For this post, I observed elementary school students in the Berkeley Carroll Primary School library, where I work as an assistant librarian, interacting with the Metis cataloging system. This is in response to our Week 3 discussion of categorization, particularly Drabinski’s assertion that classification and subject language are inherently broken. While I’m inclined to agree that there will always be flaws in cataloging, I was interested in examining how a user-driven, specialized system might take steps in the right direction.

Metis was developed by librarians in fellow New York independent school Fieldston as a cataloging system tested and honed by children to focus on the needs of young browsers; everything from simple category names (like “Making Stuff” or “Scary”) to the categories themselves (“Animals” and “Pets” are separate sections) has the child audience in mind. There are twenty-six categories, each given a letter of the alphabet and a distinct icon, but the arrangement of these categories varies by school: for instance, while the Graphic Novel section’s suggested placement is between U (Scary) and W (Memoir), Berkeley Carroll assigns this popular section down in Z to prevent its large reader base from crowding away prose readers. Subcategories are also at local discretion, allowing students and librarians to further cater to the interest of its specific audience; this is why Berkeley Carroll, which has a major unit on ocean life in third grade, has a specialized “Marine Animals” subcategory that fills a full half of the “Animals” category.

The library I observed is the size of a small classroom, and contains roughly six thousand books. It’s one of two “hubs” in the primary school building: the “Red Hub” one floor below is largely for first- and second-graders, while my “Yellow Hub” covers third and fourth, but any student is free to use any hub. Here’s a visual of how the Metis system labels book spines: students can see the assigned letter, section, and subsection, with authors not necessarily given focus.

My observation consisted of three hour-long periods of peak usage (11:45 through 12:45) over three days in the Yellow Hub. The most notable trend is the marked distinction between third- and fourth-grade browsers: third-graders are generally new to the Yellow Hub and more often ask for help locating books, while fourth-graders easily find their sections of interest. The Metis system still requires some instruction to interact with, but I take the ease with which fourth-graders search and third-graders learn as a sign of its effectiveness.

More specifically, my assessment of Metis’s intuitive nature was bolstered when two third-graders who asked for help finding books on day one were searching independently by day two or three. A fourth-grader, in recommending a book to her friend, showed her where it was in the Fantasy section, then brought her to Tales (containing mythology) and Sci-Fi to suggest further reading. When students ask for recommendation, they virtually always use the same language as Metis, specifying that they want a “scary book” or a “realistic book.” There’s a chicken-and-egg conundrum here, particularly for students more accustomed to the system: is Metis’s language capturing how these children self-categorize, or have they merely adapted to the jargon? I’m inclined to go with the latter, but regardless, it’s clear that they navigate Metis far more easily than I did in my elementary school’s Dewey Decimal library.

While Metis’s positive qualities above hold true, it’s clear even in this basic observation that its localized customization not a cure-all salve. Many students of both grades still consistently ask where certain books are located without attempting to search; while this may say more about the convenience of a librarian than the difficulty of searching Metis, it still proves that the system remains inferior to a human search engine. That much is obvious, as categories are artificial constructs that users must learn, and even a universal, simply-taught cataloging system—among its myriad problems (read: Drabinski)—can’t take something as basic as room structure into account. There’s bound to be a learning curve in every library, and the fastest option for finding books will practically always be a librarian.

My most relevant observation about the failings of Metis was in a student’s comment that Redwall, a series about warring woodland creatures, is located in T (Adventure), while Warriors, a series about warring cats, is in V (Animal Fiction). While Redwall’s animals are far more anthropomorphic than their Warriors counterparts, wearing clothes and bearing weapons and standing on two legs, this did not convince the student that the two should be separate. I’d like to say that the student-based Metis system called for a change, or a larger inquiry into the matter with more students weighing in, but the hassle of such a shift (Redwall is a physically massive series) prohibited any section changes. The ideal behind Metis is noble, but in reality it’s impossible to fulfill every demand, even understandable ones like this with little argument to be had. Perhaps if a browser’s only option for locating a book was independent searching, there would be more of an effort to further perfect cataloging, but again, a librarian on location mitigates the problem.

Still, the appeal of a specialized catalog is self-evident; students who do opt to browse can easily find what they’re looking for when the system speaks in their language, and issues like the Redwall/Warriors incident are hardly limited to Metis. There will never be a complete solution to the intrinsic flaws of cataloging, but ditching a universal standard like Dewey for a library-by-library approach, using categories and language tailored to the population of local readers, seems to be a step in the right direction.

Stray observations:

• The population I observed is obviously limited to Berkeley Carroll students, which is an deeply imperfect sample of children their age. Its small student population and a focus on independent learning (students, for instance, use a self-checkout service) is hardly the norm in American or global schools, nor is the price tag; even in an observation this basic, we should take the results with a grain of salt.
• As mentioned earlier, graphic novels are easily the most browsed section on the library, to the point where checkouts are limited to one graphic novel at a time (the book limit is normally two per day, and five total books checked out at a time). With the gradual acceptance of the medium’s value and the explosion of talent being published by imprints like First Second and Papercutz, I can imagine a future where they’re integrated with works of prose due to the sheer size a Graphics section would take up.
• Redwall is so much better than Warriors it’s not even funny.

References:

Emily Drabinksi, “Queering the catalog: queer theory and the politics of correction”