Writing History in the Digital AgeSkip other details (including permanent urls, DOI, citation information)
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License. Please contact firstname.lastname@example.org to use this work in a way not covered by the license. The print version of this book is available for sale from the University of Michigan Press. :
For more information, read Michigan Publishing's access and usage policy.
Part 5. See What I Mean? Visual, Spatial, and Game-Based History
Digital scholarship allows historians to integrate visually rich source materials and interactivity into our writing, and several of the contributors to this volume took the opportunity to demonstrate their work and reflect on how it is changing our field. In “Visualizations and Historical Arguments,” John Theibault presents a broad overview of how charts and maps have influenced historical thinking from the birth of nineteenth-century social science to today’s processor-intensive digital era. Next, Stephen Robertson’s essay “Putting Harlem on the Map” recounts how he and his colleagues used spatial history tools to reconstruct the material lives of residents in this predominantly black New York City neighborhood during the 1920s, with examples of how these maps reshaped his historical analysis and writing. Finally, in “Pox and the City: Challenges in Writing a Digital History Game,” Laura Zucconi, Ethan Watrall, Hannah Ueno, and Lisa Rosner offer an insiders’ view of the storytelling and design challenges they face in creating a role-playing historical simulation on the invention of the smallpox vaccine in nineteenth-century Scotland.
Visualizations and Historical Arguments
The popular aphorism “A picture is worth a thousand words” is a relatively recent coinage, but the idea that images can be an effective complement to or substitute for written description, narrative, or analysis is probably as old as writing itself. In the European tradition, illuminated manuscripts and incunabula incorporated images, some of which conveyed messages related to the text and some of which were mere adornments. By the late sixteenth century, linkage of image and print reached a kind of apotheosis with the publication of emblem books, in which each page consisted of an image, a motto, and a pithy verse that jointly communicated moral precepts. Western historians have always made use of visualizations in this broad sense. Reproductions of pictures of the main biographical figures referenced in a book or of other objects that figured prominently in the narrative appear in many historical works. Though the connection of these illustrations to the arguments of the book were often implicit rather than explicit, the text sometimes drew direct attention to elements of the pictures, so that the reader’s understanding was enhanced by close attention to the images.
When the term visualization is used today, it usually refers to an image that is derived from processing information—often, but not always, statistical information—and that presents the information more efficiently than regular text could. Scholars quickly recognized the potential of computers to help process information and display the results in an easily interpreted format. David Staley has argued for a sharp distinction between visualizations as “the organization of meaningful information in two- or three-dimensional spatial form intended to further a systematic inquiry” and images as a “supplement or illustration to a written account.” Staley’s definition implies two distinct uses for visualizations in the digital age: as a means of quickly identifying patterns in large datasets during the research process, which can open new lines of research and test qualitative assumptions; and as a way to enhance the presentation of arguments, moving beyond what it is possible to display in two dimensions on paper. Visualizations created for the first use may or may not appear in visual form in the final product. This essay is primarily concerned with visualizations as historical arguments in the second sense: how do we deploy the visual capabilities of the computer to show what we wish to communicate? It is slightly more ecumenical in defining visualizations than Staley is, in that it sees as forms of visual argument all uses of visual information to communicate an argument or narrative beyond the meaning of the words in text. It argues that visualizations necessarily have a rhetorical dimension and that the principal challenge facing historians who wish to use visualizations in their work is to align the rhetoric with the audience’s ability to follow it. (See the images in the web version of this essay at http://WritingHistory.trincoll.edu.)
The key dimensions of a visualization are the density and the transparency of its information. Density is the sheer amount of useful information the visualization conveys, and transparency is the ease with which that information can be understood by the reader. We have become so accustomed to the visual vocabulary of print books that we scarcely register the visual conventions on which almost all historical work relies, such as the footnote indicated by a small number or asterisk. By now, we are also perhaps so familiar with standard web-page layouts that we no longer notice most of the visual cues that indicate the site structure, especially the relation of one page to another achieved by hyperlink. But many historians still have a print mentality when it comes to information-dense graphics. When designing graphics, authors have to consider how much background information the reader brings to the visualization. The development of more complex visualizations has increased the gaps between expert and novice interpreters, which raises challenges for historians who seek the most effective visual approach.
The problem of information density in visualizations is not a new one for historians. Even the most conventional nineteenth-century political histories made use of three important visualizations: maps, time lines, and dynastic charts. It is, after all, much easier and more informative to create a chart of lines of descent to the kings of France than it is to describe the lineage in paragraphs of “begats.” Each of these forms of visualization evolved a distinct visual vocabulary, with periods of experimentation and innovation producing visualization schemes that most modern historians now find completely transparent, with earlier visual dead ends now completely forgotten. Daniel Rosenberg and Anthony Grafton have recently shown how experimentation with designs of chronologies helped produce the modern streamlined edition of the time line and information-rich variants by the end of the eighteenth century.
The emergence of the social sciences in the nineteenth century and the ability to work with large datasets created demand for new ways of visualizing information, beyond maps, time lines, and genealogical charts. Processed numerical information was best expressed in tables, charts, and graphs. Mathematics, natural sciences, and social sciences that employed statistics were at the forefront of the development of charts and graphs. History was a consumer, not a designer, of most of these new visualizations—and mostly a sparing consumer at that, since economic and social history lagged behind political history as an area of research. Simple charts and graphs, such as pie charts, line graphs, and histograms, were not difficult to interpret, and their visual conventions became part of what any ordinary reader would be expected to follow. As statistical analysis became more sophisticated, the visualizations that resulted became more and more central to the argument. In some cases, the visualization made interpretation possible. These success stories demonstrated the worth of statistical analysis and visualization. Perhaps the most notable example is John Snow’s map of the incidence of cholera in an 1854 London outbreak, which helped plot the source of the outbreak at a single water pump in the neighborhood. Snow’s cholera map showed that visualizations could serve as both narrative and analysis. Authors began to experiment with ways of using visual clues to tell complex stories about events, increasing the amount of information that could be conveyed in a small space and thereby overcoming the limitations of two dimensions in print.
A noteworthy example of innovative presentation occurs in Charles Joseph Minard’s 1869 Carte Figurative des pertes successives en hommes de l’Armée Française dans la campagne de Russie 1812–1813, which portrays the advance and retreat of French troops in Russia on a scale map, showing the changing size of the force due to death and desertion through the thickness of the line representing the force. The conditions confronting the troops during the retreat are also illustrated by a time line of winter temperatures, graphically connected to the map-based chart. Though Minard was a civil engineer, not a historian, he was able to construct a very powerful single-page narrative of the events of Napoleon’s march. Minard’s chart is often cited as a model example of information visualization because it is easy to understand, even for people with little background information on the topic or quantitative skills. The challenge for visualization is to be transparent, accurate, and rich in information. Minard’s information-rich visualizations set the standard for both transparency and accuracy in the kind of work that could be done before computerization.
As noted already, historians were mostly consumers of statistics-based visualizations from the social sciences, rather than innovators in constructing new kinds of visualizations. The advent of cliometrics and Annaliste total history in the 1960s forced more historians to become conversant with quantitative methods. Though the Annaliste approach to total history predated widespread use of the computer, much of the first wave of social scientific history relied on statistical packages like SPSS and SAS to process large amounts of data. The most determinedly quantitative works often had several pages of tables, most of which would be referenced in the text, though not always at the precise page that made the link between text and table most obvious. Instead of working as a driver of narrative, many of the tables and graphs produced in quantitative works of the 1960s, 1970s, and 1980s sat inert on the page, functioning more like the biographical pictures included in early historical works than as an integral part of the argument. Toggling between explication and evidence slowed reading considerably, so much so that readers of quantitative histories of the era sometimes broke into two broad groups: those who read the text and assumed the charts and graphs confirmed what was said there and those who read the charts and graphs while paying scant regard to the text. To be sure, many more historians developed the ability to rapidly interpret a greater variety of statistical representations. A researcher using a scatter plot with a line of best fit or a Lorenz curve comparing inequalities might reasonably expect that most readers would be persuaded by the results visible in the charts, without requiring significant textual explication. But most social histories continued to rely primarily on bar and line graphs as their most prominent visualizations. The fact that the statistical tools deployed often embedded assumptions that were inapplicable to the messiness of actual historical processes lent a false aura of scientific precision to very tentative conclusions. Many explanations have been offered for the relative decline of social history since its heyday in the 1970s. A failure of imagination in the integration of visualizations with text-based arguments may have contributed to the decline.
