Doesn't 6.2.2.10 distinguish between compilations (loosely defined)
that are works with their own known titles and compilations that are
works but lack such a title? In that case, Olds' "Stag's leap" would
fall under 6.2.2.4 and should not present the problem which Ryan
describes. The problem I see is deciding which compilations would need
to be treated under 6.2.2.10.

An editor assembles all the Shakespeare passages which mention bees or
honey and publishes them under the title "To bee or not to bee" to
modest success. The title appears on the resource embodying the work,
will undoubtedly be known to those who know it, and will be cited as
such in the writings of bee enthusiasts. The compilation may not be
one Shakespeare intended, but it is certainly not random. It passes
the test at 6.2.2.10.  Should it be entered under "Shakespeare,
William, 1564-1616. To bee or not to bee"?

Ay, there's the rub. The real world notion of works distinguishes more
sharply between canonical and non-canonical work titles than the RDA
definition.  The problem is less about the compilation's title and
more about its standing as a recognized "work" of a particular author.
 Maybe what PCC could propose is an LC-PCC-PS which would clarify that
"in reference sources" refers to reference sources that discuss or
enumerate the "works" of an author and not simply any external source
that refers to the title in hand.

I mostly agree with Bob Maxwell's point about differentiating
authorized access points, but would hate to see the collocation value
of collective titles get lost in the process. Either we'd want systems
that could provide both a list of the differentiated authorized access
points Bob describes and a list of titles proper under truncated forms
of those access points, e.g., so that I could sort the set of
"Shakespeare... Poems. Selections"  by date; or we'd want to allow
authorized access points for higher levels of abstraction, e.g.,
"Shakespeare... Poems. Selections," recognizing that they represent an
abstract level of collocation on which more precise authorized access
points can be built by extending the number of qualifiers.  In the
latter case, libraries could decide what level of differentiation
would best suit their users and collection.

What doesn't work is trying to establish an authorized access point at
the more abstract level but having its authority specify a discrete
list of titles.  In too many cases that list of titles will be
incomplete and ultimately misleading.  I fully agree with Bob's
implied point that a 400 reference for "Shakespeare ... To bee or not
to bee" should refer specifically to an authorized access point for
that compilation, and not to "Shakespeare ... Works. Selections,
2013," since my collection may need the latter access point but lack
the former title, and include a title not found in its 400s.

