This talk will describe an application of the FOIL ILP algorithm to the learning of rules for text interpretation. The system is applied to simple texts taken from an encyclopedia, and restricted to those describing chemical elements. The chemical elements ontology from Ontolingua is also an input to learning, as well as providing the target knowledge representation. A second learning approach for determining coreference in the texts will also be presented. This too uses the ontology and also WordNet. This work was done by MSc students Susan Leavy and Luc Russell in 1999-2000.