KoGraR: Standardized Statistical Analyses of Corpus Counts

Within the project “Corpus grammar” (Korpusgrammatik) at the Institute for the German Language (Institut für Deutsche Sprache, IDS) in Mannheim, techniques and tools are developed for the description of grammatical phenomena based on analyses of very large morphosyntactically annotated corpora. The goal of the project is a corpus-based grammar that captures variations of grammatical structure in presentday German. In the first project phase, pilot studies were conducted (cf. Bubenhofer et al., 2014; Fuß, 2014; Konopka, 2014) to exploit and evaluate various methodological approaches to variation phenomena. For each research question, statistical analyses were chosen and customized. From these analyses, a subset was extracted as the methodological core of the project, with the aim of supporting methodological coherence, interoperability of sub-projects and, finally, the descriptive coherence of the project result, that is, the grammar. The methodological core has been made available to project members via an easy-to-use web front-end: the results of corpus queries and other, user-defined data tables can be uploaded and analyzed automatically. The web front-end is called KoGraR.