We introduce a system called `Eliza'. Eliza is a system that is designed to respond to a user's statements. The model of the dialogue is that of a `Rogerian' psychotherapist. (A grandiose way of saying that the system questions the user!) When such a psychotherapist is told ``my dog hates me'' hs/she might reply ``what makes you think that your dog hates you?''. The system however is remarkably dumb since it is driven by a set of templates that determine how it may respond. The approach requires that we match against certain keywords then fill in the slots in some template.
Here is an example of dialogue from one version of Eliza:
This the output from a very simple Prolog program which we illustrate below.