Diverse Voices Research Strategies - Decolonizing the Library
Events in 2020 revealed the systemic racial and social disparities that affect our most vulnerable populations, from the growth of the Black Lives Matter movement to the risks essential workers face every day during the pandemic. We recognize that it is important to offer works represented by historically marginalized voices, such as people of the global majority, people with disabilities, people identifying as LBGTQIA+, etc. We also recognize that publishing, academia and libraries have traditionally centered White voices. This research guide aims to highlight some of the difficulties in locating works by diverse and historically underrepresented people, and present strategies to help student, faculty and staff find works by historically marginalized voices.
Searching for works by diverse and historically underrepresented people can be challenging, in part due to standard practices in library organization of materials outlined below. It is not necessary to understand all the complexities of metadata management, but if you are curious about implicit bias in library systems read below.
Use the subject tabs for concrete strategies for finding works by historically marginalized voices.
Metadata, Implicit Bias and Structural Inadequacies
There is implicit bias in the knowledge organization systems used by libraries. Classification systems and subject headings reflect the identities of their creators as well as the time period in which they were developed (Tomren, "Classification, Bias, and American Indian Materials" ). Academic libraries generally use Library of Congress Subject Headings (LCSH) and Classification (LCC) which were developed to serve members of the United States Congress who historically have been White, affluent, Christian men. It is not surprising that LCSH and LCC also tend to be White-centered, cis-male oriented with a Christian and Western worldview (Adler, "Race and Ethnicity in Classification Systems" 61).
Metadata intends to make finding information easier, and classification tries to facilitate shelf browsing, but for some research they can be inadequate. Subject headings and other controlled access points such as authors, can make searching for diverse authors and historically marginalized voices difficult. Here are a few examples of how library systems and schema can be biased and unhelpful.
1. It is common for author headings to have birth and/or death dates, and even occupational qualifiers, but ethnicity, race, and gender information is not typically recorded. The author heading for James Baldwin includes his birth and death dates, as well as his occupation, but not his race or sexual identity:
Baldwin, James, 1924-1987, author.
This means if you were interested in what Black or gay authors had to say in the early to mid-twentieth century, you would have know who James Baldwin was before structuring your search. Subject headings for some of James Baldwin's books MAY identify his race, religion and the social causes he championed, but which subject headings are assigned is dependent on the in-depth treatment of the topics he wrote about:
Categorizing authors into diverse classes of people can be useful for researchers, but it is also problematical. It isn't always easy to discern what an author's ethnic, racial or gender identity might be, so if authors don't self identify, we have to rely on readily available, reliable secondary sources. When information about diversity is not available, then only some authors would get the attribute. This would inevitably lead to inconsistency in how diversity attributes are assigned. Some authors prefer not to be classed apart from authors in general. Assigning diversity attributes to authors assumes that people of the global majority and historically marginalized voices are not the norm, reinforcing their otherness, and re-affirming bias in our systems.
2. There is implicit bias in Subject Headings. For example, most subject headings for occupations assume that professionals are men, and generally White (Ros, "Bias hiding in your library"). For example:
Title: Moondust : in search of the men who fell to Earth by Andrew Smith:
SUBJECT HEADING: Astronauts
Title: They had a dream : the story of African-American astronauts by Alfred Phelps:
SUBJECT HEADING: African American Astronauts
Title: Women in space : reaching the last frontier by Carole Briggs
SUBJECT HEADING: Women astronauts
Subject headings are only established when a book on a topic is published and acquired by a library. There are currently no established headings for Gay Astronauts, or Transsexual Astronauts, because there are no books yet published about gay or trans astronauts.
While the implicit bias in subject headings is bad, the fact that headings exist to find diverse people in occupations is actually a helpful feature.
3. Language evolves rapidly, but subject headings are slow to change. For example, the authorized subject headings for First Nations people is "Indians of North America (LCSH)" which is not the preferred term of our Akwesasne and Canadian neighbors. Similarly, it has recently become acceptable to use the term Enslaved people when speaking about the plight of African Americans sold into slavery, however the Library of Congress authorized term continues to be Slaves (LCSH). Here is a list of problematic Library of Congress Subject Headings to demonstrate the scope of the problem.
