25

The Differences between Digital Humanities and Digital History

Stephen Robertson

Digital humanities as it is currently constituted has not erased the differences between academic disciplines. Digital humanists share a commitment to collaboration, openness, and experimentation; a set of software tools, such as Voyant, Palladio, or Omeka; and a group of venues such as THATCamp. But disciplinary sources, questions, and approaches shape their projects, as well as their choice and use of digital tools.1 The Companion to Digital Humanities, the edited collection that has become a convenient touchstone for the emergence of the field that has succeeded humanities computing, reflects these diverse practices. It contains chapters on work in different disciplines, followed by chapters on “Principles,” “Applications,” and “Production, Dissemination, Archiving” (Schreibman, Siemens, and Unsworth). However, at the time of the Companion’s publication in 2004, work with digital technology and media still received little recognition in academic disciplines, and many practitioners held junior positions or were located in the interstices of institutions (Rockwell). In those circumstances, the attraction of an interdisciplinary digital humanities as a professional home overshadowed the disciplinary differences within it. Digital humanities became a big tent (Terras; Svensson).

Much has changed since 2004. The use of digital technology and media are now recognized, if not always embraced, by most of the major humanities professional organizations. Strands of digital sessions now appear in conference programs for major and more specialized conferences: digital history, for example, can be found in the American context not only at the annual meetings of the American Historical Association (AHA) and Organization of American Historians, but also smaller organizations such as the Southern Historical Association, the Urban History Association, and the International Congress on Medieval Studies.2 The AHA will soon adopt standards for assessing digital scholarship for promotion and tenure, joining the Modern Language Association and the College Art Association.3 Job advertisements in history, English, art history, archaeology, and media studies now include a regular sprinkling of positions requiring expertise in digital technology, some even as the primary field. Undergraduate and graduate courses, certificates, and masters programs in digital humanities are appearing in a diverse range of institutions. At the same time, in a predictable corollary to this growing acceptance, digital humanities has been the target of attacks in scholarly and more popular publications. In almost all cases, what is labeled digital humanities in those critiques is in fact digital literary studies, effectively casting the big tent as housing only a single discipline.4

In this moment, our focus needs to shift back to disciplinary difference within digital humanities.5 The audiences to which we are now seeking to explain our work are increasingly not only administrators and funding agencies whose broad remits echo the breadth of digital humanities, leaving us to elaborate only the digital. Instead, we are engaged more at the local level, with individual scholars, teachers, students, and groups beyond the academy who rarely see their interests and projects in such broad terms. As a result, to them digital humanities is doubly unintelligible, requiring us to persuade them of the credibility and relevance of both the digital and the interdisciplinary. Their willingness to engage with digital humanities is often further reduced by their association of the term with attacks on digital literary studies. A conversation instead about digital history, for example, allows explorations of the specific technologies and approaches that offer the most possibilities for the sources, questions, and teaching of historians. As Ryan Cordell perceptively notes in regards to teaching digital humanities, a “rapid introduction to as many methods and tools as possible” provides only a “glancing understanding of any aspect,” whereas “a more focused introduction to a few tools, methods, and theoretical conversations” within a disciplinary focus provides “fewer but more well-developed skills from which they can build” (see chapter 36 in this book on “How Not to Teach Digital Humanities”).

Moving forward, we would be better served by reimagining digital humanities not as single all-encompassing tent but as a house with many rooms, different spaces for disciplines that are not silos but entry points and conduits to central spaces where those from different disciplines working with particular tools and media can gather. Each of the many disciplinary rooms would have a distinctive character, reflecting a particular contribution and orientation to the field.

