Page:Fdigh-01-00001.pdf/2

Gregory however, they do increase both the opportunity and the complexity. As identified above, there are complexities centered on the implications of digitized sources and on the tools that we use to interrogate and analyze them. The opportunities are enormous centering on the ability to search, visualize, and analyze historical sources be they very large or much more modest.

Thus, doing good digital history requires a number of things: firstly, historians need to critically evaluate digital sources (whether born digital or digitized) in much the same way as they critically evaluated other sources and consider these implications in their arguments. At the moment, this is more difficult than it should be because debates around the benefits and problems of digital sources have generated much heat but little light. A more nuanced understanding of these implications is required. Secondly, there is a pressing need (and opportunity) to develop and understand new techniques to deal with the errors in these sources. One possible approach is to automatically correct errors [see, for example, Evershed and Fitch (2014)], although this will inevitably introduce new errors and adds further abstraction from the original source. An alternative is to conduct studies that help us to understand the implications of the digitizing errors.

Thirdly, there is the need (and opportunity) to develop and use methods to exploit digital sources. Close reading will always be an important component of the historian's toolbox and simple keyword searches provide a way of assisting this with very large volumes of material by allowing the passages worthy of close reading to be quickly identified; however, other approaches are also required. This presents challenges. When large quantitative databases came available, statistical techniques were available to help analyze them; however, there is no obvious equivalent for large textual sources. There are, however, a wide range of promising opportunities as diverse as: corpus linguistics, distant reading, network analysis, GIS, and so on. This presents the discipline with an opportunity – historians have always been concerned with the analysis and understanding of texts, if they can develop techniques to help with this then these should be much more broadly applicable in a world that is increasingly awash with digital texts.

Fourthly, while work on sources and methodologies is important, it is a means to an end rather than an end in itself. The work that will ultimately prove the relevance and importance of digital resources and methods will not stress the digital, it will stress the applied and make a contribution to knowledge on particular topics within history that "non-digital" historians will be interested in. This is not easy as it means effectively handling the interpretive challenges faced by traditional historians, as well as the technical and interpretive challenges presented by the digital. It also means that the discipline as a whole needs to be better at, and more receptive to, work that stresses methodological developments that help us better understand digital sources. Ultimately though, the combination of the computer's ability to manipulate and summarize large volumes of material and the human brain's ability to interpret this appropriately will provide major advances in our understanding of the past. The opportunities for those who are skilled enough, adventurous enough, and imaginative enough to do this are enormous.

Acknowledgments
The research leading to these results has received funding from the European Research Council (ERC) under the European Union's Seventh Framework Programme (FP7/2007-2013) / ERC grant "Spatial Humanities: Texts, GIS, places" (agreement number 283850).