Causation in a Physical World

Causation in a Physical World

Thomas Blanchard

The problem of causation in the physical world is the challenge of reconciling the a-causal physical picture of the world with the need for causation.

Chapter 1 describes the problem in detail and proposes a new methodology to solve it. The proper method to handle the problem isn’t conceptual analysis. Rather, solutions to the problem should be judged on how well they physically explain actual facts about effective strategies.

Chapter 2 examines the main attempts to solve the problem. I argue that they all face various problems. In particular, current attempts to locate causation in the physical world all have trouble making sense of the fact that we need causal knowledge to make rational decisions. To solve this problem, the first step is to provide a satisfactory explanation of why only those correlations that are (intuitively) causal can be exploited for the purpose of securing desired outcomes.

In chapter 3, I propose such an explanation. I argue that causal correlations are the only ones that can be exploited according to evidential decision theory(EDT). This is a surprising claim, since EDT is widely thought to recommend acting for the sake of outcomes one cannot cause. I argue that this is actually not the case, and that EDT in fact provides a plausible account of exploitable correlations.

In chapter 4, I use this account to offer a new solution to the problem of causation in a physical world. I argue that causal dependence is a matter of the cause and the effect standing in certain probabilistic relations to a third event called a probabilistic intervention. Probabilistic interventions are events that need not involve agency but nonetheless mimic certain crucial features of deliberation. I argue that this account provides a plausible solution to the problem of causation in a physical world.