In this chapter, I explore digital history’s use of the Web and computational tools as two areas that distinguish the discipline within digital humanities, and from digital literary studies in particular. In singling out digital literary studies as a counterpoint, I seek to disrupt the recent tendency in both scholarly and popular discussion to equate that field with digital humanities. Beyond that historically contingent assertion of difference, I aim to redirect attention to academic disciplines as central to what different digital humanists do in order to make clear what the field offers scholars in those disciplines. A common reaction to the argument I make here has been to point to a handful of projects that do not fit the patterns which I identify, and on that basis, dismiss the importance, if not the existence, of disciplines within digital humanities. To be clear, disciplinary difference in digital humanities is not a binary, in which approaches are the preserve of particular disciplines and entirely absent from others. Rather, difference is a spectrum of emphasis, with varying degrees of interest in methods, tools, and values. Looking for projects that do not fit the differences I identify is not the test of the argument; rather, the key is establishing which projects are exceptions and which are exemplars.6

The most common use to which digital humanists, including some historians, have put the Web has been the distribution and presentation of material to other scholarly researchers. Such a use of the Web in digital literary studies, for example, is clear in the Companion to Digital Humanities; audiences other than scholars barely receive a mention. Matthew Kirschenbaum, in exploring the user interface of the William Blake Archive, did respond to complaints that it was “not particularly easy to use” and “chock-full” of “scholarly trappings” by noting that “while we are happy to have users from many different constituencies, the site’s primary mission has always been expressly conceived as scholarly research” (“So the Colors Cover the Wires”). Perry Willett’s discussion of the audience for electronic texts noted similar tensions and likewise put “general readers” to one side (Electronic Texts). An orientation to a scholarly audience continues to characterize the newer generations of projects, such as the Shelly-Godwin Archive.7 Such uses of the Web to present both primary sources and scholarship were also the focus of the account of digital history offered by William Thomas in the Companion to Digital Humanities (“Computing and the Historical Imagination”). Like electronic editions and other TEI projects, digitized archives such as Valley of the Shadow, the project that Thomas helped lead, or my own Digital Harlem, for that matter, were designed to meet the needs and interests of the scholars who created them.8 As David Parry notes, placing them online made them publicly available, but did not expand the scope of their audience (“Be Online or Be Irrelevant”). They remained accessible, relevant, and useful primarily to those scholars and their colleagues—and to the incidental audiences of genealogists, who search historical databases for individuals, and readers seeking copies of literary texts.9 As Sheila Brennan reminds us in her contribution to this collection, your audience is those for who you design a project, not those who can find it (see chapter 32, “Public, First”).

Notwithstanding Thomas’s focus on historians using the Web to reach other scholars, since the earliest days of the Web, historians created online projects for audiences rarely addressed by other digital humanists. Many of the early practitioners of digital history were social historians and radical historians committed to democratizing the creation of the past and to collaborating with teachers to enhance and diversify the history taught in high schools and universities. Their pursuit of those goals gave digital history a distinctive emphasis on using the Web to reach classrooms and the wider public. In the American context, the availability of funding through the U.S. Department of Education’s Teaching American History program and the NEH Education Programs and Public Programs facilitated these projects.10 Professional recognition was also forthcoming: beginning in 2004, the American Historical Association’s James Harvey Robinson Award for the most outstanding contribution to the teaching of history was given to online projects on four consecutive occasions.11 In addition, the field of public history has given work with audiences beyond the classroom a place within the historical profession for which there is no equivalent in other disciplines in digital humanities—although it should be noted that that place is not in the mainstream of the profession. Public history grew from a movement in the 1970s to apply history to real-world issues, but in the United States it has only relatively recently gained support and recognition within research-oriented institutions and professional organizations.12 That digital history is recognized within public history is clear from the NCPH’s Outstanding Public History Project, which has been awarded to three online projects since 2008.13

