Language, Action and Thought

The idea of unarticulated constituents has been introduced by John Perry: the truth-conditions of many thoughts have components that the subject herself does not represent. Likewise, an object or a property may make its way into the proposition expressed by an utterance without being represented by any uttered morpheme. This happens because language is always a supplement to whatever else is available to the agents engaged in a conversation. Information needed for the agents to grasp a full propositional content comes from different sources: the frame of an action, the alignment of perspectives in a dialogue, joint attention, the memory of shared experiences and the cumulative nature of conversations. Once we have identified these different sources of information, we have to see what unarticulated constituents are constituents of. They are constituents of propositions, i.e., of classifying tools for the theorist, not representations of what is in the agent’s mind. Speakers can be sensitive to parameters that will part of the theory, but that they do not represent. I suggest the merging of two ways to represent propositions: completing the words uttered with the missing elements for a truth-evaluable item, or for the intended truth-evaluable item, and adopting Austinian propositions.