Causation in Experimental Physics

Enno Fischer (postdoctoral project)

This project explores the roles of causal reasoning in experimental physics. In particular, the project addresses the following two questions. First, what kinds of causal claims are being employed in experimentation? I examine three distinctions between kinds of causal claims. First, do physicists try to identify the actual causes of measurement results and phenomena or are they satisfied to identify potential causes? Sometimes measurement deviations or surprising phenomena motivate a systematic search for underlying actual causes such as in the discovery of X-Rays. Sometimes, however, physicists only refer to potential causes, for example, when a deviation is within the limits of the errorbars. Second, a common view is that physicists are interested in general causal claims. But there are also singular causal claims that are relevant for physicists, such as those associated with primordial cosmology. What is the difference between these claims? How, if at all, do these differences affect experimental inquiry? The third distinction concerns salient causes and background conditions. On the one hand, causal reasoning in physics is often considered to be unaffected by this distinction. On the other hand, experimental physicists rarely assume exhaustive sets of initial conditions when they model their experiments. In fact, it has also been argued that experiments are designed to enable selective causal reasoning (Hanson).

The second question is: what theories of causation capture the kinds of causal reasoning that are relevant in experimentation? On the one hand, a natural suggestion is that experimental physics employs process theories of causation as proposed by Salmon and Dowe. These are theories that define causal processes in terms of the transmission of conserved quantities. On the other hand, there are difference-making theories of causation, such as Woodward’s interventionist account. The interventionist account seems to be particularly suited because the involved notion of intervention is modelled on the paradigm of an experimental intervention. A third class of accounts that seems particularly promising are those that describe causes in terms of factors that disturb a system’s (quasi-)inertial behaviour as proposed, for example, by Maudlin. These accounts reflect the fact that experimentation often involves a systematic inquiry into the causes of deviations from a system’s predicted behaviour.