Antelope, Document
Many thanks to Lorcan Dempsey's weblog for pointing me to Michael Buckland's very interesting paper, "What is a document?" This preprint of an article published in JASIS asks whether sculpture, museum objects, and live animals could be considered "documents." This is a Xobian question if I've ever heard one.
The paper starts off by tracing the progression of the language used to talk about these issues (from "bibliography" to "document" to "documentation" and back to "document"). What I find interesting is that it doesn't make it all the way back to bibliography. It seems no more of a stretch to say "bibliographic" when talking about an antelope than it does to say "document." I wonder, perhaps, if the term's regression stopped at document because, as Buckland says (while talking about why document was originally used), "it was felt that something more than traditional 'bibliography' was needed…"
I've always been a little uncomfortable with XOBIS' name: The XML Organic Bibliographic Information Schema because of this very issue. Is 'bibliographic' really the best word to talk about events, objects, beings, etc.? What I do like, though, is that we kept it (instead of trying to find a new term). It was a nod, I think, to recognizing that these types of issues are the issues of librarianship. Computer scientists and "knowledge managers" may reframe some of the terms, but this really is the realm of library science (not that they aren't welcome to join us here).
Buckland's paper traces many cases where people used "document" to indicate the thing that is representated by a record in an information system. In particular, he cites Suzanne Briet who says a document is "any physical of symbolic sign, preserved or recorded, intended to represent, to reconstruct, or to demonstrate a physical or conceptual phenomenon." Briet goes on to list six objects and asks if each is a document:
- Star in sky — No
- Photo of star — Yes
- Stone in river — No
- Stone in museum — Yes
- Animal in wild — No
- Animal in zoo — Yes
Buckland infers, from Briet's definition and list, four rules for determining if something is a "document":
- There is materiality
- There is intentionality
- The objects have to be processed
- The object is perceived to be a document
Now to backtrack a little into XOBIS. Why "document?" The term just seems too general to be useful. What we are really talking about are beings, objects, places, and works. In XOBIS speak, these are substantive instantiations (of a concept). They are distinguished from the notional instantiations (string, language, organization, event, and time) and the purely conceptual by their, as Buckland says, materiality (e.g., they can be collected). So, in XOBIS (and this is just a stab… expect to see comments as I change my mind or refine my thoughts):
- Star in sky — Place
- role="authority" (if notable)
- [grouped in its Concept if not notable]
- Photo of star (held by library) — Work
- role="instance" or
- role="authority instance" (if notable)
- Photo of star (held by another library) — Work
- role="authority" (if notable)
- Small stone in river — Object
- role="authority" (if notable)
- [grouped in its Concept if not notable]
- Large stone in river — Place
- role="authority" (if notable)
- [grouped in its Concept if not notable]
- Stone in museum — Object
- role="instance" or
- role="authority instance" (if notable)
- Stone in another museum — Object
- role="authority" (if notable)
- Animal in wild — Being
- role="authority" (if notable)
- [grouped in its Concept if not notable]
- Animal in zoo — Being
- role="instance" or
- role="authority instance" (if notable)
- Animal in another zoo — Being
- role="authority" (if notable)
One difference between this list and Briet's is that XOBIS contains things that are not documents. So, while Briet's list says "No" if something is not a document (it only deals with material things that satisfy the conditions listed by Buckland), XOBIS tries to figure out where it might fit in the broader scheme. This is, of course, not a weakness of the original list… for it, these things are out of scope.
The things in this list that are not held in a library, zoo, or museum's collection are authorities (if they are notable enough to warrant a record) or are grouped into their Concept record (if they are not collected or notable enough to warrant an authority — this is the same thing as a "No" in Briet's list). A stone in a river is not notable unless it has some distinction (has something about it that would cause people to want to write about it, create art about it, talk about it, etc. — essentially, have a reason it would need to be referenced from another record). If it is notable (e.g., people will want to create Xobian relationships to it), it warrants an authority record. If the stone in the river doesn't have any particular distinction (and hasn't warranted a human assigned designation) then it is just a part of Concept: Stones (or something to that effect).
What is interesting is that there is a line here that is a bit hard to discern. When does a stone that is famous, not owned or collected by anyone, and not large enough to be considered a Place become an Object? Is there even such a thing? Would a stone that is large enough to not be collected but still important enough to warrant being an authority record necessarily be a Place? [Thanks to Dick for hashing out these ideas with me before I hit the 'publish' button on this weblog] While creating XOBIS (awhile ago now), we tried to reinforce every decision or statement we made with examples. Perhaps I need to find one here.
Anyway… things (beings, places, objects, and works) that can be held in a collection may either warrant "instance" records or "authority instance" records. A stone that has been removed from a river (but that does not have any significance apart from being a part of a particular collection) would be an "instance" record. The Hope Diamond, on the other hand (a rather notorious stone), would warrant an "authority instance" role for the institution that holds it and an "authority" role for those that don't.
Likewise, a random elephant out in the wild would not warrant a record (it would be handled by its Concept authority record). An elephant that is a part of a zoo's collection, but that has no significance apart from that collection, would be a role="instance" being. An elephant that belongs to an organization but, who is notable outside of the organization's particular collection (e.g. Elsa, one of Thailand's painting elephants), would warrant an "authority instance" record. If another museum were to collect Elsa's paintings, it would need a Being record with an "authority" role so that relationships from the records for the paintings could be made.
Interestingly, all Buckland's inferences seem also to apply to XOBIS. The XOBIS elements, however, have a greater granularity than just being "documents" in an information system. They are bibliographic entities that have particular characteristics (e.g., they are one of XOBIS' ten Principal Elements — four substantial, five notional, and one conceptual). Again, this is not a weakness of the original paper, its predecessors, or the organizational schemes discussed therein; it is just a difference in scope. In defining the scope here, I can't help but ask, "Why do we need two schemes… one for authorities and one for bibliographic items?" Isn't what we are talking about the same set of things that are used (or referenced) in different ways?
Anyway, this is (as always) just a draft. I expect I've overlooked things and welcome other Xobian enthusiasts to question, doubt, and point them out (I'm working on rhymes with my four year old). I guess I should also throw in the disclaimer that these are my own opinions and may not be representative of those of any of the other XOBIS creators (e.g., all deficiencies are my own and should not be taken as representative as XOBIS as a whole). Others can speak for themselves by leaving their own comments as they see fit. 
Powered by FireStats