Microtheories for Spatial Data Infrastructures - Accounting for Diversity of Local Conceptualizations at a Global Level

Abstract

The categorization of our environment into feature types is an essential prerequisite for cartography, geographic information retrieval, routing applications, spatial decision support systems, and data sharing in general. However, there is no a priori conceptualization of the world and the creation of features and types is an act of cognition. Humans conceptualize their environment based on multiple criteria such as their cultural background, knowledge, motivation, and particularly by space and time. Sharing and making these conceptualizations explicit in a formal, unambiguous way is at the core of semantic interoperability. One way to cope with semantic heterogeneities is by standardization, i.e., by agreeing on a shared conceptualization. This bears the danger of losing local diversity. In contrast, this work proposes the use of microtheories for Spatial Data Infrastructures, such as INSPIRE, to account for the diversity of local conceptualizations while maintaining their semantic interoperability at a global level. We introduce a novel methodology to structure ontologies by spatial and temporal aspects, in our case administrative boundaries, which reflect variations in feature conceptualization. A local, bottom-up approach, based on non-standard inference, is used to compute global feature definitions which are neither too broad nor too specific. Using different conceptualizations of rivers and other geographic feature types, we demonstrate how the present approach can improve the INSPIRE data model and ease its adoption by European member states.