Therefore they have propositional content after all, and conditions of satisfaction, and direction of fit.
A speech act's direction of fit characterizes the way in which acts of that type are related to the world.
Consider, for example George is happy: there is no propositional content to be satisfied in either direction; and therefore there is no direction of fit.
And if they are mental states, they are mental states that lack propositional content, direction of fit, and conditions of satisfaction.
Conative states (e.g., motivational states of want, need, and desire), also have propositional content, but they have a world-to-mind direction of fit.
And in another example, Martha is hungry, again, there's no propositional content to be satisfied, no direction of fit, and no conditions of satisfaction.
In fact, she said, we know what our intentions are without observation; and between someone observing, and someone intending and performing, an action, there is what she called a difference in the "direction of fit" (of action to thought).
Thus, in the affective state George was glad that Martha liked him there is propositional content (something about Martha liking George), but no direction of fit because the propositional content is already satisfied.
Cognitive states (e.g., beliefs, percepts, and memories) have propositional content, of course, but they have what Searle calls mind-to-world direction of fit, by which he means that the mind is describing a current reality that exists independently of it.
Given this analysis, cognition, emotion, and motivation may be irreducible after all, just as Kant asserted: they are clearly distinguished from each other by their conditions of satisfaction (emotional states don't have any) and direction of fit (cognitive states fit mind to world, and conative states fit world to mind).
That's bad for those who want to develop an independent affective (neuro)science, but it still opposes the hegemony of cognition, because something besides belief (i.e., desire) is needed, and that something is not reducible to cognition (because cognitive and conative states differ in direction of fit).
