These questions (and there are several) are all good introductions to major themes addressed in Knowing the Learner. I will give some general answers here, but also direct the reader to portions of the book where these issues are addressed more fully.
The reader begins by referring to the fact that when outcome information concerning distinct learning goals is mixed together, typically by aggregating performance scores to get a total score, that the meaning of the original goals is lost and one cannot use the aggregated score for an educational purpose. While this is a good general rule of thumb, it is necessary to recognize that there are instances where aggregation across discrete learning goals can be helpful. One of these is presented in the section of Knowing the Learner called ‘Is Aggregation Across Diverse Learning Goals Necessarily Inappropriate’ (pp. 45-50)
Regarding the question of grading, please consider that we have not chosen to oppose the idea of grading in educational programs. Rather we try to make explicit the conditions when grading can be and those were it is not educationally productive. The whole first chapter of Knowing the Learner can be considered a critique of the grading paradigm, and yet we conclude this chapter with a section called ‘The Redemption of Grading — Accomplishment Based Grading’.