Modelling Relationship between Antecedent and Consequent in Modal Conditional Statements

摘要

We consider a cognitive agent that is able to utter natural language conditional statements with modal operators of possibility and belief. Our long term aim is to simulate humans' ability to choose a statement within an agent. We are developing formal criteria on when a conditional statement can be said based on agent's knowledge. In previous works we defined a formal model for choosing a proper conditional statement. Here we focus more on a relation between antecedent and consequent. The relation between two phrases that is required to utter a conditional statement. We generalise our previous model to make it more flexible according to requirements and constraints imposed on an agent.