Feature structures and database theory

More and more often, computational linguists and language engineers need to manipulate large sets of feature structures which constitute real databases. Creating, updating, querying and in general, exploiting such linguistic databases is not an easy task, and it is desirable to have at one's disposal operators and languages for query and manipulation like (and if possible extending) those available for standard databases. The goal of this paper is to provide a database model for sets of feature structures, thus bridging the gap between the two fields, feature-structure theory and database theory, both of which are well-developed and well-formalized, but seem to have evolved in relatively separate communities. First, we show what feature structures have in common with, and how they differ from, the most classic data model, the relational model. We then formally define a database model for feature structures, and show how the standard operators of the relational data model can be generalized to feature-structure databases, and thus propose an extended relational algebra for feature structures.