Literary mapping and spatial markup
- June 10th, 2009
- amanda watson
I’ve been thinking a lot lately about the uses of digital maps in literary study, partly because I’ve been thinking about the connections between place and memory for a long time, and partly because I got interested in GIS a few years ago, while working in the UVa Library’s Scholars’ Lab along with some extremely smart geospatial data specialists. There’s been talk of a “spatial turn” in the humanities lately, and there are already models for what I’m interested in doing. Franco Moretti’s maps of literary places in Atlas of the European Novel 1800-1900 and Barbara Piatti’s in-progress Literary Atlas of Europe have helped me think about the patterns that a map can help reveal in a work of fiction. I’m very much looking forward to hearing about Barbara Hui’s LitMap project, which looks a lot like what I’d like to make: a visualization of places named in a text and stages in characters’ journeys.
Since I came to the digital humanities via a crash course in TEI markup, I tend to think first of markup languages as a way to represent places, and capture place-related metadata, within a literary text. The TEI encoding scheme includes place, location, placeName, and geogName elements, which can be used to encode a fair amount of geographic detail, which can then be keyed to a gazetteer of place names. But there are also markup languages specifically for representing geospatial information (GML, SpatialML), and for displaying it in programs like Google Earth (KML). Using some combination of a database of texts, geographic markup, and map display tools seems like a logical approach to the problem of visualizing places in literature.
But (as I’ve said on my own blog, with a different set of examples) I’m also interested in spatial information that’s harder to represent. There are a lot of ways in which literary settings don’t translate well to points on a map. Lots of authors invent fictitious, and even when one can identify more or less where they’re supposed to be, one can’t supply exact coordinates. Consider Edith Wharton’s The House of Mirth and The Age of Innocence, both set in New York at the turn of the century and in the 1870s, respectively. One of my ongoing Google Maps experiments is a map of named places in both novels, focused on New York and Newport, Rhode Island. Both novels explore an intricate, minutely-grained social world, in which a character’s address says a great deal about his or her status. In some cases, the reader can precisely identify streets, points, and landmarks. And I think you can learn quite a lot about the world of Edith Wharton’s novels by looking at the spatial relationships between high society and everyone else, or between old and new money, or between a character’s point of entry into the world of the novel and where (physically and spatially) he or she ends up.
But in other cases the locations are harder to pin down. One can surmise where Skuytercliff, the van der Luydens’ country house in The Age of Innocence, is (somewhere on the Hudson River, not far from Peekskill), but it’s a fictional house whose exact location is left vague. The blob labeled “Skuytercliff” on my map represents a conjecture. And of course geoparsing won’t work if the place names are imaginary and the coordinates are unknown. So: what do we do with unreal places that still have some connection to the geography of the real world? And what if we want to visualize a writer’s completely imaginary spaces? What if we move out of fiction and into less setting-dependent literary forms, like poetry? How would one even begin to map settings in the work of, say, Jorge Luis Borges? Are there limits to the usefulness of visualization when we use it to analyze things that are fundamentally made out of words? Are some texts “mappable” and others much less so? (I’m pretty sure the answer to that last one is “yes.” I have yet to encounter an approach to literature that works equally well for everything from all time periods.)
So what I’d like to bring to the table at THATCamp is a set of questions to bounce off of people who’ve done more work with cartographic tools than I have. In some ways, my interests resonate with Robert Nelson’s post on standards, since I’m also thinking about what to do when the objects of humanistic study (in this case, literature) turn out to be too complex for the standards and data models that we have. If we end up having that session on standards, I’d like to be in on it. But I hope there are also enough people for a session on mapping and the representation of place.
June 11th, 2009 at 2:32 am
Very short comment just to say that this sounds like it fits perfectly with what I’d like to talk about. I’m really excited to hear about your projects and your work with TEI markup, which is something that I (oddly enough) don’t have any experience with.
June 11th, 2009 at 11:06 am
Definitely think the connection to the standards post is important, and both limits and unintended consequences of this kind of tool will be essential to address. For example, if I switch to satellite view in your google map, I’d be getting a historically distorted impression from the map — there’s where I love your inclusion of historical images in the popups.
I have little experience with geo tools, but have one similar issue as I’m trying to model info about things that classes study: I can say a class studies New York City, but should I distinguish that kind of statement from the statement that a class studies MiddleEarth?
June 12th, 2009 at 8:17 am
Barbara: Yay! We must compare notes, then. TEI’s pretty easy to learn, I’ve found, especially if you have any kind of background with XML. And it’s interesting to think about because it’s really all about acts of interpretation: you’re constantly declaring “This chunk of text is a stanza. This chunk of text is a marginal note. This chunk of text is a chapter” and so on.
Patrick: re unintended consequences — exactly. Google Maps and Earth are great for mapping real-life, present-day information at a very fine-grained level. But for part-fictional, part-real settings, or for historical information, or for hazy locations, it can give the wrong impression. It would be very interesting to develop some terminology for imaginary vs. real places (imaginary gardens with real toads in them?).
(Also, Tolkien’s maps of Middle Earth were a very early factor in my interest in fictional mapping!)
June 12th, 2009 at 5:39 pm
I wonder if Google isn’t working on a virtual map of Middle Earth?
