Penn Libraries News

Diversity in the Stacks: Diversity, Equity, and Inclusion in Catalog Searching

Since 2021, a group of catalogers has been seeking ways to improve the language used in the Penn Libraries catalog to ensure it reflects principles of diversity, equity, and inclusion.

Looking down an aisle in the stacks on the Penn Museum Library. A museum gallery is visible in the background through a large glass wall.

Cultivating library collections that reflect the University of Pennsylvania’s diverse campus populations doesn’t just require library professionals to select, acquire, and store the right materials. They also must build and maintain pathways for search and discovery so that those materials can then be efficiently found, retrieved, and used. Online library search tools, like the Penn Libraries’ Franklin catalog, use descriptive metadata—information like an item’s author, the year it was published, and the topics it covers—to help library users find what they are looking for. The entire corpus of the Penn Libraries collections is vast, currently constituting over 8 million items sitting in stacks throughout campus, off-site storage, and within online databases. Robust and accurate description and careful arrangement of these resources is therefore essential to ensure that they do not become lost in this immensity. A resource with inaccurate or insufficient metadata can become effectively invisible to researchers and librarians alike. 

The experts who craft these descriptive catalog records have a daunting task, as they must grapple with the complexity of language—its constant evolution over time and differing shades of meaning and tone across diverse institutional, cultural, and even personal understandings. The meaning of words can be slippery, ephemeral, and powerful. A seemingly innocuous term in one context or era may be quite offensive, even harmful, in another. Metadata creators must remain aware of the changing nature of the language they use to describe library materials, and also must be conscientious about their own implicit bias and how that can inform their judgement and choices.  

To assist in analyzing and negotiating some of these complex issues, the Penn Libraries established a strategic working group in 2021 focused specifically on ensuring that decisions about the Libraries’ metadata were made with diversity, equity, and inclusion (DEI) in mind. The DEI and Discovery Working Group is comprised of a panel of skilled professionals from departments across the Penn Libraries and employs an ethos of transformative justice to improve the Libraries’ online discovery environments by removing or changing harmful and outdated language through new processes and system enhancements. Transformative justice is a community healing practice that seeks not only to repair specific harms but also to change the original, systemic frameworks and practices that created the problems to begin with. Consequently, this work is not only reparative but also generative and future-minded: it involves reviewing records created in the past and updating terminology to reflect current, preferred local usage, while also enhancing metadata to increase the visibility of these materials in the online catalog, and re-evaluating the ways these decisions are made and implemented moving forward. 

Critical Cataloging and Changing the Subject 

Penn Libraries’ DEI and Discovery group works from a long tradition of “critical cataloging.” This is a subset of critical librarianship, which explores the fact that libraries, like all institutions, are created and function within larger social systems of inequity and oppression, and that library users are better served if that influence is acknowledged and interrogated to work toward more ethical and equitable collections and services.  

An early practitioner of critical cataloging was Dorothy B. Porter, a librarian at Howard University in the 1930s, who built one of the world’s largest collections for Black history, culture, and thought at the Moorland-Spingarn Research Center. Porter realized that the classification system that libraries used to organize books by subject was completely inadequate for the works in her collection. Most libraries at the time were categorizing all works by Black authors under two Dewey Decimal classification numbers, either 325 for “colonization” or 326 for “slavery.” Porter rejected this practice as racially biased and instead created her own system that classified materials by author and subject (art, education, literature, medicine, economics, etc.). Her system of organization combated racist stereotypes by highlighting the true nature and influence of these works and authors.  

Decades later, Hennepin County Library librarian Sanford “Sandy” Berman became a leading critic of bias within library catalogs. In 1971, he published Prejudices and Antipathies: A Tract on the LC Subject Heads Concerning People , articulating his viewpoint. This publication offered alternatives to specific examples of bias and prejudice within the Library of Congress Subject Headings (LCSH). LCSH is a controlled vocabulary that libraries (including the Penn Libraries) regularly employ when describing resources. Using a controlled vocabulary that is shared across libraries helps with subject searching and provides consistency and interoperability within and across library catalogs. 