4. There is implicit bias in the algorithms in the software that drives search engines like Google (Reidsma, "Algorithmic Bias in Library Discovery Systems"). Try a web browser search on "African American Women" and "Asian Women" and compare the results. Google is working to fix the sexualization of brown and black women, but other search engine results still favor dating sites. Proprietary library discovery software like our online catalog (ALMA/Primo) use algorithms to provide search results as well as relevancy ranking, search result filtering, and recommended resources. While our online catalog does not present the scope of unfortunate biases we see in browser search engines, search results are not always ideal. The major vendors of library software (Ebsco, Exlibris, Proquest) are aware of bias and content neutrality in system design, so hopefully they will work to improve their search engines.
5. There is implicit bias in the Library of Congress Classification (LCC) system which serves as a shelf locator for print books, and makes browsing like items possible. When the classification structure is biased, it can make browsing the shelves uncomfortable for diverse and underrepresented library users. The classification schedules are rigid, and changing a classification requires re-assessing each book, and relabeling spine labels. Here are some examples of classification problems:
The fact that most books on Human sexuality (LCC) are classed under Social Sciences. The Family. Marriage. Woman. Human sexuality. Sex, reflects the Christian worldview of LCC, and implies that human sexuality is marriage based and not relevant to people of all ages and relationships.
Library Organizational Systems, Metadata and Why It's Complicated
The SUNY Potsdam College Libraries employs a SUNY-wide information management system called ALMA/Primo (which you might know as BearCat, QuickSearch or Book Search). ALMA/Primo stores our online catalog records (also known as metadata) along with acquisitions and circulation data. The catalog records are owned collectively and shared among 64 SUNY campuses. When several SUNY campuses own the same book, ALMA/Primo has just one copy of the catalog record which is shared amongst all the campuses and displayed in our separate online catalogs. This means we can't do a whole lot of local customization (such as add subject terms or tags to help locate book on diversity topics), as the catalog records are not owned just by SUNY Potsdam and system-wide policies on editing shared records prohibits us from making changes that would affect other campuses.
When we migrated to the new system (ALMA/Primo) in 2019, catalog records were downloaded from WorldCat, a union catalog listing billions of items held by libraries across the world. The ALMA/Primo catalog records are refreshed on a weekly basis. When the records on WorldCat are corrected or updated, they are automatically exported again into ALMA/Primo, overlaying existing records and ensuring that our catalog is up to date and somewhat accurate. This also means that custom edits we might make to a catalog record are essentially wiped clean when records are refreshed.
While creating local terms IS possible in our current system, the magnitude of the problems inherent in library systems make maintaining local practices impractical.
The catalog records have metadata to describe the items we own, and to provide access points or ways to search and find the items. For books, the metadata typically includes author, title, publisher information, content notes, and number of pages. It also includes subject headings and call numbers or shelf numbers based on a classification system.
Metadata is structured information that is searchable and allows catalog searches to yield relevant result lists. For example a subject search for books on "Impressionism" and "Painting" should yield all the books we own that have an in-depth treatment of the Impressionist movement. A book on painting in general, that only mentions Impressionism would not be assigned the the subject "Impressionism" so that search results using the "Impressionism and painting" subject terms would not be flooded with general works that are less relevant.
There are many factors at play when it comes to searching and displaying metadata. Here are a few that make our online catalog less than perfect:
Many libraries create and contribute to the catalog records in WorldCat, not just the Library of Congress. Not all libraries use LCSH and LCC, so the catalog records may use different schema and have different metadata.
Many of the catalog records for e-books have scanty metadata, and may lack subject headings. If subject search results seem sparse, try using keywords.
Other subject schemes such as Medical Subject Headings (MESH), LC Children's Subject Headings, Sears and others are included in our catalog records and may be indexed and contribute to search results.
