Throughout our research and while developing a cataloging policy that was archive specific but from the perspective of a library, numerous issues surfaced that rang true on a larger scale and extended beyond our project reaching far into the world of archive collections and cataloging those collections. We found that it was often difficult to retrieve item-level records, that the Library of Congress subject headings were often an ill fit for our items (or, arguably, for many people and subjects beyond our project) and that there was a lack of information on the practical application of linked data in an archive setting. We have touched upon some of these in the issues section of our cataloging policy but feel strongly about the need to discuss them further due to the effect they have on cataloging practices as well as the accessibility of materials in archives and in libraries.
​
“Libraries, archives, and cultural institutions hold millions of items that have never been adequately described. These items are all but unknown to, and unused by, the scholars those organizations aim to serve. …Nationally, this represents a staggering volume of items of potentially substantive intellectual value that are unknown and inaccessible to scholars.” —from the CLIR announcement of the Cataloging Hidden Special Collections and Archives program, 17 March 2008.
​
Accessibility and meeting user needs are considerations that all librarians need to grasp before attempting to catalog on either a collection level or item level. Although there are standards in place for cataloging archival materials there is no universal standard when it comes to participating in an initiative to share these materials. The ACRL/SAA Joint Statement on Access to Research Materials in Archives and Special Collections Libraries outlines responsibility and intellectual accessibility as two core values that archive and special collections libraries need to meet. Further exploration of both these values reveals that the repository is responsible for making original research materials available as well as informing researchers of a repository’s holdings (ACRL/SAA Joint Statement, 2006). While considering ways for repositories as institutions to communicate descriptions of their collections so that they will be of value to those, especially beyond a local level, one would feel that a web enabled, searchable database would be of use. Fortunately, part of OCLC’s research initiative to make archives more accessible includes free access to ArchiveGrid, which includes “over four million archival material descriptions, including MARC records from WorldCat and finding aids harvested from the web” (“ArchiveGrid”). In some circumstances we were able to utilize ArchiveGrid to locate MARC records for our items, at times it was difficult and frustrating when MARC records simply were not available, or we were unable to locate similar items to use as a guidepost. As archivists’ step into digitizing collections to maximize preservation efforts, usability and access for users sharing ideas, workflows and practices grows ever more important (Bailey, 2016). Not only will collaboration amongst catalogers help to develop better standards for sharing but it will help with record enhancements and completeness. We can already see this benefit through cataloging traditional items by using OCLC Connexion. In time when items and collections are described efficiently users will benefit from the increase in effective precision and recall.
When it came to subject headings for each of our items, we found the Library of Congress options lacking. For the most part, the headings either classified the women in the context of the men in their lives (their fathers, husbands, etc.) or they subdivided a general topic by gender in an unbalanced way (for example, there is a subject heading for women presidents but not male presidents, implying that the default is male).
As Emily Drabinski points out in “Teaching the Radical Catalog”, our catalogs and thesauruses reflect the privilege of the ruling majority, which means that the options they provide are inherently political, they are prescriptive, and they are rigid. If the terms we have available to us are not an accurate representation of not only our users but our creators, how can we expect them to be useful or accurate access points? Not only that, but the current subdivision structure reinforces a patriarchal and racist view by implying that the default is, and continues to be, what Leslie Grinner defines in her SCWAMP theory: that our dominant ideology defaults to the portrayal of the straight, Christian, white, able-bodied, property-owning person. We other everyone who doesn’t fall into this strict framework by creating a default definition that reinforces this stereotype. If descriptive terms are deemed necessary for proper cataloging, they should be used across the board, otherwise, they are not only othering, they are failing the user’s ability to accurately and holistically search.
Furthermore, these classification schemes end up stripping the power and identity from creators who should be respected in their own right. Of the seven subject headings included in the Library of Congress’s record for the Alice Roosevelt Interviews, only one of them references her, the rest either mention her father, simply “children of presidents,” or subdivide under general terms like “Washington D.C. -- Anecdotes.” We don’t contest that these terms are accurate, but we do take issue with the priority they are given over subject headings that would more accurately represent the full and independent lives of women. For the sake of this project we limited subject headings to the controlled vocabulary provided by the Library of Congress; if we were executing this in reality, we would create a repository dictionary that offered supplemental subject headings that more closely reflected the creators/featured people in the items.
A more egalitarian approach to subject headings also effectively enables searchable content in a more global-minded, community-supported way. As Kate Dohe discusses in “Care, Code, and Digital Libraries” in regards to open-source content and technological scaffolding:
“It is fitting that collections and content intended to reach the global citizenry should be available with open source software applications. Moreover, many of these applications are created, customized, and maintained by staff at the research and cultural heritage institutions that also steward the content...This should represent a shift in power dynamics from vended solutions that is nearly as significant as the shift to open access to information. To borrow an analogy from Safiya Noble’s dissertation “Searching for Black Girls: Old Traditions in New Media” (Noble 2012), open source digital library technologies are comparable to solar panels that “facilitate independent, democratic participation by citizens, and [show] that design impacts social relations at economic and political levels” in opposition to controlled and closed systems peddled as a “galaxy of knowledge” (Appleton 2019) even as they proclaim their openness and transparency.”
Though not an exact parallel to our discussion of subject headings, it does appropriately reflect the shift in power structures when the power to define is moved into a more equitably sourced arena. As she goes on to point out, many of the decisions that are made in these arenas are systematically biased in that they are centered on the voices of staff who are removed from direct interaction with the end user and who often come from a homogenous background, further skewing the results.
Another issue we ran into was the opportunity to include examples of linked data in our records. While we strongly believe in the power of linked data to increase user accessibility, create broad context for the items housed in our open archives, and increase interoperability, it’s difficult to find lessons on how to execute that strategy beyond theoretical concepts. Even W3C’s site on “Cluster Archives” references a lack of examples to follow, lack of information on creating a data model, and lack of community guidance on which tools and vocabularies to use among the problems facing linked data become more widely adopted. In “Linked Data First Steps & Catch-21,” Karen Coyle points out the MARC 21’s current inability to properly implement URI’s for items within a record in a standard compliant way, further complicating the matter of industry-wide implementation of linked data concepts. Although we would have liked to include more examples of how this linked data can be created, and further discuss its potential power, we currently find ourselves ill-equipped to be able to speak on it on a deeper level.
Overall we hope that by highlighting the issues pertaining to descriptive cataloging of archival materials that others will recognize the implications these issues create in terms of accessibility. We feel strongly about continuing discourse on these issues, which will eventually lead to concrete solutions. Those solutions will in turn increase efficiency and output amongst archivers, researchers and scholars.
​