The Nature of Attitudes
I've written extensively on belief (e.g., here, here, and here), but very little on other attitudes like desire, intention, and love (though see here on love). (I regard desire and intention as "propositional attitudes" in the philosopher's sense, but not love.) Here's the summary version of my current thinking:
(1.) Attitudes are dispositional. To believe that there's beer in the fridge is to be disposed to act, react, and cognize in certain ways. It is not a matter of what you're currently doing or thinking. So, for example, your attitudes don't change simply by virtue of your falling unconscious. To believe that there's beer in the fridge is to be disposed to go to the fridge if you decide to have a beer, to say "yes" if someone asks you if there is beer in the fridge, to feel surprise were you to open the fridge and find no beer, to conclude "I will win a million dollars!" if you're told that you will win a million dollars if there is beer in your fridge, etc. Such dispositions are, of course -- like virtually all dispositions -- subject to defeaters or excusing conditions (for example, you might not tell the truth if you want to keep the beer a secret). Similarly: To want to complete your dissertation is to be disposed to feel good if it seems that you're making progress on your dissertation, to favor actions that promote completion of the dissertation over other actions all else being equal, to think to yourself "completing that dissertation would be a good thing to do", etc. To love someone is to be disposed to prioritize that person's well-being in your choices, to feel especially bad when bad things happen to that person, to value having a connection with that person, etc.
(2.) But attitudes also have an occurrent face. In fact, they have two occurrent faces. (By "occurrent" here I mean having to do with events that transpire at particular moments of time.) Attitudes can manifest, that is, the disposition can be activated. You can actually go to the fridge to find the beer, do something to further your dissertation work, choose something reflecting your prioritization of the welfare of someone you love. And attitudes can form (or disappear). A minute ago, perhaps, you had no desire for the chicken sandwich vs. the roast beef. You hadn't started thinking about it yet; sometimes you choose one, sometimes the other. You look at the menu and make a choice, forming a desire for the chicken.
Forming an attitude can be nonconscious, if it happens underground, as it were, or it can involve dedicating attention or consciousness. The conscious formation of a belief -- or the conscious reinforcement of that belief, if it was already present -- we can call a "judgment". The conscious formation of an intention we can call a "choice". We don't have natural names for some of the other conscious attitude formation episodes. ("Falling in love" isn't quite what I have in mind as the conscious formation of love; judging that something is good isn't quite the same as forming a desire for it -- though perhaps judging good and forming/reinforcing a desire are close enough to explain why we don't have separate terms for them.) Of course, sometimes the judgments and choices don't stick and the attitude isn't actually formed. You choose to exit the freeway at the next offramp, for example, but then forget. You say to yourself, and quite sincerely judge, that death is not bad while failing to form the dispositional profile appropriate to that attitude. (I have a paper about such cases here.)
(3.) Attitudes are not discrete entities; rather they overlap. This is true both within and between attitude types.
First, within attitude types: I do not have one ball in my "belief box" for "my cat is in the tree" and another separate ball for "Waterball [the name of my cat] is in the tree" and yet another separate ball for "Waterball is in the oak" and yet another for "our only pet is in that tall plant right there", etc. I believe all those things, more or less, but not via discretely held contents. What I have is a cluster of dispositions such that all of these things are approximately true to say of me -- though possibly some are more apt than others (for example, if I'd be disposed to call the tree an elm rather than an oak if asked).
Second, between attitude types: The dispositional profile for thinking that it would be good to finish your dissertation largely overlaps with the dispositional profile for wanting to finish your dissertation, such that probably you have both or you have neither. But the dispositional profiles come apart enough that in some cases one way of putting things can be more apt: Intellectual acceptance is more central to the belief-good profile and emotionality is more central to the desire profile so that if you are prone to form pro-dissertation-completing intellectual judgments that leave you cold, it might be more apt to describe you as believing-good than as desiring (though still, in a way, you desire or half-desire), and vice versa if you're more prone to desirous emotionality than intellectual endorsement. But let me be clear: On my view this is not a matter of the attitudes being ontologically distinct and causally connected; rather it's a matter of their having partly overlapping dispositional profiles with different dispositions closer to the center. Similarly for hating vs. loathing, appreciating vs. being thankful, resentment vs. anger vs. feeling wronged vs. thinking you've been wronged, etc.
Lots of cool things follow from this approach to the attitudes, such as weak motivational internalism; the avoidance of puzzles about the countability of beliefs and of problems in drawing the line between implicit and explicit beliefs, an appealing gradualism about attitude change (including learning and forgetting), the lack of a need for a discrete type of cognitive token for every attitude verb, and a nicely moderate position about the relationship between desiring and believing good.