Does It Really Matter?
So far I have described causal realism as largely being in agreement with neo-Humeanism about the way causation appears to us; it appears as a regularity in the occurrence of entities of which the effect is counter-factually dependent on the cause, such that if the cause is manipulated the effect will come and go away. Causal realists don’t really disagree with anything in this picture, except in degree. For instance, Mumford & Anjum will accept the above description with rhe qualification that the regularity, counterfactual dependence, and reliability of intervention are not perfect, because they don’t think causation involves necessity (only a substantial tendency).
Some neo-Humeans may reject causal realism as an ill-construed attempt to figure out what objective reality is like; or, rather, an attempt to figure out what cannot be figured out. This would be an epistemi-cally motivated neo-Humeanism that opposes powers-based accounts for reasons that have nothing to do with what exactly powers-based accounts suggest about causation. We are talking about people who don’t see any appeal in Bernstein’s suggestion about a mutual benefit resulting from bridging the gap between the two systems. But they should then also accept that no realism makes any sense, not even neo-Humean metaphysics.
Others may reject causal realism, not because they think nothing can be figured out about mind-independent reality, but because they think powers-based accounts are wrong about mind-independent reality. These would be the proponents of neo-Humean metaphysics, which say that there really are no substantial connections between anything. Now, I find it difficult to discern between realist and anti-realist neo-Humeans, because very few, if anyone at all, explicitly state their position in the matter. However, David Lewis is surely one who explicitly endorses neo-Humean metaphysics, the view that the world is nothing but a mosaic of local matters of fact between which there are no necessary connections. He even explicitly commits to the reality of a plurality of worlds existing in parity.
In sum, the benefits of the powers-based approach is really in the more fine-grained account of what constitutes the cause C which is followed by E, and on which E is counterfactually dependent, and therefore offers a more fine-grained idea of what it is that we can manipulate to produce E or variations of E. The more fine-grained account has consequences for what we would consider to be good scientific practice. So, for instance, Anjum, Copeland, & Rocca (2020) argue, on the basis of Mumford & Anjum’s powers-based account of causation, that there is a fundamental flaw in the idea that only randomised controlled test studies (RTC) are of high enough quality to ground evidence-based practices. Very roughly, the idea that only RTC are good enough is based on the idea that causation boils down to a two-place relation between temporally distinct events or states, i.e. ‘if C, then E It is acknowledged that the effectiveness of C is contingent on the situation in which it occurs, but it is still assumed that there must be some particular factor C that is decisively responsible for the outcome E. Consequently, it is assumed that any investigation into the causes of some negative outcome E has to involve the comparison between two controlled settings in which background conditions are fixed, and that onsly differ with respect to the presence/absence of C. As Anjum and colleagues point out, this model is at a disadvantage in discovering, for instance, the aetiology of multifactorial diseases. In that case a more fine-grained powers-based approach has the advantage. The alluring promise of powers-based accounts resides in their potential not only for solving certain philosophical problems, but for improving actual research practices in order to be able to better understand multifactorial illnesses such as chronic fatigue and various other psychosomatic disorders.