Non-coincidental overdetermination as a solution

In document The chances of higher-level causation: an investigation into causal exclusion arguments (Page 114-118)

Chapter 2: Kim’s causal exclusion argument against non-reductivism

2.4 Exclusion and no overdetermination

2.4.1 Non-coincidental overdetermination as a solution

Unfortunately, the motivation given for this version of the principle is not satisfactory to deny overdetermination in the case of mental causation. According to some philosophers (see e.g.: Bennett 2003), inter-level overdetermination cannot be modelled and rejected based on our knowledge of intra-level overdetermination, as there are important disanalogies. The supervenience relation that holds between the physical realizer and the realized mental property makes the occurrence of overdetermination systematic. If we accept the supervenience of the mental on the physical, then in all cases of mental causation we get two

causes for physical outcomes such as my typing of this text. A mental cause (my intention) and a physical cause (my brain state). These two always take place at the same time. This means that overdetermination might not only be a highly systematic phenomenon, but a universal and necessary one, at least in this special kind of case. This is the reason why I will call it non-coincidental43. The supervenience of the mental on the physical is taken to be at

least nomologically necessary by non-reductive physicalists, but all supporters argue for stronger forms of necessity than that. This means that in worlds with the same laws as our world there is no way to separate the physical realizer and the realized mental property, but for many physicalists the separation is impossible even beyond this set of worlds. As a result, for supporters of non-reductive physicalism inter-level overdetermination is a strong necessity in the context of mental causation.

This allows for an argument against the principle of no overdetermination according to which such principles cannot be transferred from well-known cases of overdetermination to cases of overdetermination between levels. There is an important disanalogy between everyday or intra-level overdetermination scenarios and inter-level scenarios as in the case

43 In ordinary causal discourse the term „non-coincidental” means at least the following: the occurence of two

events is coordinated by a (probably not too distant) common cause. So, their co-occurence is not a coincidence, but it could be a coincidence, this is the reason why it is interesting to know which holds true in a particular case. In the case of inter-level overdetermination the connection between the mental and the physical is much stronger than that. Depending on how the supervenience relation is spelled out for mental properties the modal strength of the covariation might change, but any supervenience relation that is necessary for physicalism provide a stronger connection between the correlated causes then a common cause can. Redundant causes brought about by a common cause, as in the case of the two assassins and the dictator, are separable in our world. In some possible cases, the same kinds of causes occur in an unconnected manner. E.g. two assassins who don’t know about each other and have different motivations for killing the dictator. But mental causes always go together with physical causes. So, their co-occurence is non-coincidental in the stronger sense that in our world it cannot be a coincidence.

of mental causation. Therefore, we can’t simply model mental causation on intra-level causation.

Overdetermination might be rare in intra-level cases, but from this nothing straightforward follows for inter-level cases. Systematic overdetermination is weird in an everyday context as it involves surprising coincidences or very specific conditions that easily fail in practice. But inter-level overdetermination is a different breed. It involves no coincidences and the supervenience relation provides a neat explanation for the systematic occurrence of overdetermination(see Bennett (2003) or Sider (2003) for versions of this basic idea). As this approach provides us with a mechanism that produces systematic overdetermination it seems hard to deny the possibility.

A lot depends on the no overdetermination premise. If it proves to be a solid premise (together with causal closure and supervenience, premises never questioned by non- reductive physicalists), it is reasonable to turn the exclusion argument against the distinctness premise as suggested by Kim. If the above argument proves to be plausible it seems to be a promising way out for non-reductive physicalists. What can be put forward to answer the above argument? Why shouldn’t we accept that overdetermination can be systematic in some cases even though it is not in everyday causal scenarios?

On the one hand, the inter-level dependence between the mental and physical explains systematic overdetermination. However, on the other hand, if all supervenient causes are redundant causes then we have endless amounts of excess causes in our universe and that sounds like a design failure44. Why would God or any other kind of designer need

those redundant causes? Why would there be such an unparsimonious mechanism?

44 If there are other layers of reality between the mental and the basic physical where autonomous causation

A possible, but dubious answer would be that these redundant overdeterminers are some kind of backup causes. This means that they are there to secure certain outcomes in our universe. However, redundant higher-level causes supervening on lower-level realizers are ill-suited for such purposes. To be proper backup causes their function should be something like to make sure that certain effects take place even when something goes wrong on some basic or default causal track.

In the assassin scenario the presence of more than one shooter aims to make sure of the death of the dictator. If there is a problem with a rifle or one of the shooters gets caught by security officers, there is still a sufficient backup option activated at the right time. Redundant higher-level causes can’t serve as backup causes in this sense. A backup cause should be able to stay active when the default cause gets deactivated, otherwise it can’t serve as a backup for the default cause. But, according to the supervenience based picture of the world, tokens of higher-level causes disappear whenever no token of their physical realizers is present. Therefore, higher-level causes can’t provide backup for physical causes.

Can physical causes serve as backups for mental causes supervening on them? No, they can’t and again it is the supervenience relation that makes this impossible. When a token of their physical realizers is present, a token of mental cause is present as well. If in our world the mental supervenes on the physical, at least with nomological necessity, then in this world the mental has no chance of disappearing and leaving its realizer bare. When it disappears, the realizer disappears as well. Therefore, in worlds like our mental causes cannot be backed up by physical causes.

Redundant causation as in the case of the two assassins is predicated on the premise that overdeterminer events one and two are separate, non-overlapping existences. This is what makes overdetermination scenarios either a rare coincidence or a carefully contrived,

but still fragile situation. But exactly because of their separateness they can work as backups when the situation is contrived so.

Redundant causes in the case of lower and higher-level causes are asymmetrically dependent. Higher-level causes depend on lower-level causes. For this reason, they can’t work as backup causes for a single outcome. They are useless as redundant backup causes similarly to God’s help according to the saying: “God helps those who help themselves”. God (the mental) never causes physical outcomes directly only when some physical cause (the believer/human) causes it already. Why do we need God in that case?45

So, one has to pay a high price for accepting the idea that the non-coincidental relation between causes at different levels explains systematic overdetermination. It results in a highly unparsimonious picture of the world. In the absence of serious further positive reasons for it, we have a good reason to reject it. Simply insisting that non-coincidental overdetermination is a possibility makes the solution ad hoc as well. Nevertheless, there are further arguments for and against inter-level overdetermination.

In document The chances of higher-level causation: an investigation into causal exclusion arguments (Page 114-118)