Cross-reference terms hidden in indexing and authority records may provide results that don't include the terms you typed in your search. For example a subject search for books on "white bias" yields 194 hits, but "bias" is not a valid subject term, and is not included in the metadata for any of the records.
Not all subject terms display in the online catalog record. If you want to see all the subject terms in a record, there is a link to "Display Source Record" at the bottom of many of our online catalog records so you can view the system coding. Subject headings tags are the 6XX tags (e.g. 600, 650, 651, etc.).
Progress and the Future of Library Catalogs
There has been a longer tradition of diversifying public library holdings and the findability of diverse books for children and young adult readers. Programs like We Need Diverse Books, Reading without Walls, and the Diverse Book Finder assist librarians, teachers and parents who strive to make #OwnVoices and books about diverse populations somewhat easier to find. More work needs to be done for adult readers and in designing discovery systems that are inherently welcoming for everyone.
Progress IS being made, but is slow to take effect. The Library of Congress (LC) is open to suggestions. There is a clickable "Suggest Alternative Terminology" in each Subject Heading record on their website, and many librarians and citizens are providing LC with alternative suggestions. LC is constantly updating and changing subject headings and other metadata terms, however, the schema they use is so massive and labor intensive, the process is slow and never ending.
In 2013 the Library of Congress Demographic Group Terms project was initiated to better identify diversity in library collections. The project is "designed to describe the characteristics of the intended audiences of resources, and also the creators of, and contributors to, those resources." Group terms for gender, nationality, religion, occupation, ethnicity, etc. may be assigned to certain works if the author self identifies, or when the "demographics of the intended audience and/or the creator/contributor groups are readily available." While newer records in our online catalog may have this coding (OCLC), older records may not. The lack of consistency will make this metadata less useful. It will also be a while before our catalog is able to search and display this data.
In 2020, SUNY Library Consortium embraced the Change the Subject Project to make system changes in ALMA/Primo to no longer display the Subject Heading "Illegal Aliens." While "Illegal Aliens" is still the authorized Library of Congress Subject Heading, SUNY opted for the more inclusive term "Undocumented Immigrants." Check the SUNY Library Consortium Change the Subject Project website to see how complex managing local subject headings can be, and how such decisions can cascade to related subject headings.
Our reliance on Library of Congress Subject Headings for our cataloging is unavoidable given our staffing levels. Tracking and maintaining local practices is too labor intensive to make it practical. Employing other schema could be more feasible, however there are currently no other workable subject or classification systems that has the scope and breadth of the Library of Congress systems, and no other systems address the problem of bias adequately. The voices of many librarians and library users are needed to fuel the changes to make library catalogs more inclusive for all. There are also other projects such as BIBFRAME, RDA and other linked data projects that may provide less labor intensive solutions.
Algorithms of Oppression by Safiya Umoja NobleA revealing look at how negative biases against women of color are embedded in search engine results and algorithms Run a Google search for "black girls"--what will you find? "Big Booty" and other sexually explicit terms are likely to come up as top search terms. But, if you type in "white girls," the results are radically different. The suggested porn sites and un-moderated discussions about "why black women are so sassy" or "why black women are so angry" presents a disturbing portrait of black womanhood in modern society. In Algorithms of Oppression, Safiya Umoja Noble challenges the idea that search engines like Google offer an equal playing field for all forms of ideas, identities, and activities. Data discrimination is a real social problem; Noble argues that the combination of private interests in promoting certain sites, along with the monopoly status of a relatively small number of Internet search engines, leads to a biased set of search algorithms that privilege whiteness and discriminate against people of color, specifically women of color. Through an analysis of textual and media searches as well as extensive research on paid online advertising, Noble exposes a culture of racism and sexism in the way discoverability is created online. As search engines and their related companies grow in importance--operating as a source for email, a major vehicle for primary and secondary school learning, and beyond--understanding and reversing these disquieting trends and discriminatory practices is of utmost importance. An original, surprising and, at times, disturbing account of bias on the internet, Algorithms of Oppression contributes to our understanding of how racism is created, maintained, and disseminated in the 21st century.