Inside Digital Humanities' How to DH section welcomes esteemed researcher Lucas McMahon.
This interview is the written interview content of our How to DH concept. In this content, written questions are answered by the researcher concerned.
- Who is Lucas McMahon, please introduce yourself, your team and your project.
I’m Lucas McMahon, a Social Sciences and Humanities Council of Canada Postdoctoral Fellow at the University of Ottawa. I did my PhD in history at Princeton University, and I work on information transmission and the spatial history of empire in late antiquity and Byzantium. I’ve been involved in a number of small DH projects, all related to GIS and mostly of my own design and which are related either to my dissertation or publications. I was involved with two Center for Digital Humanities@Princeton-funded projects. I was PI alongside Abigail Sargent for a data curation grant, in which we attempted to map the find location of published Byzantine lead seals. I was also project manager for Eric Medawar’s seed grant that digitized some basic topographical data on medieval Syria.
- Can you briefly describe your background in digital humanities/computational social sciences? What are your studies and experiences in the field? What was the main research questions of your project?
On the techincal side my GIS skills are entirely self-taught. I learned by pushing buttons and breaking things. On the methods and theories of digital spatial analysis, I learned by attempting to read one summer everything published in the last decade on GIS in the Journal of Archaeological Science, which tends to be a hub for innovative methods.
The main question that I was dealing with was how to quantify space and deal with human movement through a physical environment. A lot of history, especially global history, is focused on connections between places and people. Yet the space in between needs to be traversed somehow, and the phyisical and human geography imposes constraints. All of those travellers spent a lot of time on dusty tracks and I wanted to find a way to bring that life. My first foray into this did not involve any digital techniques but looked at the logistics of supplying military forces by sea in the medieval Mediterranean. A lot of past work has assumed that if sea transport is available then logistical concerns go away. I argued that is not the case at all: ships need to be loaded and unloaded, they need ballast, they need packaging for their cargoes, they need crews, they need places to land and an environment safe for sailing. It was about getting at the everyday experiences of the people who made those connections. That piece was published in Mediterranean Historical Review. A letter that arrived at its destination might be read in a few minutes, but may have spent weeks on the road. It’s about getting at the spaces in between. Yet we do not always hear much about this in pre-modern travel accounts, and so by performing digital analysis of space we can start asking questions about the experieces of those journeys.
- What type of data do you have for the project? What characteristics does this data have?
For my work on looking at overland connections between Rome and Ravenna in the seventh and eighth centuries, I had a little bit of pre-existing data. Key to the sort of spatial analysis that I do is assessing the difficulty of human movement through the environment. For this piece I used a few specific input factors. The most important was terrain, for which I used the Italian Istituto Nazionale di Geofisica e Vulcanologia’s Digital Elevation Model (DEM). I factored in hydrology through Stanford’s DIVA-GIS’s vector data. Finally, for some of the more cosmetic parts of the mapping, I used data from the International Hydrographic Organization and the free resources on Natural Earth. It’s all quite basic geopysical data, but I am aware that at its source it was not collected for scientific or academic purposes but by government intelligence and mapping agencies. That I’m also using it for looking at military connections and communications is at least somewhat fitting, I suppose. Much of DIVA’s data, for example, is derived from publically available data from the United States’ National Imagery and Mapping Agency, now called the National Geospatial-Intelligence Agency and part of the Department of Defense.
Other parts of the data I had to make myself. One of the most difficult was establishing rough border lines between East Roman territorial holdings and the Langobard polities. Medieval borders are rarely clear, but these were particularly problematic. I wanted to test routes both within approximate Roman territory and outside. Existing maps were however a problem, and it strikes me as more than a little convenient that the Duchy of Spoleto’s seventh-century borders match those of the seventeenth-century duchy, which I doubt very much. I raised some reservations about this in my article in Studies in Late Antiquity.
- What was the process of processing and transforming raw data into structured data? Which digital tools, software and methods were used and for what?
