Alexis Dimitriadis, Adam Saulwick & Menzo Windhouwer, Utrecht institute of Linguistics OTS, University of Amsterdam


The Typological Database System: Integrated access to diverse typological databases

The Typological Database System (TDS) is an ongoing research project to develop a web-based service for unified querying of multiple independently created typological databases, supported by an integrated ontology. A prototype server is available online at: http://languagelink.let.uu.nl

The TDS currently contains information on circa 1,000 languages from five integrated databases. Its component databases contain data on a range of linguistic topics including agreement, parts of speech, word order, stress placement and predication phenomena. Some yet-to-be integrated databases also contain primary linguistic data in the form of lexicons and glossed sentences. To support a uniform means of access to this information, an ontology of linguistic concepts is used as a blueprint for bringing the heterogeneous contents of the component databases into a consistent form. Explicit links between the ontology concepts and the unified data facilitate searching through the integrated database fields.

In view of the large number of database fields in the TDS, searching is a two-step process. First, the user discovers fields relevant to the topics of interest, by searching or browsing through one of several interfaces provided by the system (which rely on the ontology in various ways). These "pre-query"fields are accumulated, forming a pre-query. In the second step, the user refines this pre-query and executes it.