True, but the 520 is also the other source of BIBFRAME Annotation types (Review, Abstract, ContentAdvice). Also, the principle Stephen raised ([publisher] bias) still holds, especially with the the Publisher description and (potentially) Contributor Bio (bias is really not a factor with Sample Text and Table of Contents).
Acknowledged - thanks both.
I think the point here about 'bias' is back to the issue of the 'who' being important. However, I'm back to my confusion on this - if the annotation approach doesn't inherently allow you to say 'this statement is made by the publisher, not the library', then why use an annotation? It may as well be just another bibframe property? (at least, knowing that a library stated this is a publisher description is no more important to me than knowing that the library stated this is the title from the title page - which I think is one of the points made by Tom Meehan)
Owen