The Naive Geography of Cyberspace Helen Couclelis Department of Geography and NCGIA University of California Santa Barbara, CA 93106 cook@geog.ucsb.edu The virtual realities produced by electronic media are most often designed, experienced, and talked about as spaces or places. Nearly always these spaces are based on those of everyday human experience: the small-scale space of objects, the larger one of buildings, the vast geographic space. These kinds of representations are dominant even though the electronic medium does not in itself presuppose a space-like (or any other specific kind) of structure, and is in no way restricted to the varieties of space that people happen to be familiar with. I propose to investigate how common spatial and geographic concepts occur in the many different manifestations of cyberspace: in interface design, in games and interactive training tools such as flight simulators, in scientific visualizations of non-spatial information, in explicit geographic reconstructions such as virtual towns and villages, among the electronic user communities forming on the Internet, in virtual-reality displays, in the cyberpunk literature. Visual representations of geographic concepts are as old as the first sketch maps and landscape drawings etched on the walls of prehistoric caves. However, cyberspace presents us with the first spatial and geographic representations that are actively functional in the sense that people interact with them in ways that are in some sense analogous to corresponding interactions with real-world entities and settings. Moreover, a large proportion of these representations have been designed or incrementally put together with maximum usability (rather than the faithful rendering of particular data bases) as the overarching criterion, emphasizing what appears natural to users over what may be scientifically correct. Thus cyberspace affords a excellent opportunity to study what aspects of the commonsense geographic world are basic enough to be maintained in these interactive representations, despite the strong technical imperatives for simplification and generalization. The purpose of this research is threefold: First, the common elements and differences across spatial and geographic representations in cyberspace will help identify the aspects of geographic concepts that appear to be the most universal and robust, as opposed to those that are contingent or imposed by the particular medium or application. Second, GIS itself generates a special kind of cyberspace that is faithful to empirical geography in some respects but not in others. It will be helpful to disentangle the intuitively obvious from the conventional or arbitrary in GIS representations against the more general background of geographic concept representations in cyberspace: this will help distinguish the representations that correspond to a strong commonsense understanding of the world from those that people may find confusing, misleading, or plain wrong. Third, many of these geographic concepts have already been formalized in software ranging from games to elaborate scientific visualizations. Examining how this was done and how well it works will facilitate appropriate formalizations for GIS. Accordingly, three groups of research questions suggest themselves: 1 Ontology and epistemology of geographic concepts in Cyberspace -How are the basic spatial concepts of distance, proximity, direction, scale, region, boundary, etc. represented in Cyberspace? What properties do they have as compared to the commonsense uses of these concepts? Which properties are consistent across applications? What non-empirical properties do these concepts often have in cyberspace? Which spatializations of non-spatial information do users find intuitively easy to use? - How are basic geographic features such as mountains, lakes, fields, towns, etc. represented in cyberspace? How are these put together to form scenes and landscapes? What representational simplifications and generalizations are found across applications (and cultures)? What empirical properties of these objects are most often maintained? How easily or successfully do people interact with these features in cyberspace? - To what extent do the principles of naive geography described in the I-21 proposal also hold in cyberspace? Are some of them more consistently adhered to than others? Are there similar principles that appear to be specific to cyberspace? What may be their cognitive justification? Are users as comfortable dealing with these virtual principles as with the ones corresponding with real-world experience? What happens when empirical and virtual principles of commonsense geography coexist in an application? How do people and objects move in these spaces? How is time treated in relation to space? 2 Geographic concepts in GIS: reality and illusion - Where does the geographic world as represented in GIS depart from commonsense experience? Are these departures characteristic of other manifestations of cyberspace? Are they in basic-level or secondary spatial/geographic concepts and categories? How do users cope with these conflicts with experience? Is the naive geography of heavy GIS users, or of people learning geography largely with the help of GIS, affected by the counter-empirical properties of GIS representations? - What spatial/geographic representations occurring in other manifestations of cyberspace could be usefully adapted for use in the next generation of GIS? What modes of interaction with these representations could be carried over to GIS to facilitate problem-solving? - Which of these representations and modes of interaction should GIS adopt to be more compatible with other manifestations of cyberspace on the Internet? 3 Appropriate formalizations of geographic concepts - How have the most useful spatial/geographic concepts of cyberspace been formalized? What concepts in GIS do they correspond to? What are the technical or practical obstacles to introducing similar formalizations in GIS? Are there any novel data models underlying robust geographic representations in cyberspace that could be adapted for use in GIS? Are there any effective data retrieval methods? - How can we provide basic interoperability between GIS and other software that generates useful representations of geographic concepts? How can the next generation of GIS be augmented, and its learning or use be made more intuitive, through a coupling with such software? All the above questions could be researched through a combination of systematic analysis of relevant software and controlled experiments with human subjects.