Connecting the Dots: Rethinking Cartographies of Connectivity
Maps are increasingly ubiquitous in the digital world. From using Yelp to locate a good pizza restaurant to deploying geospatial databases to model the historical evolution of neighborhoods, the combination of the internet, GIS software, and smart phones have made cartography accessible to more people than at any other time in human history. These innovations offer exciting new possibilities for imagining, modeling, and depicting social, economic, and political networks in geographic space. Despite the vast potential of the digital environment, the large majority of digital maps deploy visual and conceptual conventions which were developed for static displays. Most digital maps use clearly defined shapes to visualize political boundaries and cultural areas, although the borders and definitions of these regions can be highly contested and ephemeral. Such visualizations can result in beautiful maps (figure 1), but they mask and obscure the historical realities of power and governance.
Figure 1: Traditional use of shapes and colors to delineate political control. ©2014 The Institute for the Study of the Ancient World (ISAW); used by permission.
The modern cartographic approach to mapping power suggests uncontested and centralized control, which was not true of all ancient states and empires. Many large-scale ancient polities were defined by relationships between communities, not by static borders. For instance, Hellenistic kingdoms were built upon political, economic, and social connections between a number of communities and an individual monarch. Although these monarchs exercised significant military power, cities could shift their political allegiance, trade with supposed enemies, and exercise a degree of autonomy over their foreign and domestic political affairs. Instead of an area under uniform power and control, such kingdoms comprised multifaceted and variegated networks, each with its unique power dynamics.
““Most digital maps … mask and obscure the historical realities of power and governance.””
Recognizing different types of power relations requires new approaches to mapping. By emphasizing connections between places instead of drawing solid shapes and distinct borders, new digital techniques can be used to construct maps that demarcate economic, social, and historical networks along lines of known or probable travel and communication. Such graphic means provide a more nuanced depiction of early modern societies and offer new insights into historical phenomena (figure 2). This is not an entirely new concept. Charles Joseph Minard (1781-1870) pioneered this approach, most famously with his map depicting Napoleon’s invasion of Russia and retreat from Moscow (figure 3). More recently, Monica Smith performed some initial mapping of ancient geospatial networks, and the ORBIS project at Stanford University used this approach to model travel time between select communities in the Roman Empire (figure 4).[1] Apart from these examples, perhaps due to technical and data limitations, this approach has not been widely disseminated or adopted.
Figure 2: A networked view of the Ptolemaic, Attalid, and Seleucid kingdoms at their fullest extent during the Hellenistic Period.
With the advent of free and powerful desktop GIS software and the increasing availability of geospatial data, it is now possible to quickly and efficiently create such maps with minimal technical expertise. For instance, it is possible to study the places and regions that were important in epistolary networks by connecting place names mentioned by a letter writer to the writer’s location. One can also study the focus of imperial administration by the number of mentions of a place in its archive, or the economic networks of an ancient society by the relationships between minting authorities found in coin hoards. Whatever the subject area, the basic methodology for creating these maps is the same.
The starting point for this approach is the creation of a digital gazetteer--a series of tables contained in a spreadsheet or database--that lists place names, and another set of tables that describes the connections between those places. These relationships can be political, social, or even literary, depending on what is being studied. After the construction of such lists, the place names need to be associated with geospatial information; tools like Recogito can perform this matching process almost automatically.
Figure 3. Minard’s map of Napoleon’s invasion of Russia. Minard, Charles-Joseph (1781-1870), “Tableaux graphiques et cartes figuratives,” Bibliothèque numérique patrimoniale des ponts et chaussées, accessed August 29, 2019, https://patrimoine.enpc.fr/document/ENPC01_Fol_10975.
The next step is to obtain data on the possible physical routes between places, such as roads and navigable rivers. Due to the much lauded “spatial turn” in humanities, there is an abundance of such data freely available. The Ancient World Mapping Center offers a wealth of information on the roads and waterways of the ancient world; other sources, such as the China Historical GIS offer a variety of geospatial data, and volunteer projects like OpenStreetMap provide data on modern roads and rivers. If no suitable information can be found, programs like QGIS enable researchers to create their own routes.
““By eschewing the shapes and borders of traditional cartography, this mapping technique makes a conceptual shift””
After locating place names and identifying routes, a project then needs to assign a “cost” to each potential connection. Many researchers are primarily interested in the shortest routes between places, and so for them the cost is just the total length of a route; for example, a 100 km section of road will have a cost of 100. Even for thousands of routes, this can be automatically computed by a database in a matter of seconds. Cost can also be expressed as travel time or the actual expense of travel; in antiquity water travel was almost always cheaper and faster than overland routes, and so different types of cost (expense and travel time) can be associated with individual paths. Outside of the physical demands of travel, cost can also be the political or social difficulty of a particular route; perhaps a road traversed a hostile area, a sacred enclosure, or some other contested space. For these cases, cost can assume an abstract numeric scale chosen by the researcher as a reflection of the political and social implications of using these routes. After determining costs, the final step is to run a routing algorithm to find and map the least cost paths between different named locations. There are many good tutorials for this task, and one excellent example, “A Beginner’s Guide to pgRouting,” is only slightly more than 600 words in length.
Figure 4. Travel time flow map generated by the ORBIS project.
As an example of this process, I am mapping ancient Greek garrisons. My original research classified the nature of the relationships between the garrisons and physical features in the landscape, such as mountains and roads, which were entered into a database. I used data from the Pleiades project to locate nearby communities, and data from the Digital Atlas of Roman and Medieval Civilizations and ORBIS to represent the road and water network connecting these cities. I generated a cost for each route in the network by using a built-in database function to quantify the length of segments on the route. The least cost paths between each garrison and the capital city of the empire that controlled it were automatically calculated by the pgRouting software library, which allowed me to abstract and map power networks. Each path on the map was then sized according to the number of connections that traversed it, so roads or rivers where many least-cost paths overlap were displayed as thicker lines (figure 5). The result was striking: the map revealed that irrespective of city size or exact political status, communities were only permitted to keep their own garrisons along routes that were marginally connected to the main routes of travel and communication, while the major network thoroughfares were controlled by imperial forces.
Figure 5. Network map of the Attalid and Seleucid kingdoms with imperial and local garrisons indicated.
This mapping technique clearly demonstrates the importance of communication and connectivity to Hellenistic powers in a manner that is impossible with a traditional cartographic approach. It must be stressed that while powerful, this approach creates an even more abstract visualization of power than traditional maps. By eschewing the shapes and borders of traditional cartography, this mapping technique makes a conceptual shift, which may be controversial in some fields: rather than viewing state or monarchical authorities holding clearly demarcated territory, it envisions power as a network of different relationships. Outside of their possible epistemological implications, these maps are limited by the quality and quantity of available data. Routes between individual places and the monetary/political costs of those routes may never be precisely known, which may limit the effectiveness of this approach. Nevertheless, such network diagrams, when used in conjunction with other visualizations and research methodologies, offer a glimpse into one new way to use digital tools to reconceptualize and reimagine networks in geographic space.
NOTE
[1] Monica L. Smith, “Networks, Territories, and the Cartography of Ancient States,” Annals of the Association of American Geographers 95, no. 4 (2005): 832–49; Monica L. Smith, “Territories, Corridors, and Networks: A Biological Model for the Premodern State,” Complexity 12, no. 4 (March 1, 2007): 28–35.
