One of the virtues of Madole & Harden's (M&H's) approach to causation in behaviour genetics is their ability to combine their honesty in acknowledging the limits and gaps of genome-wide association study (GWAS) methodology and, more generally, in behaviour genetics with their determination to point out the way in which first-generation causal knowledge can open the door to ways in which second-generation causal knowledge can be pursued. As a consequence of the first attitude, their analysis represents a step further to overcome genetic determinism and essentialism (target article, sect. 1.2, para. 4). The second attitude leads them into the sort of pluralism that is commonly accepted in philosophy of science and medicine (see, e.g., Rocca & Anjum, Reference Rocca and Anjum2020). In their approach, M&H have recourse to some work in the metaphysics of science to deploy the theoretical framework of their investigation. The interventionist theory of causation developed by Woodward (Reference Woodward2005), which has been applied to genetics by Waters (Reference Waters2007) and by Woodward (Reference Woodward2010) himself, provides them with a useful tool to conceptualize the notion of cause behind first-generation causal knowledge in genetics. In addition, recourse to mechanisms (Craver & Darden, Reference Craver and Darden2013; Glennan, Reference Glennan1996) accounts for the kind of explanation required in second-generation causal knowledge, crucial for understanding the aetiology of complex phenomena. In this brief commentary, I intend to bring to light another metaphysical tool that can help M&H to overcome some limitations and offer a richer picture, namely, the concept of capacity or disposition (Cartwright, Reference Cartwright1989; Mumford, Reference Mumford1998). Mumford and Anjum have developed a dispositionalist theory of causation (Mumford & Anjum, Reference Mumford and Anjum2011), and have applied it to science (Anjum & Mumford, Reference Anjum and Mumford2018), medicine (Rocca & Anjum, Reference Rocca and Anjum2020), and genetics (Mumford & Anjum, Reference Mumford and Anjum2011, Ch. 10), conceiving of genes as powers or bundles of powers “coded” into the structural complexity of DNA strands.

Capacities, powers, or dispositions (I use these terms as equivalent) are “properties or potentials of things or systems, that can become manifested under certain conditions” (Rocca & Anjum, Reference Rocca and Anjum2020). Clinical randomized controlled trials (RCTs), for instance, can give rise to claims about the capacity or power of certain substances to induce specific effects in particular contexts. Such association of RCTs to capacity claims “provide a conduit from RCTs to effectiveness” (Cartwright, Reference Cartwright2011). Medicine seeks to predict in order to intervene by means of treatments, and behavioural genetics applied to social sciences seeks to predict in order to intervene too. But then, it is necessary that the sort of prediction reached is not of the kind “it works somewhere,” but rather of the kind “it will work for us” (Cartwright, Reference Cartwright2011). This requires that the treatment reliably promotes the outcome, and this demands a capacity claim (magnets power to attract metals, for instance, grounds the step from “it will attract some metal somewhere” to “it will attract some metal for us”). Capacities thus allows for extrapolation of the knowledge provided by RCTs insofar as they offer the grounding or explanation for the association reached by the RCTs (Cartwright, Reference Cartwright and Handfield2009).

Dispositions have some characteristics that make them suitable for their application in genetics. (a) Given their dispositional nature, powers might not be manifested if the triggering conditions do not occur. This feature allows for a sort of modality that matches very well with claims in genetics, because genes sometimes predispose for something, but do not necessitate it. (b) Triggering conditions make a power manifest and it is then when causation occurs; causation is thus conceived as a process rather than as a relation. Analogously, genes are causes only insofar as genetic expression is triggered. (c) Most powers need mutual manifestation powers, that is, powers that need to be met in order for them to manifest, so that causation occurs as a consequence of the joint manifestation of such powers. This characteristic accounts for phenomena such as polygeny or, in the case of behaviour genetics, heterogeneity across environments. (d) There are interfering and preventing powers, so that a power makes the effect occur differently by influencing the timing, chance, or extent to which the effect occurs (interfering powers) or prevents the effect (preventing powers). Gene silencing and mutation are paradigmatic examples in this case, and many epigenetic phenomena (histone modification and cytosine methylation) instantiate interfering powers that determine in some way gene effects. Demographic composition and environmental context are typical examples of this feature in the case of behaviour genetics. These four features account for the complexity of causation that is so pervasive in genetics. A nice illustration of some of these features is implicit in M&H's example of causal depth in cystic fibrosis, which presents shallow features (see endnote 6).

M&H characterize shallow causes as non-unitary (complex causality), non-uniform (heterogeneity of effects), and non-explanatory (cause and effect are associated without explaining why and how the effect takes place). It is quite immediate the way in which the dispositional framework captures the non-unitary and non-uniform character of shallow causes: Most dispositions manifest jointly and interfering and preventing powers account for heterogeneity of effects in different contexts. Crucially, however, powers are explanatory. It is because genes are powers that they produce the effects they do when triggering conditions occur. The picture is then one in which different powers (different genes, other biological components, environmental and cultural factors) contribute to different extent and in various ways to the effect, accounting in this way for the shallowness observed in genetic causes, but insofar as they are explanatory, powers offer a suitable road from shallowness to effective or real causation. Powers offer the way to go from whether and how often something happens to why and how something might or might not happen (Rocca & Anjum, Reference Rocca and Anjum2020). Or, in other words, “there are causes out because there are causes in.”