ABSTRACT
In trying to teach language to the ape we have made no attempt to simulate the “natural approach,” mainly because the natural approach is far from well defined and it is difficult to simulate an ill-defined condition. Instead, we attempted to devise the most efficient training procedure possible, without regard to whether it did or did not simulate the human one. Of course, the procedure we used may turn out to be, in its one essential feature—step by step mapping of the presumptive cognitive structures-far closer to the experience the child receives than was once supposed.
