# Interaction Involves Production

Instead of taking the symmetry of action and reaction to show that causation cannot be modelled in terms of interaction—because it depicts the relation between cause and effect as symmetric—I propose we take it to show that the reaction cannot really be an effect, and that the action cannot really be a cause in its own right. Really, I propose we take it to show that we must re-examine the standard conception of ‘cause’. By abandoning the standard view that causes must be understood in terms of‘extrinsic motive Agents’, an idea arguably originating from outmoded pre-Newtonian physics, we are free to conceive of the interaction as a whole as the cause, and the change in the compound whole of interacting things as the effect. According to this view, the relation between action and reaction is symmetrical, but it does not follow from this that the relation between the cause (i.e. the interaction) and the change it produces is symmetrical. Bunge overlooks this possibility because he takes causes to be by definition external to the changing entity.

It is in fact difficult to conceive of interactions without thinking of them as the production of changes in the interacting things, when, as here, interaction refers to the mutual influence of two things upon each other. The notion of force has always been understood in terms of production of changes. Newton defined it as ‘an action exerted upon a body in order to change its state, either of rest or of uniform motion in a straight line’ (1686: 13), and Hertz defined it as ‘the independently conceived effect which one of two coupled systems [...] exerts upon the motion of the other’ (1956: 185). If force, meaning the action of one thing upon another, cannot be separated from the changes it produces, then neither can interaction be separated from the production of changes, since an interaction consists in two reciprocal forces. That the forces are reciprocal means that neither can exist without the other, and that, therefore, neither can produce the other. However, together they can produce a new state of affairs.

It may be objected that static interactions are counterexamples to the claim that interactions always result in changes (see, for instance, Baltimore 2020: sect. 4.2). Two bricks leaning on each other exert a force on each other, and yet nothing changes. Baltimore suggests we instead use the term ‘result’ to describe the outcome of interactions. Now, while nothing important hinges on this point—the use of ‘result’ rather than ‘change’ still admits that interactions are productive—I still am inclined to insist that the result be understood as ‘change’. Many of the ‘results’ we would prima facie identify as non-changes, because no change turns up in our experience, seem to me to be changes. Consider the two bricks leaning on each other. According to physics, each brick is in a state of downwards motion; they are falling, just like any object resting on the surface of the Earth. However, that falling is continuously thwarted by the other brick. From a human observer’s point of view nothing appears to change, but in physical terms, their state of downwards motion is continuously being changed to not falling. Now, whether or not this account of static interactions holds true, I take it still to be the case that:

(P3): Interaction involves production.