HyperHamlet – Intricacies of data selection

HyperHamlet is a database of allusions to and quotations from Shakespeare's Hamlet, which is supported by the Swiss National Science Foundation as a joint venture between the Departments of English and German Philology, and the Image & Media Lab at the University of Basel. The compilation of a corpus, whose aim is to document the "Shakespeare phenomenon", is intricate on more than one level: the desired transdisciplinary approach between linguistics, literary and cultural studies entails data selection from a vast variety of sources; the pragmatic nature of intertextual traces, i.e. their dependence on and subordination to new contexts, further adds to formal heterogeneity. This is not only a challenge for annotation, but also for data selection. As the recognition of intertextual traces is more often than not based on intuition, this paper analyses the criteria which underlie intuition so that it can be operationalised for scholarly corpus compilation. An analogue to the pragmatic model of ostensive-inferential communication with its three constitutive parts of speaker's meaning, sentence meaning and hearer's meaning has been used for analytical heuristics. Authorial intent - in a concrete as well as in an abstract historical sense - origin and specific encyclopaedic knowledge have been found to be the basic assumptions underlying data selection, while quantitative factors provide supporting evidence.