When a logic program is processed by an answer set solver, the first task is to generate its instantiation. In a recent paper, Calimeri et el. made the idea of efficient instantiation precise for the case of disjunctive programs with function symbols, and introduced the class of “finitely ground” programs that can be efficiently instantiated. Since that class is undecidable, it is important to find its large decidable subsets. In this paper, we introduce such a subset—the class of argumentrestricted programs. It includes, in particular, all finite domain programs, ω-restricted programs, and λ-restricted programs.