Description
A central problem in the philosophy of mind is to figure out if and how mental states can ever be causes in a physically closed world. A simple example suffices to illustrate the problem.Say Sarah has the mental state of desiring a coffee, and this desire causes her to walk to the coffee shop. Given physicalism, which is the thesis that all of reality is ultimately grounded in physical states, her mental state supervenes on some physical state (for example the firing of a particular pattern of neurons in her brain). This physical state is also a cause of her walking to the coffee shop.
The non-reductive physicalist holds that the mental state supervenes on the physical state, but is not reducible to it. Therefore she seems forced to accept that Sarah’s walking to the coffee shop has two distinct sufficient causes, and is thus overdetermined. But overdetermination of an event by two causes is deemed to be highly exceptional, to the extent that it is highly implausible to be as widespread as such mundane cases of mental causation are.
This problem has been dubbed the exclusion argument, and is most famously presented by Kim (1998, 2005). The exclusion argument presents a challenge for non-reductive physicalism. By now, there is widespread consensus that this challenge takes the following form: given non- reductive physicalism, overdetermination of one event by two cases is widespread (Bennett, 2003; Hitchcock, 2012; Woodward, 2008, 2015; Weslake, Forthcoming). One type of overdetermination, namely the type that we observe in examples like that of a firing squad, is extremely rare. Therefore the non-reductive physicalist owes us an account of a second type of overdetermination that need not be rare. Concretely, the challenge lies in formulating both types of overdetermination in a manner that allows us to unambigu- ously distinguish the rare type from the widespread type. This is the challenge I will take up.
So to be clear, this paper offers no direct defense of non- reductive physicalism, but rather starts by simply assum- ing non-reductive physicalism. It does offer an indirect defense, in that if one is already inclined towards non-
reductive physicalism, then I will show that the exclusion argument doesn’t offer you reason to abandon it. Further, the assumptions here made to characterize non-reductive physicalism are non-committal, in that they are compatible with various versions of that position. In particular, I will remain silent on whether causal relata are events or properties, and on whether a mental event is distinct or not from a physical event.
The only additional assumption I will make is that causation is best captured by considering causal models. In particular, causation is best defined in terms of a relation that holds between values of variables within some model. Although I have proposed such a definition elsewhere, this paper is compatible with any definition of causation formulated in terms of causal models, as long as it considers cases of overdetermination to be cases of causation. (This is harmless assumption, since the exclusion argument rests on it.) In particular, it should consider cases of overdetermination to be cases of sufficient causation, for this is the type of causation that is at stake in the exclusion argument. Hence all talk of causation should in fact be read as talk of sufficient causation throughout.
First I will work towards defining two distinct types of overdetermination, where the one type captures cases of mental causation, like walking to the coffee shop, and the other type captures cases that are rare, like the firing squad. To get there, I present a formalization of the notion of abstraction, meaning the relation that holds between a causal model made up of high-level variables and a causal model made up of low-level variables such that the high-level variables represent properties that supervene on the properties represented by the low-level variables. Second I formalize the key assumptions of non-reductive physicalism using the vocabulary of abstraction and causal models. Finally I formalize the exclusion argument, and prove that if one ac- cepts the introduced distinction between the two types of overdetermination then it is not valid.
| Period | 9 Nov 2018 |
|---|---|
| Event title | OZSW 2018 |
| Event type | Conference |
| Degree of Recognition | International |