Natural is Uncertain, Emotional, Deceptive and Still Other. But: How to Get it? Position Statement and Questions

Linguists seem to agree since long in claiming that a ’good’ argumentation system should be keen in selecting the arguments that are ’strong’ in given circumstances, by envisaging counter-arguments and producing counter-counter arguments (in advance if needed, or on request), in order to eventually produce a well formatted, coherent and ’convincing’ message. Argumentation theories go back to the origins of our culture and AI researchers should apparently only find out appropriate techniques to produce such natural results. However, artificial argumentation systems are still far from being natural: indeed, obstacles still found in the production of a satisfying solution are due, in my view, to the unclearness of some concepts in these theories. I’ll try to list some of these problems, in the hope that they may contribute to the Workshop discussion.