In my opinion, the best way to understand Work is as the set of all 
expressions/manifestations that... belong to that Work.   "Work" is sort 
of a culturally constructed concept, but we know it when we see it. But 
I think the WEMI heirarchy is best understood as set relationships -- 
while not explained this way in the FRBR document itself, it is in no 
way _incompatible_ with the FRBR model, it's just another way to look at 
the FRBR entities.

The more traditional way to look at the FRBR entities is that more 
'platonic' approach as you say -- while I think it's ultimately 
equivalent, I think it's a lot more confusing to talk about platonic 
things that don't actually exist, then it is to talk about sets of 
things that do exist.

I think you may find this earlier essay I wrote on the same subject 
helpful:

http://bibwild.wordpress.com/2007/12/07/frbr-considered-as-set-relationships/

On 12/14/2010 2:11 PM, Susan Kane wrote:
> This is not my area of expertise ... but .... if Work in FRBR doesn't mean
> any particular manifestation, expression or item ... what does it mean?
>
> Do Works live in Plato's world of Ideas where abstracted version of things
> exist in a more real and more true sense than any shifty mimes in our world
> of sensation and change?  http://en.wikipedia.org/wiki/Theory_of_Forms
>
> Is it a bit like copyright law -- Works (ideas) can't be copyrighted but
> manifestations, expressions and items can?
>
> Or is Work actually an empty container that doesn't really exist -- even in
> the World of Ideas hovering over the Earth -- until it is filled with at
> least one (manifestation, expression, item)?
>
> And if so, who chose this particular word?
>
> Was something wrong with calling it ... Idea?  Or even ... Thing?
>
> Why work?  Work to me in plain English and even librarianese definitely
> implies a manifested thing, not the idea of thing that transcends any
> particular and specific expression.
>