The project investigates language mixing in historical documents from medieval England. Its results will contribute to the following branches of linguistic research:
On the basis of a detailed text analysis we want to assess whether current models of mixed-language structure and processing also explain language mixing patterns found in historical documents. We assume that abstract language mixing patterns and basic mechanisms of language processing have remained comparatively stable over the centuries, even though languages are subject to continuous change.
The project is investigating how far structural code-switching models which have been developed on the basis of modern language data are applicable to historical data. If historical texts turn out to show the same mixing patterns and tendencies as their modern counterparts, this would provide a strong additional argument for the predictive force of a structural model.
The analysis covers several text genres (religious, instructional, correspondence, etc.). We will employ sociolinguistic approaches to assess the speaking habits of multilingual communities in medieval England, which will shed light on pragmatic aspects of bilingual communication.
For comparison we will also analyze mixed-language data from Early German (Old High, Middle High, Middle Low and Early New High German). This comparison will show whether the mixing of a strongly inflecting Romance language (Latin) with an inflecting Germanic language (Early German) shows different mixing tendencies from the ones observed when mixing Latin with a language that has lost most of its inflectional morphology (Middle English).
In collaboration with the Institute for Natural Language Processing in Stuttgart, one sub-project focuses on the creation of a digital corpus of mixed texts. The challenge here is not only the linguistic annotation (i.e. lemmatising, tagging, parsing and providing a platform for public access) but also the development of tools for annotating mixed historical text.