A deliberative account of causation: How the evidence of deliberating agents accounts for causation and its temporal direction

Abstract

In my dissertation I develop and defend a deliberative account of causation: causal relations correspond to the evidential relations we use when we decide on one thing in order to achieve another. Tamsin’s taking her umbrella is a cause of her staying dry, for example, if and only if her deciding to take her umbrella for the sake of staying dry is adequate grounds for believing she’ll stay dry. I defend the account in the form of a biconditional that relates causal relations to evidential relations. This biconditional makes claims about causal relations, not just our causal concepts, and constrains metaphysical accounts of causation, including reductive ones. Surely we need science to investigate causal structure. But we can’t justify any particular account of causation independently of its relevance for us. This deliberative account explains why we should care about causation, why we deliberate on the future and not the past, and even why causes come prior in time to their effects. In chapter 1 I introduce the motivations for the project: to reconcile causation and our freedom as agents with the picture of the world presented by physics. Fundamental physics makes no mention of causes. And the lawlike character of the world seems to rule out freedom to decide. My dissertation offers a combined solution—I explain our freedom in epistemic terms and use this freedom to make sense of causation. In chapter 2 I draw on philosophy of action and decision theory to develop an epistemic model of deliberation, one based in requirements on belief. If we’re to deliberate, our beliefs can’t epistemically settle how we’ll decide, yet our decisions must epistemically settle what we’ll do. This combination of belief and suspension of belief explains why we rationally take ourselves to be free to decide on different options in deliberation. In chapter 3 I defend this model from near rivals that also explain freedom in terms of belief. Accounts of ‘epistemic freedom’ from David Velleman, James Joyce and Jenann Ismael appeal to our justification to form beliefs ‘unconstrained’ by evidence. Yet, I will argue, these accounts are susceptible to counterexamples and turn out to rely on a primitive ability to believe at will—one that makes the appeal to justification redundant. J. G. Fichte’s Idealist account of freedom, based in a primitive activity of the ‘I’, nicely illustrates the kind of freedom these accounts rely on. In chapter 4 I develop the epistemic model of deliberation into a deliberative account of causation. I argue that A is a type-level cause of B if and only if an agent deciding on a state of affairs of type A in ‘proper deliberation’, for the sake of a state of affairs of type B, would be good evidence of a state of affairs of type B obtaining. This biconditional explains why we should care about causal relations—they direct us to good decisions. But existing accounts of causation don’t adequately explain why causation matters. James Woodward’s interventionist account explicates ‘control’ and ‘causation’ in the very same terms—and so can’t appeal to a relation between them to explain why we should care about causal relations. David Lewis’ reductive account relies on standards for evaluating counterfactuals, but doesn’t motivate them or explain why a causal relation analysed in these terms should matter. Delivering the right verdicts is not enough. The deliberative account explains why causation matters, by relating causal relations to the evidential relations needed for deliberation. In chapter 5 I use the deliberative account to explain causal asymmetry—why, contingently, causes come before their effects. Following an approach from Huw Price, because deliberation comes prior to decision, deliberation undermines evidential relations towards the past. So an agent’s deciding for the sake of the past in proper deliberation won’t be appropriate evidence of the past, and backwards causation is not implied. To explain why deliberation comes prior to decision, I appeal to an epistemic asymmetry, one that is explained by statistical-mechanical accounts of causation in non-causal terms. But statistical- mechanical accounts still need the deliberative account to justify why the relations they pick out as causal should matter to us. The deliberative account of causation relates causal relations to the evidential relations of use to deliberating agents. It constrains metaphysical accounts, while revealing their underlying explanatory structure. And it does not rule out explanations of causal asymmetry based in physics, but complements them. Overall this project makes sense of causation by foregrounding its relevance for us.