Fluid Borders: Rethinking Historical Geography and Fixed Map Boundaries in Contested Regions

Abstract

This article introduces a quantitative methodology for analyzing contested map borders. The article applies the new analytical technique to a data set of thirty maps showing Bulgaria in ca. 800 CE, a disputed state and period in medieval historiography with relevance to modern national politics and territorial claims. Based on the data set, we generate a series of new maps that make explicit the fluid medieval boundaries and general disagreement among geographers and historiographers. Our analysis begins with a simple point-in-polygon procedure to create a majority map that depicts the points included within the borders of the Bulgarian polity in sixteen or more of the maps (>50 percent). The majority map is then combined with percentage maps, confidence interval map boundaries, and cluster maps. The confidence interval maps are created via a spatial bootstrapping procedure and measure the uncertainty in the majority map. The cluster maps are developed via a radial basis function and provide insight into the potential affectivity based on the cartographers' countries of origin. The final map reflects the general modern consensus of the borders of the Bulgarian polity around 800 CE. Besides its quantitative contribution to medieval and modern cartographic, historiographical, and political debates, this article has developed a widely applicable methodology for synthesizing map borders and territories in cases of cartographic disagreement.

本文引介分析争夺的地图疆界的量化研究方法。本文应用新的分析技术，分析三十幅展现大约在公元八百年的保加利亚之地图，该国与该时期在中世纪的历史地理中充满争议，并关乎当代的国族政治及领土宣称。我们根据该数据集，生产一系列明确标示出中世纪的流动边界、以及地理学者和历史地理学者们一般争论之处的新地图。我们的分析，起始于简单的多边形内点识别过程，以创造描绘在十六幅或十六幅以上（大于百分之五十）的地图中，保加利亚政体疆界内的涵盖点之主流地图。该主流地图接着结合百分比地图、置信区间地图边界，以及聚集地图。置信区间地图，透过空间重复抽样（bootstrapping）过程创造之，并测量主流地图中的不确定性。聚集地图，透过径向基函数建立之，并对根据製图者的国家来源可能产生的情感作用提供洞见。最终的地图，反映出当代对保加利亚政体在公元八百年左右的国界的一致认同。除了对中世纪及现代製图、历史地理学、以及政治辩论提供定量之贡献，本文亦建立可广泛应用于融合製图争议案例中的地图国界与领域的方法论。

Este artículo introduce una metodología cuantitativa para analizar mapas de fronteras en disputa. El artículo aplica una técnica analítica nueva a un conjunto de datos de treinta mapas que muestran a la Bulgaria de alrededor del 800 EC, un estado y período disputados en la historiografía medieval con relevancia para la política nacional moderna y reclamaciones territoriales. Basados en el conjunto de datos, generamos una serie de nuevos mapas que hacen explícita la fluidez de los límites medievales y el desacuerdo general entre geógrafos e historiógrafos. Nuestro análisis empieza con un simple procedimiento de punto-en-polígono para crear un mapa de mayorías que represente los puntos incluidos dentro de los límites del sistema de gobierno búlgaro en dieciséis o más de los mapas (>50 por ciento). El mapa de mayorías es luego combinado con mapas de porcentajes, mapa de límites con intervalos de confiabilidad, y mapas de agrupamiento. Los mapas de intervalos de confianza se crearon a través de un procedimiento de remuestreo [bootstrapping] espacial y miden la incertidumbre en el mapa de mayorías. Los mapas de agrupamiento [cluster] se desarrollaron a través de una función de base radial y proveen entendimiento de la afectividad potencial basada en los países de origen de los cartógrafos. El mapa final refleja el consenso general moderno sobre los límites del sistema de gobierno búlgaro alrededor del 800 EC. Además de su contribución cuantitativa a los debates cartográficos, historiográficos y políticos medievales y modernos, este artículo ha desarrollado una metodología ampliamente aplicable para sintetizar los límites de mapas y territorios en los casos de desacuerdo cartográfico.

Key Words:

• cartographic methods
• conflicting boundaries
• map boundary analysis
• map confidence intervals
• medieval geography

关键词:

• 製图方法，冲突的边界，地图边界分析，地图置信区间，中世纪地理。

Palabras clave:

• métodos cartográficos
• límites en conflicto
• análisis de límites del mapa
• intervalos de confianza del mapa
• geografía medieval

Acknowledgments

We thank Ian Reese of Land Information New Zealand for design assistance in constructing the supplemental poster. Our gratitude also goes to students of ENSP 204 at the College of William and Mary and especially to Kyle Titlow for refining the methodology through repeated testing. Christos Nüssli from Euratlas kindly shared several maps from his personal collection. Kostadin Sokolov from the University of Mainz also retrieved and sent us digital maps from libraries and data sets, as did Stella Tkatchov. At the University of Michigan, Karl Longstreth helped us to find some earlier maps. Martin Gallivan, Gregory Hancock, Matthias Leu, John Lovette, and Liam Hamilton attended the pilot presentation of our project and gave constructive feedback. The GIS Stack Exchange members assisted with the Python code to bulk convert our bootstrapping output tables into database format and suggested improvements for the iterative portion of our spatial bootstrapping procedure.

Supplemental Material

Supplemental data for this article can be accessed on the publisher's website at http://dx.doi.org/10.1080/00330124.2015.1054017

Appendix

Additional information

Notes on contributors

Stuart E. Hamilton

STUART E. HAMILTON is an Assistant Professor in the Department of Geography and Geosciences at Salisbury University, Salisbury, MD 21801. E-mail: [email protected]. His research interests include novel applications of GIS, mangrove mapping, and GIS for development.

Alexander Angelov

ALEXANDER ANGELOV is an Assistant Professor in the Department of Religious Studies and a member of the interdisciplinary programs in Russian and Post-Soviet Studies and Medieval and Renaissance Studies at the College of William and Mary, Williamsburg, VA 23187. E-mail: [email protected]. His primary research fields are medieval Christianity, Byzantine, and Balkan history.

Vladimir Atanasov

VLADIMIR ATANASOV is an Associate Professor at the Raymond A. Mason School of Business, College of William and Mary, Williamsburg, VA 23187. E-mail: [email protected]. His research interests include statistical methods for causal inference, corporate governance, and financial market frictions.

Marco Millones

MARCO MILLONES is acting Director of the Center for Geospatial Analysis at the College of William and Mary, Williamsburg, VA 23187. E-mail: [email protected]. His research interests include human–environment geography, spatial analysis, and GIS applications to public policy in Latin America.