Interventions on causal exclusion

Abstract

Two strains of interventionist responses to the causal exclusion argument are reviewed and critically assessed. On the one hand, one can argue that manipulating supervenient mental states is an effective strategy for manipulating the subvenient physical states, and hence should count as genuine causes to the subvenient physical states. But unless the supervenient and subvenient states manifest some difference in their manipulability conditions, there is no reason to treat them as distinct, which in turn goes against the basic assumption of nonreductive physicalism. On the other hand, one can preserve the distinction between the two by introducing asymmetric manipulability conditions that the supervenience thesis entails. But this response can be used to argue that mental causes never have physical effects. However, this argumentation can also be used to show that mental causes can have mental effects.

Keywords:

• downward causation
• interactionism
• multiple realization
• nonreductive physicalism
• parallelism

Acknowledgements

I am grateful to the editorial board and an anonymous referee of Philosophical Explorations for their critical comments on an earlier version of this paper. This study has been financially supported by the Emil Aaltonen Foundation, the Ella and Georg Ehrnrooth Foundation, the Finnish Cultural Foundation, the Waldemar von Frenckell Foundation, the Alfred Kordelin Foundation, the Otto A. Malm Foundation and the Oskar Öflund Foundation.

Notes on contributor

Tuomas K. Pernu is at the Department of Biosciences, University of Helsinki. He has published articles on philosophy of mind, epistemology and philosophy of science.

Notes

1. These counterfactuals should be interpreted manipulatively: given V, if you would set the value of P₁ to 1, you would also set the value of to 1, whereas setting the value of M to 1 would not necessarily have the same result. There is also a way to interpret these counterfactuals in terms of possible world semantics. Following List and Menzies (2009) we can say that the causal relationship between M and M* is “realization-insensitive”, since M* is present under some small perturbations in the realization of M. More specifically, if M is realized by P₁, then given the assumption of realization-insensitivity, according to how List and Menzies (2009) define the notion, M* is present in some closest non-P₁-worlds (i.e. P₂-worlds) that are M-worlds. Now extend the original model proposed by List and Menzies (2009) to take into account the realization-insensitivity with respect to effect as well: if M* is realized by , then given the assumption of realization-insensitivity, M is present in some closest (i.e. ) that are M*-worlds. Suppose that both M and M* occur in the actual world and the relationship between them is realization-insensitive, both with respect to M and with respect to M*, in the sense that M is realized by P₁ and M* is realized by , but they could have been equally well been realized by P₂ or by , respectively. Now the set of relevant alternative possible worlds would seem to contain, not only (M & P₁)-worlds and (M & P₂)-worlds, but also (M* & )-worlds and (M* & )-worlds. What this means is that an inference for the counterfactual “if M would hold, then would hold” is not licensed in the actual world, even though both M and occur in it, since there is a relevant alternative world where both M and M* occur but fails to occur. The details of this reasoning would fall out of the scope of this paper.

2. I am indebted to an anonymous referee of this journal for presenting this caveat.