Gaming

This weeks readings focused on historical gaming–that is the use of games to aid in historical understanding. One might ask in the vein of Carl Smith, can you do serious history in a video game? This would probably be the wrong question. The goal of the historians and programmers producing the games is not to produce new historical insights that interact with the historiography, but as an educational tool. As the creators of Pox in the City state, their goal is to “support a key facet of history pedagogy, teaching student to make inferences about the past based on historical evidence.” I wonder, however, how successful this is. The question being what are the students taking away from the game. How can we know if students are indeed “making inferences about the past?”

Without actually doing research in what students take away from historical games, the best we can do is place the issue within pedagogical knowledge we have of how students learn about history. Games present the possibility of posing questions to the player and forcing the player to use historical evidence to figure answer them.  This is a clear goal in the creator of Pox in the City, and it is also demonstrated in The Lost Museum, in which the player tries to evaluate who burnt down P.T. Barnum’s museum based off of evidence that demonstrates the complex relationship between Barnum and various groups in anti-bellum  New York. Such a question-based investigation allows the student to actively participate in a discovery process.

Having said this, I still wonder how players are approaching and understanding the material in the game. As Sam Wineburg has demonstrated professional historians approach the past and especially historical texts much differently than our students do.[1] Why should we expect that students are getting out of the video game what we want them to? The game is inherently indirect in its presentation of material. It is completely contingent on the decisions the player makes in the game itself.

 

[1] Sam Wineburg, Historical Thinking and Other Unnatural Act: Charting the Future of Teaching the Past (Philadelphia: Temple University Press, 2001).

Networking the topics in an ambassador’s letters (part 1)

For my paper, I am reading the letters of a French ambassador, Jacques de Germigny,  to the Ottoman Empire between 1579 and 1584 and listing all of the proper nouns that he refers to as a means of finding the topics that he was concerned with. Currently, I am reading the letters and placing the proper nouns (for example, Espagne, Aly Aga, etc.) into a spread sheet. I am trying to figure out the best way to format the spread sheet to make the information easier for me to put into RAW. I often have multiple topics per letter, so I figure that the spread sheet will become very long. The paleography of the letters will probably pose some problems, so I may have to not include some letters that I am not fully able to make out or whose encoding are not deciphered. So far I have not had that issue, but I am also not too deep into the manuscript yet, and I know some of the more difficult letters are to come.

Crowdsourced History

How productive is the public in processes of historical production? All of our readings in digital history for this week dealt with this question in one way or another. In terms of crowd-production of historical discussions, I was surprised to see how few factual inaccuracies Roy Rosenzweig found in Wikipedia relative to other more authoritative encyclopedias. I find myself frequently impressed by Wikipedia, actually. This is primarily because I keep the bar fairly low for the website, however. Moreover, I am more impressed with the variety of articles on the site. For instance, Wikipedia has articles on various French ambassadors to the Ottoman Empire that one will never find in the Encyclopedia Britannica. Nevertheless, one must not forget the gender bias in Wikipedia that Leslie Madson-Brooks discusses, or the problems that academics have in contributing to Wikipedia exemplified by Timothy Messer-Kruse’s experience (reading the comments should truly frighten anyone who considers the chasm between what historians do and much of the public’s perception of what historians do). These issues aside, I think both Roy and Leslie make very valid points that we ought to learn from the Wiki experience and use the Wiki platform to our own ends: producing wikis for primary sources as well as open-source text books that professors can actively manipulate for the uses of their classes.

Such Wikis, especially those of the primary source variety that Madsen-Brooks discusses could utilize the the “amateur volunteers” that Trevor Owens discusses as well as professional volunteers to fill the Wikis. The long-term possibilities of such a project would allow the wiki to be filled over time as the few volunteers who actually care can participate at their leisure. This would avoid some of the time limitations that the Hurricane Digital Memory Bank ran into. Although such a primary source wiki might open up such primary sources to the sort of mal-use that are pervasive in the “history_pic” issues that so upset Rebecca Onion, I find the positives much outweigh the negatives. Moreover, if one wants to discuss the issues in how the public interact with history, I think that un-contextualized or miss-contextualized pictures are the least of our worries. I think there is much more pertinent problems related to Wikipedia (which actually informs much of the public on historical issues) than decontextualized photographs.

