Comparing Geospatial Ontologies with Indigenous Conceptualizations of Time

Abstract

GIScience 2016 Short Paper Proceedings Comparing Geospatial Ontologies with Indigenous Conceptualizations of Time Genevieve Reid 1 , Renee Sieber 1 McGill University, Department of Geography, Burnside Hall Building, Room 705, 805 Sherbrook Street West, Montreal, Quebec H3A 0B9 Email: genevieve.reid@mail.mcgill.ca renee.sieber@mcgill.ca Abstract The geographic domain has widely been studied in ontology research. However, integrating the conceptualization of time and temporal referencing of geographic concepts in data models is a complex task that has by no means been “solved”. Existing geospatial ontologies have adopted a space-time model that, for example, distinguishes endurant entities (lasting through time, e.g., fixed natural features) from perdurant entities (e.g., processes or events). Such a model might exclude indigenous conceptualizations of time that are far more sophisticated. We find that conventional ontologies make assumptions about time that fail to take into consideration indigenous notions including: 1. Time is not linear; 2. Nothing is completely fixed in time; 3. Time has agency; and 4. Time is not temporal but social. 1. Time and Geospatial Ontologies In GIScience, indigenous conceptualizations of space and time have been depicted as being in direct opposition to those used to design geospatial technologies (Rundstrom 1995; Veland et al. 2014). Indeed, geospatial technologies emphasize a more static view of the world that is often inconsistent with indigenous perspectives on space and time. Compared with geographic features, notions of time have received less attention both in geospatial ontologies and in indigenous ontologies research. Time and temporal referencing of geographic concepts are nonetheless challenging to geospatial ontologies applied in indigenous contexts. Including indigenous conceptualizations in geospatial ontologies and in the Geospatial Semantic Web is crucial. Wellen and Sieber (2013) argue that developing an inclusive semantic interoperability is not only possible, but also critical to ensure future accessibility of geospatial technologies for indigenous communities, and minimize loss and misinterpretation of information when geospatial ontologies are used to record indigenous knowledge. 1.1 Endurant/Perdurant Model An important contribution to the conceptualization of time in geospatial ontology development, was the distinction between endurant objects that endure through time and perdurant objects that happen in a certain time (e.g., processes or events) (Agarwal 2005). Grenon and Smith (2004) propose a spatio-temporal ontology of change and processes called SNAP/SPAN based on the duality endurant/perdurant. The SNAP/SPAN model distinguishes endurant entities, which have spatial properties, from perdurant entities, which have temporal properties. Temporal intervals and instants describe perdurant entities through linear time. Even though the SNAP/SPAN distinction is widely adopted in geospatial ontologies (Agarwal 2005), philosophical assumptions about time behind this model can fundamentally differ from indigenous conceptualizations.