Stephen

On Fri, Feb 22, 2013 at 7:21 PM, Robert Maxwell <[log in to unmask]> wrote:
> I am completely sympathetic to what you’re saying, Ryan. I want to point one
> thing out, however. We shouldn’t be creating undifferentiated work access
> points. Leaving aside the issue of the title of author-originated
> compilations such as Stag’s leap—I agree, that’s the commonly known name for
> this aggregate work and should therefore be the preferred title—let’s
> suppose we have some examples of the “after the fact” compilations by
> someone other than the author (e.g. “Selected poems of Shakespeare”). The
> conventional collective title for these would be “Poems. Selections”—at
> least, that is how the preferred title would begin. But RDA tells us to add
> things if the title of a work is the same as the title of another work by
> the same creator. This is brought out in the core statements at 6.3 (form of
> work), 6.4 (date of work), 6.5 (place of origin of the work), and 6.6 (other
> distinguishing characteristic of the work), together with 6.27.1.9
> (additions to access points representing works).
>
>
>
> Now a collection of poems by a particular author is a work in and of itself,
> an aggregate work, but a work nonetheless. RDA 5.1.2 makes this clear — “the
> terms work and expression should be read … to include … aggregates of such
> entities.” This is true whether the collection was put together by the
> author or by somebody else. It follows, then, that a different collection of
> poems by that author is a different aggregate work. If both collections are
> of the “after the fact, not collected or planned by the author” type, we
> would use the conventional collective title “Poems. Selections” as the
> preferred access point. But that title is not enough—in our hypothetical
> situation we have at least two works with the same title “Poems. Selections”
> and so the core statements with 6.27.1.9 kick in and we should be adding one
> of the four differentiating elements, whichever makes most sense (or in
> RDA’s language “as appropriate”). For instance, if a collection is well
> known by the name of its editor, we’d add that, possibly:
>
>
>
> [Poet’s authorized access point]. Poems. Selections (surname of editor A)
>
> [Poet’s authorized access point]. Poems. Selections (surname of editor B)
>
>
>
> Alternately, if it makes more sense, we could add form, or date, or place of
> origin, or something else.
>
>
>
> I just wanted to bring this up to help people to think this through.
> Different collections (aggregate works) are different works but if their
> preferred title is a conventional collective title then they share the same
> title and the title needs something added to it to differentiate between
> those works.
>
>
>
> Bob
>
>
>
>
>
> Robert L. Maxwell
> Head, Special Collections and Formats Catalog Dept.
> 6728 Harold B. Lee Library
> Brigham Young University
> Provo, UT 84602
> (801)422-5568
>
> "We should set an example for all the world, rather than confine ourselves
> to the course which has been heretofore pursued"--Eliza R. Snow, 1842.
>
>
>
> From: Program for Cooperative Cataloging [mailto:[log in to unmask]]
> On Behalf Of Finnerty, Ryan
> Sent: Friday, February 22, 2013 5:21 PM
> To: [log in to unmask]
> Subject: Applying collective titles to works of poetry, etc.
>
>
>
> Hello PCC List,
>
>
>
> We here at UC San Diego are gearing up to begin RDA for bibliographic
> records in early March so we’ve been going through the LC training materials
> with our staff. The refresher training for compilations (available at
> http://www.loc.gov/aba/rda/Refresher_training_dec_2011.html), plus the RDA
> records coming into our catalog, have us concerned. We hope to open a
> dialogue here and see what other people think. We welcome your thoughts!
>
>
>
> Ryan Finnerty
>
> For the UCSD RDA Planning and Implementation Group
>
>
>
> **************
>
>
>
>
>
> We are concerned about LC’s interpretation and application of RDA 6.2.2.10
> to books of poetry, short stories, essays, etc., in that they use a
> conventional collective title with Selections as the preferred title for the
> work. Note: For the rest of this message, we will use poetry as an example,
> but the same applies to essays, short stories, etc.
>
>
>
> The instruction reads, “If a compilation of works is known by a title that
> is used in resources embodying that compilation or in reference sources,
> apply the instructions at 6.2.2.4-6.2.2.5. For other compilations, apply the
> instructions at 6.2.2.10.1–6.2.2.10.3, as applicable.”
>
>
>
> So in order to apply a conventional collective title you have to believe
> that an original book of poetry is both: 1) A compilation,  and; 2) Not
> known by its own title. We believe that neither condition applies.
>
>
>
> In regards to the first point, we are not talking about books of poetry
> whose contents are mash-ups of poems previously published in disparate
> editions. These type of resources are compilations and should get a
> conventional collective title.
>
>
>
> We are talking about books of poetry, often issued for the first time, whose
> contents are intended by their creators to exist together as one work. Many
> books of poetry are thematically intertwined and are meant to be experienced
> as a collective whole by the reader. And while yes, any one poem can be
> extracted and exist on its own, the same can be said for chapters from a
> novel.
>
>
>
> For example, take this work by Loren Erdich “I take back the sponge cake.”
> Based on the Amazon blurb, this cannot be considered a mere compilation of
> poems, but a poetic work intended to exist as a whole:
>
>
>
> I Take Back the Sponge Cake is a choose-your-own-adventure collaboration
> between poet Sierra Nelson and visual artist Loren Erdrich. Each turn of the
> page features an ink and watercolor drawing, a poem, and a choice between
> two sound-alike words that create a variety of paths through the book. The
> adventure always begins in the same place, but depending on your choices
> your reading experience moves by emotional meander--leaping, looping, and
> surprising until it finally reaches one of the possible endings.
>
>
>
>
>
> Or Sharon Olds latest work, “Stag’s leap” a poetic work that tells a
> continuing story from start to finish of her husband leaving her to her
> finding new freedom. You can isolate a single poem  and consider it a work,
> but it’s the larger work that tells the story or creates the mood that the
> author intended.
>
>
>
> For the second point, how could the two above works not be known by their
> own titles? Anyone looking for these no matter what the context (in a
> bookstore, online marketplace, library, etc.) would be looking for them with
> the title the author gave them.
>
>
>
> In a list of titles under an author, how could a bunch of entries such as
> “Poems. Selections” be considered helpful? Users would have to go through
> every record with the collective title to find what they want rather than
> just selecting the one desired title from the list in the first place. This
> seems to place the convenience of the cataloger above the convenience of the
> user.
>
>
>
> We see the following practical problems with this practice:
>
>
>
> 1) When it comes to authority work, we are, in essence, creating
> undifferentiated work level authority records. One of the major objectives
> of moving to RDA is to lay the groundwork for linked data. According to LC’s
> own RDA training slides showing the WEMI relationships, a work is linked to
> an expression, is linked to a manifestation, is linked to an item. If a work
> record represents more than one work, it introduces chaos into this linked
> model. See n2012075841 and no98041871 as examples (the latter even conflates
> two authors on the same authority record—very confusing when putting more
> than one work in a record).
>
>
>
> 2) What do we do when a work becomes famous at a later point and its title
> now meets this nebulous notion of being “known”? Do we go back and scrub the
> collective title off the records and revise and/or cancel authority records?
> I imagine LC and others have no plans to do this, but for current works that
> will someday enter the canon, this poses a big problem.
>
>
>
> 3) How do we represent the work in subject access for works of secondary
> literature? Will we have the following subject field for a work of criticism
> written expressly about “I take back the sponge cake”:
>
>
>
> 600 10 Erdich Loren. $t Poems. $k Selections $x Criticism and interpretation
>
>
>
> We very much understand LC/PCC’s desire to leave behind LCRI 25.10. We want
> to leave it behind as well. A lot of people never understood this rule, and
> those that did had differing ideas of adequate and distinctive.  We simply
> propose that for works of poetry (etc.) you give the preferred title as
> instructed in 6.2.2.4-6.2.2.5 unless it is evident from what you have in
> hand that is consists of a combination of different poems from previously
> published works, or a combination of old and new. For example, Barry
> Wallenstein has a work entitled “Drastic dislocations : new and selected
> poems.” Giving a collective title here would be appropriate and expected.
> However, we would make a distinct and differentiated authority record for
> this work with proper additions to the authorized access point so if anyone
> comes to the catalog searching for “Drastic dislocations” they would be
> referred directly to the correct record. And additionally, if a new work by
> one author came in called simply “20 poems” and there was no evidence this
> was a combination of poems previously appearing elsewhere, then let “20
> poems” be its preferred title.
>
>
>
> At UC San Diego, this is not just a philosophical issue. We collect
> extensively in modern poetry and these conventional collective titles are
> already causing confusion in our catalog. Our literature librarian has
> noticed them and is asking us to remove them from the records. Since we want
> to (and currently do) accept PCC records as they are without review, we
> would love to see this practice changed.
>
>



-- 
Stephen Hearn, Metadata Strategist
Technical Services, University Libraries
University of Minnesota
160 Wilson Library
309 19th Avenue South
Minneapolis, MN 55455
Ph: 612-625-2328
Fx: 612-625-3428