My Omeka Exhibit

Beginning to create an online exhibit is rather simple with Omeka. The most time consuming aspect of creating a basic exhibit is through inputting the meta data, which is extensive. However, it ought to be since one is producing the exhibit for others to use the information, and presumably the user of the provided data would like to know from where it came. Some of the meta date in the Dublincore system makes you think about your document in terms of what information should be provided. For example, what information should should one enter to the “identifier” area.  It states that the identifier should be “an unambiguous reference to the resource within a given context,” but what exactly does that mean? This is really the only data set I could not figure out how to fill.

It does not take much playing to learn how to organize various pages and exhibits. One simply builds from the bottom up–adding items, then combining them into collections and exhibits, and adding simple pages to explain what is actually going on in the page. In creating my exhibit, I focused on the complexity of Franco-Ottoman diplomacy. By this, I was trying to demonstrate the multidimensional aspects of French relations with the Ottoman Empire. To do this, I chose documents that had in their root the various issues that surrounded the relationship in the early 1580s–the invitation to Henri II to attend the circumcision of the Sultan’s son, the negotiations of the capitulations, issues of the Persian and Ottoman conflict that always interested the French, as well as letters directly from each monarch to the other. Since this was a fairly straightforward mission, I chose the “Thanks, Roy” theme since it looked the cleanest and provided very little clutter between the viewer and the documents in the exhibit itself.

My goal was to introduce the Franco-Ottoman diplomacy with the documents themselves, so I provided a brief introduction into the complexity the diplomacy and the multiple activities and concerns that went on at the same time. This is followed directly by a page that provides the sources themselves. The one problem I had on this page was being able to change the order in which the documents appeared. I never figured out what I did wrong, but they are not presented in any logical order unfortunately. I would have preferred to order them chronologically, but it was to no avail. Either way, I was able to do much in a short time, I think, because of the general usability of Omeka. My exhibit, admittedly, is fairly boring. I think this is partly due to the fact that I have not generally thought about how to present such historical information  in a manner that privileges the document rather than the analysis of the document. Nevertheless, it is a rather informative process to go through. It truly demonstrates how conditioned some of us have become to following prescribed formulas in how to present historical information.

Public History

The digital age has had an impact on public history in a much more fundamental fashion than it has on traditional historical scholarship. Up until our readings this week, I had not given much consideration to the impact of the internet on public history, perhaps because I have focused on traditional mediums of disseminating historical knowledge. In such traditional mediums, the internet age has not changed that much other than easing the access to our sources as well as producing new methods that one can choose to use or not.  For public history, the internet has produced new expectations among the consumers of public history projects. As Bruce Wyman indicates, not only do we expect technology to be present in exhibits, but we expect that technology to be current. Thus one is forced to think about how technology can aid in the presentation of historical material. The presentation is not as simple as adding technology wherever one can. It must actually aid in the popular consumption of history.

Museums do not just have to deal with issues of technological interaction in the brick and mortar buildings, but also the various issues of meeting the history consumer who expects to interact with institutions on the web far before they ever make it to the actual exhibit–that is if they ever choose to go. How does one provide a historical experience online that can either act as a preview for an exhibit or perhaps stand in its place altogether for what Anne Lindsay refers to as the “virtual tourist.” Both Mark Tebeau and Bruce Wyman have focused on holding fast to the story, the narrative, that people are interested in while presenting the material. I think it is important also to remember that central to the historical pursuit is answering questions. By making questions central to the presentation of material, it can help to draw people in. People go to museums not just to see or hear stories, but to have questions answered–often times vague questions, but nevertheless, questions. Moreover, by making questions central, it can also cause history consumers to think differently about the topic at hand.

