The Promises and Perils of the Digital Humanities

Musicologists have long sought to create extensive catalogues that capture various aspects of musical life, including works written by a composer or group of composers, performances given at an institution or in a city, editions produced by a publisher or around a single work, and many others. The only difference between our project and the many other catalogues or databases that have come before is that ours was conceived as digital-from-the-start, and – perhaps more importantly – as a combination scholarly-pedagogical-public research tool.

A digital database makes it possible to search for specific terms or combinations of terms; to organize and reorganize along multiple axes of research; and to link to digitized primary and secondary sources, including images, film clips, recordings, and other websites. Similarly, a digital database lends itself well to digital visualizations of the kinds we offer here: maps first and foremost, but also graphs, charts, slideshows, playlists, and more. Finally, digital databases are not only easier to make public; they are also easier to develop collaboratively. Students and faculty, multiple departments, even different institutions and the general public can work together to generate content for a database that lives online. In this way, working digitally not only democratizes access to information, but it also democratizes the production of knowledge.

Through search and visualization, a traditional collection of information on any topic becomes a potential source of what digital humanists refer to as “big data.” Applying computational methods to “big data” collections within the humanities affords the possibility of adding “distant reading” to the traditional technique of “close reading.” What broader trends are detectable within hundreds of data points that are undetectable within the individual data points that so often constitute the object of study for humanities research? While close reading remains an invaluable technique – one that is central to our project – we are excited to explore the insights gained through distant readings of large data sets encompassing individuals, institutions, discourses, and geographical markers.

The promises of the digital humanities, then, are many: accessible research; collaboration; new evaluation techniques; visualization.

The perils of the digital humanities, of course, are just as many. The democratization of the production of knowledge exists in tension with a loss of authority and the potential for error-filled work. Our own digital database – produced via collaboration between students and faculty – surely contains errors, for which we take full responsibility. We will need to fact-check our own work again and again to correct typos, confirm contested information, and ensure consistent spelling. The significance of big data-facilitated, “distant reading” remains questionable, as even the pioneer of literary computational analysis Franco Moretti admits.

The greatest peril lies not in methods but in the technology itself. The rapid obsolescence of internet-based platforms means that digital humanities tools or experiments often have a shelf life of years, as opposed to the decades or centuries boasted by printed books. Broken links, removed videos, the withdrawal of developer support for a web-based app like Google Earth – all can cripple a project just as its impact seems to be expanding (case in point: UCLA’s Hypercities).

And that’s if you can even get your project off the ground: the greatest challenge we’ve faced has been locating the right software or platform to visualize our data. The web version of Google Earth will no longer be supported as of December, 2015, which means it will no longer be possible to embed Google Earth maps in websites. ArcGIS – the go-to tool for spatial analysis – works better with numbers than with qualitative data, and seems ill at ease with Parisian addresses. Moreover, the cost and level of expertise required to work with ArcGIS constitutes a hefty obstacle to democratic access to and production of knowledge. Unlike Google Earth and ArcGIS, Google MyMaps only allows 10 layers per map and does not allow the superimposition of historical maps. It sacrifices rigor for accessibility.

Our solution, for now: use multiple GIS platforms to make different kinds of arguments or create different kinds of resources. In addition to those mentioned, we experiment with Carto, a home-made TimeMaps add-on to Google Maps, and Omeka’s Neatline plugin. A brief survey of other digital mapping projects in the humanities shows that we are not alone in choosing not to put all of our eggs in one basket, and we may ultimately need to create our own platform to accommodate the specific research questions and digital materials central to our work. Stay tuned…