Levels of Abstraction in Textual Representations of Geoinformation – Example: Cadastral Descriptions in Brazil

Transformations between textual descriptions of spatial information and other types of representation, e.g. maps, require a deep understanding of the actual information given in the texts. The level of detail used in such descriptions depends on various aspects, e.g.: the purpose of the description and thus the necessary information for the application, the author’s education and availability of tools for measurements, or the author’s assumptions about additional knowledge that a possible reader of the text might have. In this paper, the variability of descriptions is demonstrated by cadastral texts from Brazil. Analysing this data, the original hypothesis had to be rejected that an abstraction hierarchy based on the content of spatial information can be established so that a textual description as a whole can be assigned to one of those abstraction levels. However, by adjusting the idea of the first hypothesis to single spatial attributes and by introducing a hierarchy model of “missing information sources” and uncertainty, it is possible to represent explicitly the desired information content of the texts. The necessary flexibility is achieved by an extendible approach.