Networks, Dynamics and Explanation

This paper explores some issues having to do with the use of networks in scientific explanations. It focuses on the very common case in which what is of interest is the spread of some process (a disease, a neural signal etc.) along a network. In such cases, the use of a network in explanation requires the specification of a dynamics governing this process in addition to and independent of the network structure. Such a dynamics will incorporate causal information. This is one of several reasons why it is a mistake to think of network explanations, at least in typical applications, as entirely non-causal. In addition the independence of the network structure and the dynamics of the process occuring on it provides the key to the "directional" features of such explanations. Other topics discussed include the circumstances in which use of networks is most likely to be fruitful and the difference between networks in the form of undirected graphs and the use of directed graphs to represent causal relationships