Although mechanistic explanation as developed by Craver & Bechtel (2007; Craver 2008) does seem to accurately characterize one kind of explanation in neuroscience, and one kind of functional arrangement in neural systems, I’ve argued here that the formulation is not wide enough to capture the variety of mechanisms in the brain. When we formally distinguish the system S from the mechanism M in virtue of which S exhibits the explanandum phenomenon ψ, we see that although it seems correct to describe the relationship between M and its parts {X} in terms of constitution, it will only sometimes be the case that S is (partially) constituted by {X}.
As an alternative to the relationship of componential constitution, I have offered the notion of an enabling constraint that can exist between a system and the mechanism(s) in virtue of which it has its various functions. SAC dendrites appear to have their function in virtue of the enabling constraints imposed by entities at the same and higher levels of organization; and DSGC function is enabled by the constraints imposed by the SAC dendrites. In neither case is it appropriate to describe the relationship between the mechanism M and the relevant system S in terms of constitution, nor are all (or, in the case of DSGCs arguably any) of the parts {X} of M components of S.
Overall, I hope to have made the case that moving beyond level-restricted mechanistic explanation will allow us to better capture the variety of neural systems that emerge from the constant, constraining, biasing interplay between feed-forward, feedback, bottom-up, and top-down processes in the dynamic brain.