How to present material does not provide the only difficulties, but what material to present and to whom. Here I am referring to the digitization of archival/museum collections and who has access to them. I agree with Tim Sherratt that the fear of misuse or losing context overlooks the amount of people who will treat the documents appropriately and within the appropriate context. Nevertheless, I think this is a rather ridiculous discussion. As long as the documents are not part of private collections, I do not fully understand why they would not be provided to the public.

Mapping the First Michigan Cavalry

Mapping the battles of the First Michigan Cavalry from the beginning of 1864 to the end of the Civil War posed various problems that I had to overcome. Some of these are particular to me. For instance, my lack of familiarity with both Civil War and more importantly the geography of Virginia made recognizing the places at which the battles took place particularly problematic. Places such as “Mallory’s Crossroads” are not easily searched through Google and often return multiple results. While this is a complication that all must overcome, a better familiarity with both topic matters would facilitate more accurate judgments.  I tried to find the coordinates of battles wherever possible in order to be more precise, but once again, this is a pursuit easier said than done. Where this was not possible, I simply leaned on GoogleMaps’s ability to recognize the place for which I was searching. There were a few battle places that I could not find, such as Duguidsville. In these cases, they were simply left off of the map.

When mapping the actual series of battles, the large quantities of battles on a single map provided an added complication. How can one make sense of a map that has forty-two icons representing battles that do not present a clear linearity? In order to make the map more readable, I collected the various battles into groups according to the campaign they were connected to or chronological connections. Thus all the battles that were part of Sheridan’s raid from Winchester share the same icon and icon color. The same goes for the Appomattox courthouse march, Sheridan’s Shenandoah Valley campaign, and the others. Moreover, since the first Michigan cavalry unit did not simply process in a single direct line, but also made numerous turns and back tracked, where the path either stretched over an extensive distance or followed a complicated route, a line is provided (in the same color as the battle icons connected to that particular campaign) to connect provide an aid in recognizing the general route taken by the cavalry unit. Although this provides a multicolored map, hopefully what is loses in eye appeal, it gains in readability.

Mapping History

The mixture of textmining and new open-source software has made possible new ways of presenting and interpreting data. As Tim Hitchcock argued in his article, digitization of sources have turned texts into data pieces that can be set against other data pieces: for our concerns, maps. Perhaps Hitchcock overemphasizes the importance of textmining in this endeavor, but the emphasis on attaching the products of textual analysis (through either close reading or distant reading) to geo-coordinates and representing it in map format can present new ways of thinking about our historical topics.

Edward Ayers and Scott Nesbit as well as Stephen Robertson have demonstrated how using these mapping methods can re-orient the way we think about historical problems. Visualizing the proximity of large numbers of union armies to large numbers of emancipation events led Ayers and Nesbitt to be able demonstrate that individual agency and that of armies (or scales, as they prefer to think of it) combined to emancipate slaves during the civil war. Union armies disrupted the regular channels that kept slaves from fleeing as well as provided a sort of indirect protection to slaves who chose to run away during such times of chaos. Nevertheless, the slaves were still emancipating themselves, even if the nearby armies provided an incentive for them to do take such a chance. Ayers and Nesbit accomplish this feat by applying the small scale of individual emancipation events as well as army movements against the scale of the national map to see a much larger patter. One can find their mapping here. Robertson uses mapping to see new patterns in Harlem, albeit on a much smaller scale. By applying every event in 1920s and 1930s Harlem recorded in his sources to the digital Harlem map, Robertson was able to see the neighborhood in a much more local fashion. This showed him that there was a greater white presence in the Harlem than most appreciate. Moreover, this presence was made most obvious in traffic accidents, which were a particularly acute area of interracial interaction, and at times conflict, since most drivers were white. Interestingly, even on opposite scalse, mapping allowed Ayers, Nesbit, and Robertson to see new patterns in their historical topics.

Mapping has also been used as a way of simulating the historical past in the form of ORBIS, and interactive scholarly work depicting transportation networks in the Roman Empire.  While ORBIS provides a means by which individuals can interact with the trade/transportation networks of the ancient world, it only provides calculations based off of generalized weather patterns (for the seasons) mixed with the time it took to traverse geographic features with technology from the time. Unfortunately, the simulation (which has its own historical challenges in terms of modeling that I will not discuss here) avoids any presentation of change over time. The features that were particularly problematic for pre-modern transportation (corsairs, ongoing conflicts, political crises, plague outbreaks etc.) that might have influenced (and were at least in the minds of the travelers in determining their route) have no place in this simulation.

