Kim and Downward Causation
What underlies Kim’s argument for the epiphenomenalism of higher-level mental properties is a view of reality as constituted by a hierarchy of different strata, starting from a ‘lowest’ level of fundamental particles and going upwards through levels of progressively more complex phenomena. This ‘layered’ model of reality, Kim thinks, has formed “a constant”, though mostly implicit, “backdrop” to most contemporary discussions about reductionism and emergence.10 In this model, the higher-level phenomena are taken to ‘arise’ from and depend on the processes of the comparatively lower levels and ultimately, on processes at the fundamental level. This dependence is usually not taken to be of a causal kind, but to be a form of supervenience together with some form of asymmetric determination. Namely, it is not only held that there can be no difference in the mental phenomena without some difference in the physical phenomena, but also that the latter phenomena are responsible for the former being what they are, and not the other way round.
How the different layers in this hierarchy are meant to be distinguished from one another is a vexed issue, and it is highly doubtful whether we can expect anything as neat and tidy as a hierarchy where phenomena can be clearly attributed to such layers (rather than the layers’ blending into one another, as it were).11 But these difficulties can be bracketed for the context of our present argument. To use a rough and workable distinction between these layers, we will just assume that they more or less correspond to the subject-areas of different special sciences dealing with increasingly complex phenomena (such as microphysics of different degrees of complexity, inorganic chemistry, organic chemistry and biology etc.).
According to Kim, a ‘layered’ model of reality combines with two further naturalist theses to create a fundamental difficulty for the idea that mental properties and phenomena can both be distinct from physical ones and possess a genuine causal role. First, the thesis that the level of physical phenomena is ‘causally closed’, i.e., that it is governed by laws which only allow for causation by other phenomena at the same (i.e., physical) level. This ‘causal closure’ is often derived from a combination of the assumption that every physical event must have a sufficient physical cause and a rejection of causal overdetermination in cases of mind-body interaction. These two assumptions are meant to follow, on the one hand, from a basic commitment to physicalism, and, on the other hand, from the way we naturally view the relation between mental phenomena and the underlying physical phenomena when both are (supposedly) involved in the production of one and the same physical effect, e.g., a movement of my arm. Since every physical event already has a sufficient physical cause, attributing causal influence to a distinct mental event, such as my desire to raise my arm, would lead to causal overdetermination, and we do not usually assume that when I raise my arm because I want to raise it, my arm’s movement is overdetermined by a physical and a mental cause. So it seems that mental events are precluded from exerting any ‘downward’ causal influence on physical ones.
As I have just stated it, this objection to mental-to-physical causation only gets a grip if we focus on cases of deterministic causation, since only then can the effect have a ‘sufficient’ physical cause. When an effect has only indeterministic causes, by contrast, there will be no sufficient cause (physical or other) at all. The objection can be modified, though, to accommodate cases of indeterministic causation as well (e.g., one can argue that, insofar as in indeterministic causation the probabilities of a physical event’s occurrence are fixed, they must already be fixed by the occurrence of antecedent physical events).12 However, for ease of presentation, I will just stick to the case of deterministic causation here.
The second additional thesis Kim relies on leads to a generalisation of the causal impotence result to all forms of causal influence. It is the thesis that a higher-level (mental) phenomenon can only be caused by causing the lower-level (physical) processes from which it arises and on which it supervenes. Since the supervenience base is meant to be responsible for the occurrence of the higher-level phenomenon, you can cause the latter only by causing the former. Together with the consequence from physical closure that mental phenomena cannot cause physical ones, this rules out mental-to-mental causation, too, and makes the mental phenomena completely ‘causally impotent’. All the ‘real’ causal work is done at the physical level and the higher-level changes are just the epiphenomenal side-effects of what happens there.
As several philosophers have noted, Kim’s argument, if successful, not only applies to the relation between mental and underlying physical events. It is a completely general argument-scheme, which can be applied to the relation between any two layers of reality, whenever one is held to be more fundamental or basic than the other, and the latter supervenes on the former.13 For, with regard to any ‘lower’ or more fundamental level, one can argue for a principle of ‘causal closure’, which shields off the lower-level phenomena from causal interference from events ‘higher up’. If every lower-level phenomenon already has a sufficient cause at the same level, then, barring causal overdetermination, there remains ‘nothing left to do’ for higher-level phenomena to causally influence what happens at the lower levels. This would make higher-level phenomena causally impotent with respect to lower-level ones. And, if the only way to causally influence the occurrence of higher-level phenomena is to cause those lower-level phenomena upon which the higher-level ones supervene, this will mean that higher-level phenomena are generally epiphenomenal.
Once it is applied generally, though, this argument threatens to make any phenomena which are not situated on a fundamental level of reality causally impotent. Causality thus threatens to ‘drain away’, down to the lowest, most fundamental level of reality—and if there is no most fundamental level of microphysics at which to stop, it will ‘drain away’ indefinitely. 1 4 As a result, even on an optimistic assessment, practically all causal explanations we de facto use will turn out to be false (since we give extremely few—if any—such explanations in terms of phenomena at a fundamental microphysical level, if there is any such level at all). This very radicalism is one chief reason why many philosophers have felt that this argument must be flawed. Some have taken this consequence to amount to nothing less than a reductio of the argument, since they have not considered it to be a genuine option for us to give up our well-established explanatory practices wholesale.15
But even when they haven’t shared this reductio assessment, most philosophers have tried to resist the argument in other ways. There is an impressive list of arguments offered in the literature which try to show that we can admit mental (or macrophysical) causes without bringing in widespread causal overdetermination—e.g., because we should adopt a different understanding of causation than the one Kim relies on.16 But while many of these answers do significantly depart from Kim’s background picture of reality, most of the proposed solutions share with Kim the original set-up of a general Humean background, or at least do not hinge on rejecting it. They also usually do not assume that powers, construed realistically, play any essential role when it comes to explaining ‘cross-level’ causal influence, or that processes, as opposed to events, do. What I want to examine in the following sections is whether rejecting this background will change the prospects of Kim’s argument as well.