ABSTRACT
Correspondence analysis (CA), as a method to analyze linguistic data, was proposed by Jean-Paul Benzecri who stated that it was able “to address any type of issue related to the form, meaning or style of texts”. This chapter utilises a small lexical table constructed from the Life UK corpus to set out the principles of CA. The interpretation of axes and planes is based on standard real-valued indices, which is important in CA since the contributions of points to the inertia of the axes depend not only on their coordinate values, but also their weights. In textual CA, such indices are used to select the words and documents on which interpretations have to be based, as well as those which need to be represented, and filter out the rest. Ludovic Lebart has extended the scope of this research to free text answers, leading him to conceive of the joint analysis of textual and contextual data.
