Last week I started looking around MIT to connect with people who might help me develop my idea of doing retrospective GPS tracking of l9th-century writers. Like almost any university these days, MIT is full of people interested in and working on problems of what is commonly called “digital humanities.”
Many of them, predictably, are in MIT’s Comparative Media Studies, which is in the School of Humanities, Arts, and Social Sciences. Others, less predictably (but this is MIT!), are found in engineering or computer science—for example, the Macro Connections research group organized around Professor César Hidalgo in the Media Lab, which develops methods to represent and analyze networks of historical and cultural events and processes.
I looked up Shahar Ronen, who has been part of this group: he has an undergraduate degree and considerable work experience in computer science, a master’s degree in history, and is just completing his master’s degree at MIT in Media Arts and Communications. We were introduced by Brian Keegan, a former undergraduate student of mine, now back in the Boston area as a postdoc at Northeastern University and actively connecting people like Shahar and me who share interests if not disciplines.
I give these details to make this point: of all the search engines on the planet, the best by far is a university. It collects an enormous amount of information but does so in a way that makes it efficient, and often enjoyable, to find what you need, and to be assured of its quality.
I explained to Shahar my vague ideas about tracking a cluster of 19th century writers, following them in space on different time scales—days, weeks, months, years, decades—to develop a better understanding of how they moved around the planet, responding to changes in its air, land, and water, and how they were motivated by quests, at once profoundly social and profoundly bodily, for health, stimulation, relaxation, solitude, and society.
Shahar quickly took me to a much more pragmatic level: what data could I mine to find reasonably precise and continuous geographic information about these writers? I said that one obvious source would be a diary. Shahar suggested as a model the website for the diary of Samuel Pepys, which links to maps so that events are coordinated with locations:
http://www.pepysdiary.com/diary/1660/05/23/
He followed up later with a sample from the diary showing how named entities (like locations) could be identified and classified using the Stanford named-entity recognizer:
http://nlp.stanford.edu/software/CRF-NER.shtml
The next task would be to assign specific GPS coordinates for each location mentioned in a text such as a diary, as well as to ascertain how long the person stayed at each location. This additional information would be needed to end up with a dynamic representation of someone’s movements over time in the past.
This is tedious, but certainly doable, if you have a rich text to work with. None of the late l9th century writers of interest to me kept a diary, except occasionally (such as William Morris’s diary of his camping trips in Iceland). But they were all assiduous letter writers. The quantity, length, and complexity of their letters should provide a wealth of information regarding their whereabouts and movements. Sharar and I discussed how published collections of letters, many of which should be available in digital form, could be mined for this information. In these collections, each letter has a caption telling where it was written, as well as when and to whom. The contents of each letter could be searched for words relating to place and movement.
As Shahar and I discussed the coding involved, the more I realized that such a project in digital humanities is possible only because of a much deeper foundation of scholarship in the non-digital humanities. It depends, first, on the habit of daily letter-writing, so engrained by education and socialization in the educated classes of the l9th century. It also depends on the habit of saving letters, on both ends of the correspondence, so that they end up preserved in an archive somewhere.
Finally, the project depends on the labor, often over decades, of producing edited volumes of collected letters, which give such rich information and pleasure to countless scholars and readers. For Morris and Stevenson, this work has been done, and so the digital humanist is set to go. This is not the case for Verne, for whom there is no comprehensive edition of letters, among other obstacles in accessing archival information about him.
I also began to think about what is lost, as well as what is gained, by the whole process of using digital techniques to extract geographical data from a textual mine. What ends up on the slag heap? The mind of the writer, for starters. We know where he is, or where he is headed, but we don’t know what he is thinking. Information about geographical position omits elements of the text that might provide, for example, rich and vivid descriptions suggesting why the writer cares about that place.
Also omitted would be the crucial information that in many cases a writer is thinking not about where he is, but about somewhere else. When William Morris was camping out in the northern reaches of Iceland, he was often desperately homesick, thinking about his wife and daughters left behind at Kelmscott Manor in the company of his wife’s lover Dante Gabriel Rossetti. When Jules Verne was lying on his stomach on a small fishing vessel sailing from port to port in the English Channel, he was daydreaming about a submarine below the ocean’s surface in oceans far away, on the other side of the planet, hidden from all tracking by a captain who had pledged never to have anything more to do with terrestrial humanity.
And when Robert Louis Stevenson was climbing a hill in Apemama, in the Gilbert Islands in the South Pacific, during a several-week stay there a guest of its thuggish ruler, he suddenly imagined himself back in Scotland, on a hill overlooking Edinburgh:
The tropics vanish, and meseems that I,
From Halkerside, from topmost Allermuir,
Or steep Caerketton, dreaming gaze again.
Far set in fields and woods, the town I see
Spring gallant from the shallows of her smoke, . .
So he wrote to a friend back home, in a poem overflowing with Stevenson’s longing for his homeland. At that time he thought he would return there someday, but he never did, and now lies buried at the top of another hill, in Samoa.
I want to find a way to track writers as if they were migratory animals, but I want to go beyond that. They are animals with consciousness, which connects them in habit and dream and hope to places far from where their bodies are. Letters can be used to track not only physical journeys but also imaginative ones. It should be possible to combine these tracks in a single map. To do this requires reading the letters with attention to feeling, nuance, allusion, and tone. The full map might end up nearly as large as the territory, but that is how it is with imaginative writing.
