Specifying Events and their Effects in Controlled Natural Language

This paper shows how a controlled natural language can be used to construct precise formal representations for reasoning about events and their effects. Specifications written in PENG Light are translated with the help of discourse representation structures into the input language of the Simplified Event Calculus. This logic-based formalism is declarative and can be used for various reasoning tasks, among them for question answering. Taking a simple scenario from the transport domain as a starting point, we illustrate how PENG Light can be used to specify the narrative part of the scenario and the commonsense knowledge that is required to reason about direct and indirect effects of events as well as about continuous change. There is no need to encode this scenario and the background axioms in a formal notation, since the relevant information can be expressed directly on the level of the controlled natural language.