Full text loading...
In literary studies, intertextual theory stresses that works are co-dependent. Because the concept of intertextuality is generally defined only vaguely, it covers a variety of phenomena, ranging from exact quotations over subtle allusions to higher-level genre constraints. Philologists have been studying textual correspondences between medieval works long before these connections were called intertexts. Most famously, K.H. Heeroma uncovered many parallel verse groups between Middle Dutch epics and argued that these could be indicative of shared authorship. Other scholars, however, argued that these intertexts more likely resulted from a genre-level convergence of the epic discourse. Here, I describe a computational system that deploys established methods from the field of information retrieval to automatically detect intertexts in Middle Dutch narrative fiction. The calibrated system is empirically evaluated on a sample of intertextual connections which were taken inventory of in prior work. We apply the system for the detection of micro-intertexts (correspondences of just two verse lines). Statistical analysis suggests that micro-intertexts indeed prominently occur within epic (sub)genres, but that they also feature commonly in the oeuvres of single authors. Finally, the system detects a wide array of previously undetected intertexts between epic works (such as the Lorreinen continuation and Limborch) that deserve a careful examination in future work, especially in the context of authorship attribution.
Article metrics loading...
Full text loading...
References
Data & Media loading...