A BiOT Account of Gricean Reasoning Simon Pauw Abstract: - Tea or Coffee? - Yes - Huh? No, I mean, would you like tea or coffee? - Yes - Christ! Do you want tea? - Yes - So...no coffee? - I’d prefer coffee. - But...alright, so coffee it is. Milk or sugar? - Yes - ..., milk? - no - One lump or two? - Yes - No, I mean..., how many do you want? - Three - Would you like some biscuits or choc...never mind... One could wonder if it’s such a good idea to use logic to study natural language. Actually one could wonder if it is such a good idea to study logic at all. But, let’s suppose it is. Many people with a more practical disposition in life consider it a lost cause to say anything logical about language because it’s too intuitive. They’re probably right. But, let’s suppose they aren’t. The art would be to capture the intuitive part of language in logic. Clearly language doesn’t quite behave as strictly logical as, well, logic does. If I want coffee, logically the sentence “I want tea or coffee” is true. Somehow“Yes” doesn’t quite do it for us as a response to the question: “Would you like tea or coffee?”. There is a difference between the semantical content of an utterance and the conversational intentions of the speaker of an utterance. When I ask you if you want tea or coffee, I give you three options to choose from namely tea, coffee or nothing. I expect you to choose precisely one of those options. This is the conversational intention but it’s not part of the semantical content. In the field of linguistics these conversational intentions are commonly referred to as conversational implicatures or just implicatures. There are many different approaches to give a formal account for implicatures. Paul Grice has been most influential in this area (if not only for coining the term implicatures).1 Much contemporary research is based on the findings of Grice. Grice suggested that a speaker obeys certain conversational conventions. For instance, someone wants to say something he believes is true and relevant. The speaker has an intended meaning and by means of these conversational conventions this intended meaning is mapped to an utterance. The hearer of the utterance can infer the intended meaning by taking the speakers perspective into account. Grice proposed a list of four conversational conventions (maxims). In [1] Maria Aloni proposes to formalize these maxims in bidirectional optimality theory (BiOT). With this paper I try to investigate how well Aloni’s approach works and how to improve Aloni’s findings. This paper contains nine sections. The first three set out the field. Discussing some common ways of dealing with implicatures. Section four and five build up to Aloni’s approach. In section six the problems with Aloni’s approach are discussed. This leads to some suggestions to improve this approach (worked out in section seven). The last two chapter are reserved for the conclusion and discussion of the suggested improvements.