So you could have:
A is a performance of Z
C is an arrangement of Z
F is a performance of C
All of these seem to be legitimate statements, although they may not map precisely to the FRBR Group 1 entities. I think this is an example of what I’ve been arguing — that it’s more important to establish the relationships than it is to fit everything into a set number of levels.
They can be legitimate statements despite those relationships not being the ‘canonical’ Group 1 to Group 1 relationships! FRBR does not limit you to one and only one kind of relationship only between one Group 1 entity and it’s immediate hiearchical superior!
In the above list, A is an expression. Z is a work. C is an expression. F is also an expression. Expression F may be related to Expression C with a “is a performance of” relationships—-that relationship may or may not be in FRBR right now, but should be! FRBR certainly allows you to draw relationships between one expression and another, or one work and another, or one exprsesion and a work other than it’s parent work, etc.
The specific list of relationships in the FRBR document (which may not include ‘is a performance of’, but may include some more generic category of this), the FRBR document all but admits was just an initial sketch of what relationships were possible, and was not meant to be definitive or complete. That we haven’t actually done work to polish off this list of relationships in the past ten years—is a pity.
But if “A is a performance of Z” means that Z is a work, doesn’t “F is a performance of C” mean that C could also be a work? As some of our music-enabled colleagues replied, by some definitions, an arrangement is a separate work. Could C be simultaneously an expression and a work?
I don’t know the answers, but I think we need to explore this further before we declare that we have four possible boxes (WEMI) and that everything fits neatly into one or the other. That’s why I think it is more important to state that C is an arrangement of Z than to say what box it fits into. It can remain an arrangement of Z and find itself in different boxes based on different contexts or functions.
A picture is worth a thousand words…
… and attached this great diagram showing how the whole shebang can be described in the FRBR model using the RDF syntax (197 KB PDF). (Image belongs to Rob Styles of Talis.) Have a look, it’s boss stuff.