Adventures in ontology
A number of recent events have come together to force me to write a post on the subject of ontology. Firstly, there’s the ongoing debate about the classification of the stuff in the solar system, which has been fascinating to observe. Secondly, there’s the terrifying fact that I’m about to begin my librarianship qualifications, which will immerse me yet deeper in the murky world of cataloguing. And finally, I came across a link to an article called Ontology is Overrated, which caused my thinking on the subject to step up a gear.
But what is ontology? In a philosophical sense, ontology is the study of the nature of existence, or of ‘being’. In an information science sense, it concerns classification and categorisation. The Dewey Decimal System is the most famous ontological system for books, assigning them to a place on the shelf by determining what they are about. Within the bounds of the library world, I find this to be a most intriguing subject, and am developing a quite alarming degree of interest in cataloguing – to an extent which must surely lead to a bulk purchase of grey cardigans in my near future.
The naming of Pluto and other bits of rock flying around the sun has been under a great deal of discussion recently and the back-and-forth between the scientists has been fascinating. It appears that there has been no formal scientific definition of a planet before now, hence the long debates on the issue. Calling it a ‘dwarf planet’ or a ‘pluton’ will not change its essential nature, only the way it is defined. But Shakespeare was (in a certain way) wrong when he made Juliet say ‘that which we call a rose by any other name would smell as sweet’, because psychology is more complicated than that. The names and categories we give to things are essentially arbitrary, and do not change the things themselves, but they change the way we look at them. If a new variety of rose was named ‘stinkwort’, I can’t imagine anyone would want to take a sniff. In the same way, Pluto continues to spin around the sun regardless of its ontological description, but we will think of it differently if it gets re-classified. Will the British have a secret liking for the dwarf planets due to our celebrated love of the underdog? Will everyone other than astronomers forget that it exists after a while? Will everyone blame Charon, its twin planet/moon/friend, for being a likely cause of its ‘demotion’? I don’t know, but it’s an interesting topic to ponder.
The ontology article I linked to makes a good case for the inadequacy of ontological classification in the electronic world, since on the web and elsewhere there is no ‘shelf’, and information can be accessed through any number of routes. Since reading it, I have made a conscious effort to add new tags to my posts, but at the same time not let them get out of control – too many categories is just as bad as too few. A post can be ‘about’ any number of things, but there must be a cut-off point somewhere. And since this blog is visible to search engines, I’m sure that anyone who really needs to find it (now there’s a disturbing thought!) will do so without my clumsy ontology.
In the physical world of the library, however, some more rigid form of ontology is important, in order to allow people to find books on a shelf. In some ways, it doesn’t matter how they are ordered as long as they *are* ordered somehow. Whether it is by subject, size, colour (sometimes tempting given the number of people who ask for ‘that blue book my tutor mentioned’) or date of acquisition (which is the chosen method of the British Library’s document supply centre). As long as they have a shelfmark, they can be located.
I prefer the subject method as used by Dewey or the Library of Congress cataloguing system, as it makes browsing much easier. It does have its drawbacks, of course. For example, books on Picasso can be found in a different place depending on their focus (with Dewey, anyway). If the book examines his paintings, then 759.6 (Spanish painters) would be a good place to look for it, but a book on his drawings would be at 741.946 (drawings by Spaniards) and his lithographs in a third location. This is all very well if you only want to investigate one aspect of his work, but for the student taking a holistic view of Picasso, it must be a tad frustrating. William Blake is even worse – do you classify him as a writer or an artist? Or both?
The problems don’t end there, of course. Sometimes it is hard to make a hard and fast decision concerning what a book is ‘about’. You’re looking for a book on the social impact of sport and the library catalogue has had a tantrum and refuses to cooperate. Where do you begin your search? Under sport, which tends to live at 796, or under sociology, specifically 306 (Culture and Institutions)? Or how about someone who wants to know how to care for their pet rabbits? 599 for mammals, or 636 for animal husbandry? Different libraries will shelve books in slightly different ways according to the needs of their users (in our case, the decision sometimes hinges on which academic department requested the book) and local practice. Some people think a particular book is ‘about’ one thing, but others would disagree and say it’s ‘about’ something else entirely. The more general a book, the more this becomes a problem. Still, thank goodness for the catalogue (whether a card index or a whizzy computerised one), which allows the library users to find the book, wherever a librarian’s whimsy has placed it.
