Modular Generation of Relational Query Paraphrases

This article proposes a novel technique to generate natural language descriptions for a wide class of relational database queries. The approach to describing queries is phrasal and is restricted to a class of queries that return only whole schema tuples as answers. Query containment and equivalence are decidable for this class and this property is exploited in the maintenance and use of a phrasal lexicon. The query description mechanism is implemented within the Schema Tuple Query Processor (STEP) system (http://www.cs.umu.se/~mjm/step). Because the said query class is also closed over elementary set operations, it may be reasoned with in a relatively unrestricted manner. This enables a modular separation between a reasoning component and a ‘tactical’ realization component. To demonstrate this modularity, this fragment is shown to be adequate for several cooperative reasoning techniques. Thus the cooperative information system serves as the ‘strategic’ component, deciding what to say, while the generation system acts as the ‘tactical’ component, deciding how to say it. Naturally expressions within the said query language are the interchange language between these two components.