International Workshop on Historical GIS

about Shanghai

Digital Gazetteer Development

Ruth Mostern
University of California, Berkeley, USA

China has the world's oldest continuous tradition when it comes to structured representation of geographical places. Long before the political unification of the empire in the third century BC, the Yugong ( ^ ), named and described nine regions of the realm. It included short essays about each region that defined its boundaries and listed the commodities that it produced. Beginning with the Han Dynasty (206 BCE-220 CE), each era's official history (zhengshi v ) included a geography monograph (dilizhi a z ) that listed every administrative unit in the empire using a hierarchical modelall of the prefectural (zhou { ) level jurisdictions were listed under the entry for each provincial (dao D , lu , sheng ) jurisdiction, and all of the county (xian ) level jurisdictions were listed under each prefecture. For each unit, the monograph listed any changes during the dynasty: its founding, abolition, or change in governing jurisdiction, name, or administrative status. In addition to the geography monographs, a variety of other texts about places emerged by the Southern Song (12-13th centuries CE): military geographies that listed passes, roads, and distances near the frontier; taxation rolls organized by jurisdiction; local encyclopedias that included information and georefrences for every kind of place from tourist attractions to irrigation ditches, and encyclopedias of sacred mountains that associated each peak and spring with its governing deity.

I describe this tradition in order to make the point that text-based records about places were valuable to a variety of constituencies (and not only in China - genres such as these developed in Europe and elsewhere as well). It is not that China did not have maps, for there was a flourishing graphical map tradition that developed alongside the text-based tradition of spatial description. Nevertheless, geographical encyclopedias (fangzhi ) served purposes that graphical maps did not. They were able to depict geographical change over time, indicate relationships among places, include all levels of resolution from the entire known world to a single fish pond in one unified text, and incorporate a wide range of attribute information, including qualitative material and excerpts from other texts (essays about places, biographies of people associated with places, etc.).

The development of GIS technology as a means for generating graphical maps has removed some of the unique advantages of text descriptions over maps. Maps, after all, have always been more intuitively capable of illustrating the spatial relationships among features. Now, a GIS can link any place on a map with extensive attribute data, permit visualization and analysis of places with vastly varying scale, and, with some difficulty, can show change over time.

Nevertheless, digital gazetteers (diming chaxun xitong a W d t ) remain crucial to digital historical geography projects. The essential attributes of a gazetteer entry are a geographical name, a spatial footprint, and a feature type. It may also include information about hierarchical relationships among places, the dates that a place existed, or other features. Although a gazetteer is not the same as a GIS (relying less on deriving precise geospatial footprints for places, but including more different kinds of information about a place, it can be linked to a graphical implementation, allowing the results of user queries to be displayed on maps, and allowing map-based searches of records. Gazetteers can do many things, and linked gazetteer services can do even more. A gazetteer on its own is a powerful reference tool, as bureaucrats, armies and scholars discovered over 2,000 years ago. Users can find information about individual places, like what alternate names a place had in different languages. They can mount complex queries, like what jurisdiction a given river passed through in a particular year.

For a networked cultural geography project like the Electronic Cultural Atlas Initiative (ECAI), the ability for scholars to reference a gazetteer means that art historians, literary scholars, and others can use an existing authority file for registering places, rather than having to become geographers themselves and produce their own, possibly inconsistent, georeferences. If various gazetteers are linked together, users can compare different, perhaps contested or contradictory, records about places. If gazetteers are linked to other sources about places, users can find historical maps that reference places in the gazetteer, library catalogue records for books about places, or biographical dictionaries that list people associated with places.

Numerous gazetteers, gazetteer content standards, and gazetteer service implementations have already been developed. However, few existing gazetteers include historical places or confront the issues of documentation and temporal change that they present. Likewise, they do not aspire to embrace all of the global variations in conceptualizations of space and territory, or the range of culturally specific feature types. Few existing gazetteers link gazetteer records with texts that mention places, or with historical maps, websites and other networked resources. Therefore, ECAI, the Academia Sinica Computing Center (ASCC), and the Alexandria Digital Library (ADL) are collaborating on a project to create digital gazetteer standards and solutions that meet the needs of global communities of historians and other scholars who study culture.

The first goal of this collaboration is to develop a content standard and data structure for gazetteer entries and feature type thesauri that solve as many of the problems of multilingual, historical, global cultural data as possible. Our long-term goal is to allow users to employ gazetteers to move seamlessly between maps and texts, since place names in most sources are not linked to spatial coordinates. Given a mature gazetteer service, users who select a spot on a map and time in history will be able to find passages in texts, entries in bibliographies, visual images, and other registered distributed resources that relate to that place and time. Users will also be able to select a name that they find in a text, a historical map, or another resource, and see it visualized on a map. In this way, gazetteers will become an important element of interoperability among historical GIS projects and other digital historical projects.

The International Historical GIS Workshop will be immediately preceded by a meeting of the ECAI/ASCC/ADL gazetteer group. A number of important questions remain to be solved: what kind of coding system to use for complex entries, how to deal with fuzzy or incomplete information about places, what a place actually is. My paper will expand on the issues raised in this abstract and will also report on the activities of the gazetteer group.

Related URLs:
ADL gazetteer project:

ASCC gazetteer project:


International Workshop on Historical GIS Fudan University, Shanghai, August 23-25, 2001