While historians debated how to incorporate statistical methods into scholarship, statisticians were becoming more self-conscious about how the results of analysis were being used. This concern with how quantitative information was presented first came to the attention of most humanities scholars with the publications of Edward Tufte in the 1980s. His three key works—The Visual Display of Quantitative Information (1983), Envisioning Information (1990), and Visual Explanations (1997)—placed the aesthetics and explanatory power of graphs and charts under closer scrutiny. Tufte was most responsible for renewing attention to Minard’s Carte Figurative. Tufte’s main target in his books was what he called “chartjunk,” unnecessary clutter and contrived images that made visualizations confusing and sometimes deliberately misleading. Chartjunk was mostly associated with news and business publications in what today are called “infographics.” Historians, like social scientists writing for scholarly publications, tended to avoid visual embellishments of charts and graphs, but they did have to be attentive to ensuring that legends were clear enough that readers were not deceived by how information was displayed. Even accurate information can mislead if it is presented in a way that creates false visual cues. For example, inexperienced readers might need some guidance with a logarithmic chart so that they do not mistake an exponential change for a linear one. Sometimes, visualizations can be shaped to seem more conclusive than the underlying data actually warrant. For example, if a line graph showing differences ranging from 65 percent to 85 percent has its baseline set at 50 percent rather than zero, it leads people to see the differences in values as starker than they actually are. These issues of what one might call rhetorical honesty in the formulation of visualizations were compounded for historians and other humanists by the hard choices that were required to generate the data to be processed in the first place. As Gibbs and Owens note in their contribution to this volume, historians have traditionally been told to mask the twists and turns of the research process in their finished work, to make their argument as strong as possible. This traditional approach can make any visualization seem like the product of a black box. Their proposal to share both data and methods in as transparent a manner as possible can have the additional benefit of making those visualizations easier to understand, because the logic of how and why they were generated is visible.
Information-rich maps are a particularly good example of the challenge of balancing honesty in visual rhetoric and clarity and persuasiveness. One often has to come to a map visualization with sufficient background information to “read through” peculiarities of delivery. For example, maps of presidential elections that appear in almost any textbook of American history typically color in each state according to which candidate received the electoral votes. It is understood that the visual impact of the color contrast might over- or understate how close the vote actually was. Some states are large in area but small in population; others are the other way around. There are ways to make maps that account for those differences. For example, a cartogram adjusts the size of geographical areas to make them proportional to their populations, while a choropleth map uses shadings of color to indicate the strength of the victory in a given area. The two adjustments, cartogram and choropleth, can be combined to further the information density. Unfortunately, the distortions of the cartogram when combined with the choropleth can also make the information harder, rather than easier, to interpret, without a very high level of prior knowledge. This raises the question whether it is hard to interpret such a visualization because the format of combining cartogram and choropleth is unfamiliar or because it makes unreasonable demands on the background information of the reader. Most historians have probably encountered a choropleth map and a cartogram in a print history book or contemporary source in the course of their research but probably have not encountered the two combined.
The question whether a visualization is hard to interpret because it is unfamiliar or because it relies on unrealistic expectations of background information assumes greater importance because digitization allows for even greater information density and novelty of form. Geospatial locating of information has been one of the richest areas of development in digital humanities. Complex visualizations based on maps are emerging as part of a “geospatial turn” in the humanities. One particular way that geospatial information density can increase is by animating it, adding time as another dimension of visualization. Just as a map can make one inch equal one mile, an animated time line can make one second equal one year. A simple combination of an animated map and a time line can create a powerful narrative without any text at all. A brilliant example of this is Isao Hashimoto’s animated map of the 2,053 nuclear explosions between 1945 and 1998, which dramatically narrates the contours of the nuclear age. Aside from the title, there is no background information associated with the animation. The only text in the piece is in the legend, which emerges as each new nuclear power first explodes a device. Sound, not words, is used as a second way of highlighting the data points. Yet, despite the absence of background information text, almost anyone watching the animation will come away with a deep understanding of the key features of the nuclear age. Only a modest background knowledge (such as knowing who the main antagonists in the Cold War were) makes the presentation of what might seem dry factoids not only informative but moving.
Hashimoto’s animation of nuclear testing cannot be manipulated by the user, aside from pausing and resuming the animation. Edward Ayers has coined the term cinematic maps to describe map-based animations that show the process of change over time. The University of Richmond’s Digital Scholarship Lab has taken the traditional maps of presidential elections from 1840 to 2008 and turned them into an animated sequence. These maps try to overcome the information distortions caused by population differences and the electoral college, by providing not only county-level votes but a dot density map that shows the aggregated votes of five hundred voters. Instead of adjusting the size of the geographical area to make it fit the voting pattern, the dot pattern reflects the density of votes in each region. Still, it is easy to imagine how this information could be converted to cartograms and choropleth maps of presidential elections to tell yet another story about changing voting patterns over the decades.
As part of Stanford University’s Spatial History Project, online visualizations have been created to accompany Richard White’s recent book on the development of the transcontinental railroad. This project is particularly interesting for understanding the impact of digital humanities on current historical practice, because it is directly associated with a print work and seems likely to serve as a template for future hybrid productions of print and digital. It is also closely aligned with a still more expansive set of visualizations from the Stanford Spatial History Project that relate to the themes of the book, collected under the heading Shaping the West. There are twenty-six different visualizations included at the site, sixteen of which are animated. Not unexpectedly, several of the animations simply plot space and time, like the animations previously described. But others complicate the visualization by layering information in innovative ways. For example, one visualization reframes shipping distances in California in terms of not just track length but also time to delivery and cost. Even though the animations have been created as an accompaniment to an academic work, they offer an interactive opportunity to the reader that those other animations do not. Readers can customize the presentation of data to isolate issues of particular interest to them, rather than depending on the author to frame the question being answered. Interactive engagement with a visualization is yet another innovation made possible by digitization.
As historians ask more complex questions about the data they have assembled, the problem of how best to present the information requires more thought. On the Shaping the West site, each visualization has an “About” or “Help” tab that functions as a legend and guide to the information contained in the site. The visualizations are not self-explanatory. A particularly complex visualization links the geography of the railroads with a network diagram of the boards of directors and sources of capital for each. The “About” tab for that visualization includes a “How to Read” statement for the graphic. Such “How to Read” statements recognize that the visual vocabulary of innovative sites may not be familiar enough to make an argument without further explication of methods.
Undoubtedly the biggest advocate for the rhetorical power of statistical animations that incorporate interactive features is the Swedish statistician Hans Rosling. Using a tool called Gapminder, he has created an animation of life expectancy at birth and per capita GDP since 1800 for all countries, to demonstrate the evolution of world health. One can “play” Rosling’s animation in a noninteractive mode to see the story he tells. Color coding differentiates countries in different parts of the world. If one scrolls over the circles on the chart, one can see which country each represents. Circles vary in size depending on the population of the country and on change in size over time in response to population growth, so the reasonably well-informed viewer can quickly locate major countries like China, India, and the United States even without scrolling over the circles. The animation contains deep layering of information that is easy to interpret, even without an extensive background. In videos where he talks about the data, Rosling shows that the information illustrates a dramatic narrative of the convergence of the world on higher levels of health and wealth, but the point comes across perfectly clearly even without verbal accompaniment. Rosling draws on an extremely rich database, and readers are able to customize the display of information according to their own interests. One tab allows readers to orient the circles on a map of the world rather than on two axes of a chart. Another allows readers to choose which country’s data to include or not include in the animation. If one wants to isolate countries from a single continent or countries that start out a similar size, one can do so. One can also adjust the time line to focus on narrower periods where crucial changes might be taking place, instead of having to go through the entire time span for “big picture” changes. Because of this option for customization, Rosling’s project both makes an argument that is explicit in the first animation one sees and provides the basis for a reader’s further exploration of his or her own interests. While the kind of graphs used by Rosling were developed prior to the web, they posed real challenges of presentation and interpretation in the two-dimensional format of print. Animation increases their interpretive force dramatically.