The work of the Roy Rosenzweig Center for History and New Media14 (RRCHNM), founded in 1994, provides examples of how digital historians used the Web to reach both audiences in classrooms and among the wider public. Roy Rosenzweig established the center with the mission of using digital technology and media “to democratize the past—to incorporate multiple voices, reach diverse audiences, and encourage popular participation in presenting and preserving the past.”15 RRCHNM’s earliest online projects were collaborations with the American Social History Project created with and for teachers. History Matters: The U.S. Survey Course on the Web, launched in 1998, included historical sources like those found in digital archives, but annotated by historians and packaged with online forums with leading historians and teachers, syllabi annotated by teachers, projects created by history students, and interviews with distinguished teachers who shared their strategies and techniques for using the historical sources in the classroom.16 Other projects explored the French Revolution and highlighted the new fields of world history and the history of childhood and youth, and recent history, such as the events of 1989.17 Teachinghistory.org pursued this work on a particularly large scale, building on and disseminating the results of more than 1,000 Teaching American History projects, providing both history content and teaching materials and bringing together different communities concerned with history education.18 The audience for these sites continues to grow: well into its second decade, and lacking design features now expected on the Web, History Matters nonetheless attracts increasing numbers of visitors each year: 2.1 million unique visitors and just over 3 million visits in 2014. That same year, Teachinghistory.org saw its audience grow to 1.6 million unique visitors and just over 3 million visits.

RRCHNM also used the Web to develop public history projects involving audiences beyond the classroom. Early projects focused on collecting and preserving the past, in collaboration with a variety of communities. ECHO fostered communication and dialogue among historians, scientists, engineers, doctors, and technologists in order to develop and disseminate methods to collect and preserve the recent history of science, technology, and industry.19 The September 11 Digital Archive,20 a collaboration with the American Social History Project at the CUNY Graduate Center, collected digital material related to the attacks both through a web interface and direct collecting. The Hurricane Digital Memory Bank further developed that approach, adding the geolocation of contributions and using extensive outreach activities and strategic partnerships with more than twenty universities, museums, state humanities councils, media outlets, nonprofit relief groups, and grassroots empowerment groups.21 In the Bracero History Archive, a web-accessible repository facilitated many parties working together to build a collection of oral histories with Bracero guest workers and other types of historical evidence of their experiences, and then present it with supporting materials and teaching resources.22 Gulag: Many Days, Many Lives and Objects of History involved collaborations with Russian museums and the National Park Service, and with the National Museum of American History, respectively, to build and present exhibits.23 While historians also use the Web to distribute and present material to other scholars, the practice of placing material online to reach and collaborate with the wider public, and to reach teachers and students in classroom settings, has become established and professionally recognized within history to an extent that distinguishes digital history within digital humanities.

Digital history is distinctive within digital humanities in a second area—its use of computational tools. This distinction is less intense than that in the use of the Web. Historians can be found using mapping, text analysis, and network software, the three types of computational tools favored in the digital humanities, as well as 3D modeling, image analysis, and social media. However, digital historians have turned to digital mapping to a greater extent than other disciplines in the digital humanities, adopting it as their favored computational tool. For example, of the list of digital humanities projects maintained by John Levin and adopted by the GeoHumanities Special Interest Group of ADHO, almost two-thirds are historical projects.24 It is certainly the case that digital mapping is not as central to history as it is to archaeology, but mapping has long been a part of record keeping the practices of archaeology. The turn to digital mapping in history is not continuous with older practices to the same extent. If maps have appeared in historical scholarship, historians have until recently written about the past with little attention to space, notwithstanding the discipline’s definition around nations and cities.

The distinctive prominence of mapping in digital history is a product of both the spur to mapping that the spatial turn provided to historians and of the more limited availability, until recently, of historical material to which text analysis tools could be applied. In the last decade, historians have given renewed attention to space and place as part of spatial turn in the humanities—which began in the 1970s with a focus on perceptions and representations, that, as Karen Haltunen put it, “tended to the metaphorical” and employed the “idiom of borders and boundaries, frontiers and crossroads, centers and margins,” and extended in the early 2000s to a concern with “spatial issues more materially” (2).25 It was this more recent development that provided a spur to mapping; visualization was not as necessary a part of exploring spatial perceptions and language.