A renewed emphasis on geography is an unanticipated consequence of the digital turn. Making even a simple map was a huge pain in the ass even a few years ago. Now the maps generate themselves–scroll down the “About This Book” page at the Google Books scan of Voyage Round the World by Charles Wilkes (1849). Every modern place name in the book automatically identified and plotted on a Google Map, with links back to the specific page of the book where the place is mentioned.
But you guys are absolutely right that not everything can be mapped, and that imprecise place names (“I was a day’s ride outside of Dodge when Black Bart and his gang rode up…”) are difficult cases.
June 17th, 2009 at 7:12 pm
Geez, there are a lot of great ideas coming into THATCamp!
It sounds to me like there may need to be a lot of discussion in various sessions about “fuzzy data”–though even that term doesn’t really seem to capture the uncertainties in these entities. Maybe “fluffy data?” I know it has certainly been talked about by lots of folks before, so hopefully we can build on some of that.
Whatever we call it, there are clearly descriptions of real and imagined places and dates whose precise limits are uncertain. How do we represent “antebellum” or “the Gilded Age” or “a day’s ride outside of Dodge” or “the industrial north?” Let alone the imagined and semi-imagined places that Amanda mentions in her posts. I’d love to see how Middle Earth and Narnia could get mapped–and the idea of blending both the real and imagined in one resource is fantastic.
My simple sense is that if we can each individually grasp (and share) what is meant by these places or times, we ought to be able to represent them somehow digitally, but the how doesn’t come easy . . .
June 18th, 2009 at 12:40 pm
I too have become very interested in mapping literary spaces since I found myself having to use Google Earth to understand the layout of Camden Town better when writing about Gibson’s _Pattern Recognition_ in the final chapter of my dissertation. The questions that have been raised here are all interesting ones. Let me add a few more thoughts.
1. One of the valuable things about mapping fiction is that it helps us get a better sense of the context that the characters move within. For example, Hemingway’s novels are often very precise about where his characters are in Paris, Spain, or wherever. This makes sense given that Hemingway was a person who cared very much about being in the right places (and drinking the right things) at the right times. When he tells us where his characters are, then, we have the opportunity to place them on a map. But this placement should stimulate us to think about what is surrounding them at a particular moment. What shops/buildings are in the area? If there are shops, who are the patrons and what goods are being sold? Which of Hemingway’s real contemporaries might have lived in the area that he is talking about? How does the fictional story overlay with his real one? One can go on. If we approach mapping the literary text in this way, then mapping isn’t actually the end itself. Rather, it opens the way to new research questions. We gain a richer sense of the text and perhaps–especially in the case of Hemingway–how this integrates with the author’s own life.
2. Along with mapping a particular text, I think that it can be equally useful to map a particular place with a multitude of texts. Take, for example, the NYT “Literary Map of Manhattan” (www.nytimes.com/packages/html/books/20050605_BOOKMAP_GRAPHIC/). If we think about everything imaginary that inhabits a real physical space, we might start to see new connections between works that are otherwise unconnected. For example, imagine a film atlas that shows scenes from every film made in LA’s Griffith Park. What is the BatCave butting up against?
In both of these situations, what I see as being most important is the fact that the map is being used to make an argument. While mapping can be fun, I think it needs to be used to help us make an argument about a text or texts. Otherwise, we’ve moved from literary studies (albeit in a DH vein) to geography.
Having said that, I’ll admit that we don’t always know the argument that the map will help us make until we have made it. I experienced that myself when I made a map of Jhumpa Lahiri’s short story “Sexy” for use in my survey of American Lit this semester (KMZ available at briancroxall.pbworks.com/f/Lahiri%27sSexy.kmz). Once I could see the story spatially, a whole new aspect of it revealed itself to me.
In any case, I’ll look forward to our revolving conversations about mapping and the humanities.
June 24th, 2009 at 5:51 pm
Golly, THATcamp’s gonna be rad.
How bout tying this to a historical question: what are the Detroits and New Orleans of history? What are the places loaded with symbolism, being somewhat forgotten, just out of the circle of awareness of mass culture?
How to use social networks among writers to show the invisible boundary which different social networks don’t cross? Average the map of colonial administrators’ diaries for Bombay, and then find the exceptions who travel beyond the pale…
June 24th, 2009 at 9:58 pm
I suspect that fuzzy (or fluffy) data are one of the hallmarks of the humanities. I can see the potential for a conversation on the role of imprecision in the decision to adopt sciencey or non-sciencey approaches to humanities data sets.
Brian, I think your question gets at the reasons why one would build a visualization tool, or any other kind of tool for that matter: for discovery of new knowledge, i.e. seeing patterns that wouldn’t have been visible otherwise? or for developing evidence for an existing idea? Research or writing? Or teaching? (Would any of this be useful to students, is another question I’ve been asking myself. Since we’ve got the makings of a bunch of pedagogy-oriented sessions, it may end up being a recurring theme.)
And social network visualization! I don’t know a whole lot about social network analysis, but I think it would be a terrific way to explore the connectedness of any group of people — historical figures or writers or characters in Shakespeare plays or even the “invisible college” of scholars. If anyone who’s worked on social networks wants to talk about that, I’d love to be in on the conversation.
June 25th, 2009 at 11:38 am
[…] of subjectivity in spatial and temporal visualization. I’m picking up here on ideas by Amanda and Brian, and also on a series of conversations I’ve been having this week at the annual […]