Context Free, Context Sensitive and Type 0 grammars are based on the Chomsky hierarchy
A programming language is based on a strong foundation of a context-free grammar.
Imagine a text-based AI program that would be able to learn something simple through trial and error.
It is feasible to teach the trial and error AI program simple programming grammars, like simple data commands. Therefore it is feasible to teach the program Type 0 grammars.
By relying on the simple relational data commands that it had learned it would be possible to teach the program to learn to use simple context-sensitive grammars.
By designing the AI program to try to acquire simple (context-free) grammars during the early stages, it should be possible to teach the program parts of a very simple language. Then, by relying on those simple grammars that it had learned, further learning could be enhanced so that it could construct fragments of a slightly more complicated language that was based on the more complicated (context-sensitive) grammars. These grammars would be implicit; the program would not understand the grammatical rules that it would use, they would be derived as the program searched for commonalities and categories in the letters, words and phrases used in the text-based exchanges with human users. And the responsive behaviors (as expressed through the exchanges of text) would also guide the program once a meaningful rapport was established between person and computer program.