The present article is the second part of a longer paper in which we outline a model of (scientific) method as a system of instructions aimed at a certain kind of (cognitively interesting) goal. The article offers a detailed explication of the notion of instruction in terms of binary relations between certain kinds of states. Instructions are taken as imperatives and their role is associating input states with output ones. In particular, instruction φ! associates an input state in which it is not the case that φ with an output state in which it is the case that φ. An important distinction between instruction and its occurrence is introduced. It enables us to recognize certain kinds of transitions from input states to output states and vice versa, namely the derivate transition and the postulate transition, as well as to define certain kinds of relations between occurrences, such as their continuity or mutual independence. A classification of instructions by their logical forms (namely, categorical and hypothetical ones) and by their subject-matters (namely objectual, conceptual and propositional ones) is proposed as well.