ABSTRACT
Cognitive science is a form of "reverse engineering". We are trying to explain the mind by building systems that have minds. A "Turing" hierarchy of empirical constraints can be applied to this task, from Tl, toy models that capture only an arbitrary fragment of the performance capacity, to T2, the standard "pen-pal" Turing Test, to T3, the Total Turing Test, to T4 (T3 plus internal indistinguishability). The analog "shape" constraint is the grounding inherited by the arbitrarily shaped symbol that names the category and by all the symbol combinations that it enters into. As items in the general cognitive armoury, neural nets are naturals for the important toy task of learning the invariants in the sensory projections of objects that allow us to sort them and assign them symbolic category names.
