An intentional action springs from an agent's beliefs and desires such that, given the specifics of the situation, if the beliefs are held constant and the desires are held constant, then the action would not change.
Consequently, in prescribing a different action - when saying that an agent “ought to have done something differently,” this has to imply that the agent ought to have had different beliefs, or ought to have had different desires, or both.
The former - prescribing an action based on a different belief - covers the realm of "practical ought". The latter - prescribing an action based on a different desire - is covered by "moral ought".
A woman, on a long run in the hot sun, takes a drink of what looks like clean water. Another person who knows that the glass did not contain water but a colorless, transparent, tasteless liquid that causes nausea and cramps tells the runner, “You should not have done that.”
This is not a moral judgment. He is not condemning the runner for drinking out of the glass - not in this case, anyways. Instead, he is telling the runner, “Given your actual desires, an agent with true and complete beliefs would not have drank from that glass. Your desires, all things considered, are not going to be objectively satisfied by that action.”
When a statement that an agent ought to have done something else is reduced to a statement that the agent ought to have believed something else (or an agent with true and complete beliefs but the same desires would have done something else) the prescription is not moral - it is practical.
The hallmark of beliefs is truth. An agent chooses the action that would fulfill the most and strongest of her desires in a world where her beliefs and true and complete. The runner decided to drink from the glass because, in a world where her beliefs were true and complete, drinking from the glass would have quenched her thirst and produced no ill effects. Whether an action is actually successful depends substantially on whether the beliefs are accurate. If they are, then the action will produce the intended results. If not, the agent risks an unsuccessful action.
A woman on a long run in the hot sun, sees a wallet sitting on a table. She picks up the wallet, takes out some of the cash, and puts it in her pocket, returning the wallet to the table. Another person, who had been watching the wallet from a hidden location, and tells the runner, “You should not have taken the money.”
In this case, we may assume that a person with the same desires, but true and complete beliefs, would have still taken the money. Those true and complete beliefs would include the belief that the watcher is simply doing research and has no intention of thwarting the agent's desires in any way. She has no aversion to taking the money. Her only deterrence, let us assume, would be an interest in avoiding punishment. So, let us imagine a situation in which she will not be punished.
Still, it makes sense to make the moral claim, "She should not take the money." She is guilty of theft, and theft is a moral crime.
In this case, the claim that she should have done differently is not a claim that a person with true and complete beliefs would have done something else. It is a claim that a person with the right desires would have done something else.
What counts as a "right desire?"
Desires, unlike beliefs, do not have anything like a truth correspondence. There is nothing like a correct or incorrect desire. Instead, our evolutionary past has disposed us to desire that which pursuing or avoiding had allowed our ancestors to be biologically successful.
Our tastes for food, for example, is not tuned to any type of intrinsic goodness. We have a taste for food that kept our ancestors alive long enough to have children and raise them to the point that they can have children. Broccoli is good for us, but it was not as good for our ancestors as lots of calories when one could get them. Alcohol - unlike stagnate water - is substantially germ free.
The fact that we are disposed to desire that which kept our ancestors alive does not imply that these desires are in any way correct or right. Indeed, as we are seeing, the desires that kept our ancestors alive in their environment might not be too healthy for us in our current environment. What this points out is the fact that there is no correct or incorrect thing to desire. We like what we like. Some of us will have offspring who will grow up to have offspring, and some of us will not. There is no "intrinsic value" to one option or another.
So, in the absence of intrinsic value or an inherent correctness to desire, how can we make sense of talking about the "right desire"?
There is still the fact that there are some desires that people generally have reason to promote, and some desires that people generally have reason to inhibit. Furthermore, desires are malleable, and can be strengthened or weakend through praise, condemnation, reward, and punishment. So, there are desires that people generally have reason to respond to with praise and rewards, and desires that people generally have reason to respond to with condemnation and punishment.
To say, "You ought not to have taken the money" is to say, "You ought not to have the desire set that would have caused you to take the money." This does not mean that the desires are incorrect in the same way that beliefs can be incorrect. Instead, it means that the desire set is one that people generally have reason to respond to with condemnation and punishment. In fact, the claim, "you should not have taken the money" is not only a statement that the agent has a desire set the people have reason to respond to with condemnation. It is - at the same time - an act of condemnation.
In saying that the agent ought not to have had the desire set that would have motived her to take the money, it is important to keep in mind the fact that desires are persistent entities. You cannot simply turn them on and off like a switch. Even if you could, the next question to ask is: what character traits would be sitting in the background to motivate turning a particular desire on or off - and can that desire be turned on and off?
Consequently, when we evaluate a desire, we must evaluate its effects on the wide range of regular every-day circumstances a person may find themselves in. It may well be the case that a desire set that objectively satisfy other desires in certain rare circumstances would thwart other desires in a wide range of every-day circumstances in which people routinely find themselves. It this case, people generally have little reason to praise the act that objectively satisfies other desires. Instead, they have reason to condemn it - to avoid the desire-thwarting effects that would arise in the every-day circumstances that people often find themselves.
To say that a person ought to have done something different implies either that a person with true and accurate beliefs would have done something different, or that a person with good desires would have done something different, or both.
The first option relates to practical ought - it prescribes a course of action that will better allow the agent to realize that agent's desires. It aims to direct the agent to potentially more successful actions.
The second option relates to moral ought. It aims to use social tools such as praise and condemnation to promote desires that people generally have reason to promote, and inhibit desires that people generally have reason to inhibit. One of the aspects of doing this is that it embeds the praise or the condemnation in the moral statement itself. Consequently, a moral claim not only prescribes a particular desire set, it employs the tools of praise and condemnation that aim to bring about the prescribed desire set.