Animation and reader control of the data stream are not the only ways in which digitization affects interpretation. Websites are new enough that there are still opportunities to subvert standard expectations and make readers more attentive to how visual cues structure an argument (in ways that are much less costly than trying to subvert visual cues in print media). A good example is Whitney Trettien’s “Computers, Cut-Ups, and Combinatory Volvelles: An Archaeology of Text-Generating Mechanisms,” which encourages a nonlinear reading of her argument about nonlinear texts. The front page of the site does not offer a table of contents or an obvious sequential path through the material. It has a grid of fourteen by twenty-one white squares that light up and change color either when scrolled over or when specific pages of text are clicked. Color coding allows readers to see which sections of the website deal with specific themes, creating a second way of envisioning the argument. Physical proximity of squares and color groupings work together to create a structure to the argument that is as easy to see as it is to read. Interestingly, the idea of color coding in a grid format was presaged in Elizabeth Peabody’s nineteenth-century Universal History: Arranged to Illustrate Bem’s Charts of Chronology. Trettien’s experiment with a visual “breadcrumb trail” makes it possible to reimagine how arguments can be presented in an environment where the reader controls what page to turn to next.
As Peabody’s work shows, it is possible to use color as a visual cue in print texts, but it is generally prohibitively expensive. Online, color is both efficient and cost-free. A superb example of using color to highlight relationships in text is Ben Fry’s concordance of the six editions of Charles Darwin’s Origin of Species completed in his lifetime. The original text is represented by each sentence being compressed into a single line. The reader can scroll over each line to get a text box of the sentence. The additions in subsequent editions are represented by different colored lines. The colors allow one to quickly grasp, for example, that chapter 4 was most extensively revised in the third and fifth editions, while chapter 6 was most extensively revised in the fourth edition. Dramatic changes are visible because of tiny lines of color. This same principle is taken up in the WordSeer project at the University of California, Berkeley. The WordSeer researchers have digitized a corpus of printed slave narratives and compressed each narrative to a single bar of a heat map. One can search for words across all of the narratives and see how frequently they turn up in each of the paragraphs, represented by lines within each bar. Color, brightness, and lines and bars thus become powerful ways of making interpretive leaps about texts, as long as one has the background knowledge to understand the implications of the visualization.
Much historical writing is implicitly or explicitly about network connections, but historians are less familiar with how social scientists have been visualizing networks than they are with standard statistical visualizations. The Mapping the Republic of Letters project at Stanford University shows how network visualizations can be used in historical work. It overlays a networked map of correspondents on the actual map of Europe, with each link in the visualization representing a letter sent between an author at one location and a reader at another. The network described at Mapping the Republic of Letters is personal rather than conceptual, as was the network of railroad boards of directors previously mentioned. The railroad visualization was more complicated, because the nodes of the network were not linked to a map but were a pure visualization of relationships. The online prototype of visualizations of network relationships is Thinkmap’s Visual Thesaurus, which allows readers to move from node to node in pursuit of related concepts. Visualizing networks poses several dangers for historians. First of all, a network theory of graphs adheres to mathematical principles that have little relation to lived human experiences. In an effective network visualization, the location of nodes is not predetermined but is specified by the nature of the links between them. If you remove one source of links from the analysis, the location of nodes may become different. As one undertakes a more complicated kind of network analysis in the mode of Visual Thesaurus, it is more likely that following any single link trail can quickly get one lost in the thicket of concepts. Unless one understands the algorithms being used to create the nodes, it is extraordinarily difficult to understand why nodes are in a specific relationship to one another. Thus network analysis demands the kind of “hermeneutics of data” advocated by Gibbs and Owens in this volume. But even when the concepts and relationships being illustrated are relatively straightforward, the task of visualizing them can prove complicated by the volume of connections being analyzed. The sheer density of nodes can make it hard to single out factors that might interest the reader. For example, in the network visualization of those scholars who make up the “vizosphere” (the leading edge of discussion about the future of visualizations), there is a barely differentiated blob of circles instead of a clear pattern of lines between sites.
Innovative visualizations have entered the mainstream of online user experience in the professions and social sciences. Just as SPSS, SAS, and, later, R were created to enable basic statistical analysis, programs like Gephi have been created to undertake network analysis. In the wake of Tufte’s work, numerous authors now write about information design, though again mostly targeting a business and journalism audience. Every day, sites like FlowingData highlight innovative uses of visualization to make new arguments, such as PeopleMovin, which illustrates migration flows between countries. It is clear from these sites that people are still expanding the realm of the possible in visualizing information. Looking over these visualizations, even when they are not explicitly historical, will give historians strategies for making more powerful arguments to complement and sometimes even substitute for text. But the task of building those arguments will have to include educating fellow historians about how to interpret visualizations. As we have already noted, it can be very difficult for the uninitiated historian to intuit relationships between entities in a network analysis when they are put into a visualization scheme. Yet networks are often at the center of questions of greatest interest to historians. To the extent that the difficulties in interpreting innovative visualizations like interactive network diagrams are caused by a simple lack of familiarity with them, they can be overcome by building more such sites. To the extent that they are caused by a lack of background knowledge to understand the cues, creators of such sites will have to learn to build new ways of incorporating that background information as economically, in the use of text, as possible. In either case, at some point, historians will have to accustom themselves to “reading” network diagrams as adeptly as they read maps or scatter plots.
1. The English Emblem Book Project, Pennsylvania State University, http://emblem.libraries.psu.edu/.
6. One can overstate the degree of that decline. Social history was always something of a grab-bag term that included nonquantitative approaches. Nevertheless, the relative decline of social history in comparison to cultural history is unmistakable. See Robert B. Townsend, “Decline of the West or Rise of the Rest?,” Perspectives on History, September 2011, http://www.historians.org/perspectives/issues/2011/1109/1109pro1.cfm.
7. Edward Tufte, The Visual Display of Quantitative Information (Cheshire, CT: Graphics Press, 1983), Edward Tufte, Envisioning Information (Cheshire, CT: Graphics Press, 1990), Edward Tufte, Visual Explanations (Cheshire, CT: Graphics Press, 1997).
9. Mark Newman, a physicist at the University of Michigan, gives a good overview of the uses of cartograms and choropleth maps for analyzing the 2008 election returns at the level of state and county voting, at http://www-personal.umich.edu/~mejn/election/2008/. Examples include a simple cartogram of county-level results and a choropleth cartogram of the same results.
10. Joanna Guldi, “What Is the Spatial Turn?” Spatial Humanities, Institute for Enabling Geospatial Scholarship, Scholars’ Lab, University of Virginia, http://spatial.scholarslab.org/spatial-turn/.
11. Isao Hashimoto, “1945–1998,” Preparatory Commission for the Comprehensive Nuclear-Test-Ban Treaty Organization (CTBCO), 2003, http://www.ctbto.org/specials/1945-1998-by-isao-hashimoto/.
12. Edward Ayers, “Turning towards Place, Space, and Time,” in The Spatial Humanities: GIS and the Future of Humanities Scholarship, ed. David J. Bodenhamer, John Corrigan, and Trevor M. Harris (Bloomington: Indiana University Press, 2010), 1–13.
13. “Elections, 1840–2008,” Voting America: United States Politics, 1840–2008, Digital Scholarship Lab, University of Richmond, 2010, http://dsl.richmond.edu/voting/elections.html.
14. Richard White, Railroaded (New York: W. W. Norton, 2011); Toral Patel et al., “Visualization: Transcontinental Railroad Development, 1879–1893,” Spatial History Project, Stanford University, http://www.stanford.edu/group/spatialhistory/cgi-bin/site/viz.php?id=341&project_id=0.
15. Shaping the West, Spatial History Project, Stanford University, http://www.stanford.edu/group/spatialhistory/cgi-bin/site/project.php?id=997.
16. Hans Rosling, “Wealth and Health of Nations,” Gapminder, http://www.gapminder.org/world/.
17. “Hans Rosling Shows the Best Stats You’ve Ever Seen,” TED Talk, 2006, http://www.ted.com/talks/hans_rosling_shows_the_best_stats_you_ve_ever_seen.html.
18. Whitney Trettien, “Computers, Cut-Ups, and Combinatory Volvelles: An Archaeology of Text-Generating Mechanisms,” http://whitneyannetrettien.com/thesis/.
19. Elizabeth Peabody, Universal History: Arranged to Illustrate Bem’s Charts of Chronology (New York, 1859), referred from http://beineckeroom26.library.yale.edu/2011/09/02/visualizing-history/.
20. Ben Fry, “On the Origin of Species: The Preservation of Favoured Traces,” http://benfry.com/traces/.
21. WordSeer, “Search Slave Narratives,” University of California, Berkeley, http://wordseer.berkeley.edu/slave_narratives/heatmap.php.
22. Mapping the Republic of Letters, Stanford University, https://republicofletters.stanford.edu/#maps.
23. Thinkmap, Visual Thesaurus, http://www.visualthesaurus.com/.
24. Moritz Stefaner, “Vizosphere,” http://www.visualizing.org/full-screen/29391.
25. Nathan Yau, FlowingData, http://flowingdata.com; PeopleMovin: Migration Flows Across the World, http://peoplemov.in/.
