Text representation, dictionary structure, and lexical knowledge

Text encoding, when applied to complex text types, requires principled separation of (visual) presentation properties of text from (conceptual) functional organization of content. This is particularly pressing when machine-readable dictionaries (MRDs) are being analysed for implicit lexical semantic content. We argue that a current proposal for textencoding guidelines falls somewhat short of making such a separation; we furthermore argue for such a separation, by demonstrating the benefits it brings to a methodology for extracting lexical data from MRDs. © 1992 Oxford University Press.