The Irreducibility of Downward Causation
Kim (1999) presents an argument against the irreducibility and novelty of downward causation that will be examined in some of the contributions. It rests on the idea that an emergent property and its lower-level, emergence base are both nomologically sufficient (i.e., sufficient given the actual laws of nature) for the production of the lower-level effect. In turn, this idea is motivated in two different ways. In Kim (1999), nomological sufficiency is taken to be a transitive relation: if a lower-level emergence base is nomo- logically sufficient for a higher-level emergent property (given that the latter emerges from the former according to certain laws of nature) and the higher-level emergent property is in turn nomologically sufficient for the lower-level effect (given that causation implies nomological sufficiency), then the emergence base is nomologically sufficient for the lower-level effect too. In other places (e.g., Kim 2005), the emergence base is nomologi- cally sufficient for the production of the lower-level effect if they are both (micro-)physical and the principle of the causal closure of the (micro-)physical realm is true. This principle asserts that every (micro-)physical effect has a wholly sufficient (micro-)physical cause, which can be identified here with the emergence base. At any rate, in both cases, downward causation turns out to be “redundant”: the lower-level entities are causally sufficient for all the lower-level effects.
There are three ways to comply with Kim’s arguments: (i) accepting them and holding that downward causation is actually reducible; (ii) rejecting their assumptions (first and foremost, the principle of the causal closure of the (micro-)physical); (iii) accepting their assumptions and finding a different role for downward causation.
If we choose (i), downward causation turns out to be a rather uninteresting phenomenon. Yet, if we choose (ii), we need to demonstrate that there are causal gaps at the (micro-)physical level of the universe: gaps which can only be filled by non-(micro-)physical entities. More radically, we also need to ask: why should we accept the principle of the causal closure of the (micro-)physical realm? Brian McLaughlin (1992) and David Papineau (2000) hold that, if there are non-(micro-)physical causes operating at the (micro-)physical level, such causes must introduce special physical forces and/or violate the principles of the conservation of energy and momentum. Thus, it is reasonable to hold that there are only (micro-)physical causes of (micro-)physical effects. Among others, Sophie C. Gibb (2010) has tried to demonstrate that this is not necessarily the case. Scott Sturgeon (2003) has argued that quantum physics is actually compatible with there being causal gaps at the (micro-)physical level. But even if this is so, the next question is: can all cases of downward causation be accounted for by simply acknowledging such (micro-)physical causal gaps?4 Eventually, those who adopt (iii) either claim that downward causation actually is a non-causal relation (see section 4) or that downward causation has special sorts of relata as causes and/or effects. For example: the causal closure principle is taken to concern (micro-)physical events, both qua causes and qua effects. Yet, in irreducible cases of downward causation, there could be different sorts of causes and/or effects at work. For example: agents qua substances could be responsible for downward, mental causation; facts instead of events could be the relevant lower-level effects (Lowe 2000). Two problems are left open. First, these assumptions about special causes and/ or effects should be accommodated within a general metaphysical theory of causation. Inter alia, such a theory should try to settle whether there are many different forms of causation (e.g., event causation, fact causation, etc.). Secondly, since the causal closure principle can be rephrased so as to encompass these forms of causation and the special entities that they involve, it must be shown that the relevant reformulations of the principle are less justified than the original formulation in terms of events.
In addition to Kim’s arguments, there are also some empirical concerns about the irreducibility of downward causation, due to the idea that downward causation can be replaced by special, lower-level phenomena. For example: mechanisms are taken to be lower-level structures of entities and activities that are meant to explain certain peculiar effects in biology and the neurosciences—phenomena that one might have otherwise interpreted as due to downward causation.5 More generally, reductionists could grant that certain lower-level entities cause specific lower-level effects only if they stand in peculiar arrangements. Such arrangements could then replace higher-level causes. However, the reductions provided by such strategies are successful only if the lower-level “surrogates” of higher-level causes depend in no way on the higher level. Otherwise, the reductionist project would fail. If we still needed the higher-level entities in order to single out their relevant “surrogates” (e.g., if we still needed mental properties in order to single out their relevant neural “surrogates”), there would be at least a conceptual problem for the reductionist project: the higher level would not actually turn out to be redundant.6