Putting Harlem on the Map
Beginning in 1904, black New Yorkers relocated their residences, churches, and businesses to the streets around the new subway station at 135th Street and Lenox Avenue in Harlem. Waves of African American migrants from the South and immigrants from the Caribbean joined them, creating a community in which blacks resided segregated from whites. By 1920, the area occupied almost exclusively by blacks stretched south to 130th Street, north to 144th Street, and from Fifth Avenue across to Eighth Avenue, encompassing a population of some seventy-three thousand people. In the next decade, Harlem became the “Negro Mecca” (a more cosmopolitan place than America’s other “black metropolis,” Chicago), deserving of the title “the world’s black capital.” By 1930, the black population, now numbering around two hundred thousand, had spilled over 8th Avenue to Amsterdam Avenue, and blacks were living as far north as 160th Street and approaching 110th Street to the south. Of course, Harlem’s status was about more than simply size or numbers. Harlem was home to the political and cultural leaders of black America: the New York Urban League and Garvey’s Universal Negro Improvement Association (UNIA) had their headquarters there; writers Langston Hughes, Wallace Thurman, and Zora Neale Hurston all lodged in the same building; and Duke Ellington and Ethel Waters performed in the neighborhood’s clubs and theaters.
The picture of Harlem that I have presented so far is in line with those you would find setting the scene for most studies of the neighborhood or offering a snapshot in some broader account. After six years of using digital geospatial tools to study Harlem, I am struck by how much that picture omits and how little sense of the place it conveys. If urban history is defined by a concern with particular places and spaces, it has long been satisfied with a large-scale or very selective treatment of physical locations. Certainly, when three colleagues and I conceived a collaborative study of everyday life in Harlem in the 1920s, we largely took the place as a given. Our central concern was to identify ordinary blacks, residents other than those familiar from the ever-expanding literature on the Harlem Renaissance and the beginnings of the civil rights movement. In fact, only when we headed to the archives did we even confront having to precisely define the neighborhood’s boundaries. (See additional images in the web version of this essay at http://WritingHistory.trincoll.edu.)
The digital tools we used to manage what we found in the archives changed how I thought about Harlem. Our core sources consisted of almost three thousand case files containing accounts of crimes by or involving blacks, as well as hundreds of pages from black newspapers. I had used a simple FileMaker database to organize similar material in a previous project and proposed that we develop a more sophisticated online database for this material, so that all the collaborators could share and use it. In addition, since our research concerned a neighborhood and since legal records almost invariably include information on locations, I envisioned our database as linked to maps. A geographic information system (GIS) offered that combination of a database and a mapping system, and with only a general idea of how such software worked and what it could do, I imagined using it to re-create several key blocks in Harlem. I was imagining this in 2002, when the existing historical GIS projects employed desktop ArcGIS software developed for social scientists, relied on quantitative data, and could not easily be shared online. The designers and programmers who we employed, from the University of Sydney’s Archaeological Computing Laboratory (ACL, now Arts eResearch), offered us a different approach. When they constructed our database, in 2006, they linked it to Google Maps, made available online only a year earlier, as a temporary fix to allow me to create preliminary maps of our evidence. As we grappled with a raft of design and technical challenges, it became clear that this simplified, web-based form of GIS would serve the purposes of the project, and Digital Harlem became one of the first scholarly sites to employ what has become known as the geospatial web.
GIS organizes and integrates sources on the basis of their shared geographic location—in the case of an urban setting, their street address. Working with addresses involved thinking about Harlem on a much smaller scale than had other scholars. Maps in existing historical studies focused on showing the boundaries of the area dominated by black residents and the location of a handful of landmarks, at a scale that located Harlem in Upper Manhattan, the area above Central Park. With a street map as the background, the area around the boundaries and landmarks appeared only as undifferentiated space. To fill the spaces in these maps, to re-create Harlem at the scale of individual addresses, I turned to a real estate atlas from 1930, from which the ACL eventually created a map to overlay on Google Maps. Real estate maps include the footprints of buildings, with information on how many floors high they stood, on the materials from which they were constructed, and on the presence of elevators and stores. In other words, these maps literally fill in the blocks of the street map, helping, as Ian Gregory and Paul Ell put it, to “subdivide the place under study into multiple smaller places and give some indication of how these places interact.” Thinking about Harlem at this scale, you are immediately confronted with how imprecisely most historical scholarship treats location: events and buildings are not given an address or are given only a partial or incorrect address, and little attention is paid to how that location is related to other places, to what is proximate or distant. It is enough to say that the places mentioned are somewhere in Harlem.
Examining Harlem at a smaller scale involved me in what Karen Halttunen identified as the second wave of the spatial turn in humanities scholarship, the move from constructing a spatial analysis that “tended to the metaphorical” and employed the “idiom of borders and boundaries, frontiers and crossroads, centers and margins,” to a concern with “spatial issues more materially.” While you could argue that my work is spatial history and not digital history, this would ignore the extent to which such mapping is not just enabled by digital tools but, really, only possible when you use such tools. Real estate maps are so small in scale that they cannot be reproduced in print publications, with those covering Harlem amounting to almost an entire atlas volume. However, digitized and overlaid on Google Maps, real estate maps become scalable, making it possible to zoom out from individual buildings to the neighborhood view favored in historical maps of Harlem and to an even larger scale that situates Upper Manhattan in the larger city.
Geospatial tools involve not only maps but also databases. The power of such tools is that they use geographic location to integrate material from a wide range of disparate sources. “What is important about assigning a geographic reference to data,” Karen Kemp points out, “is that it then becomes possible to compare that characteristic, event, phenomenon, etc. with others that exist or have existed in the same geographic space. What were previously unrelated facts become integrated and correlated.” The sources that feature in Digital Harlem are qualitative records, rather than the quantitative data traditionally used in GIS. Few other attempts have been made to use the technology to analyze such sources. The most prominent example is The Valley of the Shadow, which maps railroads and roads, agricultural production and farm values, households owning slaves, and voting and the location of churches, in order to compare two communities in the era of the Civil War.
What distinguishes Digital Harlem from The Valley of the Shadow is not only its sources—legal records and newspapers—but also that the database contains everything in the sources that is associated with an address. The geospatial database allowed us to incorporate and organize a range of material that historians typically treat as ephemera or pass over as too sparse or fragmentary to support an analysis. From the newspapers, we took not just the news stories on which scholars typically focus but also the society columns, sports reports, news from churches and fraternal organizations, and advertisements. From the legal records, we took every offense (not just a particular crime or group of crimes) and information on the victim and witnesses as well as the offender, on the nature of the crime’s location, and on the circumstances in which it occurred, which ranged from card games to shopping trips. The range of activity captured by Digital Harlem can be seen in the list of event types in the database. (Crimes constitute a minority of the events in the database but are prominent in that list because the variety of different offenses in the law effectively disaggregates crime more than the categories we have used for other events.) Such recurrent events as plays, movies, church services, and street speakers are not included as event types but, instead, can be located by searching for their venues by location type. At the same time, the need for a location excluded some material; every newspaper issue included some stories that did not include an address and therefore could not be included in the database. So, as much as the database offered a way to bring together a wide range of material, it was not a means of creating a “total history,” an all-encompassing picture of everyday life.
In making it possible to place the contents of a database on an online map, GIS takes advantage of one of the core properties of the digital medium, that it is visual—an aspect that historians have been slow to develop, exploit, and make integral to historical analysis. As Trevor Harris, Jesse Rouse, and Susan Bergeron argue, “The visual display of information creates a visceral connection to the content that goes beyond what is possible through traditional text documents.” (That contrast is evident in a comparison of how traffic accidents appear in the database and displayed on a map.) Not only is mapped data seen in its geographical context, but layers of different data and, hence, large quantities of data can be combined on a single map, providing an image of the complexity of the past. You can examine maps of sources at different scales and “discover relationships . . . by visually detecting spatial patterns that remain hidden in texts and tables.” Those spatial relationships prompted questions I might otherwise have ignored, and they facilitated comparisons that I would not have considered. There are definite limits to the visualizations created by Digital Harlem and by GIS in general. As Ian Gregory notes, “They provide a simple summary of the distribution of a variable in ways that attempt to stress the spatial by simplifying attribute [the characteristics recorded about entities in the database] and fixing time.” The patterns evident on these maps reveal the “where” involved in an inquiry. In the process, they ask, though they do not explain, “Why there?” Pursuing answers to such questions changed the way I thought about Harlem’s past and gave me a different perspective on the neighborhood.
