4.12721 With an object that falls under it, a formal concept is already given. Thus one cannot introduce as primitive ideas the objects of a formal concept and the formal concept itself. Thus one cannot (like Russell) introduce, e.g., the concept of a function and also special functions as primitive ideas; or the concept of number and specific numbers.

Yes, Russell seems to be making a mistake here about what it makes sense to call “primitive ideas.” So what is (or should be) primitive? The concept under which certain objects fall, or the objects themselves, which perhaps somehow bring the relevant concept with them? Or can there be no true primitiveness or foundation here?