Beliefs, desires, and directions of fit

According to one popular account, beliefs and desires should be distinguished by their respective directions of fit. Beliefs aim to have their content fit the world, desires aim to have the world fit their content.

Now I have no idea how to make this less metaphorical, and nor can I make much sense of the idea that my mental states, and not me, have aims. But putting that aside, there's still a problem.

A desire can't be anything that represents something, and aims to have the world fit that representation. Instruction manuals seem to have those properties, but I don't understand instruction manuals to be kinds of desire.

So perhaps this direction of fit is necessary, but not sufficient, for something to be a desire. Then we might ask: what else is necessary? Here's my hunch: The presence of the reverse direction of fit as well.

Precisely what seems to distinguish rational creatures from the less rational and non-rational is the more reflective nature of our impact on the world. It's that our drives and aims are complicated responses to features of the world. The way to capture that kind of increased complexity seems to me to require that we think of our responses to the world as belief dependent. What makes a desire a desire, and not a compulsion, is precisely that it is partnered with some representations of how the world is, and beliefs about the responses it calls for.

(It might be that both directions of fit are still only jointly necessary but not sufficient for some state to be a desire. I also haven't said much here about whether the direction of fit account is correct for belief: I doubt it, but that I need to think about that some more.)