In other words, the most interesting aspects of Ayers, Nesbit, and Robertson to the historian–the ability to cross reference events presented in sources with geo-spatial awareness over the duration of time on a single map–is lost in ORBIS because the concern of change over time is lost. Thus the historical patterns–what mapping makes most visible and I think where mapping provides the greatest opportunities for historians–are lost in ORBIS. Instead, it presents a static picture of transportation only influenced by nature and geography, not the ebb and flow of political, social, demographic, and biological events that drove and often interrupted that transportation.

Mapping seems to me to be most productive, then, when we can map events across time. Imagine if ORBIS could calculate travel duration and routes during different epochs. For instance, during the third-century crisis. one can imagine that travel times to some areas were expanded, some routes avoided, some places essentially cut of from others. This is certainly asking a lot, especially since much work went into the creation of ORBIS, but the patterns that the time variable would helps produce would truly transform ORBIS from a tool into a piece of scholarship, which Meeks and Grossner already contend that it is.

Visualizing Networks

This week in our Clio wired class, we are using the network analysis tools Gephi, RAW, and Palladio. After perusing Gephi, I think I will avoid the use of it since one must download it, and it seems needlessly complicated in comparison to RAW and Palladio. The construction of multiple excel sheets to produce the same visualizations that Palladio and RAW produces seems to me to be excessive.  Of course, I understand that the extra complication means that Gelphi is probably more manipulable. At the moment, the extra control over Gelphi is not necessary for my current interests. In comparing Palladio and RAW, Palladio seems much more useful for questions that are more interested in the geographic representation of networks since it has only one visualization option outside of the mapping feature: the “graph” feature that can be seen below

.Screenshot (4)

As one can see the graph feature clusters the data sets with those groups they are connected to. It is easy to see that most of the regiments were not involved in the same battles, but there was some crossover, which connect the regiments together in the graph: 136th New York, 44th New York, and 29th New York all participated in the battle of Chancellorsville, for instance. The clustering aspect of the graph, however, makes it tough to actually read the the names of the battles.

RAW Provides more possibilities for visualizing the data. In the circular dendrogram the overlapping regiments participating in various battles are more clearly shown. However, one must sacrifice the ability to view the differences in the quantity of battles that each regiment participated in.

