Text Visualization for the Support of Lexicography-Based Scholarly Work

We discuss three visualisation techniques for corpus analysis, Concordance Mosaic, Metafacet and ComFre, and explore the design rationale based on a characterization of the corpus linguistic domain. The Concordance Mosaic visualization is designed for the investigation of collocation patterns. It encodes word positions in a concordance list in a manner that emphasizes quantitative analysis of frequency or collocation statistics. Metafacet provides an interface for investigating concordance lists through the lens of meta-data. When combined with the Mosaic it provides a powerful technique for investigating collocations in the context of meta-data. ComFre can be used to compare word frequencies between two corpora of different size, it has potential use as a technique for identifying terms which are representative of the corpora under investigation. The domain characterization shows how the visualizations were designed with corpus linguistic methodologies at the core. It consists of a task analysis based on the methodology outlined in Sinclairs’ Reading Concordances: An Introduction, and the analysis of methodology case studies from language scholars.