AbstractA common assumption for argumentation-based dialogues is that any argument exchanged is complete, in the sense that its premises entail its claim. However, in real world dialogues, agents commonly exchange enthymemes - arguments with incomplete logical structure. Existing literature on formalising the use of enthymemes usually concentrates on how agents may construct enthymemes from an intended argument and reconstruct intended arguments from received enthymemes, based on assumptions about shared knowledge and context. However, little attention has been given to what happens when
the agents employ enthymemes during dialogues, and those that do focus mainly on the ‘backward extension’ of enthymemes.
In our thesis we, first, formalise the dialogical exchange of enthymemes that are missing some constituent elements, such that it is not possible to directly entail the claim of the intended argument from the premises of the enthymeme exchanged. Secondly, we introduce a dialogue system that handles the backward extension of enthymemes as well as the ‘forward extension’ of them and how the agents can deal with any misunderstandings regarding what they revealed and what their counterpart thought was intended. For both dialogue systems we show that, under certain conditions, the status of moves made during a dialogue conforming to either one of the systems, corresponds with the status of arguments in the Dung argument framework instantiated by the contents of the moves made at that stage in the dialogue.
By developing these dialogue systems, we try to close the gap between formal logic-based models of dialogue and the kinds of dialogue studied by the informal logic community, which focus on more human oriented models of dialogue. Moreover, our results verify that when enthymemes are implemented in a dialogue, participants can still reach the same outcome as they would have done if they used their complete intended arguments. This is important as we therefore show that there is no disadvantage to the use of enthymemes in dialogues, a common real-world feature of dialogues that supports efficient inter-agent communication.
|Date of Award||1 Dec 2022|
|Supervisor||Sanjay Modgil (Supervisor), Elizabeth Black (Supervisor) & Christopher Hampson (Supervisor)|