Over the past few months, I have been reclassifying the “Popular Movie Collection” at the small academic library where I work. This is the second of two blog posts about the project. In the first entry, I wrote about finding a way to fit movies into the Library of Congress classification system. In this post, I am going to talk about implementing that system, and in particular, my attempts to make sense of movie genres.
As I explained in Part 1, the classification scheme I am using categorizes movies from the U.S. by genre. I created a list of genres to use by synthesizing the lists given in the Library of Congress classification system and the Library of Congress Genre/Form Thesaurus (LCGFT). I used only “broad” genres, hoping to have a small set of terms that could describe all of the movies in our collection. I assumed I might make some adjustments to my genre list as I processed the first few dozens movies, but I didn’t expect too much difficulty.
Very quickly, however, I realized that assigning genres to movies is HARD. More specifically, deciding on one single genre that is best to describe a movie can be extremely challenging. The LCGFT is designed as a faceted thesaurus, with catalogers encouraged to assign multiple terms, and at different levels of specificity, to individual movies. That way, for example, Skyfall is described in its catalog record as an “action and adventure film,” “spy film,” and “crime film.” This makes perfect sense from a search perspective — if you search for any one of those terms in our catalog, Skyfall will be in your results. But since I am using genre to decide where to shelve the movie, each film has to be assigned to one and only one category, since it can only sit in one place on the shelf.
Compounding the challenge was my desire for as short a list of genres as possible. For example, I opted to not include “Romantic Comedy” as one of our genres, although it is a valid LCGFT term. But what, then, to do with romantic comedies? Are they romances, or comedies? The answer, of course, is both (hence the name) — unless it is neither. Is a romantic comedy really the same as a pure romance, or a pure comedy? To some degree, such philosophizing is beside the point. For me, what matters most is where our users are most likely to look for a movie. But it was hard for me to not get caught up in trying to make logical sense of genre structures.
It turns out that defining genres is notoriously difficult. Where is the line drawn between fantasy and science fiction? Or between horror and suspense? When I started working on this project, I decided to do some exploring into the literature about genre in film & media studies. From my extremely cursory foray into that literature, there is no accepted way of delimiting the boundaries of a genre. According to genre theorist Rick Altman, some approaches create a short list of archetypal movies that define a genre, while others draw up more expansive lists to show what a genre could be. This means that when we think about what a “fantasy film” or a “horror film” should be, we often imagine one of the defining archetypal examples. But those examples are obviously only a small subset of all movies; many other movies can be considered fantasies or horror films, or at least incorporate some of the elements of those genres. There were some movies in our collection that were easily identifiable as typical of a given genre (The Hobbit is obviously fantasy; Cabin in the Woods is obviously horror), but many movies seem to fall into the interstices.
Wolverine is frustrated that he doesn’t know his genre.
X-Men was one that I struggled with in this regard. Is it action and adventure? Fantasy? Science fiction? It has elements of all of these, but doesn’t easily fit within the typical definition of any of them. Again, my goal was to put it in the category that would make the most sense to our students. With titles that I found particularly difficult (like X-Men), I did a little bit of informal user testing by asking our student workers what made the most sense to them (I also asked friends on Facebook, launching what turned out to be a very lively conversation). This sort of crowd-sourcing seems particularly appropriate for movie genres, since as some film theorists argue, genres are defined in part by the popular understanding of them. The results of these informal surveys were sometimes useful, but often they had as many conflicting opinions as I did (no one, in fact, could agree on where X-Men belonged, and it ended up in science fiction). In part due to genres’ reliance on cultural convention for definition, absolute certainty and agreement is often impossible.
