Representations of Geographic Space: Toward a Conceptual Synthesis

Abstract The need for a conceptually high-level, unifying representational scheme for geographic phenomena was recognized long ago. A number of attempts to address this issue have been made in the past, and each has usually been centered around a specific theoretical point of view (Bartels 1982). This topic has reappeared recently within the context of geographical information systems (GIS). Recognizing that the representational scheme employed in large part determines the efficiency and ease of use within a given application context, there has been much activity in developing better methods for representing geographic data in digital form. Nevertheless, progress has been slow, with such activity usually focused on narrowly viewed implementational issues with no clear answers or overall insights toward solving the overall problem. This vacuum suggests the need for a more unified approach to both research in methods of representing geographic data and practical geographic database design based on a common and unified framework. Representational theories for spatial and non-spatial phenomena developed within other fields, particularly cognitive and perceptual psychology, computer vision and database management systems, are examined within the context of the geographic literature. Drawing on and combining these concepts with a top-down approach, I suggest a set of unifying principles and an overall frame-work for representing geographical phenomena based on these principles. This effort includes an enumeration of basic types of spatial relationships and their characteristics. Although the framework as presented is very general, it seems to represent a complete blending of previous "geometric" and "perceptual" approaches, demonstrating that these approaches are not only compatible but complementary. The theories developed in other disciplines provide insight into the functional relationship between the image-based and object-based views that have existed implicitly in Geography for many years.