The point is that the paper does not (and can not) distinguish one situation from the other.

Edit: your diagrams are too simplistic anyway. Papers can be read elsewhere. Citations have more causes than readership and its incremental effect will be heterogeneous.