At the same time, mapping tools became more accessible and useful to historians seeking to extend the spatial turn to the material. The first computational mapping tools, GIS software, which appeared in the early 1960s and became more widely available in the early 1980s, were, and still are, expensive, complex, and cumbersome and operate only on computers running Windows. Moreover, this software was designed for precise, quantitative data that could be parsed in highly structured tabular databases and cartographic maps that emphasized generalization. It fit uneasily with the sources that most historians use, which are characterized by ambiguity, uncertainty, uniqueness, and with the discipline’s move away from statistical analysis after the 1970s. A small group of historians did adopt GIS software and continue to work within its limitations, under the umbrella of the Social Science History Association and its journal, Social Science History, and largely disconnected from digital history. In the last decade, the development of the geospatial web has made mapping available and accessible to digital historians, with web mash-ups allowing maps to be created and iterated with unprecedented ease. Web mapping platforms such as Google Maps, MapQuest, CartoDB, Palladio, and Neatline do not have the same quantitative orientation as GIS software, allowing a diversity of spatial data to be visualized.26 More advanced and customized web mapping is also now possible using open-source tools such Leaflet, GeoServer, d3.js, and OpenLayers. With these tools, historians explored a range of questions and topics: urban development, housing, ghettos, moviegoing, language use, sound, industries ranging from iron to cutlery, events such as bombing, military recruitment, the slave trade, slave revolts, and gaining freedom from slavery, the spread of disease, witchcraft accusations, feminist activism, and the circulation of correspondence.27 These tools are obviously available to all digital humanists, and other disciplines have also experienced the spatial turn. So, while those factors help explain the prominence of mapping in digital history, they do not explain its relatively lesser place in the practice of other disciplines within digital humanities.

Digital historians have often used the geospatial web as a platform for presenting historical sources. Maps allow the visual display of information, and for the organization and integration of different sources on the basis of their shared geographic location. The first two winners of the AHA’s Roy Rosenzweig Prize for Innovation in Digital History28 were mapping projects built on Google Maps: Digital Harlem: Everyday Life, 1915–1930 (which I created with collaborators at the University of Sydney) and Bobby Allen’s Going to the Show. Digital Harlem offered a visualization of everyday life through maps of locations, events and individuals’ lives, using information extracted from legal records, newspapers, and other published and archival texts.29 Going to the Show mapped the locations of movie venues in North Carolina between 1896 and 1922, combined with an archive of related material.30 More recent examples include: Mapping the Republic of Letters,31 which traces the circulation of correspondence among enlightenment intellectuals using prototype browser-based tools now being developed as an open-source platform, Palladio; The Roaring Twenties, which displays the location of sounds and noise complaints in 1920s New York City, initially using Google Maps, before shifting to MapQuest; and Photogrammar, which located 170,000 photographs taken by staff of the Farm Security Administration in the 1930s and 1940s, using Leaflet and CartoDB.32

Historians have also used web maps that display historical data as part of their analysis of those sources. The interactive, iterative features of digital maps make them research tools, a means of discovering as well as displaying knowledge. Not only are mapped sources placed in their geographic contexts, but selections of those sources can be mapped, different layers of sources can be juxtaposed, and the scale can be zoomed from the level of individual buildings out to neighborhoods, cities, and regions. Viewing these maps can reveal spatial patterns not evident from reading the texts, relationships that facilitate comparisons and prompt questions. The answers to those questions are not on the maps, but in the sources from which they are derived. So mapping a range of sources using Digital Harlem,33 for example, highlighted a more extensive and expansive white presence in the neighborhood, leading to questions about the nature of a range of interracial encounters, when they occurred, and the extent to which life in this black neighborhood could be lived apart from whites (Robertson, White, and Garton). Similarly, Visualizing Emancipation mapped the position of Union troops and a range of emancipation events drawn from Official Records of the War of the Rebellion and other sources to trace the spread of emancipation during the Civil War and explore the relationship between the presence of troops and enslaved men and women’s escape from bondage (Ayers and Nesbit).34

