When Slave Biographies: The Atlantic Database Network launched in 2011, it aggregated data on slavery and enslaved people from three scholarly sources. Nearly 10 years later, Enslaved.org: Peoples of the Historic Slave Trade—built on the original project and using linked open data technology for a new, more comprehensive iteration—launched in December 2020 as a partnership between Matrix: The Center for Digital Humanities and Social Sciences at Michigan State University (MSU), the MSU Department of History, the University of Maryland, and scholars at other institutions. In addition to serving as a hub for multiple databases on enslaved individuals, the new site hosts the digital peer-reviewed Journal of Slavery and Data Preservation.

The platform currently holds some 750,000 records of people, events, places, and sources that span slavery in North and South America, Africa, and Western Europe, from the early 15th century to final slave emancipation in Brazil in 1888.

Users—from scholars to members of the public—can explore records by categories such as ethnicity, age, or source; search places; browse biographies; or assemble a data visualization based on a range of parameters. Those with an interest in genealogy can search family histories; scholars can download entire datasets. Data can also be exported so users can use their own tools.

TECH CATCHES UP

Slave Biographies originated as a collaboration between Gwendolyn Midlo Hall, an adjunct professor in the MSU Department of History, and Walter Hawthorne, professor of African history and the associate dean of academic and student affairs in the College of Social Science at MSU. The two were interested in combining their databases—Hall’s Louisiana Slave Database and Hawthorne’s Maranhão Inventories Slave Database—plus University of Arkansas at Little Rock Assistant Professor of History Brian Mitchell’s Free Black Database, which held records from 19th-century New Orleans.

The team behind Slave Biographies eventually expanded to include Daryle Williams, associate professor of history and associate dean for faculty affairs in the College of Arts and Humanities at the University of Maryland, who was researching the slave trade in 19th-century Brazil; Dean Rehberger, associate professor of history and director of MSU’s Matrix; and Ethan Watrall, assistant professor of anthropology at MSU. Funded through a National Endowment for the Humanities (NEH) grant and hosted at MSU, Slave Biographies served essentially as a repository, with no real ability to cross-reference information.

But “we were able to view and analyze the data that historians typically collect in this domain around historical slavery, and we could see certain types of information regularly being collected,” project manager Catherine Foley told LJ. During the course of the Slave Biographies project, team members saw that there was a lot of overlap in what scholars were interested in sharing and collecting: people’s names, their connections and relationships to family members, and their origins.” The stumbling block the developers encountered, she added, was “trying to get our hands around everything.”

The tools they had at the time weren’t up to the task, said Rehberger—“the notion that somehow I could take all these databases and make one database out of them. It wasn't sustainable. It was possible, but it took so much work to normalize all that data and bring it together.”

Over the years, however, the available technology caught up to their vision. Rehberger and Foley began hearing about the semantic web and linked open data, and realized that this strategy could accomplish what they wanted. “If we could just extract the linked open data from all these different datasets and put them together, and be able to model them using an ontology, it would make a lot more sense,” Rehberger told LJ. “And that's exactly what we ended up doing.”

In 2017, the team brought an idea for a new iteration to Earl Lewis, then president of the Andrew W. Mellon Foundation. Mellon was looking to invest in new areas of slavery studies and agreed to fund their proposal to expand on Slave Biographies. Four principal investigators—Rehberger, Hawthorne, Watrall, and Williams—joined forces under the guidance of MSU data scientists, programmers, librarians, and researchers to begin work on Phase 1 of Enslaved.org. Lewis brought in Henry Louis Gates Jr., director of the Hutchins Center for African and African American Research at Harvard, as a partner, and a number of historians from across the country joined the Enslaved.org advisory board.

With a second implementation grant from Mellon—for a total of $2 million—the work of populating Enslaved.org began in 2019. In addition to the three original collections, Enslaved.org integrated several other legacy projects, including Voyages: The Trans-Atlantic Slave Trade Database, Legacies of British Slave-ownership, African American National Biography, and Williams’s suite of datasets about Free Africans and others in 19th-century Brazilian slave society.

“This project was the confluence of lots of different types of technologies coming together at the same time,” said Rehberger. “We had a great community of scholars who really wanted to share this notion that data about enslaved people shouldn't be owned, that it should be accessible.”

ACCESSIBLE AND SCHOLARLY

1863 petition submitted on behalf of a Free African named Timoteo, from Brazilian National Archives (top), includes data extracted for a forthcoming dataset at emancipados.enslaved.org as well as slavevoyages.org (Voyages ID: 1870) (bottom), among others.

