The objective of this session is to distinguish between textitdeclarative and textitprocedural criteria for creating dictionaries, and to understand the principles of implementation of this distinction in lexical databases.
A lexical database is an implementation of a lexicon as a computer database. For this purpose the microstructure, mesostructure and macrostructure of the lexicon needs to be very explicitly designed. On the other hand, the speed, size and flexibility of database systems permits a variety of different uses for a given lexical database.
In order to achieve this flexibility, an important point needs to be made more explicit than is usually the case. Traditionally, dictionaries are purpose-built. That is, all aspects of the dictionary are specifically designed around a specific functionality for the intended dictionary user. This is the context in which distinctions such as onomosiological vs. semasiological are made.
A lexical database, however, enables the lexicographer to abstract a general declarative, factual structure out of different lexicon deployment contexts, and concentrate on organising the lexical information in the most compact, clear and convenient fashion. The different procedural requirements which underlie traditional types can then be defined as different views on the database, i.e. different output filters which permit different dictionaries to be generated from the same database.
The practical tasks in this session involve