Another reason it is so difficult to assign every movie to a genre category is that the idea of genre as it is sometimes defined is not meant to encompass all movies. Film theorists often treat “genre movies” as a distinct type of movie, one that follows a predefined set of norms regarding narrative, theme, and tone. Often, “genre movies” are associated with a particular historical time period, that of the Hollywood studio system in the mid-20th century. This definition of genre is at odds with the one I was using in my classification, in which I treated genre as a characteristic that all movies possess. To iron out this contradiction, I created a genre category for “Drama,” a term that is notably absent in both the LCGFT and LC classification system. I essentially treat this genre as a placeholder for otherwise “non-genre” films, those that don’t conform to the kind of subject matter or narratives that belong to an established genre. In practice, this works fairly well, since I think most people do think of “dramas” as a distinct genre. Intellectually, however, I remain somewhat unsatisfied with “Drama” as a genre, since its definition does not have any real content of its own, but instead relies on being opposed to other existing genres.
Also complicating matters is the fact that a genre’s definition can depend on various elements of a film, including content, theme, style, and so on. In fact, different sources emphasize some elements over others, but without consistency. The Library of Congress’ 1998 Moving Image Genre-form Guide states that genres “are recognizable primarily by content, and to a lesser degree by style.” A more recent source, OLAC’s Best Practices Guide for LCGFT for Moving Images describes genres as consisting of a “packaging of various topical and stylistic elements.” One thing that is clear is that a genre is more than subject matter; genre terms themselves are distinct from subject terms for that exact reason. To put it differently, subject terms describe what a film is about, while genres describe what a film is. Catalogers are familiar with this distinction, but putting into practice is often more challenging that we would like to think. For one thing, as Martha Yee has pointed out, many genre terms are in fact defined by subject matter. The LCGFT, for example, defines “Crime Films” as “fictional films that feature the commission and investigation of crimes” – clearly a definition that relies on what the film is about.
Nope, not a “gangster film.”
This sort of content-based ascription of genres, though, leads to imperfect categorizations. Transformers is a movie about robots from outer space, which would lead to thinking it is a science fiction movie. However, in discussions with students at my library, I found that many people think of it primarily as an action & adventure film. Similarly, Some Like It Hot features gangsters prominently in its plot, but no one thinks of it in the same genre as The Godfather. Genres describe more than just a movie’s content; defining precisely what that is, however, can be tricky. From a cataloger’s standpoint, the challenge is even greater when trying to assign a single genre to a movie that I have not seen. It’s easy to determine what a movie is about, from the container and the description on IMDb, but it isn’t always possible to get at the subtler elements of tone and theme that can distinguish, say, a thriller from an action movie. Assigning genres based on a subset of plot elements is often the easiest route, and one that I found myself taking for many movies, especially those that I was not familiar with, even while I had to acknowledge that it was an imperfect way to go about the task.
Ultimately, the process of determining a single genre for each movie involved a combination of looking at what other cataloger’s had put in the catalog record; checking IMDb (including user reviews to get a sense of how viewers understood the film); reading the movie’s case, to see how it described itself; and asking students and co-workers on tough cases. None of these sources were perfect, and they often disagreed with one another, but together they helped me get a sense of the movie. I have come to accept that genres are, at best, a very inconsistent science, and that I am trying to put them to a use that does not exactly match their inherently fuzzy nature. In a great many cases, there is no one “right” answer. However, as the project is nearing its completion, I do think that basing call numbers on genres has helped group similar movies together and has given the collection an overall structure.
My next steps are to finish the reclassification, then create some signs or labels that will help make the new structure understandable to users as they approach the Popular Movies shelf. Is this the perfect solution for classifying and shelving movies? Probably not. But for this collection, it works well enough. I hope it makes the shelves seem approachable for a casual user — if you see a movie that catches your eye, the other movies close to it should be similar. It also helps librarians find movies easily, because if I know the genre, I can guess approximately where on the shelf to find the case. The system isn’t perfect, but it is at least a starting point and an improvement on shelving by title.
Plus, now I know a lot about movie genres, which should come in handy at a cocktail party at some point…
My readings about genre in film & media studies came from Film Genre Reader III, edited by Barry Keith Grant (2003). Some of the pieces are a bit unapproachable for non-experts, but the collection as a whole was an extremely interesting look into a field I had never explored before.