I used least-cost path techniques in ArcGIS Pro. I did this for two reasons: first, I wanted to see how closely the new historical route matched the least-cost path, and secondly, it was part of building data that could then be used for time-distance analysis. In the Roman Empire, and even under the Ostrogoths, travelling from Rome to the Adriatic or north of the Apennines meant taking a route that was not very direct. If one wanted to go to Milan or Gaul or Hispania overland, one took the Via Flaminia to the east to the Adriatic, and then backtracked northwest. This route was only eclipsed by the Via Francigena much later. But for a brief moment under East Roman control in the seventh and eighth centuries, a new route is attested. An itinerary written in Ravenna sometime in the seventh or eighth centuries often follows earlier Roman itineraries, but has one notable change in how one travelled from Ravenna to Rome. It has an entirely new route but lacks the Via Flaminia. The Via Flaminia was one of the main roads seized by the Langobardi in their attempt to put pressure on the city of Rome during the opening phase of their revolt, but the road was never re-opened to East Roman communications. The new route ran through Perugia and Gubbio and was an extension of the old Etruscan road that the Romans had named the Via Amerina. Intriguingly, the least-cost path bore a strong similarity to this route, preferring it to the longer but much more popular Via Flaminia. It goes to show that how once a route is established it becomes the “least-cost path” itself simply by virtue of being known and travelled. But controlling a slice of central Italy after the Langobard revolt, the East Roman government had to establish a new route to keep their capitals connected by land, and it turned out to be a pretty good one. Despite Gregory the Great complaining about the Langobardi cutting communications with Ravenna at times and how exposed that tendril of territory looks on a map, it seems to have worked the other way around in making it difficult for Langobardi on either side of the corridor to communicate. Not all of these conclusions would have been possible without the least-cost path analysis.
Secondly, I wanted to get a sense of how long it took on various forms of transport to get between Rome and Ravenna. The idea was to come to a general idea of the rhythms of communication between the two capitals along the corridor. For this I computed daily travel times and distances, and contextualized it through the report of a revolt during which a number of messengers moved between Rome and Ravenna in a short period of time. This really helped get a sense of how news about unrest moved around and how both the rebel and the government tried to use this to their advantage. One could make back-of-the-napkin calculations for this sort of thing, or use GIS. Unfortunately the GIS is fairly complicated and existing tools like ORBIS cannot answer this question because the Via Amerina doesn’t exist in it!
- What kind of outcomes did your work yield or not yield (Databases, Visualizations, Events, Publications etc.)? Were you able to answer your initial research questions or were your expectations met?
This started life as a proof-of-concept that became a dissertation chapter that became an article in Studies in Late Antiquity. It was a success because it taught me a lot, both in terms of how to do the GIS analysis but also how to refine the historical questions that I am attempting to answer.
- What were the biggest challenges you or your team faced during the project? Were these challenges related to the data or the tools used? What suggestions do you have for the process (such as aspects of the software and programs that need to be improved)?
The biggest challenge was ESRI’s proprietary algorithms. I did the work, wrote the article, submitted it for publication, went through a very positive peer review process, and then just as I was about to submit the final text for copyediting I decided to run everything one more time to check the numbers. In the meantime, several major ArcGIS Pro updates had come out since I received the original results, and none of my data worked any more. I sorted it all out in the end, but the main lesson was always to read the patch notes and check what exactly is going on with the math when ESRI updates a tool.
- What were the differences between using conventional methods and digital methods? How would this work have been done without digital methods?
It’s mixed. I could have come to a few of the main conclusions of the article without digital techniques. The Via Amerina might have been pretty close to the least-cost path, but it was also path that did not require a major military expedition against Spoleto. It’s not like the east Romans conducted a survey of Umbria and built a new road: they just started using the one that they still controlled, which one exarch happened to punch through from Rome to Ravenna once and which they then managed to retain control over. Similarly, the changing travel patterns between the eastern Mediterranean and Ravenna did not need any digital analysis to determine, and in any case is a matter of maritime mobility which requries very different analytical tools. The travel time analysis required this sort of GIS work, though, as the algorithm considers the difficulty in moving up and down slopes as well as crossing rivers.
- Does your project include open source data, database or application? Do you share your data with linked-data methods or open accessed web platforms (eg. Github)?
It does not. All the results are published in the article. I’d like to make the various polygons I drew of zones of political control available for others to use, but probably as part of a larger dataset of similar things.
- What will be a next step for your project?
I’m presently using a variation of the technique to look at how communications changed in Asia Minor in the twelfth century when the establishment of the Turks at Konya severed the main Roman military road to Cilicia and Syria. I make use of the article’s results for a chapter on commications within east Roman Italy and with the eastern Mediterranean in the book I am writing on the medieval Roman information state.
GIS movement models offer near endless iterations. The goal is both to answer a historical question while also finding some way to improve the quality of the model. What I’d like to do next is find a way to factor in movement in wetland areas. For example, the data on this model sees the area around Ravenna as flat and dry. In late antiquity and the early middle ages, it was a hydrologically fluid zone, filled with marshes. The algorithm sees this as easy to move through when in reality it is anything but.
As we come to the end of the interview, we would like to thank dear Lucas McMahon for his kind participation and wonderful answers. Stay tuned to see you in the next interviews.
Комментарии