Our 'GIS and Society' research is concerned with how socially differentiated geographic information can be produced and represented within a GIS, and the implications that arise for the variety of GIS receptors. In rural South Africa and industrial West Virginia, GIS are being developed which incorporate grassroots perspectives on natural resource access and techno-environmental risk. These GIS combine conventional socio-economic, environmental, and infrastructural data with non-conventional behavioral and cognitive information. A regional political ecology conceptual framework informs the GIS production process. The research is intended to contribute towards more democratic decision-making processes while also exploring the constraints and possibilities associated with alternative GIS production and use.
In both case study localities, a diversity of geographic information has historically been produced and consumed within a particular set of power relations. Kiepersol (South Africa) is a rural area with a long history of struggle over ownership and access to land, water and biomass resources. (The South Africa component of the research is being conducted in collaboration with Tim Warner (West Virginia University) and Richard Levin (University of the Witwater Johannesburg)). The apartheid regime forcibly removed of millions of African people in the geographical project of bantustan -- or "homeland" -- creation. In our research local knowledge about apartheid forced removals, agro-ecological potential, and access to land, biomass, and hydrological resources are used to complement more traditional top-down agency-driven data.
The Kanawha Valley, near Charleston (West Virginia), is one of
the largest industrial chemical complexes in the world. In
'chemical valley', as it is known locally, the potential for
environmental catastrophe combined with numerous more chronic
health risks, are very much part of peoples' everyday lives. Risk
management and access to information are, therefore, of major
importance. Here the research focus includes an analysis of how
people gain access to geographical information and how the
representation of that information impacts the perception and
management of risk. Specifically this includes investigating the
ways in which the GIS production process transforms existing
power relations and how access to chemical hazard information
influences risk perception and management.
The West Virginia and South Africa case studies are linked conceptually through (regional) political ecology. Political ecology is a diverse collection of theoretically informed case studies. Within geography, most political ecology research has been concerned with rural areas in peripheral regions, although some high quality research has been conducted in core regions as well. The relationship between rural relations of production and natural resource ownership, access, and use is a central political ecology theme. Land degradation is an important theme too. More recently, political ecology has broadened to include research on the cultural construction of nature, environmental hazards, epistemologies of environment, and new social movements.
Regional political ecology (RPE) is a useful conceptual
framework for conducting "GIS and Society" field work.
RPE is concerned with geographic scale and competition over
environmental resources. In seeking to represent many, and often
differing, representations of reality, the linking of RPE with
GIS stimulates a number of important questions. This research
employs a regional political ecology conceptual framework for GIS
production which involves involves an analysis of how
differing social groups construct and compete for natural
resource and information access and how techno-environmental
risks are perceived, spatially distributed, and socially
regulated.
Four broad objectives guide this research. They are:
The 'success' of GIS to date may be due in part to the belief
in non-contradictory and 'objective' perceptions of reality. We
examine in this research how a GIS containing potentially
conflicting information from multiple sources is politically
embedded and articulated within existing policy-making, politics,
and institutions. In so doing we seek to contribute to the
democratization of land use planning in South Africa and risk
assessment and mitigation in West Virginia. Within the GIS, local
resource utilization and land use practices are examined as
politically contested and the product of processes operating at
varying scales of analysis. The inclusion and representation of
local knowledge in a digital form and the generation of meta-data
are significant challenges to the GIS production process.
The research objectives articulated above necessitate an innovative and experimental research methodology. Our methodology builds on experimental research already completed and involves the integration of the following methods:
In Kiepersol, the first phase of GIS database development included traditional environmental, physical and infrastructural data describing the human and physical geography of the area. The "Kiepersol GIS" was developed with community participation and incorporates non-conventional behavioral and cognitive information associated with 'multiple realities' of highly contested rural territory. Data were derived from community workshops, individual surveys, and mental mapping exercises. To date we have experimented with mental mapping as a means of capturing local knowledge from socially differentiated groups. This has been based on the field interpretation of hard copy topographic 1:50,000 scale maps. Base maps for use in conjunction with GPS capability in the field, will be generated from the GIS to support the mental map generation and field survey work. Our intent is to develop culture-sensitive approaches to data capture and digital representation which do not place as great a dependency on Euclidean or 'First-World' interpretations of space.
