Causality, computing, and complexity

I discuss two categories of causal relationships: primitive causal interactions of the sort characterized by Phil Dowe and the more general manipulable causal relationships as defined by James Woodward. All primitive causal interactions are manipulable causal relationships, but there are manipulable causal relationships that are not primitive causal interactions. I’ll call the latter constructed causal relationships, and I’ll argue that constructed causal relationships serve as a foundation for both computing and complex systems. Perhaps even more interesting are autonomous causal relationships. These are constructed causal relationships in which the causal mechanism resides primarily in the effect. A typical example is a software execution engine. Software execution engines are on the effect side of a cause-effect relationship in which software is the cause and the behavior of the execution engine is the effect. The mechanism responsible for that causal relationship resides in the execution engine. Introduction Causality has been a longstanding and controversial topic in philosophy. I think it’s fair to say, though, that Phil Dowe’s primitive causal interactions and James Woodward’s manipulable causality currently reign as the primary ways to think about philosophical causality. See the section below on actions and events for some additional background. Primitive causal interactions Primitive causal interactions correspond to direct interactions—the sorts of interactions one associates with billiard balls hitting one another. (Dowe 2000) defines a [primitive1] causal interaction as follows. An object is anything found in the ontology of science (such as particles, waves or fields), or common sense (such as chairs, buildings, or people). A world line is the collection of points on a space-time (Minkowski) diagram which represents the history of an object. A causal interaction is an intersection of world lines which involves [an] exchange of a conserved quantity. For our purposes one can think of a primitive causal interaction as occurring at a single moment in time and at a single point in space.2 Manipulable causal relationships A manipulable causal relationship is defined by Woodward (2003, 2012/forthcoming), as follows.3 1 Dowe uses the term causal interaction rather than primitive causal interaction. He entitles his book Physical Causality reflecting his commitment to fundamental physical interactions. Hence my use of the term primitive. 2 Quantum interactions do not occur at a single moment in time or at a single point in space. They occur through an exchange of virtual particles. Emission and absorption occur at different moments and different locations. But this issue will not be relevant for this paper, which will be concerned with macro-level interactions. 3 Woodward’s general definition is expressed in terms of probability distributions rather than just values.