The bottom-up approach advocated here to understand higher-level phenomena may be considered a non-Platonic approach that aims to construct artificial, but strongly biologically inspired systems. These systems should be able to simulate complex behavioral tasks, but do so by application of simple elements, artificial neurons, and a simple decentralized neuronal architecture. If successful one could then study whether more abstract concepts introduced in psychology or philosophy, for example, could sensibly be applied to such a system. We claim to have shown an example supporting this approach.