Abstract

This contribution discusses the function and use of a new corpus linguistic freeware (under Creative Commons Licence), called "LDA-Toolkit". After a brief introduction, the tools and their functionality are presented, followed by two examples of analyses. The LDA-Toolkit has been especially developed for linguistic discourse and image analysis for use in research and teaching. The handling is easy (GUI) and does not require any programming skills. Its tools contain a token list, a concordance tool, keyword analysis, contrastive key-cluster, key-ngram and key-pos-gram analysis, contrastive co-occurrence analysis and easy auxiliary devices to visualize results as word nets or word clouds. All tools as well as all results of the analyses are concentrated and recorded in a virtual working platform called "LDA tree" or "LDA explorer", because the results are visualized in a hierarchic tree model of co(n)texts. Furhtermore, the records of this main tool can easily be saved as status of work and reloaded for later analysis. The functionality of the LDA-toolkit is exemplified by two examples: the first example illustrates a corpus-driven method of linguistic discourse study using keyword and contrastive ngram analysis. It focuses on the legal debate about euthanasia and advanced health care directives in Germany. The second example outlines a corpus-assisted approach to analyzing public media images and stereotypes. It examines the German pirate party's image ("Piratenpartei") in German newspapers. © 2012 De Gruyter Mouton. All rights reserved.