In the Kanawha Valley, we are exploring the definitions, perceptions, responses, and mediations of risk associated with the four social categories of capital, labor, community, and the state. An epidemiological component of the research focuses on health risks and the spatial distribution of cancer clusters. Multiple representations of information are essential to such a pursuit. Local knowledge from the community will be obtained through mental maps, oral histories, and workshops within an ethnographic methodological framework. So-called 'expert' knowledge will be obtained from existing spatial information and environmental legislation and regulations as well as primary data collection from persons associated with the (local) state and capital.
Participatory workshops will be a central method for incorporating local knowledge into the GIS production process. Two sets of workshops will be held during the course of the project in each of the areas where interviews take place. In Kiepersol, these involve large-scale commercial farms; farmworker compounds; and the within bantustans. The first set of workshops will introduce communities to the project and involve a mental mapping exercise. They will also provide a context to discuss the quantitative and qualitative research. The difficult process of conceptualizing social differentiation in rural South Africa will draw upon previous research undertaken. In bantustan workshops, participants will be divided into groups on the basis of their gender and access to agricultural land. Four groups will be established: women with land; women without land; men with land; men without land. Chiefs and their patrons will constitute a fifth group which will be interviewed separately and in a different time and place. The area has a history of bitter conflict between some residents and the institution of the chieftaincy. These five groups will construct resource and resource access mental maps of the area.
Capturing and encoding local knowledge, which is often aspatial and qualitative, represents a significant challenge to GIS production. From our work to date we are aware that much local knowledge is spatially fuzzy and does not conform easily to the spatial primitive paradigm of point, line, and polygon employed by GIS. Oral histories and narrative provide some of the most compelling and informative knowledge to come from the interview-survey process. Not least, the anticipated variety, and possibly conflicting responses, from the socially differentiated groups will provide additional complications for incorporation within the GIS. Cognitive maps, for example, deal with the spatial information about an environment and encompass internalized perceptions of knowledge and experience. Cognitive mapping functions can be divided into cognitive maps which are based on routes and involve judgments about location, direction, or distance; cognitive collages, which are like thematic overlays of multimedia from differing perspectives; and spatial mental models which represent the relations among different elements. These views of cognitive mapping appear to support Golledge's description of cognitive maps as 'internalized GIS'. Cognitive maps generated from route learning are different from those obtained from map reading. This project utilizes both approaches in the form of tracing paper overlaid over local 1:50,000 topographic maps and GIS produced shaded relief maps, and boundary-transect walks utilizing GPS. More accurate judgments for map-based mental map construction can be obtained by an iterative process in which explanation, questioning, and field walks are included in the mental map production process.
The development of GIS-interactive multimedia (IM) linkages promise to advance current GIS capabilities considerably. In order to include the variety of narratives, oral histories, anecdotal information, sound, text, photographs, sketches, maps, and video clips which are the tangible materials of local knowledge, we seek to develop icon-driven capabilities within the GIS-IM system to access the full range of traditional and local knowledge available for interpretation. The many relationships between geographical location, data, and the several media modes will be established using an authoring system. Data inquiries will be handled using 'hot link' icons from the GIS based on the hypertext concept. This model describes a set of nodes connected by undifferentiated links, where the nodes can be abstractions made up from any kind of text or graphical information elements. The nodes and the associations between them, the links, form semantic units which may express a single idea or simple data element, or a complex unit such as a map, table, or image. The links tie together the various semantic units and provide a means of navigating through the data.
The development of Hypermaps moves beyond the establishment of
links between semantic nodes to include links between spatial
location and nodes. The ability of GIS to undertake spatial
search functions would be linked to the identification of
multiple media objects found within the search parameters. Once
identified these objects can be retrieved, displayed, or used as
signposts to other sources of information contained within the
GIS or multimedia database. This logical movement through the
information base utilizes the power of GIS and the flexible
nature of multimedia to incorporate information in various media.
While these linkages will provide significant freedom to explore
the informational relationships contained in the database, one of
the main issues involved in the design of these systems revolves
around the actual organization, management, and content of these
nodes and links within the computer environment.