(TL;DR: I’m starting to implement services and publish data to support searching across library collections that use customized subject headings, such as the increasingly-adopted substitutes for LCSH terms like “Illegal aliens”. Read on for what I’m doing, why, and where I would value advice and discussion on how to proceed.)
I’ve run the Forward to Libraries service for a few years now. As I’ve noted in earlier posts here, it’s currently used on The Online Books Page and in some Wikipedia articles to search for resources in your local library (or any other library you’re interested in) on a subject you’re exploring. One of the key pieces of infrastructure that makes it work is the Library of Congress Subject Headings (LCSH) system, which many research libraries use to describe their holdings. Using the headings in the system, along with mappings between it and other systems for describing subjects (such as the English Wikipedia article titles that Forward to Libraries knows how to relate to LCSH) allows researchers to find materials on the same subjects across multiple collections, using common terminology.
There are limitations to relying on LCSH for cross-collection subject searches, though. First of all, many libraries, particularly those outside the US, do not use LCSH. Some use other subject vocabularies. If a mapping has been defined between LCSH and another subject vocabulary (as has been done, for example, with MeSH) one can use that mapping to determine search terms to use in libraries that use that subject vocabulary. We don’t yet have that capability in Forward to Libraries, but I’m hoping to add it eventually.
Changing the subjects
I’m now also seeing more libraries that use LCSH, but that also use different terms for certain subjects that they find more appropriate for their users. While there is a process for updating LCSH terms (and its terms get updated on a monthly basis) the process can be slow, hard for non-specialists to participate in, and contentious, particularly for larger-scale subject heading changes. It can also be subject to pressure by non-librarians. The Library of Congress ultimately answers to Congress (as its name suggests), and members of Congress have used funding bills to block changes in subject headings that the librarian-run process had approved. They did that in 2016 for the subject heading “Illegal aliens”, where librarians had recommended using other terms to cover subjects related to unauthorized immigration. The documentary film “Change the Subject” (linked with context in this article) has a detailed report on this controversy.
Four years after the immigration subject changes were blocked, some libraries have decided not to wait for LCSH to change, and are introducing their own subject terms. The University of Colorado Boulder, for example, announced in 2018 that they would use the term “Undocumented immigrants” where the Library of Congress had “Illegal aliens”. Other libraries have recently announced similar changes. Some library consortia have organized systematic programs to supersede outdated and offensive terms in LCSH in their catalogs. Some groups now maintain specialized subject vocabularies that can both supplement and supersede LCSH terms, such as Homosaurus for LGBT+-related subjects. And there’s also been increasing interest in using subject terms and classifications adapted to local communities. For instance, the Brian Deer Classification System is intended to be both used and shaped by local indigenous communities, and therefore libraries in different locations that use it may well use different terms for some subjects, depending on local usage and interests.
Supporting cross-collection search in a community of localized catalogs
We can still search across collections that use local terms, as long as we know what those terms are and how to translate between them. Forward to Libraries already uses a data file indicating Wikipedia article titles that correspond closely to LCSH subjects, and vice versa. By extension, we can also create a data file indicating terms to use at a given library that correspond to terms in LCSH and other vocabularies, so we can see what resources are available at different places on a given topics.
You can see how that works in practice at The Online Books Page. As I write this, we’re still using the unaltered LCSH subjects (updated to October 2020), so we have a subject page showing free online books on “Illegal aliens”. You can follow links from there to see what other libraries have. If you select the “elsewhere” link in the upper left column and choose the Library of Congress as the library to search, you’ll see what they hold under that subject heading. But if you instead choose the University of Colorado Boulder, you’ll see what they have under “Undocumented immigrants”, the subject term they’ve adopted there.
Similar routing happens from Wikipedia. The closest related Wikipedia article at present is “Illegal immigration”, and if you go down to the Further Reading section and select links in the Library Resources box, selecting “Online books” or most libraries will currently take you to their “Illegal aliens” subject search. But selecting University of Colorado Boulder (from “Resources in other libraries” if you don’t already have it specified as your preferred library in Wikipedia) will take you to their “Undocumented immigrants” search. This routing applies two mappings, one from Wikipedia terms to LCSH terms, and another from LCSH terms to local library terms.
A common data resource
These sorts of transformations are fundamentally data-driven. My Forward to Libraries Github repository now includes a data file listing local subject terms that different libraries use, and how they relate to LCSH subject terms. (The library codes used in the file are the same ones that are used in my libraries data file, and are based on OCLC and/or ISIL identifiers.) The local subject terms file is very short for now– as I write this, it only has enough data for the examples I’ve described above, but I’ll be adding more data shortly for other libraries that have announced and implemented subject headings changes. (And I’ll be glad to hear about more so I can add them.)
As with other data in this repository, the data in this file is CC0, so it can be used by anyone for any purpose. In particular, it could be be used by services other than my Forward to Libraries tool, such as by aggregated catalogs that incorporate data from multiple libraries, some of which might use localized subject terms that have LCSH analogues.
Where to go next
What I’ve shown so far is not far removed from a proof-of-concept demo, but I hope it suggests ways that services can be developed to support searches among and across library collections with diverse subject headings. As I mentioned, I’ll be adding more data on localized subject headings as I hear about it, as well as adding more functionality to the Forward to Libraries service (such as the ability to link from a collection with localized subject headings, so I can support them in The Online Books Page, or in other libraries that have such headings and want to use to the service). There are some extensions that could be done to the basic data model to support scaling up these sorts of localizations, such as customizations used by all the libraries in a given consortium, or ones that adopt wholesale an alternative set of subjects, whether that be MeSH, Homosaurus, or the subject thesaurus of a national library outside the US.
Even with data declarations supporting those sorts of “bulk” subject mappings, a universal subject mapping knowledge base could get large over time. I’ve created my own mapping file for my services, and for now I’m happy to grow it as needed and share the data freely. But if there is another suitable mapping hub already available or in the works, I’m happy to consider using that instead.
It’s important to support exploration across a community of diverse libraries with a diverse array of subject terms and descriptions. I hope the tools and data I’ve described here will help advance us towards that goal, and that I can help grow them from their current nascent state to make them more broadly useful.