Less often, historians have created maps to illustrate arguments or present the answer to questions, forgoing some interactivity and ability to query the data in order to highlight a particular analysis. Slave Revolt in Jamaica, 1760–1761, for example, offers an animated narrative map of a slave revolt built using Leaflet. That map allows the user to move within the timeline, but in no other way to affect the map. Mapping Occupation35 presented a spatial narrative that coupled a text with a series of maps to explore the Union Army’s occupation of the Reconstruction South. Icons on some of maps can be clicked to reveal the underlying data and the scale can be adjusted, but otherwise the maps are limited to illustrating a particular argument.36

The relative prominence of mapping in digital history is also a product of historians’ limited use of the textual analysis tools. It is not the case that historians have been unable to see ways to use text mining and topic modeling in their research. To the contrary, historians were among the first digital humanists to experiment with such tools. Dan Cohen37 made early use of data from Google books, and he went on to be part of the With Criminal Intent38 project that data-mined the Old Bailey trials (Searching for the Victorians; Data Mining with Criminal Intent). Sharon Block39 and Rob Nelson employed topic modeling. Block explored the eighteenth-century Pennsylvania Gazette (“Doing More with Digitization”). Nelson explored the Confederacy’s paper of record during the Civil War (Mining the Dispatch40). But until recently, little work has built on those experiments.

One explanation for the relatively limited use of text analysis tools by historians is a lack of digitized sources. As Dan Cohen noted in 2008, “Unless we can have machines scan, sort, and apply digital techniques to the full texts of documents, we can’t do sophisticated digital scholarship” (Cohen et al., “Interchange”). Sources that are central to historical scholarship have been digitized and made machine-readable on a far lesser scale than those on which literary studies relies. In part, this reflects the fact that historians rely more on documents that are not published or printed. Making what is digitized accessible to computational tools is a more expensive and time-consuming task in the case of those sources than with published sources. The optical character recognition (OCR) software that extracts text from images currently only works with printed texts, meaning that handwriting must be transcribed. Although the Proceedings of the Old Bailey, one of the best-known and most widely used digitized collections, used manual keying as part of its digitization process, transcription is beyond the resources of most institutions and employed largely only through crowdsourcing.41 Much of the early digitization of historical sources was highly selective, chosen from larger collections to answer specific research questions or to make available well-known or popular documents.42 Both forms of collection have limited attraction to a wider audience of researchers, especially given the premium the profession places on original research.43 Contrast the use of the Valley of the Shadow44 and the Proceedings of the Old Bailey:45 the former is a pioneering virtual archive of sources selected by Edward Ayers for a comparative investigation of two communities in American Civil War; the latter is a collection of all the surviving editions of the published accounts of trials that took place at the Old Bailey. Although the Valley project produced a path-breaking digital article in the American Historical Review and a prize-winning book, and has been widely used in teaching, I could identify only one other scholar who used it for research.46 As of July 2015, by contrast, 401 publications cite the Proceedings.47

In addition, many of the historical sources that are digitized are only accessible by search, in databases that do not have an API or other means of accessing the complete data. This is particularly true of the commercial databases that contain digitized newspapers, a published source widely relevant to historical research and less so to other fields in digital humanities. Sixteen of the most important American newspapers and nine of the leading African American newspapers, for example, can only be found in Proquest Historical Newspapers. While those databases allow researchers to search text generated by OCR, the results that are returned are only images of the page, containing no machine-readable text. Not only is there no means of bulk export, but the licensing agreements expressly prohibit data mining. Both Proquest and Gale Cengage have announced plans to open up their databases to computational analysis, but it remains unclear at what cost and by what means.48 Peter Leonard and Lindsay King did obtain access to Proquest’s Vogue Archive and their project, Robots Reading Vogue, shows what would be possible with access to these collections, combining an Ngram viewer, topic modeling, a frequency analysis for advertising, and a visual analysis of magazine covers. Those visualizations link to articles in Vogue, providing an alternative to search as a gateway into the archive—assuming that you are affiliated with an institution that subscribes to the Proquest database.49

