Information About a Given Entity: From Semantics Towards Automated Deduction

摘要

The standard method to retrieve information can be formally defined as follows. To ask a query, one gives the properties of the entities to be retrieved, and the answer is the set of all the entities that satisfy the query. Another method, is to ask the overall information about a given entity, and the answer is the corresponding information. An example of the first kind of query is: 'who are the persons who have had a car accident?', an example of the second kind is: 'what is the overall information about a given person?'. This latter method has deserved very few researches though it has great potential practical applications. However, it raises many non-trivial issues that are investigated here. The first one is to find a precise definition of the fact that a piece of information 'is about' a given entity. We propose a new formal definition of this notion of aboutness for query languages in first-order logic with function symbols, and the main properties that follow from this definition are presented. The second one is to define a bridge between this abstract semantic definition and automated deduction methods based on Resolution Principle. Deduction strategies are defined in this direction and it is proved that they are complete.