Ontology. It doesn’t matter a jot, but is very important. It doesn’t change anything, but it changes our pecerptions of things. It gives librarians something to do, but it also gives them headaches. Can’t deal with it, can’t live without it. Now, can someone give me a definition of the boundaries of ontology? Anyone…
The Singing Librarian Talks 
What we call things certainly has a major impact on our perception of the world. Samuel R. Delany investigated this concept in his novel Babel-17, envisioning an entire language so efficiently constructed that it could be used as a weapon against poor people like us limited to ‘normal’ language. Umberto Eco has also played with this, both in his scholarly papers and his novels.
The other point you bring up possibly has more immediate relevance to everyone today – how items are filed and cross-indexed. This is the prime area that I think library science needs to attack, as the better such indexing is (especially for those items on the net), the more useful the information is to anyone trying to make sense of the world. Algorithms need to be developed that not only show that, say, “rose” is a term Shakespeare used, and where he used it, but ranks this usage in terms of multiple purposes (relevance to science, literature, horticulture, philology, etc). Books that cover multiple subjects are really a problem; perhaps in terms of where to shelve such a book, such an algorithm would pinpoint the area that the book is mostly about. I think there will always be some subjectivity to such classifications (after all, people use ‘fuzzy’ logic most of the time, not that stuff from the symbolic logic class), but even a small improvement in this area might give the world (and researchers) some large benefits.
Wow, it’s years since I read Babel-17, but I remember finding it fascinating. Science fiction plus linguistics – pure joy!
Algorithms are not my field, but something I’d like to investigate is the cataloguing and classification of recorded and printed music. This is an area where most libraries are sorely lacking, and there must be ways to improve it.
Happy to hear you’ve read it! Very few people outside of science fiction fans are aware of this book, which is sad, as it’s an excellent book.
Music is a language in its own right, and how we go about classifying it would seem to me to be something that we would need a totally different way of doing it, something that could possibly catch the nuances of various arrangements of a single piece.
Sorry, I’m afraid I am a bit of a science fiction fan. I’m quite an omnivore in my reading, though!
Fortunately I work in a small public library rather than a large research one, so my classification needs are filled quite nicely by Dewey.
Overclassification is also confusing and can happen even in small collections. It is common in libraries in the US is to separate fiction into genres for shelving. Our adult fiction was shelved this way when I started work here. Separate sections for mysteries, westerns, science fiction and fantasy, horror, romance and large print. Many of our patrons liked this system. Our mystery readers for example could go straight to their section and not be bothered looking through books of no interest to them to find what they wanted.
Take an author like Andrew Greeley. We had some of his books in general fiction, some in mystery and some in large print. Deciding whether a book is a mystery or a suspense novel was sometimes a challenge. Is something like The Time Traveler’s Wife general fiction or SF? When is a book set in the American west a western and when is it an historical novel? Legal thrillers went into general fiction, medical thrillers went into mystery. It was just too confusing.
Then there is the problem of funds to purchase novels. I can’t always afford both regular print and large print of the same title. So it is practical for me to purchase the large print since anyone can read it whereas some people can only read the large print. We were constantly having to redirect people to different areas.
Finally I decided to interfile all of the adult fiction. About 8 months in advance of doing this I begtan using brightly colored label protectors for the different genres. Bright red for mystery; green for western; etc. People got used to seeing their mystery section awash in red labels. This way when I interfiled the books they were less distressed because now it is easy to spot the mysteries when they walk through the stacks. A much better solution than the genre stickers.
I still have people complaining about the simplified system. I personally like it much better. And so do many of my readers who have to admit that they now find authors they enjoy who are not technically in their favorite genre.
What a great idea, Hypatia. I wish our library here in Lebanon would do that interfiling thing. I think it lends itself to browsing more.
I worked in a library that used the Library of Congress system of cataloguing. I never got to classify material, only put it on the shelf. Loved that job, but was always getting stuck in the stacks because there was some interesting book that grabbed my attention and needed to be looked at. “Oops, my shift is almost up and I have only shelved 12 books rather than the 3 carts they expected me to. Dag nab it.” Eventually I developed some discipline and because of all that “emergency high speed shelving” practice became one of the quickest and most accurate shelvers in the library. I also developed an uncanny ability to locate lost books that had been put away in the wrong place. What do those people thing the blue shelves are for and all the signs that say “Please do not reshelve books” mean anyway? A book that is shelved in the wrong place is in a Limbo whose existence can drive any nice librarian completely mad.
Even in our allegedly academic library, we shelve all adult fiction together, alphabetically by author. Genre classification would be pointless, and although we could shelve the books with the works about their authors (which, I believe, is what Dewey would like us to do), it just seems better to have the works of fiction together. Very handy for library staff who realise that they need something to read in their lunch break!