TEXTELSEM - Texts, Tells and Semantics

Finished
TEXTELSEM i3mainz, CC BY SA 4.0

Modelling the historical geography of northern Mesopotamia in the 2nd millennium BC. Using the integrated analysis of texts, archaeology and space, geoinformatics and the semantic network

Motivation

TEXTELSEM is an interdisciplinary project at the interface of Near Eastern archaeology, Ancient Oriental Studies and information technology that is developing a new range of methods and visualisations for research on historical geography. Archaeological and spatial data generate information on settlement structures and terrain characteristics. Texts contain a dense network of explicit and implicit information on places, landscapes and routes. The quantity and complexity of philological and archaeological data make the use of IT-supported forms of knowledge management necessary and sensible. TEXTELSEM aims to connect these information groups by developing information integration based on concepts of semantic data description. The project focuses on the joint development of suitable ontologies for the description of the disciplinary data sets. TEXTELSEM builds on the work done in HIGEOMES and aims to further develop not only the methods prototypically developed there but also the content.

Spatially, the project is not limited to Upper Mesopotamia, but now also includes the Eastern Tigris region - an important area as a settlement chamber and important link in the traffic route network of the Ancient Near East. Recently, it has been systematically archaeologically opened up through intensive field research in Iraqi Kurdistan. This offers a unique opportunity to correlate current geoarchaeological data with information from the written sources of the 2nd century for the chronotope and thus to gain new insights into the development of the historical geography of this region, which is enormously important in terms of cultural history. The temporal limitation of the project results from the fact that said northern Mesopotamian region is known in particular detail on the basis of the cuneiform corpora, especially the Old Babylonian texts from Mari and Middle Assyrian sources.

The cooperation between the French and German partners combines complementary competences that enable this project to combine historical-systematic basic research with the development of new methods of analysis and visualisation. The basis is a jointly developed, knowledge-supported information system that integrates scientific data in such a way that, in addition to research and the interoperable provision of information, a partially automated generation of hypotheses is also possible. In this way, knowledge discovery technologies can decisively narrow down large-scale vague localisation hypotheses of the type “somewhere in region XY” and generate identification hypotheses on the basis of existing knowledge. The tools developed within the framework of the proposed project (data structures and visualisation methods) can be transferred to other time periods and spaces.

Activities

In the past year, i3mainz developed a web platform to provide easy access to the linked archaeological and philological data without in-depth database knowledge (Fig. 2). Thanks to the simultaneous retrieval and analysis, scholars can gain cross-disciplinary insights. For more complex analyses and cartographic processing, an export as KML is available. To this end, i3mainz staff organised a workshop in Paris in April 2017.

Research assistant Magdalena Scherl has developed a search function with “fuzzy-search” techniques for improved retrieval of differently transliterated or transcribed place names (Fig. 3). She has also brought the ontology for the archaeological data from Munich up to the current OWL2 standard, so that not only a link with other linked data is possible, but also automated inference by means of inference and integrity rules. Martin Unold is working on his dissertation on the processing of uncertain and vague information in the Semantic Web in order to further improve the evaluability of the data in the coming year.

Results

For more current information on TEXTELSEM, please visit the project page. There you will also find literature references and links.