In the absence of such interfaces, however, the closed and sometimes shrouded character of commercial products has encouraged historians and other humanities scholars to be simply consumers of digital content, to accept the search interfaces provided by vendors as the means by which to conduct research in digitized sources rather than looking to tools like topic modeling software. Studies of research practices have shown that many humanities scholars rely on search, beginning with Google search to identify sources, and proceeding to full text keyword searching to research within digitized collections (Rutner and Schonfeld; Kemman, Kleppe, and Scagliola; Chassanoff). So naturalized has search become that few recognize that it is not, as Ted Underwood points out, “a finding aid analogous to a card catalog . . . [but] a name for a large family of algorithms that humanists have been using for several decades to test hypotheses and sort documents by relevance to their hypothesis” (65). It can be a powerful method, disrupting the hierarchies and categories of information established in the past. Full-text search examines every publication in a database, not simply those that have become canonical, and every word in a publication, working from the bottom up rather than from the top down—for example, from the journal to the issue to the article, as a researcher in a library would. In the case of newspapers, full-text search checks many words that researchers likely would not have read when browsing, where their focus is on particular sections and headlines, not the advertisements, schedules, and notices that make up the bulk of most papers. In the case of archival collections, search can remove information from the context of the institution that structured the collection; it can, in Tim Hitchcock’s words, de-center institutions in favor of individuals (“Digital Searching”). But search is also a limited method. Search struggles to deal with what lies outside a set of results. In returning only the terms one enters, a search filters out any alternative hypotheses. For historians, this poses particular challenges, as the language and ways of organizing knowledge in the past often differ significantly from contemporary terms and patterns of thought. If we use the wrong search terms, we literally misread our sources. And working with interfaces that tell us how many results we found without reference to how many results were possible, it is not always clear just how significant those results might be. In the case of most newspaper databases, search also returns individual stories, removed from their place on the page and in newspaper as a whole. In other words, search radically decontextualizes the results it produces.50 Text mining and topic modeling can offer a different perspective that addresses those limits and complements search and other approaches to historical research.

Recognition of the value of topic modeling as a complement to search is one of several factors that have led to more use of text analysis by historians. Digitized historical sources open to text mining and topic modeling are also beginning to become more available. In the U.S. context, a key development is the Library of Congress’ ongoing Chronicling America project.51 Now approaching 9.5 million pages, this collection of newspapers from across the nation is not just freely available, but is also equipped with an API, allowing researchers to extract large datasets. However, although Chronicling America contains a rich selection of newspapers, it does not include the major papers in Proquest’s databases. Nor does the collection cover the same date range as Proquest Historical Newspapers, being limited to the years before 1923 that are out of copyright. And the more limited resources of this project means that the OCR-generated text is uncorrected and consequently marred by numerous errors.52 OCR accuracy is greater in commercial products, but just how much more accurate is impossible to know, as the companies do not release that information, notwithstanding how important it is to assessing the significance of search results.

A small number of historians have already text mined Chronicling America. As part of their Mapping Texts project, Andrew Torget and Jon Christensen used word counts, named entity extraction, and topic modeling to examine language patterns in Texas newspapers.53 In a more focused study, Cameron Blevins used named-entity extraction with two of those Texas newspapers to explore what they revealed about the imagined geography of the United States in a period understood as one of integration and incorporation. What he found and reported in the first digital history article to appear in the Journal of American History (“Space, Nation, and the Triumph of Region”54) was that, at odds with the prevailing view, the newspaper was focused on region, not nation.55 In An Epidemiology of Information: Datamining the 1918 Flu Pandemic,56 E. Thomas Ewing and others from Virginia Tech employed topic modeling and segmentation and tone classification on a set of twenty newspapers to explore the transmission of disease-related information (Ewing et al.). Chronicling America is also serving as a basis for research projects in literary studies. Ryan Cordell’s Viral Texts project traces networks reprinting texts in nineteenth-century newspapers.57 Elizabeth Lorang and others are developing an image classifier to identify poetry in newspapers, in order to explore the magnitude of poetic content in the press (Lorange et al.). And Lauren Klein and her collaborators at Georgia Tech are using abolitionist newspapers as the case study in their project to build TOME, a tool to visualize topic-modeling results (Klein).

