Summary: | My aim in this article is to demonstrate that natural language processing, especially automatic analysis of the content of a textual document, may benefit from a greater understanding of how the visual properties of texts intervene in the construction of meaning. It is particularly worth studying the role of layout in meaning insofar as layout features may be explicitly mentioned in xml documents. I adopt the point of view of the model of text architecture, which provides a theoretical framework to analyze layout and visual features. I focus on the analysis of the varied functions of headings in a text and I show how to exploit headings to characterize the content of the section they head.
|