Report: Early Modern Linked Pasts
Attending the Linked Pasts workshop in Madrid was a great way to connect with this community who uses the technologies of the semantic web in geohumanities research. Both days were useful for a newcomer like me to learn about Commons-supported projects and shared challenges.
I normally spend a good deal of my conference time with other historians of eighteenth-century France. Because we do not have well-developed infrastructure for geo-oriented digital research in early modern studies (ca. 1400-1790), conversations about these resources, their impact and possibilities are less likely to happen at, say, the Society for French Historical Studies. Finding common ground over methodological concerns at Linked Pasts was especially welcome.
With the Digital Humanities Research Group at Western Sydney University, I have been exploring ways to improve the way we define and describe historical places that are already part of our databases on the book trade in Enlightenment Europe (for the Mapping Print, Charting Enlightenment [MPCE] project). Here, I am interested in expanding the opportunities that early modernists have to link to infrastructure resources (gazetteers and other authority records). For example, we are working towards linking metadata from data.bnf.fr with our records of authors, booksellers, books, and places. One advantage of this is that the BnF data is already linked to Geonames, VIAF, ISNI, and the internal BnF authority records (RAMEAU).
My interests also extend to finding ways to represent connections (e.g. roads) between early modern places. Before I can map roads, however, I need to map the communities they pass through, by, around, etc. Relying on modern data about village and town locations in rural France, for example, misrepresents the political geography and place name histories for the eighteenth century. Identifying places as they existed or were understood during the early modern period has, for me, translated into documenting attestations of place names in the corpus of geographic reference works printed in French before the “nationalization” of place names and administrative jurisdictions in the early years of the French Revolution.
Moving beyond its origins as a community focused on the classical period, Linked Pasts 2016 emphasized that this challenge of fashioning ways to create digital records about historical places cuts across chronological and geographical boundaries. The format of the meeting, including formal presentations and brainstorming groups, communal meals and poster sessions (my poster), gave us multiple ways to interact with colleagues. I was grateful to learn more about the Kima project (Sinai Rusinek, Glauco Mantegari, Dimid Duchovny) during the poster session, for example, talking with Sinai and Glauco about the complexities of documenting/locating sacred, mythical, and “extinct” place names. Johan Åhlfeldt’s work on Regnum Francorum is a clear model/point of connection for early modern French places. I would not have had the opportunity to meet him outside of the Linked Pasts setting. From Maxim Romanov’s presentation on Classical Arabic Texts, I took to heart the power of embedding analytical frameworks into our entity-identification processes. Other presentations had useful project management and tools development ideas for the early modern European geo-related challenges I’m most familiar with. I admired the Github documentation for the CroAla index locorum project (Neven Jovanović), making choices, questions, and methods transparent to all. And for someone working with NER on French texts, the POSTDATA on Medieval Iberia (Gimena del Rio Riande, Elena González Blanco, Clara Martínez Cantón, and María Luisa Díez Platas) presentation was an impressive example of how to build and adapt NLP tools to suit the needs of non-English historical texts.
— Elton Barker (@eltonteb) December 15, 2016
Discussions with the Early Modern Letters Online (EMLO) team (Arno Bosse, Miranda Lewis, and their new collaborator Alexandre Tessier) suggested how useful it would be to compare methodologies for how early modernists design workarounds to collect and manage geodata. (I suspect these conversations were even more memorable thanks to the paprika-covered octopus we shared). As projects like EMLO and MPCE create in-house tools to manage, improve, and add to our datasets are there ways we can learn from each other? To what degree can we communicate about work-in-progress? More generally, these questions are concerned with reducing duplication in early modern DH projects. In this sense, they are related to Nicole Coleman’s presentation on FIBRA, an initiative to create a digital workspace for humanists to access, manipulate, and interpret digital evidence (or, as the H+D team calls it, a “data creator and editor” for humanists). I would also highlight Karl Grossner‘s fantastic work to get us all thinking beyond just places, people, and objects in linked data: how, he asks, do we model events? journeys? (Karl’s work, from ORBIS to the new world-historical gazetteer project, always helps me think differently about how to do digital things with early modern evidence.)
— Nicole Coleman (@cncoleman) December 15, 2016
Fantastic! I get to spend the next few years helping design & build a World-Historical Gazetteer w/great U of Pittsburgh team (me in Denver) https://t.co/9t1McEV9dm
I left Madrid with two wishes:
- to collect and share early modern digital project documentation about existing workarounds and
- to brainstorm across projects/institutions on the steps early modernists could take to align our metadata. (I have incorporated these into my 2017 Resource Development Grant proposal.)
…and one (big) question: What are the best practices to define the scope of an enriched historical gazetteer?
If I am a believer in the need for an early modern gazetteer that “enables connectivity through common references” (Simon, Isaksen, Barker, and de Soto Canamares, Placing Names, 99), I remain unsure about the best way to go about defining the limits of such a gazetteer. Many Linked Pasts participants and enriched gazetteer makers, myself included, start by identifying and annotating place names in one or several digitized historical documents (now possible with Recogito, among other tools). My own research starts with Diderot’s Encyclopédie, kindly provided by ARTFL. This allows us to establish a “link between the specialized geo-web and the wider web of words and hyperlinks” (Southall, Mostern, and Berman, “On Historical Gazetteers,” 2011, p. 128). The selected documents implicitly determine the eventual users, a.k.a. the “communicative community” that “need[s] to refer to some thing” (Ryan Shaw referring to Strawson in Placing Names, 56).
In a gazetteer-weak field like Europe ca. 1400-1790, it is tempting to have an expansive gazetteer that accounts for a long “history of how a name has been used to communicate” (Shaw, Placing Names, 61). For example, we could include references from geographic reference texts printed after 1790 date (to demonstrate the relationship between pre- and post-Revolutionary political geography). Time constraints, however, can limit the “depth” of a gazetteer. How do we balance including attestations with realistic project timelines? Or, when do you stop linking to documents and start hoping that future gazetteers will pick up where you left off?
Like many digital humanities endeavors, most researchers (myself included) don’t know that they could need something until they have no time to learn about/build it. Improving infrastructure resources and training will improve this over the next few generations, but, for now, building a gazetteer feels like a risky endeavor. It might be a resource for me and my immediate colleagues studying the Encyclopédie, or it might be something useful to hundreds of early modernists who accept that places named in the Encyclopédie reflect the same place in their seventeenth-century source. Already limited by time and the availability of sources that exist in sufficiently clean digital text files, we are also dependent on researchers’ willingness to identify a place reference in one context with a place in relatively close chronological/cultural proximity. How researchers make these decisions – how they use gazetteers – will change over time as new ones become available. Coming back to my early modern “wishes” above, it seems useful to survey who uses gazetteers, for what purposes, and with what expectations. We can then get a sense of the extent of historical gazetteer user communities, and how far they stretch beyond the community of makers.
A big thanks to the organizers and all participants. Looking forward to staying in touch.
— Katherine McDonough (@khetiwe24) December 16, 2016