Beyond work with newspapers, other noteworthy historical text analysis projects include two explorations of the Digital National Security Archive that are grappling with an even more fragmentary source and bringing digital methods to the field of diplomatic history, one of the discipline’s most methodologically conservative fields. Micki Kaufman’s multi-award-winning “‘Everything on Paper Will Be Used Against Me’: Quantifying Kissinger,” employs a variety of text analysis methods to explore Henry Kissinger’s memoranda and teleconference transcripts.58 The Declassification Engine project, led by Matthew Connelly and David Madigan, looks at the full range of documents in the archive to explore official secrecy through analysis of redacted text and declassified documents.59 Michelle Moravec’s work using text analysis tools in women’s history is important not only for bringing digital history to another field where it has been slow to find a place, women’s history, but also as an example of the kind of work an individual scholar can do without a formal technical background or extensive institutional or grant support. Among other projects, Moravec has used corpus linguistics to explore gender in the History of Women’s Suffrage —Elizabeth Cady Stanton, Susan B. Anthony, Matilda Joslyn Gage, and Ida Husted Harper’s multivolume record of the suffrage movement (“‘Under this name she is fitly described’”). Moravec’s current project, The Politics of Women’s Culture, also features text analyses of Women’s Liberation Movement periodicals to explore networks within the movement.60

Building on this work, responding to the emerging opportunities, and working to obtain more access and additional digitized sources will require historians to confront their reliance on search as a research method and its limits as an approach to understanding digitized sources. Scholars that are simply searching collections with an API like the Proceedings of the Old Bailey or Chronicling America as part of their research—and many are, judging by the character of publications citing the Proceedings of the Old Bailey and Chronicling America—are missing an opportunity to put those search results into context.61 The need to contextualize searches applies even to the use of digitized collections to recover the history of ordinary individuals, an important dimension of what the era of “big data” offers historians (Hitchcock, “Big Data, Small Data”). Recent text analysis projects, together with these imperatives, make it likely that the limited use of text analysis tools that currently distinguishes digital historians will become less marked in the future.

Signs of a trend toward text analysis are a reminder that disciplinary differences in digital humanities are not fixed or immutable. They grow from the interests and questions that characterize a discipline at a given time, and which attract awards and recognition, and from the sources used to explore those topics, and which have been digitized and made machine-readable and available to researchers. All those elements are subject to change. Notwithstanding that fluidity, at any given time disciplinary differences are present. Discussions of digital humanities that do not recognize the different sources, approaches, and questions of humanities disciplines limit the audience they will engage and can even alienate humanities scholars who do not see their field represented. In so doing, they create a barrier to interest and to entry to the field. It does not matter how big the tent is if the entrance is narrow. To engage more scholars in digital humanities requires approaching them in disciplinary terms and showing how the values and practices of digital humanities can help them pursue their disciplinary projects. Discussions with historians, for example, need to begin with digital mapping and to present text analysis in relation to the historical sources accessible to computational tools, and to the other methods, such as search, that scholars have been using to explore those sources. Once scholars can see the value of digital humanities to their discipline, once they have entered not a big tent, but a room devoted to their field within a house shared with other disciplines, then they are in a position to engage with digital humanities as an interdisciplinary field. As Tom Scheinfeldt argues, “Understanding what makes us distinctive will help us better see what in our practices may be of use to our colleagues in other disciplines and to see more clearly what they have to offer us” (“The Dividends of Difference”). This chapter represents an effort to understand the distinctive features of digital history, arguing that historians are different in the way they use the Web and in the computational tools they currently favor. To test and sharpen that analysis requires other scholars to take up the task of understanding what makes their disciplines distinctive within the digital humanities.

Notes

Bibliography