LCSH consists of words and phrases for categorized subjects, their variant spellings, synonyms, and related cross-references. This curated list of subject terms has been maintained, expanded upon, and revised to reflect current usage since 1898 by the United States Library of Congress, with input from libraries throughout the world. Librarians submit formal proposals for changes and additions to the LCSH, which the Library of Congress then evaluates through an extensive review process. Many changes made over the years have directly addressed concerns related to diversity, equity, and inclusion. For example, offensive terms have been removed from LCSH, such as “Yellow Peril” in 1989, and other terms have been updated (“Cripples” was changed to “Handicapped,” then most recently to “People with disabilities” in 2002).  

Not all decisions to change or remove terms are clear-cut and not all proposals are accepted. In 2014, a group of Dartmouth students started a campaign to “Drop the I-word” after they noticed their library catalog contained the anti-immigrant LCSH subject heading “Illegal aliens.” Librarians invited them to help submit a proposal to change the heading to “Undocumented immigrants” which the Library of Congress initially rejected, but later accepted after an outcry from librarians. This discussion eventually made national headlines. Congress itself became involved when Representative Diane Black (R-TN) introduced HR 4926, “Stopping Partisan Policy at the Library of Congress Act,” which instructed the Library of Congress to retain the original “Illegal aliens” heading. The bill never passed, but the Library of Congress complied with its edict, nonetheless. Backlash from the library community persisted—many forums were convened on the matter, and a documentary, Change the Subject, was produced featuring the Dartmouth activists. In November 2021, the Library of Congress announced that the heading would be split into two: “Illegal immigration” and “Noncitizens.” 

The controversy over this terminology, along with greater imperatives to advance social justice and anti-racism growing out of the protests over the murder of George Floyd in the summer of 2020, sparked renewed and wider interest in critical cataloging and efforts to “decolonize” the catalog. Many libraries have started to use their own locally-chosen headings or terms from alternative vocabularies to better reflect those values and goals. Librarians, interested library users, and the general public can also now discuss terminology at the online forum cataloginglab.org or request to attend an LCSH editorial meeting.  

Restoring Justice, Moving Forward  

Since its establishment in the spring of 2021, the DEI and Discovery group at the Penn Libraries has focused on a variety of initiatives to improve existing and new metadata, especially in terms of how it is experienced by users of the Franklin catalog. To start, the group addressed the infamous “I-word” problem by reviewing and expanding upon a feature that in-house computer programmers have developed to flip all instances of “Illegal aliens” to “Undocumented immigrants.” Even if records are imported with the undesirable term, all that users will see is the preferred term, while still getting the same results by searching for either phrase. This term-flipping mechanism provides an automated solution to updating terms throughout the entire catalog instantly while library staff go through the detailed work of modifying the underlying metadata behind the scenes. The Penn Libraries has made a number of other changes to subject headings recently, and more will soon follow. 

In addition to changing harmful, inaccurate, or undesirable language in LCSH-provided headings, the group also promotes the use and visibility of alternative vocabularies such as the Homosaurus, a resource that compiles terms related to the LGBT+ community. Catalogers can use these terms instead of, or in addition to, LCSH. 

Most recently the group has been working on recommendations for how to consistently describe library resources that clearly contain prejudiced content. These are often older, primary source materials—for example, antisemitic propaganda or archival works containing depictions of blackface minstrelsy. In addition, the group has provided recommendations for tackling the “’straight white American man’ assumption”—the problem of general subject terms for descriptors like occupation (writer, artist, astronaut, etc.) consistently defaulting to white men while more specialized headings are used for minoritized groups (for example, African American women artists). This bias in the metadata can result in the erasure of the more specialized group resources in searches for the generalized term. 

Ultimately, the group must analyze an extremely large, constantly changing dataset, containing catalog records that were originally created many decades ago when the materials were first acquired. Making matters more complicated, Penn Libraries staff do not create all the metadata in the Libraries’ catalog records. Rather, they pull much of it from external sources like publishers or Worldcat. Hundreds, if not thousands, of catalog records are created or imported almost every day, both manually and automatically, to ensure that the dataset is current and represents the library collections as fully as possible. Consequently, while Libraries staff are working hard to improve metadata in the catalog, it is a constantly moving target and users can still encounter problematic language.  

For more information about the DEI and Discovery group as well as media resources and updates on the group’s work, consult the Penn Libraries’ guide. Users who encounter problematic language in the Penn Libraries online catalog and would like to report it can feel free to send a message to the DEI and Discovery working group or contact the relevant library subject liaison.