In this paper a control scheme wich linearizes the system is discussed. The idea here is to integrate recurrent neural networks and the linearizing control scheme proposed by Kravaris and Chung. A straightforward approach would have been to identify the non-linear plant using a recurrent neural network, and then synthesize the control law using this network. However, this particular methodology is eschewed here, for this would mean tedious calculations of the varios Lie derivatives of the network and the exact cancellation of non-linear terms. Rather than go through a process of first identifying the plant an then evaluating the various parameters for linearizing the plant, a more interesting scheme would be one where the network designs the linearizing laws for the system. This means that the network provides us with the linearizing parameters as outputs, rather than the outputs of the system.