Over the last few months, I have been working on reclassifying the “Popular Movie Collection” at the small academic library where I work. This is the first of two blog posts I am planning about the project, describing the process of deciding how to classify the movies.
The “Popular Movie Collection” includes DVDs, Blu-rays, and videogames, and as the name suggests, it is provided primarily for students’ entertainment, rather than as direct support for coursework. The collection has been built up piecemeal through purchases and donations, and consequently has a rather eclectic range of new movies, classic films, and documentaries. It also includes our library’s small collection of Xbox and Playstation games, which students can check out for home use or to use with the gaming consoles in one of our group rooms.
The collection had been shelved by title, I presume because of its initially small size. As the collection grew, however, the decision was made to begin classifying and shelving the collection using Library of Congress call numbers. Some months before I began working here, new movies began being given LC call numbers as they were entered to the collection, but no effort was made towards retroactively re-classifying the entire collection. Now that I have somewhat settled into my role and have the time and energy to devote to larger projects, I decided to tackle the DVDs. We also have a large donation of DVDs waiting to be added to the collection, and I decided it was wise to handle the existing collection before adding to it.
Doing some initial research, however, I quickly realized a problem: the Library of Congress Classification system isn’t designed to handle movies. Documentaries would be no problem; they could easily be applied LC numbers based on subject. The small number of fiction television shows and videogames in the collection meant that they could be given one number each (PN1992.55 and GV1436, respectively). But LC has no easy way of accommodating fiction movies. There are ranges of numbers associated with motion pictures in the PN range, but they are clearly designed for books about movies, rather than for movies themselves. One number, PN1997 (1997.2 for movies produced after 2000), is designated for “Individual motion pictures,” and while it was originally intended for scripts, it is often assigned to recordings of movies. PN1995.9 covers “Motion pictures – Other special topics,” and is sometimes used, since it includes numbers for genres, countries, topics, and other topics. Local guidelines on adapting the LC system for this collection were clearly needed, and in the process of creating those, I had to think through how the shelving system could best support our goals for the collection.
This presentation from Maryke Barber at Hollins University (PDF) helped me think through some of the options available. I considered and quickly discarded putting all fiction movies under PN1997 and PN1997.2, then sub-arranging them by title. This had the appeal of a simple solution, but was not really scalable, even to the needs of my small collection. Things would get very complicated very fast, leading to unhelpfully long Cutter numbers, which in turn would lead to shelving errors. It also didn’t add any value to our shelving system — if I was going to do all this work, I wanted the end result to be more useful to our users than what we started with.
I also thought about following the rules for creating call numbers for literature, in which the author’s last name determines the number, and substituting the director for the author. One of the big draws of this approach was that it would let me utilize the complete run of LC numbers in the PR-PZ range. This makes sense in some cases — it might be useful to have all of Alfred Hitchcock’s movies together, for example — but doesn’t really make sense for most movies. I didn’t want to break up the Star Wars films, for example, because the entire series wasn’t directed by the same person. Shelving by director might make more sense for a film studies collection, in which the users have more specific needs, but for a popular collection, it seems a poor fit. I also wasn’t crazy about assigning so much importance to the director, since movies are a much more collaborative form than novels.
The solution (or at least our solution)
Ultimately, I decided to divide the movies by country of origin, and then further divide U.S. movies by genre. Movies from countries other than the U.S. are assigned to PN1997 and subdivided by country of origin; U.S. movies are classed PN1995.9 and subdivided by genre. My university has a very international student body, so providing access by country seems appropriate to me, since presumably some users would want to easily find movies from a particular country. I do have some qualms about separating international movies from U.S. movies, since I worry that it creates a sense of “other-ness” and flattens those movies to only their nationality, discouraging some users from checking them out. Part of me thinks, for example, that putting a South Korean action movie alongside Fast & Furious and Skyfall might encourage more users to look at it, but then it is much harder for a user who specifically wants a South Korean film to find it. I decided that the value of being able to group all of the movies from one country together outweighted those concerns, but I do still have some doubts about it.
I chose to subdivide U.S. movies by genre again based on my perception of users’ needs. Partly, the idea came from the fact that the PN1995.9 range in LC Classification already includes subdivisions for some genres, among other subjects (the list of subjects in that range is actually quite hilarious, including everything from Colors to Existentialism to Postage stamps). Arranging by genre, I hoped, would also seem familiar to our users, since that is how most commercial video rental outlets organize movies, from Blockbuster to Netflix. Of course, I feel rather uncomfortable suggesting that libraries should act more like commercial rental places, but for this collection, for these users, I think it makes sense. My main goal is to increase browsability on the shelves, and imitating the contexts that people are used to finding movies in made sense for that goal.
I based my genre list on the Library of Congress Genre/Form Terms list, specifically on the sub-list for movies created by Scott M. Dutkiewicz for OLAC (PDF). Immediately, I realized that I needed to narrow the list down, or I would have scores of genres, many with only one or two movies in them. That would obviously not serve my purpose of co-locating similar movies. I created a my short list by comparing the LCGFT list with those genres given LC Cutter numbers. I opted for genres that are at the top of the LCGFT hierarchy (i.e. those with no “broader terms” listed). I also used my knowledge of the collection and its goals to decide on what terms to include. I then created a spreadsheet with Cutter numbers for each genre that I have used to apply this new system to each movie.
I also made one last local decision: to use the year of the movie’s original release in the call number, rather than the year of the DVD’s publication. Again, my goal was for the call number to include as much information about the movie as possible. Because DVD is a recent format (and Blu-ray even more so), the publication year has no necessary relationship to when the movie came out and, in most cases, is meaningless. The exceptions are DVDs that are released as “special editions” with lots of new material and collections of movies released as a set, for which the publication year is ued.
The final product:
Le Jour Se Leve
.F73 (Cutter number for France)
J68 (Cutter number for title)
1939 (Date of original release)
Lady Sings the Blues
.B55 (Cutter number for Biography)
L33 (Cutter number for title)
1972 (Date of original release)
With the system for creating call numbers established, I thought the hardest part was over and began the process of retrospectively applying it to our collection. Little did I know, however, that genres would become an enormous source of confusion. In my next post, I’ll explain why and go into the details about creating call numbers for the collection.