One example of the new perspective that Digital Harlem offers is the map of Harlem’s nightlife (fig. 8), which includes layers showing the neighborhood’s nightclubs, the speakeasies that became ubiquitous during Prohibition, and the buffet flats that black residents set up as an alternative to those venues. The nightclubs are the most familiar of Harlem’s attractions, described in a range of sources. The presence of speakeasies is also well known, but they and their locations are discussed only in general terms. This map uses lists of locations regularly published in the New York Age as part of the editor’s efforts to get Prohibition authorities to take action. The third venue, buffet flats operated in residences, attract only a passing mention by scholars, but they became a concern for reformers at the end of the 1920s and the target of a 1928 undercover investigation by the Committee of Fourteen, a white antiprostitution organization. Combining these sources shows that these nightlife venues had very different geographies. Nightclubs, many of which predated Prohibition, clustered around 135th Street and on and east of Seventh Avenue, locations that stretched from the core of black settlement toward the areas of white population. Speakeasies could be found far more widely in storefronts throughout Harlem, mostly on the avenues (including on Eighth Avenue, which had no other nightlife), with clusters in the vicinity of nightclubs. Buffet flats likewise operated at more widely scattered locations than nightclubs, but most were on cross streets, not avenues, with many above 140th Street, a largely residential district distant from whites. Whereas whites owned most nightclubs and speakeasies and catered to white or racially mixed crowds, blacks operated the buffet flats for black patrons, without publicizing their locations, thereby extending some of the privacy of a residence to their customers. Mapping nightlife thus helps identify an unrecognized black response to Prohibition’s impact on Harlem: the creation of places where blacks could be apart from the whites who appeared in increasing numbers in the neighborhood’s nightlife.
Data on traffic accidents offer another example of how Digital Harlem changed my thinking. I would not have paid attention to these incidents if I had not been entering every event that had a location into the database. Reports of accidents appeared regularly in black newspapers but usually amounted to no more than a few sentences. A map of that information showed accidents throughout the neighborhood, concentrated on Seventh and Lenox Avenues, which ran north-south the length of Harlem. These avenues have long been recognized as occupying a central place in Harlem life. Christened the “Black Broadway” by writer Wallace Thurman, Seventh Avenue featured the nightlife visited by thousands each evening in the late 1920s, as well as many of Harlem’s major churches, which drew large crowds each Sunday. On the sidewalks of both avenues, individuals went “strolling,” donning their best clothes to display their style, to socialize with friends and to meet strangers. Men and women collecting bets on the numbers occupied the corners each weekday morning, replaced in the evenings by prostitutes. On weekends, residents paraded in the avenues themselves, with their fraternal lodges, as members of the UNIA, or in funeral processions. In addition to these well-known activities, the two avenues saw more traffic than any other roadway north of 59th Street, most of it traveling through the neighborhood en route into and out of the city, joined in the evening by fleets of taxis bringing visitors to Harlem’s nightlife. Public transport also ran on both avenues: street cars traveled on Lenox from 116th to 148th Street, and double-decker buses ran on Seventh from 110th to 155th Street.
Accidents revealed an overlooked facet of life in the neighborhood and recast the context and meaning of what happened on the avenues. Whites made up most of those driving on Harlem’s streets, behind the wheel of private cars, buses, and streetcars. As a result, traffic accidents often produced interracial encounters, conflict, and occasionally violence. In some cases, efforts to better control the traffic exacerbated that conflict. The traffic police eventually stationed at the most dangerous intersections included black officers, whose direction of white drivers often attracted crowds of blacks and enough controversy that the police department refused to appoint them to posts on 125th Street, where many pedestrians as well as drivers would have been white. In the context of the white presence in the street and the white ownership of the businesses lining the avenues, when blacks paraded, strolled the sidewalks, and spoke on the street corners, they effectively contested white dominance of the avenues and claimed those places for themselves. That such racial contests for space took place within Harlem made me pay more attention to the white places in what I had thought of as simply a black neighborhood. Not only did there prove to be a more extensive white presence than the passing mentions in the historiography implied, but as I argue in a forthcoming article, it was more contentious, introducing the racial negotiation, resistance, and accommodation that characterized the rest of the city into some places in Harlem itself, taking away from the respite that the neighborhood offered blacks.
While mapping traffic accidents made me aware of the white presence in Harlem, visualizing the lives of individuals gave me a new sense of how blacks moved around the neighborhood and the wider city. Digital Harlem uses lines to mark two kinds of temporal relations, to link sequences of locations at which events like parades took place and to link an individual’s residence with locations where she or he spent time while living at that address. The richest pictures of individual lives are contained in probation files, whose subjects reported their activities on a weekly basis for up to five years. Maps of their lives highlight the distance they had to travel to work and how often many changed their residence. Morgan Thompson, a West Indian on probation after he lost his temper and stabbed a man who had confronted his seventeen-year-old son, worked as a laborer for construction contractors. Between 1928 and 1933, that work took him to fifteen different construction sites, in downtown Manhattan, on the Upper East Side, and in the outer boroughs of Brooklyn, Queens, and the Bronx. Only once did he work in Harlem, on the construction of the YMCA on 135th Street. Maps also identify other largely unexplored occasions when blacks had to move through the city, including to attend sporting events. While basketball games took place in Harlem’s dance halls and church community houses, cricket and baseball games involving black teams took place in Washington Heights, the Bronx, and Brooklyn.
Within Harlem, while Morgan Thompson and his family lived at the same address on West 144th Street for over a decade, many black families regularly relocated. Perry Brown, a forty-five-year-old on probation for stealing coats to pay for his wife’s medical care, relocated five times in three years, not simply to get better housing, but sometimes to get rooms to lease to boarders, to obtain premises easier for his wife to maintain, or when he could not pay the rent. The variety of apartments that Brown and his wife occupied highlights the range of housing that existed in Harlem beyond the overcrowded tenements and middle-class dwellings that typically draw attention. Moreover, for all the upheaval attending relocation, mapping the Browns’ homes highlights that moving generally only involved shifting a few blocks for Harlemites, with little likely disruption of their relationships and involvement in the community.
Working with Digital Harlem also led me to write in a different form, a blog. This online writing is typically seen primarily in terms of length, as a short form, but what shaped my blogging was its relationship to other forms of digital history and historical writing. In the first instance, blogging provided a way to supplement Digital Harlem. By their very nature, the maps created on the site raise questions rather than answering them, and they could not simply stand on their own online. Digital Harlem required additional context, and a blog linked to the site provided a medium in which to tell stories about the maps on the site. In addition, Digital Harlem’s maps are difficult to incorporate into print publications, where dynamic maps have to become static, without scalability or links to the database of sources, and where colored visualizations have to be reproduced in black and white, limiting the layers and quantity of data that can be included. When presented in a blog, screenshots of the maps retain their color and can be linked to the site. They also can be supplemented with images, which are not currently in the database and for which space is limited in print publications. Thus the blog posts can also provide a form of footnote to traditional published writing, where I can elaborate in more detail on a topic that there is only space to touch on in a print article. Relative to a published footnote, the blog is a longer form. Obviously, the blog can serve both as a context for Digital Harlem and as a reference for a print publication.
The blog post “Traffic Accidents in 1920s Harlem,” for example, provides both context for a map and an extension of an argument in a forthcoming article. The 784-word post describes the broad pattern of accidents, the character of the traffic on Lenox and Seventh Avenues, traffic police posts and the reactions to black officers, and an example of the racial conflict that some accidents produced. It includes screenshots of a map of accidents, that map with traffic police posts added (with instructions on how to create the maps in Digital Harlem), and photographs of children crossing the street and of Officer Reuben Carter directing traffic. The article that will link to the post includes two paragraphs focused on the racial conflicts that occurred as a result of accidents, but it has no maps or photographs (only brief mentions of the broader patterns of accidents) and no discussion of the origins of the traffic police posts. The posts on individual lives take a slightly different approach to elaborating on what can be gleaned from the map. In writing about Perry Brown, I drew on his probation file to craft a narrative of this slice of his life and used his membership of a fraternal lodge as a way into a broader discussion of lodges in Harlem, with a map of their locations, a photograph of members of one Elks lodge, and an image of the Temple of Imperial Elks. In this case, the post does not serve as an elaborated footnote; our published analysis of Brown’s life is considerably more detailed, but it lacks the maps and images incorporated into the post. Approached from the published article (or via a search engine), posts thus serve a mediating function, offering static screenshots that link to Digital Harlem, where dynamic maps can be re-created and explored.