Dallas136th New York InfantryBentonville136th New York InfantryStone Mountain136th New York InfantryAtlanta136th New York InfantryPeach tree Creek136th New York InfantryKenesaw Mountain136th New York InfantryTurner’s Ferry136th New York InfantryCassville136th New York InfantryResaca136th New York InfantryChattanooga136th New York InfantryWauhatchie136th New York InfantryGettysburg136th New York Infantry1st Michigan Cavalry29th New York Infantry44th New York InfantryChancellorsville136th New York Infantry29th New York Infantry44th New York InfantryAverasboro136th New York InfantryHagerstown1st Michigan CavalryOld Church1st Michigan CavalryTrevilian Station1st Michigan Cavalry4th New York CavalryWinchester1st Michigan CavalryFront Royal1st Michigan Cavalry4th New York CavalryShepherdstown1st Michigan Cavalry44th New York InfantrySmithfield1st Michigan Cavalry4th New York CavalryOpequon1st Michigan Cavalry4th New York CavalryCedar Creek1st Michigan CavalryPicket1st Michigan CavalryHawes’s Shop1st Michigan CavalryFive Forks1st Michigan CavalryWillow Springs1st Michigan CavalryBrentsville1st Michigan Cavalry4th New York CavalryFort Scott1st Michigan CavalryMonterey1st Michigan CavalryCold Harbor1st Michigan Cavalry44th New York InfantryFalling Waters1st Michigan CavalryRapidan1st Michigan CavalryRobertson’s River1st Michigan CavalryBrandy Station1st Michigan CavalryCentreville1st Michigan CavalryTodd’s Tavern1st Michigan CavalryBeaver Dam1st Michigan CavalryYellow Tavern1st Michigan CavalryMilford Station1st Michigan CavalryDinwiddle1st Michigan CavalryBull Run29th New York Infantry44th New York Infantry4th New York CavalryCross Keys29th New York Infantry4th New York CavalryGroveton29th New York InfantryFredericksburg44th New York InfantryPiney Branch Church44th New York InfantryWilderness44th New York Infantry4th New York CavalryMine Run44th New York InfantryRappahanock Station44th New York Infantry4th New York CavalryMiddleburg44th New York Infantry4th New York CavalryTotopotomoy44th New York InfantryMalvern Hill44th New York InfantryGaines Mill44th New York InfantryHanover Court House44th New York InfantryYorktown44th New York InfantryLaurel Hill44th New York InfantryBethesda Church44th New York InfantryPetersburg44th New York InfantryWeldon Railroad44th New York InfantryPoplar Springs44th New York InfantryNorth Anna44th New York InfantryBeverly Ford4th New York CavalryNew Creek Station4th New York CavalryStrasburg4th New York CavalryHarrisonburg4th New York CavalryPort Republic4th New York CavalryNew Market4th New York CavalryMiddletown4th New York CavalryLuray4th New York CavalryFairfax Courthouse4th New York CavalryGrove Church4th New York CavalryJefferson4th New York CavalryHartwood Church4th New York CavalryHope Landing4th New York CavalryKelly’s Ford4th New York CavalrySnicker’s Gap4th New York CavalryAldie4th New York CavalryUpperville4th New York CavalryJones Cross Roads4th New York CavalryCulpepper Court House4th New York CavalryRacoon Ford4th New York CavalryRapidan Station4th New York CavalryPiedmont4th New York CavalryBealton Station4th New York CavalryRobertson’s Tavern4th New York CavalryRichmond4th New York CavalryAylett’s4th New York CavalryWhite House4th New York CavalryJones’ Bridge4th New York CavalryCharles City Courthouse4th New York CavalryPrince George Court House4th New York CavalryDeep Bottom4th New York CavalryWhite Post4th New York CavalryBerryville4th New York CavalryCharlestown4th New York CavalryHalltown4th New York CavalryLeetown4th New York CavalryFisher’s Hill4th New York CavalryTom’s Brook4th New York CavalryRood’s Hill4th New York CavalryLiberty Mills4th New York Cavalry

This is just one of the various visualizations that RAW provides. Nevertheless, which program one uses is probably determined by the question that one is hoping these visualizations will contribute to. Adding coordinates to these battles as well as dates could illuminate the overlap between regiments and the battles they participated in over time and space in a way that might help show how the North dealt with logistical concerns in the civil war. This question would probably be better focused toward Palladio that is built for such a mapping project. Although, more data would better demonstrate what network analysis can actually do, these examples provide some suggestions in that direction.

Network Analysis

After this week’s readings, I find myself somewhat on the fence concerning network analysis. The concept seems incredibly interesting. Being able to visualize the networks between French ambassadors, for instance, would be very helpful in my own research. I wonder if some, such as Theibault, might be overestimating the utility of these visualizations when he argues that visualizations may at time “substitute for” text. This seems difficult if this includes “educating fellow historians about how to interpret visualizations.” How can these visualizations “substitute for” text if one must receive some education in how to interpret those visualizations? To what degree are we insulating ourselves more than we already are if we are trying to use visualization in that way.

The question seems particularly astute when one considers Johanna Drucker’s argument that we “rethink the foundation of the way data are conceived as capta by shifting its terms from cetainty to ambiguity and find graphical means of expressing interpretative complexity.” I certainly agree with her argument. We are constantly making assumptions and assertions when we, for instance, graph the population of a country–do we count transients or not in this population, etc. Nevertheless, the graphs she presents that intentionally demonstrate the ambiguity of the “interpretative complexity” look to be much more difficult to understand.  I fear that the use of these interpretative tools in a way that seeks to substitute them for text undermines one of the goals of historical research: making it clear and available to the public. While one might argue that the historical methods lend themselves already to a more turgid prose that the public would prefer not to interact with. I assert that interested non-historian readers can nevertheless still understand the essential arguments posited and evidence used in a historical journal in a way that one probably cannot  if one opens an economic journal (that is to say if he or she has not learned higher level math and econometrics).

