Abstract

The Acting Entity-characterization of mechanisms, defended in the last chapter, is rather broad. It allows for almost all causal goings-on to be mechanisms. Let us call the AE-characterization of mechanisms as formulated in the previous chapter the minimal notion of a mechanism (Glennan 2017). (Minimal Notion) A mechanism for a phenomenon consists of entities and activities organized in such a way that they are responsible for the phenomenon. Something is a mechanism in the minimal sense if it consists of more than one entity, at least one activity, and a certain organization that is crucial for the phenomenon to be produced. An ion channel sitting on the axon membrane is not a mechanism—it consists of only one entity and does not produce any phenomenon. An ion diffusing through this ion channel and thereby producing a voltage gradient is a mechanism in the minimal sense. (...) One problem with the minimal notion is that, since it renders almost all causal goings-on mechanisms, it cannot make sense of the intuitive and theoretically relevant difference between mechanisms such as that leading to Roland Barthes’s death, and mechanisms like the neurotransmitter release mechanism or the action potential mechanism. The difference is of theoretical importance because ephemeral mechanisms cannot be used for the various tasks that the notion of a mechanism is supposed to perform according to the new mechanists (Krickel 2018): first, the minimal notion of a mechanism cannot account for the normativity that is often implied in mechanism-talk. Many mechanisms are said to be able to fail or succeed in bringing about a phenomenon. (...) Second, mechanisms in the minimal sense cannot be used to justify type-level mechanistic explanations. Like the mechanism that led to Roland Barthes’s death, minimal mechanisms might occur only once and, therefore, cannot ground explanations that have general phenomena as their explananda. For example, the mechanism for neurotransmitter release is supposed to explain neurotransmitter release in general rather than in one particular instance. Again, this suggests that the minimal notion of a mechanism is insufficient to account for relevant kinds of mechanisms. In the following sections I introduce a taxonomy of mechanisms that goes beyond the minimal notion. First, I introduce the notion of a functional mechanism: one can distinguish between those mechanisms that fulfill a (biological) function, and those that do not (Garson 2013; Piccinini 2015; Maley and Piccinini 2017). Indeed, combining the notion of a mechanism with that of a function seems to be promising with regard to making sense of the normativity of mechanism-talk: a mechanism that has a certain function is supposed to fulfill that function and might fail to fulfill it. In what follows, I discuss different suggestions for how to characterize functional mechanisms. It will turn out that neither of these notions successfully accounts for the normativity of mechanism-talk unless the second and third sub-types of mechanisms are taken into account. I will call the second type regular mechanism; the third type I will call reversely regular mechanism (Krickel 2018). Both notions rest on the idea that one can distinguish between one-off mechanisms and mechanisms that establish some kind of regularity (Andersen 2012). Regular mechanisms, as I will show, have to be understood as mechanisms that bring about a particular phenomenon more often than they bring about any other phenomenon. Reversely regular mechanisms are mechanisms that bring about a particular phenomenon that is more often brought about by that mechanism than by any other mechanism. I will show how these two notions of regularity together are necessary and sufficient for grounding type-level mechanistic explanations (see also Krickel 2018), and when combined with the functional notion of a mechanism, can solve the problem of accidental goal contributions, which afflicts the most promising account of functions as discussed in the next section.