The dictionary evolved from medieval glosses that explained fragments of discourse in their contexts. Those fragments were later collected, then classified and reduced to their simplest forms, ie words. The most important aspect of that evolution from the gloss to the dictionary is that the fragment to be explained was decontextualized, extracted from discourse. The main objective of the dictionary is to give an image of the system. It is now possible to improve the dictionary in its role as a tool for explaining discourse. It cannot provide explanations that would be adapted to every single context, but it can give to the user a huge quantity of discourse, and provide explanations that would be more closely adapted to every occurrence or type of occurrence. Lexicographers would be well advised to investigate those new possibilities.

This paper compares two French dictionaries of collocations (the Dictionnaire des cooccurrents de Beauchesne (2001) and the Lexique Actif du Français (Mel'cuk et Polguère, in preparation)) with electronic versions of two well known French dictionaries: the Petit Robert Electronique (version 2.1) and the Trésor de la Langue Française Informatisé. Our goal is to evaluate to what extent specialised dictionaries are really better suited than general monolingual dictionaries in representing collocations for educational purposes. Three parameters are examined: the access to collocations in dictionaries, the quantity of data and the linguistic treatment of collocations.

This paper briefly surveys three of the main resources for word sense disambiguation that are currently in use - WordNet, FrameNet, and Levin classes - and proposes an alternative approach, focusing on verbs and their valencies. This new approach does not attempt to account for all possible uses of a verb, but rather all its normal uses ("norms"). By corpus pattern analysis (CPA), the normal patterns of use of verbs are established. A meaning ("primary implicature") is associated with each pattern. The patterns are then available as benchmarks against which the probable meaning of any sentence can be measured. The status of abnormal or unusual uses ("exploitations") is also briefly discussed. Also, three kinds of alternation are recognized: syntactic diathesis alternations, semantic-type alternations, and lexical alternations.

Computers can be used in lexicography to support the analysis of the language, and to support the synthesis of the dictionary text. There are of course many other interactions between computing and lexicography, including the preparation and presentation of electronic dictionaries, the use of dictionaries in language technology systems, and the automatic acquisition of lexical information. In technologically advanced dictionary-making, the lexicographer works with two main systems on their computer: the Corpus Query System for analysis and the Dictionary Writing System for synthesis. Currently, these are always independent, with communication between the two via cut-and-paste. We describe requirements for state-of-the-art dictionary writing systems.