Mathew Wilkens’s “The Geographic Imagination of Civil War-Era American Fiction” addresses an interesting concept: how often are specific places, at the city, state, or country level, mentioned explicitly in United States in pre-, concurrent, and post-civil war era literature? This seems to make use of a similar information generation strategy as the network analysis modeling, but within the specific context of geography. Overall, the results are not hard delineations, in that this sort of study does not require or ask for generalizable rules for trends in place references, but makes a point of indicating exceptions to each trend the article discusses. The methodology does, as previously mentioned, share common traits with network analysis, but it differs in the specificity of the topic, in that it requires TEI encoding with text recognition software, and a significant amount of human correction of unavoidable computing errors, such as the decision to exclude the name “Charlotte” as it was being mistaken for the city.
The evidence, as one might expect, is presented globally in the form of maps, graphs, and charts, with explanations of the data and information gathering strategies presented in textual form. The strategies of compilation were, interestingly, an amalgamation of several of the digital methods we have discussed throughout the semester, including TEI and XML encoding, variations on topic modeling (where the topics are used, in a way, to systematically exclude certain names or colloquialisms deemed irrelevant or non-useful), and finally to be compiled in a fashion very similar to network analyses. The presence of the appendix was especially useful, as it meticulously describes the methods of corpus selection, the types of data to be recorded, the possible errors present in such methodology, and most importantly, it discusses the questions such studies might be efficient at answering and those it would not. Furthermore, throughout the entire article, Wilkens is painstakingly accurate in describing the data presented, especially to the effect of discouraging generalized rules for place publication. For instance, Wilkens makes it clear that while population shifts and growth in established cities versus new cities explain some of the phenomena, it is not “the whole story”(826). And it is this area where the raw numbers might be lacking in the way of explanation that Wilkens brings in possible cultural topics of study. Wilkens describes the ways in which literary attention often lags behind population shifts, and there are certain intangibles, such as connotative associations with a given location, such as New York City.
As a possible exemplar for the entire study, Wilkens presents the example of Chicago versus New Orleans in terms of literary reference. Here, New Orleans is the clear winner, though Chicago was growing exponentially during the time period discussed. Wilkens offers by way of explanation the conservative trend in literary movement to retell, that is, to write about people, places, and events that have been written about before. As a result, New Orleans, the significantly older and more established city of the pair, would have enjoyed the status of having been discussed at greater length for a much longer time, although Wilkens does note that the gap between the literary representations diminishes as the timeline for his study progresses towards 1870. This demonstrates the lag Wilkens discusses, and it acts to strengthen his suppositions about the trends he describes.
Finally, Wilkens concludes that the population and age of a given location are a fair indication of literary reference likelihood, but notes the important exceptions, such as the significant percentage of references to locations outside of the United States, both in reference to contextually contemporary cities like London and Paris and references to ancient or biblical locations, like Babylon. Ultimately, Wilkens is able to demonstrate that New England, which is credited by Mathiessen as being the root location for the American Renaissance, is in fact not a completely dominating epicenter, as the data demonstrates. While still significant, the New England is only representative of just over 15 percent of the place names referenced in the study, which as Wilkens points out, is “hardly overwhelming” (830). And to this end, I suggest that Wilkens methodologies thoroughly support his argument, and therefore the article and the study effectively address the question of what factors affect literary place references and how that might describe a locale for the Renaissance.