Causality as Constraint Introduction

Suppose you are trying to arrange furniture you already own in your new house. "Each piece of furniture has some weak constraints associated with it. For example, the bed must go against a wall because the headboard is rickety; two small tables go on either side of the bed; the couch is missing its back legs and so sits on books, which means it must go against a wall so that the back is not visible, and so on. The constraints associated with each piece of furniture are weak because each one can be satisfied in many ways; the bed, for example, can be put against any of several walls. However, once the bed is placed, the entire arrangement may be determined: there may be only one other wall large enough for the couch, the tables must go next to the bed, and so forth" (Kosslyn & Koenig, 1992, p. 111). How are we to understand the relationship between the constraints and the particular furniture arrangement that results? Despite deliberately sidestepping causal language in formulating that question, it is difficult to avoid the sense that "simultaneous constraint satisfaction" as occurs in Kosslyn and Koenig's example is somehow responsible for the way the furniture gets arranged. This example suggests that constraints satisfy at least one oft-touted requirement of causal relationships: constraints support what philosophers call "counterfactual conditionals", that is, if it were not for the constraints the resulting furniture arrangement would not be what it is. And yet it is clear that the constraints don't cause the furniture arrangement the same way cue sticks cause cue balls to move. As Salthe says, although constraints carry information that has meaning for the entities involved in natural processes, the information is not dynamically involved in the processes (Salthe, 1985, p. 71). The concept of constraint was first used formally in the Physical Mechanics. Although never actually defining it, Lindsay uses the term in his introductory textbook first to refer to the way in which the motion of a simple pendulum or a particle on an inclined plane (Lindsay, 1961, p. 35) is "compelled by the geometry of its environment to move on some specified curve or surface" (ibid., p. 239, my emphasis). Later on, in his explanation of oscillations, Lindsay states that "some of the most important cases of constrained motion are those in which particles are connected by rods and strings", and cannot, therefore, move any which way. In his explanation of D' Alembert's Principle, Lindsay states that "If the masses were subject to no constraints (i.e. if they were not connected in any way or forced to move along certain curves or surfaces) ... " (ibid., p. 251); and in his discussion of Gauss' Principle of Least Constraint Lindsay states, "The system being subjected to certain constraints (i.e. the masses being perhaps connected to each other by rods or cords, or constrained to move along certain curves or surfaces) ... " (ibid., p. 254). I assume that in the last clause the word "constrained" means, as in the earlier quotes, either