Recently, as I realized I had considered the blog in relationship to my project only and not to the larger Internet, Digital Harlem has led me to another form of online writing. When I began this project, I implicitly subscribed to the “build it and they will come” assumption that underlies much digital history, the idea that going online immediately delivers an audience. It does not. Search engines bring some users, but attracting more requires connecting your work to the rest of the web. Following the lead of other digital projects, I have turned to Wikipedia. I had hoped to merely add links to existing articles, but the originality of Digital Harlem means that few of the topics it deals with were previously mentioned. As a result, I have had to write contributions to articles, grappling with Wikipedia’s wariness about primary sources, blogs, and citing yourself and with the existing organization of the articles related to 1920s Harlem. The challenges of writing for Wikipedia are well documented elsewhere in this collection; to that discussion, my experience adds an example of why such engagement is more meaningful to digital historians than to other scholars and is hence something not to be left to students.
However I write about it, approaching 1920s Harlem through the maps created by Digital Harlem, using the tools offered by the geospatial web, has literally caused the way I think about the neighborhood to go through a spatial turn. Maps of the variety of sources my colleagues and I have gathered confront me with the multiple places that made up Harlem and the wide variety of events that took place there. Trying to understand those maps draws me down to the level of individual places and to the relations between them, into the web of locations in which individuals lived their lives—where they resided, worked, and spent their leisure time. Used in this way, the geospatial web can capture “the confluence of multiple rhythms” that Henri Lefebvre argued make up everyday life, offering a new perspective on what it was like to live in Harlem.
Acknowledgments: Digital Harlem exists thanks to an Australian Research Council Discovery Grant (DP0343148), on which I collaborated with Shane White, Stephen Garton, and Graham White, all of whom are due thanks for gathering the sources used in the project and for supporting the development of the site.
1. Gilbert Osofsky, Harlem: The Making of a Ghetto, 2nd ed. (New York: HarperCollins, 1971), xvi; James Weldon Johnson, Black Manhattan (1930; repr., New York: Da Capo, 1991), 147; David Levering Lewis, When Harlem Was in Vogue (New York: Penguin, 1981), 193–94; Jervis Anderson, This Was Harlem: A Cultural Portrait, 1900–1950 (New York: Farrar, Straus and Giroux, 1981), 128, 174, 236.
2. Shane White, Stephen Robertson, Stephen Garton, and Graham White, “Black Metropolis: Harlem 1915–1930,” Department of History, University of Sydney, http://sydney.edu.au/arts/history/research/projects/harlem.shtml.
3. Ian Gregory and Paul Ell, Historical GIS: Technologies, Methodologies, and Scholarship (Cambridge: Cambridge University Press, 2007), 1–19, 89–90; Anne Kelly Knowles, “GIS and History,” in Placing History: How Maps, Spatial Data, and GIS Are Changing Historical Scholarship, ed. Amy Hillier and Anne Kelly Knowles (Redlands, CA: Esri, 2008), 2–7; David Bodenhamer, “History and GIS: Implications for the Discipline,” in Hillier and Knowles, Placing History, 222–30.
4. Digital Harlem: Everyday Life, 1915–1930, http://www.acl.arts.usyd.edu.au/harlem/; “Arts eResearch,” http://sydney.edu.au/arts/eresearch/; Muki Haklay, Alex Singleton, and Chris Parker, “Web Mapping 2.0: The Neogeography of the GeoWeb,” Geography Compass 2, no. 6 (2008): 2011–39; Trevor Harris, Jesse Rouse, and Susan Bergeron, “The Geospatial Semantic Web, Pareto GIS, and the Humanities,” in The Spatial Humanities: GIS and the Future of Humanities Scholarship, ed. David Bodenhamer, John Corrigan, and Trevor Harris (Bloomington: Indiana University Press, 2010), 124–42. Damien Evans took the lead in designing our database and site, and Andrew Wilson later created the overlay map.
5. Irma Watkins-Owens does include an analysis of the residents of a single block in her study of West Indians in Harlem, Blood Relations: Caribbean Immigrants and the Harlem Community, 1900–1930 (Bloomington: Indiana University Press, 1996).
6. Johnson, Black Manhattan, 147; Osofsky, Harlem, xv–xvii; Cheryl Lynn Greenberg, “Or Does It Explode?” Black Harlem in the Great Depression (New York: Oxford University Press, 1991), i; Stephen Watson, The Harlem Renaissance: Hub of African-American Culture, 1920–1930 (New York: Pantheon, 1995).
7. “The Map,” Digital Harlem Blog, http://digitalharlemblog.wordpress.com/digital-harlem-the-site/the-map/.
11. Edward Ayers et al., “Augusta County Maps” and “Franklin County Maps,” The Valley of the Shadow, http://valley.lib.virginia.edu/VoS/maps/augustamaps.html and http://valley.lib.virginia.edu/VoS/maps/franklinmaps.html. For discussions of this site, see Gregory and Ell, Historical GIS, 195–98; and Knowles, “GIS and History,” 5–6.
16. Stephen Robertson, “Harlem Undercover: Vice Investigators, Race, and Prostitution in the 1920s,” Journal of Urban History 35, no. 4 (May 2009): 486–504; Stephen Robertson, “Harlem Undercover—the Maps,” Digital Harlem Blog, http://digitalharlemblog.wordpress.com/2009/04/17/harlem-undercover-the-maps/.
17. Wallace Thurman, “Negro Life in New York’s Harlem: A Lively Picture of a Popular and Interesting Section” (1927), in The Collected Writings of Wallace Thurman, ed. Amritjit and Daniel M. Scott III (New Brunswick, NJ: Rutgers University Press, 2003), 40–41; Anderson, This Was Harlem, 319–23.
18. Stephen Robertson, “Traffic Accidents in 1920s Harlem,” Digital Harlem Blog, http://digitalharlemblog.wordpress.com/2010/04/01/traffic-accidents-in-1920s-harlem/.
20. Lines are only one of the ways that Digital Harlem can visualize time. Layers can also represent different time periods, providing an image of change over time. A fairly straightforward example, a map with layers for prostitution arrests in 1925 and 1930, can be found in Stephen Robertson, “Prostitution Arrests,” Digital Harlem Blog, http://digitalharlemblog.wordpress.com/2009/10/05/prostitution-arrests/
21. Court of General Sessions Probation Department Case File 11076 (1928), Municipal Archives, New York City Department of Records. We have employed pseudonyms for individuals on probation, as required by the Municipal Archives. See also Stephen Robertson, “Morgan Thompson—a West Indian Laborer’s Life in Harlem,” Digital Harlem Blog, http://digitalharlemblog.wordpress.com/2009/12/01/morganthompson/; Stephen Robertson, Shane White, Stephen Garton, and Graham White, “This Harlem Life: Black Families and Everyday Life in the 1920s and 1930s,” Journal of Social History 44, no. 1 (2010): 97–98, 100–104.
22. Stephen Robertson, “Harlem and Baseball in the 1920s,” Digital Harlem Blog, http://digitalharlemblog.wordpress.com/2011/07/27/baseball-1920s-harlem/
23. Court of General Sessions Probation Department Case File 15872 (1930), Municipal Archives, New York City Department of Records; Stephen Robertson, “Perry Brown: A Lodge Member’s Life in Harlem,” Digital Harlem Blog, http://digitalharlemblog.wordpress.com/2010/07/15/perry-brown-lodge-member/; Robertson et al., “This Harlem Life,” 104–8.
24. Stephen Robertson, Digital Harlem Blog, http://digitalharlemblog.wordpress.com.
27. Stephen Robertson, “Digital Harlem and Wikipedia,” Digital Harlem Blog, http://digitalharlemblog.wordpress.com/2012/01/11/digital-harlem-and-wikipedia/.
28. The description of Lefebvre’s argument is from Stefan Kipfer, Kaniska Goonwardena, Christian Schimd, and Richard Milgrom, “On the Production of Henri Lefebvre,” in Space, Difference, Everyday Life: Reading Henri Lefebvre, ed. Kaniska Goonwardena, Stefan Kipfer, Richard Milgrom, and Christian Schimd (New York: Routledge, 2008), 15.
Pox and the City: Challenges in Writing a Digital History Game
Real event or plausible scenario? First-person shooter or third-person isometric perspective? These are some of the questions we confronted as we began the collaborative digital history project Pox and the City, a role-playing game funded by a start-up grant from the Office of Digital Humanities. How do we adapt the content into a playable scenario that retains educational and research value? What restrictions do pedagogical concerns place on the actual programming? How are these concepts visually represented in a digital world?
