[This blog post was co-authored by Mercy Procaccini]

“Language is powerful. It conveys meaning, framing, and sets intentions.” – Reimagine Descriptive Workflows: A Community-informed Agenda for Reparative and Inclusive Descriptive Practice, 2022

In 2020, in response to the murder of George Floyd, OCLC leadership and staff charted a path as an organization, pledging; “We will think critically and implement actions to advance racial equity. We will not forget. We will continue to do more.”

An immediate outcome of this commitment was Reimagine Descriptive Workflows (RDW), a community-informed project that took place in 2021. RDW was structured to surface issues and to identify opportunities to effect lasting change in descriptive practices.

Among the goals set by the RDW participants were to:

• Acknowledge a need to change the current system
• Connect with others doing similar work
• Identify opportunities to engage in collaborative problem-solving
• Develop concrete approaches to enable reimagined descriptive metadata practices.

A report that documents the convening and its findings, Reimagine Descriptive Workflows: A Community-informed Agenda for Reparative and Inclusive Descriptive Practice reflects that “power and bias in collections is hard-coded from the beginning of the descriptive workflow process,” and that powerful naming and labeling systems which include content standards and data communication formats “can create systemic imbalances beyond the inherent problems of labeling and description.”

Acknowledging the need to enact change is only a first step. The biggest opportunity is in taking a hard look at existing structures, acknowledging the work that must be done, and charting a course forward. Acting on racist and oppressive structures requires stepping back to reframe and relearn, and working in community with others.

OCLC has been reflecting on our own role in repairing systems and workflows. OCLC takes seriously the responsibility to “dream and enact change,” and has been actively considering opportunities to recognize historical and current exclusions in the library industry and do our part to disrupt a cycle of harm in library descriptive practices. We can utilize learnings and practices derived from Reimagine Descriptive Workflows in doing this work.

WorldCat ontology

During the development of a fundamental component of OCLC’s linked-data ecosystem, the WorldCat ontology, the OCLC product management team recognized the need to pause standard production workflows to conduct an internal diversity, equity, and inclusion review of the ontology prior to release. The WorldCat ontology contains class and property labels and definitions that are used for production metadata services such as WorldCat Entities and additionally underpin OCLC’s entity editor, OCLC Meridian. The WorldCat ontology was developed to model extant bibliographic descriptions and authority data and builds on earlier OCLC projects, which benefited from the involvement and participation of community partners.

Seeking not to replicate harm that is rooted in legacy knowledge structures and systems, an internal diversity, equity, and inclusion review of the property labels and definitions surfaced opportunities for revision. In response, we sought a broader range of perspectives and expertise by engaging community members in reviewing the ontology.

Dreaming together: planning and carrying out the work

The OCLC team (comprised of staff from the OCLC Research Library Partnership, Global Product Management, and Global Technologies) invited a small group of professionals with a demonstrated commitment to reparative description and who work with organizations that have prioritized this work. In many cases these participants also drew from their lived experiences.

The project team took to heart recommendations from Reimagine Descriptive Workflows that involvement from community members be undertaken with care to minimize extractive practices. In this spirit, we offered an honorarium to participants.*

Given our relatively tight timeline, we were overjoyed that these skilled library and archival professionals made room in their busy schedules for this work. We provided drafts of the ontology and supporting primer document and invited participants to share written comments prior to our virtual discussion. Being able to review participants’ comments, suggestions, and concerns in advance allowed the OCLC team to synthesize common themes and to approach our conversations with care.

We organized participants into small discussion groups (ranging from 2-4 people), a format that has worked in previous projects. This provided the opportunity for multiple voices to be in conversation with us and with one another in each session. From the interview team’s perspective, this gave discussants a chance to uphold one another’s ideas, learn from each other’s expertise, and it allowed us to see where topics gained weight.

Learnings: some examples

Making space for communal ownership and creation

The comments and discussion reflected an interrogation of bibliographic ontology specifications that embed a Western understanding of works as individual acts of creation. Indigenous communities, among others, consider the inception of works (such as authorship, composing, and building) as community—rather than individualistic—endeavors. Our recommendations suggested revisions to acknowledge and account for shared creation and ownership.

Additionally, assertions about individual creation can at times erase or obscure the creation of knowledge or intellectual content by Indigenous and non-Western communities, instead incorrectly attributing the work to a colonial or settler individual/author. One respondent explained:

[T]here has been a lot of knowledge and cultural extraction from Tribal communities. … There is a perfect example… There is a community [and] their Indigenous language has been around for a very long time, but an anthropologist came in and documented it, wrote it down and now they [the anthropologist] are the author. Literally now based on Western copyright law, [the anthropologist is] the owner of the Indigenous language so now the Tribe has to go through that person’s estate to get actual rights to their own language… So that’s … where this term [author] could be problematic. [O]ur language has been documented by anthropologists …but we don’t consider them the creator of that … so, author here is “person responsible for the content” but they’re not, they’re just the person documenting it. 

Discussions of this topic can be found in a variety of spaces (including community-informed publishing and data collection) and around Indigenous data sovereignty. We found this to be a powerful example of the necessity and importance of breaking away from existing frameworks that exclude other ways of describing and organizing information.

The inherent non-neutrality of language

Respondents’ feedback also illuminated problems with descriptive structures and terms that imply neutrality related to the acquisition or ownership of materials, property, and land. The draft ontology terms and definitions in some cases could uphold a narrative that obscures histories of seizure, colonization, and violence. Participants explained that their concerns were not about finding a better term, but an acknowledgment that these actions are not neutral. The recommendation here is to consider language that invites critical reflection regarding the circumstances of ownership and acquisition.

