The systematic revision of practices is challenged by their observer dependent nature. Unlike natural sytems, systems of practice do not exist independently of the conscious attitudes of human observation. For this reason descriptions of their causes appear multiple and indeterminate by comparison with the systematic descriptions of natural entities. Features of the world such as money, property, marriage, government and football games require conscious observers and agents in order for them to exist as such. To this can be added the domains of art, design and architecture. These observer dependent domains are nevertheless complex systems that are open to subtle explanation and revision. Unlike observer independent objects, however, the relationships among the agencies which bear down upon practice are governed by ethical rules and conventions of what is good, what is best, and what works. These principles are given to the agencies of practice not as causes but as meaningful functions.The paper critiques action research and similar models of investigation into practice. It argues that action research is compromised on the one hand by being too scholastic while, on the other hand, by being too caught up in practical ends. Action research understands practice as a "project" with a tendency to project scholastic values into its understanding of practice. Scholastic values privilege the "truth" of rational over irrational transactions underlying the processes of practical production.This paper advances a functional account of artistic practice. The model integrates the theoretical work of Cornell realist Richard Boyd with Gilles Deleuze's systematic interpretation of Foucault's theory of knowledge-power. Rather than a linear process that proceeds from inception to outcomes, artistic practice is portrayed as a set of interdependent agencies held in homeostatic relation. Using examples from practice in the fine arts and design the paper goes on to identify various agencies and the terms of their functional relations. Outcomes of research into practice are explained as the disclosure and evaluation of the systematic relations among its functions.