When completed, Pox and the City will allow students to explore the interplay of disease, patients, healers, and social institutions in medical history. Set in early nineteenth-century Edinburgh, Scotland, the game is designed to allow players to adopt one of three roles: a newly graduated physician, intent on setting up a paying medical practice by using the recent discovery of vaccination for smallpox; an Irish immigrant, just arrived in Edinburgh’s immigrant district and hoping to establish himself in a market stall in the city’s central district; and a smallpox virus, “intent” on replicating and spreading throughout the city. Each role has a home base in the city and a distinct set of tasks he/it must perform in order to move to the next level.
The game is a collaboration among scholars with different specialties and different approaches to writing history. Lisa Rosner, a historian of medicine with a long-standing interest in web design as a digital humanities tool, is the content specialist for Edinburgh medicine and for visual representations of the city. Laura Zucconi is both a historian of medicine and an avid gamer, with a previous incarnation as a programmer. Both are based at Stockton College, New Jersey. Ethan Watrall, an anthropologist and a designer of serious games, from Michigan State University, oversees the design team, Adventure Club, also based in East Lansing. Hannah Ueno, a visual artist and 3D graphic designer from Stockton, is creating a virtual world of 1800s Edinburgh as a complement to the game. The project is also a collaboration between the project designers and the staff of the Historical Collections of the College of Physicians of Philadelphia, who provide many of the primary sources on which the game is based. Each of the many partners brings a specific area of expertise to the project, but what may appear as a simple decision in one area suddenly becomes problematic when it intersects with another area. For the authors of this essay, working on Pox and the City has transformed the writing of history from a process designed and carried out by a single individual, firmly in control, to an exhilarating, surprising, and, above all, collaborative effort, akin to completing a giant jigsaw puzzle when we are not sure who has all the pieces.
The best way to illustrate the “jigsaw” analogy is to follow the way our collaboration evolved. This project began with Rosner, as the content specialist, firmly in control of the narrative and debating the issue of where to situate the game in time and space. One option was to re-create an actual historical event, Edward Jenner’s 1798 research establishing that vaccinating a patient with cowpox matter resulted in his or her immunity to the deadly disease smallpox. Another option was to create a plausible historical scenario, situated in early nineteenth-century Edinburgh, based on Rosner’s expertise on the medical history of the city. Since the former is based on historical fact, while the latter would be a kind of historical fiction, Rosner’s choice initially seemed clear: we should develop a role-playing game based on Jenner’s actual medical research. Such a game would appear to offer the best support for a key facet of history pedagogy, teaching students to make inferences about the past based on historical evidence.
Once Rosner and Zucconi began to collaborate on gamifying the content, though, this choice appeared less clear-cut. We investigated the precise pedagogical purpose served by a role-playing activity, digital or otherwise, and what kinds of assumptions might be embedded within the choice of specific topics to enact. One common purpose is to convey a “you are there” sense of a specific historical moment, to allow students to re-create an event or series of events. Common scenarios used for this are the 1787 debates over the Constitution or the set of alliances leading to the outbreak of World War I. This type of scenario can be easily paired with another pedagogical imperative, getting students to read primary sources. The underlying assumption is that the more students read the sources, the more they will re-create the actual historical events, and the better they will understand how those events took place.
This assumption is deeply ingrained in the pedagogy of the history of science. Many high school and college science courses re-create classic experiments, such as Robert Millikan’s oil-drop experiment, in student labs. The goal is not just to teach the scientific facts but also to provide examples of analytical reasoning and scientific creativity. James Conant’s seminal work Harvard Case Histories in Experimental Science explicitly incorporated this goal into the teaching of the history of science. Like the Harvard Business School case histories, specific episodes in the history of science were to serve as exemplars, negative and positive, of rational inquiry, innovation, and problem solving. This view of the history of science privileged major scientists and has made a very successful transition to new media, such as the Public Broadcasting Service’s Nova.
But the best-laid plans go awry in role-playing activities as often as elsewhere. The group dynamics of role playing in the classroom are complicated: particularly charismatic or present-minded students may skew the results, so that the Founding Fathers end up abolishing slavery in 1787 or so that Italy succeeds in negotiating a peace settlement that heads off World War I. Within the history of science, the complexities of re-creating even a single individual’s process of discovery in the classroom is time-consuming and requires a level of engagement with primary source materials—including lab apparatuses—that is difficult to achieve in a standard semester-length course. The laboratory program at St. John’s College, which does use texts based directly on classic works in the sciences, like Isaac Newton’s Principia, requires a full three years of intensive immersion. In the history of medicine, any kind of hands-on re-creation of historical events—for example, Jenner’s inoculation practice—is out of the question.
As our research into serious games told us, adding the game component to role-playing activities creates even more problems. A historical simulation, or case study, is not the same as a game. How can students play at being Edward Jenner? Would they be rewarded—for example, earn points or collect digital tokens—for reading about his life and work and choosing correctly among a set of online scenarios? What would make this a game, rather than an online test, perhaps (but only perhaps) more appealing than a traditional written test? Unlike Nova episodes, games do not lend themselves to re-creation of the lives and ideas of specific, well-known individuals, because then players are constrained, not empowered, by their knowledge of real historical events. Instead, games work best when they are open-ended, allowing players a set of choices without predetermined outcomes.
Games also work best when they are visually interesting. One of the appeals of historical games, like the Caesar series, is the opportunity to move through a world that no longer exists. This is also an important pedagogical point: not only can students learn about the history of, for example, architecture or urban conditions, but they can also learn to “read” visual as well as text-based historical sources. Yet this was a complicating factor if we made Edward Jenner the focus of the game. We know very little about Jenner’s physical environment—what his house looked like, where he performed his vaccinations, how his village was situated in the landscape. It might be possible to re-create the environment by using contemporary images, but such images would only be approximations. They would be, in effect, a kind of historical fiction rather than historical fact.
So we returned again to historical pedagogy to look for alternatives to the traditional role-playing reenactment of a specific historical event. An obvious model is the use of films to teach history. Few historians would argue that films, even those based on real events, are entirely accurate. But they can be ideal media for conveying an understanding of the past, and their use is supported by a growing body of scholarly literature. Games, like films, can be based on serious scholarly research and are as well suited to visual as textual sources. Moreover, the scholarly literature continues to expand through both print and born-digital publications.
We redesigned our jigsaw puzzle: instead of starting with the content and then somehow forcing it into a game, we started with the structure of the game and asked ourselves what kind of content would best serve its purposes. We wanted to create a dynamic and engaging game in a role-playing multiuser format that allowed an exploration of the social history of medicine, rather than just a recapitulation of accepted theories. In this environment, students could more freely explore the medical culture of the nineteenth century by asking themselves questions when presented with problems, such as how a doctor might convince a wealthy patron to be vaccinated and whether the doctor would act the same way toward a patient from the laboring class. The student must figure out on his or her own which documents to access in the archive and how to synthesize that data with the game mechanics. This process of question and discovery makes for a greater impact in terms of active learning.
For the high school or undergraduate student, the plausible scenario helps them learn how to do research. For graduate students and other researchers, the plausible scenario approach can aid them in what to look for when working in the archives. A re-creation of nineteenth-century Edinburgh that permits free-form movement of the players allows them to interact in ways that our current models of historical narrative may not address. As players try to solve game problems, such as where to get money to set up a medical practice, they may devise a novel solution. Their task would then become combing the archives to see if pertinent economic data relevant to their theory had been overlooked by previous histories. Even if such data cannot be found, the researcher would, at the very least, develop a better understanding of nineteenth-century economic values, avoiding the pitfall of accidentally imparting anachronistic perceptions.
A plausible scenario format comes with the difficulties of character creation and developing quests that highlight historically important data. There is also the danger of losing the historical narrative as players create a new environment through their actions. We resolved the issue of what types of characters to create as player characters (PCs) and nonplayer characters (NPCs) by returning to the content, in this case, a central concept for the history of medicine: the interaction of disease, patient, and healer. Thus the PCs would be the doctor, an immigrant laborer, and the smallpox virus itself. This choice of PCs would work well with either a research or pedagogical approach to the game. NPCs would account for any other people that would normally interact with our three standard PCs, such as a wealthy patron for the doctor.
