How Language Could Have Evolved

A bi-directional graph neural network is developed that is posited to generate language phenomena in Mammals.  A first order network is developed that is posited to be common to all Mammals.    The network is then extended to include several ubiquitous Human language phenomena, such as the conjuntion, direct object, and tenses.   Computationally the neural network is implemeneted with only a single c++ class, Node{},  which is  approximately 500 lines.    A single function, touch(), is used for all runtime interactions from main() in order to invoke network operation and to produce the described language phenomena.  The main() function is replicated  as a series of functions: main1, main2, 3, 4... as the Node class is evolved to produce increasingly complex language phenomena.