Text Segmentation methods try to break up a text into thematically meaningful segments. MorphAdorner implements two linear segmentation methods which use measures of lexical cohesion to produce segments: Marti Hearst's TextTiler and Freddy Choi's C99. Both of these try to find those portions of a text in which the vocabulary changes from one subtopic to another. These change points mark the boundaries of the text segments.

Segmentation methods have been traditionally been applied to non-fiction discursive texts. We are interested in investigating whether segmentation methods illuminate the thematic structure of a wider span of genres in both fiction and non-fiction.

You can try MorphAdorner's linear text segmenters online.