“The point of departure is the measurement problem, as it appears in physics; the manner in which measurements allow us to characterize subsystems; the role of such subsystems as tools in system analysis; and the relationships existing between different ways of perceiving or interacting with the same system. Our conclusions are: (1) there exists no universal family of of analytic units appropriate for the treatment of all interactions; (2) there are on the contrary many such families of analytic units, all of which are equally “real” and entitled to be treated on the same footing; (3) the appropriate use of natural interactions can enormously extend the class of physical observables accessible to us; (4) the concept of a model must be formulated, in its most general terms, as the sharing of a subsystem by two otherwise distinct systems, capable of imposing the same dynamic on an appropriate system with which they can both interact. We establish these results through a variety of terminologies which turn out to be equivalent: stability, invariance, symmetry, homeostasis.”

Introduction; Quoted in: &quot; Fundamentals of Measurement and Representation of Natural Systems by Robert Rosen http://www.panmere.com/?page_id=15&quot; at panmere.com. <br class="br">Fundamentals of measurement and representation of natural systems. (1978)