After settling on the basic ideas of both research and pedagogy played in a plausible scenario format, we turned to the look of the game, in consultation with designer Ethan Watrall. He added a new set of pieces for our jigsaw puzzle, and we once again had to shake some out and start rebuilding. Our initial conception was that the game would follow the style made popular by such first-person shooters as Planet Wolfenstein and Doom. We felt that this would give the player the best feel for nineteenth-century Edinburgh, as they would move through various scenes of closes, markets, and buildings based on contemporary maps, etchings, and watercolors. Such immersive environments have proven effective in simulation-based training games for pilots and military personnel.
Watrall pointed out several technical problems, though, with this popular style of game perspective. The first is that we expect students to play the game on a standard computer using a mouse and keyboard, rather than a joystick or gamepad. The first-person perspective simply does not work well with a mouse interface. The second problem had to do with the design decision, prevalent in the genre of serious games, to program in Flash. Knowledge of Flash is common among game designers and allows comparatively rapid game development. Since students can be competent Flash designers, it helps keep design costs within the grant budget. In addition, because Adobe Flash is both ubiquitous and cross-platform, there are very few issues with compatibility or accessibility. Although Flash-based applications do not work on Apple devices like the iPad and iPhone, that is not a drawback for Pox and the City, because we expect students to play it through a web browser on a standard computer. However, Flash programming does not lend itself well to games played in a browser in first-person perspective, because of the variation in download rates.
Watrall suggested that the game design use a third-person isometric perspective, with the “camera shot” above and at an angle to the player character. This perspective plays well with a browser deployment. Watrall pointed out that recent research indicates that a third-person perspective allows for greater immersion in a role-playing game, because the player can see his character within the environment. Additionally, this perspective permits a wider view of the environment, thus richer detail can be built into the game and absorbed by the player. The following illustration of an Edinburgh map adapted for game play demonstrates these points. (See the additional image in the web version of this essay at http://WritingHistory.trincoll.edu.)
The issue of immersion is not limited to just resolving a player’s perspective. As our fourth collaborator, Hannah Ueno, pointed out, the style of graphics equally affects how well a player feels connected and interacts with the visual features of the game. A photorealistic quality to the graphics is the ideal environment, but if this is not done well, it will actually detract from player interaction and negatively impact the overall learning outcomes. Studies have shown that a more stylized art approach that is illustrative or “cartoony” creates a certain level of suspended disbelief that allows the player to feel more connected to the game. At the time of writing, we have yet to fit in all the jigsaw pieces associated with a particular art style, but we are leaning toward a graphic look that imitates nineteenth-century watercolors and line drawings depicting Edinburgh’s Old Town. From the 1860s through the 1920s, Edinburgh undertook a series of urban construction projects that eventually eliminated the unhygienic alleys and courtyards of previous centuries. Local artists, concerned to record their rich architectural heritage, went street by street through the city, creating a wealth of visual imagery. Of course, they were not merely recording what they saw but, rather, interpreting it as a record of a bygone era—a vanished past, once great, now fallen into decay. We expect these illustrations to work very well in evoking the era and drawing players into the game. That many of the streets can be located on historical and contemporary maps, available online from the National Library of Scotland, adds another layer to the pedagogical goal of the game.
Pox and the City is still in its initial stages, with much work to do even in developing the basic design for the game, let alone working out art assets or detailed scenarios. It is fair to say that our ideas of writing a history game as an isolated process, carried out by an individual in a book-lined study, have been permanently transformed. We hope the game, once completed, will prove to be a similar vehicle for transformation for our students and colleagues.
Update: Since this chapter was written, the game has been publicly released at http://poxandthecity.blogspot.com
1. “Three Cheers for Pox and the City!” School of Arts and Humanities Grants and News, South Jersey Center for Digital Humanities, Stockton College, http://intraweb.stockton.edu/eyos/page.cfm?siteID=69&pageID=246.
2. “NEH Announces 22 New Start-Up Grants,” Office of Digital Humanities, National Endowment for the Humanities, April 2011, http://www.neh.gov/ODH/ODHHome/tabid/36/EntryId/161/NEH-Announces-22-New-Digital-Humanities-Start-Up-Grants-April-2011.aspx.
3. Edward Jenner, An Inquiry into the Causes and Effects of the Variolae Vaccinae (London, 1798), http://babel.hathitrust.org/cgi/pt?id=ucm.5325107182.
4. A recent example is discussed in Lauren Kientz Anderson, “Exciting New Pedagogy Based in History of Ideas,” U.S. Intellectual History, January 12, 2010, http://us-intellectual-history.blogspot.com/2010/01/exciting-new-pedagogy-based-in-history.html.
5. For one of many examples, see Ryan McAlister, “The Millikan Oil Drop Experiment,” Fall 2003, http://ffden-2.phys.uaf.edu/212_fall2003.web.dir/ryan_mcallister/slide3.htm.
7. See, for example, “Galileo’s Battle for the Heavens,” PBS Nova, 2002, http://www.pbs.org/wgbh/nova/galileo/
8. “Academic Program: Laboratory,” St. John’s College, 2008, http://www.stjohnscollege.edu/academic/laboratory.shtml.
10. Jason Ocampo, “Caesar IV: Review,” Gamespot, October 5, 2006, http://www.gamespot.com/pc/strategy/caesar4/review.html?tag=summary%3Bread-review.
11. An excellent teaching resource is Mark C. Carnes, ed., Past Imperfect: History according to the Movies (New York: Henry Holt, 1996); see also Robert Brent Toplin, “The Historian and Film: Challenges Ahead,” Perspectives 34, no. 4 (April 1996), http://www.historians.org/perspectives/issues/1996/9604/9604FIL.CFM. The Internet Modern History Sourcebook, maintained by Fordham University, includes a section titled “Modern History in the Movies,” http://www.fordham.edu/halsall/mod/modsbookmovies.asp.
12. Much of the current discussion highlights the use of commercial games in teaching, but that can be expected to change as the field of serious games expands. See Russell Francis, “Towards a Theory of a Games-Based Pedagogy” (paper presented at the conference “Innovating E-Learning 2006: Transforming Learning Experiences,” JISC online conference, March 2006), http://gu-se.academia.edu/RussellFrancis/Papers/409557/Towards_a_Theory_of_a_Games-Based_Pedagogy; see also Ethan Watrall, “Interactive Entertainment as Public Archaeology,” Archaeological Record 2, no. 2 (2002): 37–39, http://www.saa.org/Portals/0/SAA/Publications/thesaaarchrec/mar02.pdf; Brian Winn et al., “What Should Higher Education Learn from Games?” (presentation at the National Learning Infrastructure Initiative, San Diego, CA, January 2004, http://www.educause.edu/library/resources/what-should-higher-education-learn-games).
14. In 2008, Edward Castronova of Indiana University designed a version of the Neverwinter Nights multiuser game called Arden: The World of Shakespeare for experimental economics, as he described in “A Test of the Law of Demand in a Virtual World: Exploring the Petri Dish Approach to Social Science” (CESifo Working Paper 2355, Department of Telecommunications, Indiana University, Bloomington), http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1173642.
17. S. Bangay and L. Preston, “An Investigation into Factors Influencing Immersion in Interactive Virtual Reality Environments,” in Virtual Environments in Clinical Psychology and Neuroscience, ed. G. Riva, B. K. Wiederhold, and E. Molinary (Amsterdam: Ios, 1998); Charlene Jennet et al., “Measuring and Defining the Experience of Immersion in Games,” International Journal of Human-Computer Studies 66, no. 9 (September 2008): 641–61.
18. R. Wages, S. M. Grunvogel, and B. Grutzmacher, “How Realistic Is Realism? Considerations on the Aesthetics of Computer Games,” Lecture Notes in Computer Science 3166 (2004): 83–92; J. Seyama and R. S. Nagayama, “The Uncanny Valley: The Effect of Realism on the Impression of Artificial Human Faces,” Presence: Teleoperators and Virtual Environments 16, no. 4 (August 2007): 337–51; Babak Kaveh, “A Fresh Look at the Concept of Immersion,” Game Design Ideas, March 10, 2010, http://www.gamedesignideas.com/video-games/a-fresh-look-at-the-concept-of-immersion.html.
19. Thomas Shepherd, Modern Athens, Displayed in a Series of Views (London: Jones, 1830); James Drummond, Old Edinburgh (Edinburgh: G. Waterston Sons and Stewart, 1879); Bruce Home, Old Houses in Edinburgh (Edinburgh: Hay and Bagster, 1905).
20. Maps of Scotland, National Library of Scotland, http://maps.nls.uk/.