I am not asserting that we avoid network analysis, but that we not overestimate its explanatory power.  I certainly accept that its use can provide a very productive complimentary role (and to be fair Theibault does call for its use in this way, not solely a substitutive one). Moreover, I would also agree with Scott Weingart that we ought not to include network analysis because it is a really cool looking graph.

Nevertheless, similarly to topic modeling and text mining, I imagine we can use network analysis not solely as an explanatory or evidentiary tool, but also as an exploratory tool. We can use it to help generate questions, find new connections that need to be investigated. One can see these possibilities through only a cursory glance at the case studies page of “Mapping the Republic of Letters.” But in doing so we must take care to remember that some of the connections that some central figures and a network map may in fact be red herrings and not actually hold the central position in a letter network that the map shows (see Weingart). As I get into more hands-on interaction with network analysis and map interpretation, my position may change, and I may see how these maps can easily substitute themselves for text. At the moment I am skeptical. This said, I do see a productive role for network mapping in historical research–especially as a means of illuminating connections that deserve fuller investigation.

nGram and Text Mining

Using the nGram viewer can provide interesting trends that one might otherwise not expect. Searching the terms “France,” “turcs,” “Ottoman,” “Angleterre” (England), “Espagne” (Spain), “Tyran” (tyrant), and “Allemagne” (Germany) in the Google nGram viewer, searching only French texts between 1500 and 1700.

The nGram produced very interesting results. Between 1550 and 1556, “Turcs” was referenced more than the other countries including France. One would certainly not expect such an increase. Even more interesting, however, is the general increase in the reference to all of the countries between 1567 and 1584. Such an increase is striking because it is in the middle of a particularly caustic domestic crisis in the French Wars of Religion.  One certainly expects an increasing in references to “France” during the domestic crisis, especially since it inspired a movement for wholesale reform within the kingdom during the period. [1] The increase in references to external countries, however, is rather unexpected. Another striking feature of the graph demonstrates that in general the waxing and waning of numerical references to the search terms is that the countries as well as “turcs” tend to flow along with one another–increasing and decreasing at the same periods. One notable difference is that references to “turcs” increases between 1670 and 1675 while references to all other countries are decreasing during that period. Nevertheless, there seems to be an overall interest in foreign countries during the same periods rather than growing and decreasing interest in particular areas at different periods.

Interestingly enough, searching the New York Times nGram viewer, “chronicle,” shows a similar parallel waxing and waning references to foreign states in a much different period in the United States. I searched “Saudi Arabia,” “France,” “England,” “Germany,” and “Turkey.”

Except for the 1860s, when there is an increase in France and England, but not the others. By the 1870s, all the countries except Saudi Arabia begin moving parallel with one another.

Using the text mining tools in Voyant was not quite as productive as I wished. Finding full text searchable texts proved quite difficult (those that lacked a PDF layered in front), especially since the Gallica website had various errors loading texts this week. Thus, I used the seven New York Gettysburg texts. Not being familiar with the documents makes it more difficult to understand what voyant returns. The most interesting features to me are the word trends feature and the keywords in context. Quickly being able to compare the differences in the frequency of a charged word such as “honor” between the different texts can be very illuminating.


As shown above, documents three and four are much more interested in honor than documents one or seven. This might demonstrate differences in the overall concern of the different documents, but if we are interested in how honor was understood in these documents a quick look at the “keywords in context” feature will provide a brief overview of how the documents used honor.

Certainly these are just snippets, but they can provide at least a beginning process of where to search.

[1] Mark Greengrass, Governing Passions: Peace and Reform in the French Kingdom, 1576-1584 (New York: Oxford University Press, 2007)