“The idea is to create a resource where someone comes in—no paywalls, free of charge, open source,” said Williams. “A lot of people are going to go there and just type in a name. And then that name would lead toward the records that we have linked.” Others may not have a name—they may be tracking down people through events, family lore, or other information such as the “runaway ads” in a site like Freedom on the Move.

At Enslaved.org, scholarly material lives alongside tools that are “accessible, friendly, not a lot of text at the beginning, not a lot of history,” said Williams. “It does not try to tell you the entire sweep of the history of enslavement, nor does it try to tell you the entire sweep of every technology that's related to linked open data.”

The Journal of Slavery and Data Preservation, launched in conjunction with the site, publishes original, peer-reviewed research and datasets about the lives of enslaved individuals and their descendants. The journal, which publishes through Harvard’s open source Dataverse repository, will be a source of new datasets for readers, also serving as an entry point for them to explore the original sources and “the context in which the data was collected, the decision-making that took place,” said Foley. “There's a very important emphasis on sources of documents—where are these documents held, making sure that those archives and libraries are being recognized.”

Foley hopes the journal will also bring in dataset creators through its call for submissions. “The journal as an entry point allows us to do some review of their data, then tease out the pieces that work for enslaved.org, and then allow us to encourage the scholars to do good documentation and put it into Dataverse for preservation and other uses.”

The Enslaved.org website will also provide educational resources for K–12 teachers. Several hundred detailed accounts of enslaved individuals’ lives can be found in three biographical dictionaries produced by Harvard’s Hutchins Center. Adapted versions of more than 75 of these are available on Enslaved.org’s “Stories” page and include timelines, a bibliography, and links to other resources, including the full, original biographies.

More records will be added as the database grows and partnerships are made with academics, archives, museums, and other sources. Enslaved.org has an NEH grant submission in the works; the team is working on community engagement as well, and figuring out how to accommodate volunteers. They hope to add about 30 more databases in 2021, including collections from the Library of Virginia, Maryland State Archives, and Oatlands Historic House and Gardens in Leesburg, VA. Williams and a graduate student are working podcasts as part of a multiplatform series called “Behind the Data.”

This project, added Rehberger, shines a light on much of the behind-the-scenes work in humanities data. “The typical model is, the historian goes off to the archive, does all this really hard work, and comes out with this monograph. What's often ignored is all the work of the librarians and archivists, the really hard intellectual work that goes into putting all the information together.”

QUESTIONS OF SEMANTICS, ETHICS

In developing Enslaved.org, researchers and developers worked together to create ontologies and controlled vocabularies for a range of fields used to link, ingest, and structure the data. While the project drew on prior work done on Slave Biographies and a number of other databases, the ability to use semantic web technology greatly broadened its mandate. Information from the site’s component datasets is entered in “semantic triples”—three-part sentences with a subject, predicate, and object. Triples (e.g. “Maria born 1830”) can be pulled from any article, register, or biography, and then linked with other information in the network so that names, places, events, and dates can be cross-referenced.

Developing the ontologies has involved a great degree of semantic work, Williams told LJ, especially when it comes to ethnography. “We have a whole range of registries, records, that talk about race and color,” he noted. “One question is, are race and color the same thing? Or is one one thing and one another? Do [developers] create two different columns in the same dataset? Does the entire structure of the hub, and the discovery process, disambiguate or distinguish between race and color, or do we conflate them together?”

That disambiguation is further complicated by nuances of languages and dialects. When certain terms are taken from documents, are they part of the enslaved person’s identity, or an invented description on the part of the enslaver? Those challenges, said Williams, center important questions: “What is the nature of enslaved societies? What is the nature of Black and Blackness, and Black people’s lives? For whom do the historical records speak?”

Those present in records but not named represent challenge of their own. Linking between datasets means that information listing enslaved people as property—bills of sale, insurance records, etc.—can be linked to data contained in other documents, such as baptismal certificates, that bear their names. But in building a dataset before those connections are made, “Do we create a line in our record for an unnamed or unknown person because hopefully, maybe at some point, someone else will figure out who that person is?” asked Williams. “Or do the unnamed and unknown not get included? Do we erase them?”

Yet another consideration is the human trauma encoded in this data. “The experiences that those records reflect are about trauma, and about the traumatization of lives, families, bodies, and peoples,” Williams noted. While the site was built in the spirit of discovery, “you're talking about celebration and discovery of trauma. Is it our ethical responsibility, is it our role, to encode that trauma? Particularly when we think about that some of the people in our dataset—their descendants, if they work at it, can find their own [ancestors]…. That's been a challenge, some of the ethics.”

He added, “We need to at least recognize a human existence there, somewhere between knowledge and empathy, and that has to be a foundation for what we do in data work. I do feel that there is some possibility there, and that's where I'd like us to be.”