Why do you care about "desires that tend to fulfill other desires" if "desire fulfillment has no value"? A goal to realise a "state of affairs" without an accompanying desire is not possible.
The first thing I want to note is that I did not say that desire fulfillment has no value. I said that desire fulfillment has no intrinsic value. More broadly, I said that intrinsic value does not exist. Nothing has intrinsic value - not even desire fulfillment.
However, there are two types of value that do exist.
An agent with a desire that P has a motivating reason to realize a state of affairs S where P is true in S. Another way of saying this is to say that the agent values S, or S had value for this agent. This type of value exists. It is real.
This is consistent with Geoff's claim that the goal of realizing a state of affairs without an accompanying desire is not possible. In a broad sense, this is true - a desire gives goals to the person that has the desire. A desire that P gives the agent a goal of realizing states of affairs in which P is true.
The second type of value is instrumental value. Let us say that some tool T is useful in realizing S where P is true in S. The agent with the desire that P has reason to value T as a tool - an instrument - useful for bringing about S.
If the agent wants to eat the contents of a can of tunafish, he has reason to value a can opener - not because the can-opener itself fulfills any desire, but because it us useful in creating a state of affairs (an open can of tuna fish) where the agent's desire can be fulfilled.
Desire fulfillment is like everything else in the universe. It has value when desire fulfillment is the object if a desire (Sam desires that Julie's desires are fulfilled), or when desire fulfillment is a useful tool. (If Sam fulfills Julie's desire and Julie knows it, then Julie will perform some action that will fulfill Sam's desire.)
However, the value of desire fulfillment is of little interest in this theory (except to deny that it has intrinsic value). What desirism focuses on is the value of various desires - particularly malleable desires subject to modification using social forces such as praise, condemnation, reward, and punishment.
The two ways in which all things can have value are the two ways in which desires can have value. A desire can fulfill another desire directly (Sam desires that Julie desires to spend time with Sam). A desire can have instrumental value. (Given that Sam desires to have a car delivered to San Francisco, if Julie desires to drive to San Francisco, then Sam will find that desire useful in getting his car delivered there.)
I have a desire to retire - to quit my mundane job and research and write blog posts full time. Towards this end, I put money in a retirement account. It would do me no good if others routinely went into my account and took that money for their own purposes. Consequently, I would find a widespread aversion to going into other people's accounts and taking their money to be useful. I have reason to promote such an aversion.
At the same time, I find myself surrounded by people who would also benefit from a widespread aversion to taking what belongs to other people. This not only applies to retirement accounts. They benefit by others having an aversion to taking the car that they use to get back and forth to work, or taking the contents of their refrigerator, or taking the tools that they use in the fulfillment of their desires. At the same time that I have reason to promote a widespread aversion to taking the property of others, I am joined by countless others who also have reason to promote this aversion.
In our common language, we come up with a set of terms to use to identify desires that people generally have the most and strongest reasons to promote. We also come up with a language that uses praise and condemnation to promote the desires we have reason to promote, and inhibit the desires we have reason to inhibit.
We include in these desires we have reason to promote an aversion rape and, in fact, an aversion to acts without acquiring the consent of those who would be directly affected. We add an aversion (with some exceptions) to killing and assault and an aversion to lying or "bearing false witness". We promote a desire to help those in immediate need of assistance and an aversion to breaking the rules for those institutions that have rules (e.g., games - where "cheating" is a moral crime and "cheater" an accusation of blame and condemnation). We include an aversion to causing unintended harm and condemn those who lack this aversion as negligent or reckless, and we add an aversion to taking more than one's share.
When we do all of these things, we end up with a moral system.
It is a system that substantially evaluates desires according to their instrumental value and uses social tools to promote those that are useful while inhibiting or blocking those that tend to be harmful or dangerous. The motivation for doing this does not come from any sort of intrinsic value property. It comes from the desiresthat would be fulfilled by a desire that tends to fulfill other desires, and from desires that would be thwarted by desires that tend to thwart other desires.
There is no God. There is no natural law. There are no intrinsic values, categorical imperatives, social contracts, impartial observers, or committees hiding behind a vail of ignorance deciding what is just.
There are desires. A desire that P gives an agent a motivating reason to realize states of affairs where P is true. In doing so, the agent is not responding to any type of value property. The motivational force is programmed directly into the desire. A "desire that P" simply is a way of organizing brain matter such that the agent is motivated to realize states of affairs in which P is true.
Evolutionary forces have molded these desires over hundreds of millions of years. We tend to have desires that motivate us to act in ways that, in the past, have resulted in genetic replication. (Note: reproduction is only one form of replication. Many animals - e.g., most ants in an ant colony - replicate without reproduction.)
The evidence is clear that in humans (and most complex animals) evolution has favored a malleable brain - a brain that can fit into various environments by learning a set of desires and beliefs appropriate for that environment. In other words, some of these desires are malleable, depending on the interactions we have with our environment.
Of these malleable desires, some tend to fulfill other desires while some tend to thwart other desires. The desires that tend to fulfill other desires are those that people generally have reason to promote. Those desires that tend to thwart other desires are those that people generally have reason to inhibit.
The tools for promoting and inhibiting desires include praise, condemnation, reward, and punishment.
Read the paragraphs above again and let me know, if you will, if there is any need to insert anywhere any type of intrinsic value property - or any of the items such as categorical imperatives or a God that I mentioned in my list of fictions? We can see, as we go through the list, why any of these might be tempting theories. Social contracts and categorical imperatives, for example, certainly appear to be going in the right direction. Intrinsic value theories that put the intrinsic value in "pleasure", "happiness", "preference satisfaction", or "the well-being of conscious creatures" go in the direction of the truth. However, this account will show where these alternatives fall short, and how to cover the rest of the distance without them.
Why do (should?) we care about desires that tend to fulfill other desires?
Because they are useful. Because they 'tend to fulfill other desires'. Those 'other desires' they tend to fulfill provide the reason to care.