I've been thinking about the collapse objection to rule-consequentialism because there's an objection in Maitzen's PPR article from 1995 (?) where he suggests that Goldman's 1986 account of justification might be susceptible to a similar style objection.
Step One: Suppose our epistemic aim is something in the neighborhood of this: maximize true belief while minimize false belief.
Step Two: If this aim, the "nominal aim", is the reason for having/pursuing justification, it ought to follow that:
(i) beliefs are justified if they serve the aim;
(ii) beliefs are unjustified if they don't serve the aim.
Step Three: However, any true belief will serve the aim and will thereby count as justified by (i) and any false belief will fail to serve the aim and will thereby count as unjustified by (ii).
So, Maitzen recommends revising the standard view about our epistemic aim.
I find this talk of the the reason for having justification strange, but let's set that aside for now. I'm more bothered by the idea that all TBs are JBs than all FBs are not JBs, but it seems to me that there's a way for Goldman to deal with these sorts of worries. Following Hooker's lead, can't Goldman formulate his rule-consequentialist approach to justification in terms of internalized rules rather than rule conformity? I thought the way that the collapse argument typically went was something like this. Suppose the rule-consequentialist says that an act is justified if it is permitted by a set of rules, R1, such that there's no other set of rules, R2, such that conforming to R2 is better from the point of view of utility than conforming to R1 is. Then, the objection continues, that conforming to the rule that the act-utilitarian sets down will always be at least as good or better than conforming to any alternative set of rules and so the rule-consequentialist view is either incoherent or extensionally equivalent to act-consequentialism. Suppose we say instead that an act is justified if permitted by a set of rules, R1, such that there's no other set of rules, R2, such that if R2 were internalized rather than R1, agents who internalized these rules would bring about more utility than those who internalized R1. Then the idea might be that we bring about more utility by internalizing some set of rules that is distinct from the act-utilitarian rule. Perhaps Goldman can avoid Maitzen's objection by opting for something along the lines of Hooker's account.
I confess that I still have a persisting worry about Hooker's account, which is that the view seems hard to motivate. I guess I can't say that the view is incoherent, but I worry that on purely value-theoretic grounds there is little that recommends it and I don't see any alternative way of arguing that this kind of rule-consequentialism is preferable to alternative consequentialist views.