One objection to Humean views about motivation, and to the 'Standard Model' of intention on which intentions are complexes of desire and belief, is that these views don't allow agents to choose their reasons for doing some action. In Reasons Without Rationalism, Kieran Setiya presents the Standard Model as unable to explain how "Our reasons are in some sense 'up to us' — we decide why to do something, as well as what to do – and we seem to recognize our reasons, as such" (39). Similarly, in Rationality in Action John Searle objects to desire-belief views of motivation, writing that "when one has several reasons for performing an action, one may act on only one of them; one may select which reason one acts on" (65).
I can see where these objections are coming from. On a traditional Humean picture, reason is the slave of the passions, and it doesn't have the ability to hold one desire back so that another can motivate action. Neither can it noninstrumentally create or strengthen a desire so as to make it and not another desire decisive in motivating action. Nevertheless, I think a traditional Humean approach on which all motivating reasons are desire-belief pairs and all practical reasoning is instrumental can explain reason-choosing. Let me show how this works.
The key point to keep in mind is that the desire-belief pair that motivates an action and thus gives the reasons for doing it need not be the strongest one favoring that action. Sometimes the strongest pair in favor of A can't get me to do A because we have an even stronger pair inhibiting A-ing for those reasons. But this stronger inhibiting pair might not exert any force in opposition to a weaker pair that favors A, because acting on this weaker pair would involve A-ing for other reasons. Then the weaker pair can motivate action and be the sole source of my reasons.
That's kind of abstract, so let's deal with a concrete example. Here's Setiya's example:
I've never met Setiya, so I don't know his real reasons for writing Reasons Without Rationalism. But let me make up a story. Maybe his strongest desire to write the book was for a fragment of immortality. But maybe he also desired even more strongly not to write the book for reasons of immortality. Now the only way he could write the book for a fragment of immortality would involve some kind of self-deception about his own motives. If he believed that he was about to write for immortality, his strong desire not to write for reasons of immortality would come into play and inhibit writing.
But suppose he had no similar desire not to write the book for personal satisfaction. The desire not to write for immortality doesn't do anything to inhibit action motivated by the desire for personal satisfaction (again, barring some unfortunate self-deception where he thinks he's about to write for immortality when he's actually writing for satisfaction). Even if the desire to write for personal satisfaction is weaker than the desire to write for immortality, it's able to be the sole motivator, because the desire to write for immortality is specifically inhibited from causing action by the desire not to write for reasons of immortality. So the desire for personal satisfaction could motivate him by itself and be the sole source of his reasons.
This is how reason-choosing works — desires not to act on some reason inhibit action that we recognize as being for that reason, so that we can only act out of other desires. It requires a bit of self-knowledge, and that's something that I think the Humean view gets right. I'm sure there are plenty of well-meaning people who don't want to make decisions for racist reasons, and would be inhibited from acting if they knew they were about to act for racist reasons, but nevertheless act for racist reasons because they're unaware of the role of racist beliefs or desires in driving their actions.
Reason-choosing requires a somewhat complex motivational structure where we have multiple first-order motivational states favoring some action and a second-order motivational state to inhibit acting out of some of them. (The second-order states against acting for D1 might themselves be derived in traditional instrumental style by a more fundamental desire to act for D2 alone.) Often we act without having anything this complex going on. This would mean that we often act without choosing our reasons.
And I think that's right! For the most part, we cook and cross streets and make casual conversation without choosing our reasons for doing what we do. Choosing reasons is something that a theory of motivation has to explain, but it's a special and complex case rather than a ubiquitous phenomenon.