Principles for Assumptions Generation in Enthymeme-Based Dialogue

ORBE LEIVA, DIEGO S.; GARCIA, ALEJANDRO J.; GOTTIFREDI, SEBASTIAN

JOURNAL OF ARTIFICIAL INTELLIGENCE RESEARCH, THE

AI ACCESS FOUNDATION

Año: 2025 vol. 83

1076-9757

In enthymeme-based dialogues, involved participants create assumptions in order to decode arguments from the exchangedenthymemes. This work introduces the concept of assumptions operator, which formalizes the mechanism for generatingthese assumptions, and proposes a set of principles to guide the construction of these operators. Said principles are inspiredby Grice’s Maxims of Conversation, as well as Govier’s ARG conditions for cogent arguments. Then, in order to analyzehow the used operator influences the dialogue and how that dialogue differs from the one in which the original argument issent, we propose a framework to compare both scenarios, the former being the enthymemic one and the latter the completeone. Finally, we formally show that if the used assumptions operator complies with a set of the aforementioned principles,then most arguments in the complete dialogue have their counterpart in the enthymemic one. Furthermore, we show thatunder certain conditions, the enthymemic dialogue preserves some semantic properties from the complete one, specifically:conflict-freeness, acceptability and admissibility.