For example, the draft definition of “acquired/acquired by” included the explanatory phrase “otherwise obtained ownership over the agent or place.” One respondent commented:

Consider well-established shift away from validating “ownership” in context of slavery. Acquisition (through purchase or not) in hindsight might not amount to accepted ownership but more to unjust violence, mastery, coercion, detainment, theft, loot. Emphasizing “ownership” might therefore erase those potential transgressions. I would recommend qualifying ownership further.  

This example helped the project team understand how “otherwise obtained” essentially functions as a euphemism here—we understand what it means but use language that obscures the true nature of the activity and presents it as seemingly neutral.

Similarly, in response to “founded by,” one respondent explained that their Tribal community considered words such as “founded, discovered” to be “very colonized, offensive words… it makes it sound like it [place/location] wasn’t in existence before the founding date, or whoever founded it.” 

Kinship relationships are important and expansive

The original definitions for a subset of properties that can be understood as kinship-based relationships (including parent, child, and spouse) emphasized legal and biological bases in defining them. Respondent feedback uniformly suggested that element description should be expanded to relationships that are self-defined. As one respondent explained:

Partner and spouse are not equivalent, even though conceptually, a lot of people put those together. …for many in the [LGBTQ+] community, partner was a poor equivalent to spouse— “This person is my spouse; he’s not my partner” (we don’t have a business relationship)…. Going back to Indigenous communities, spousal relationships did not necessarily have legal status. [Enslaved people], for example, couldn’t get married or have a spousal relationship, unless it was recognized by the state, and of course, those relationships existed, in spite of those prohibitions. 

For the ontology properties “given name” and “family name” respondents indicated that the way that these properties are currently structured feels exclusionary to many cultural practices and groups. In advising the project team of the shortcomings of the current structures, they acknowledged the challenges and provided resources rather than specific guidance.  

There [are] just widespread issues with how names are represented, not just “is this a family name versus a given name?” but, which names count, which names are used in what circumstances? Because they’re often really contextually specific. And also, what’s the normative or what is the form of the name that the person uses in terms of both the names and the sequence? In my experience … this disproportionately tends to affect Asians and also scholars who work in Asian languages, because people adopt … their American name, and then they have their birth name and then they have different transliterations that they prefer. Sometimes it’s just, this is how you render it. But it often crosses over the other way where you have Western scholars who are working in Asian languages and they have a very deliberately chosen preferred transliteration that also has a semantic meaning. So it goes a lot of different ways….I think that even if you just make a choice out of product or pragmatic reasons, it needs a lot of scoping and a lot of description … to be used consistently at the very least. 

Moving forward and enacting change

As a result of this process, OCLC is making improvements to the WorldCat ontology. Some revisions have already been implemented, while other issues await better solutions. This work will be ongoing; updates and revisions will continue to be made in dialogue with users and community.

In much of the respondents’ feedback, they acknowledged that the issues were not necessarily with the ontology and the descriptions themselves, but in how their usage might cause harm or obscure necessary truths. They invited OCLC to consider how—through scope notes, primer documents, and programming—to support appropriate and inclusive usage of the ontology. So while OCLC must effect change in the technology systems and structures we develop and maintain, we know that future steps will include working collaboratively with community to illuminate the challenge areas and transform descriptive practice.

In this blog post we’ve shared just a few of our key learnings. We are grateful for the opportunity to learn so much. We (and the WorldCat ontology) have truly benefited from the insights and expertise of our reviewers. The process echoes one of the seeming contradictions that must be balanced in working towards anti-racist and less harmful descriptive practices drawn from the RDW report:

“Language must be precise to demonstrate respect and inclusivity / In a diverse world, there will never be full agreement on the same words.”

This effort, working in community to ensure that the WorldCat ontology shifts away from systems and structures that were developed during the nineteenth century, demonstrates that this work is complex, nuanced—and worth undertaking.

Closing with gratitude

We offer thanks and appreciation to those who contributed to this effort, not only for their time and expertise, but for their generosity in explaining concepts, care for one another, and commitment to speaking up for communities.

• Pardaad Chamsaz, Metadata Lead for Equity and Inclusion, British Library
• Iman Dagher, Arabic & Islamic Studies Metadata Librarian, UCLA
• Christine Fernsebner Eslao, Metadata Technologies Program Manager, Harvard Library Information & Technical Services
• Selena Ortega-Chiolero, Museum Specialist, Chickaloon Village Traditional Council, the governing body of Nay’dini’aa Na’ Kayax (Chickaloon Native Village)
• Adolfo R. Tarango, Cataloging and Metadata Librarian, University of British Columbia
• Thurstan Young, Collection Metadata Standards Manager, British Library

Thanks are also due to the cross divisional team that undertook this work within OCLC.

• Rebecca Dean, Lead Data Analyst, OCLC Global Product – ontology development team
• Jeff Mixter, Senior Product Manager, OCLC Global Product – project co-lead, linked data product lead
• Charlene Morrison, Senior Data Analyst, OCLC Global Technology – ontology development team
• Michael Phillips, Vocabulary Specialist, OCLC Global Product project – co-lead, ontology development team
• Mercy Procaccini, Senior Program Officer, OCLC Research Library Partnership – interview design and reporting team
• Merrilee Proffitt, Senior Manager, OCLC Research Library Partnership – interview design and reporting team
• Richard Urban, Senior Program Officer, OCLC Research Library Partnership –  project co-lead, interview design and reporting team
• Anne Washington, Product Analyst, OCLC Global Product – ontology development team
• Gina Winkler, OCLC Global Product – Executive Director of Metadata and Digital Services – project sponsor

Would you like updates on OCLC’s linked data products and services? Sign up here!

* While offering an honorarium is intended to minimize extractive practices, it is not a panacea, and responses may vary. Some participants may be unable to accept payment for legal or practical reasons; others may choose to contribute without compensation.