In my target article (Anderson this collection) I suggested that once one distinguishes between the system S that ψs and the mechanism M in virtue of which it does so, it is easy to see that there are three possible relationships between M and S. First, the components of M can all also be components of S, such that M is a relevant sub-component of S. Let’s call this relationship R1. A relationship of type R1 obtains between the drive-train of an automobile and the automobile as a whole. Second, (R2), M and S can be identical. I can’t think of an uncontroversial example of this relationship, and imagine that such a case is relatively rare. Third and finally, (R3), M and S can cross-cut in various ways, sharing some but not all of their parts. In my view, for instance, it is the neuron the fires an action potential, but not all of the entities that comprise the mechanism for generating action potentials are also part of the neuron. For example, the ions in the extracellular fluid that are crucial for establishing the membrane potential are not part of the neuron, although they are clearly part of the mechanism. Similarly, I argued in my target article that in the case of direction-selectivity in SAC dendrites, although it is the dendrite itself that is directionally selective, many of the parts of the relevant mechanism are not in fact parts of the dendrite. Moreover, in the case of DSGCs, the cell and the mechanism in virtue of which it is direction-selective share at most one part: the synapse between the SAC dendrite and the DSGC.
One advantage of making these distinctions, I believe, is that it allows one to see quite clearly when top-down constraints are responsible for function, as I argued is the case for direction selectivity in SAC dendrites. But Kohler suggests that appearances may be misleading here. In fact, he argues, we should “reconstitute the phenomenon” by recognizing that the relevant direction-selective system is not the SAC dendrite, but is rather the dendrite + the non-dendritic elements of the mechanism, including other SACs. This larger system can be then be treated within the standard framework of componential mechanism. We can call this approach to addressing these sorts of cases “the Kohler strategy”.
As I noted in my target article, the Kohler strategy is certainly open to the mechanist. It does, however, have the following effects. First, it tends to make the systems of the brain to which functions are attributed relatively larger and more diffuse, which arguably reduces precision. Second, it would in effect turn all apparent instances of R3 into instances of R2.[1] I noted above that I thought the class of R2 would be small. If I am right about the prevalence of R3 functional relationships in the brain, then this strategy would make R2 very large. But it would do so essentially by legislation, as a way of preserving the universal applicability of the componential mechanist framework. How forced this appears will depend on how closely one believes the guiding assumptions of that framework match the architectural facts of the brain. We will return to this last point after reviewing some of the considerations that appear to favor the Kohler strategy.