess the output thing never got a fork stuck in it (will it ever?):
We have learned that lots of input is necessary before students can begin to produce output.  It is also commonly accepted  that some students will be ready before others to produce output .  We give lip service to the idea that it is okay for students to not speak until they are ready.  It feels right, from a pedagogical and also human point of view, to assess students solely on comprehension when they are starting out in a language.
At what point, then, and to what degree, do we begin to hold students accountable for output?  To what degree are they accountable for accuracy?  If we do not teach grammar rules with worksheets and quizzes, can grammatical mistakes ever affect their grades in a negative way?  If so, when?  And to what degree?
So the question is asking at what point can we reasonably expect and assess output, and in what form?  I am hoping that someone has been doing this for longer than I have and can actually pin point a time, in the 4 years, when requiring output is reasonable and appropriate.
I am striving for integrity in my program.  I want to teach only what is important, in a way that is effective and enjoyable.  I want to assess only the skills and knowledge that are important and that I have spent the most time working on in class.  I want that grade to reflect the degree to which the student has met the reasonable and attainable expectations that I have set for him/her. 
Right now I can’t say that that is the case.