Thanks for all of the responses. I do see where my third example wouldn't work under this rule. A better example would have been
Seymour Lubetzky : writings on the classical art of cataloging / compiled and edited by Elaine Svenonius, Dorothy McGarry.
That also means every new David Sedaris book will have "Short stories. Selections" on it. I just don't see how "X. Selections" adds anything of value to the description.
And while these two examples above would be an easier connection to make to this rule, I can easily go find any number of nonfiction works by single authors on the shelves where each chapter can be taken as its own work and stand independently of the other chapters. Are we to then evaluate every nonfiction book to decide if the chapters can stand on their own as works and apply this? Because the way this rule and its PS are written, it seems pretty matter of fact.
Also, the only reason that Silverstein's "Falling Up" or Whitman's "Leaves of Grass" have "become known by a single title" is time and fame. If this rule had been in place when those books were originally published then they too would have "Poems. Selections" squished between the authorized access point and the title proper. So nothing new will ever have time or fame in its favor to NOT get "Selections".
For every book of poetry or short stories or papers that bears a collective title to also get these ambiguous work titles seems to be overkill.
Sorry, I just find this to be the wrong direction and for RDA cataloging and while I see that it was intended to perhaps make decision making easier, it also seems to take the "work" concept beyond the way most people think about it.
Metadata & Emerging Technologies Librarian
Professor, Library Services
Minnesota State University, Mankato