Self-reference, far from being just a logician's and a philosopher's puzzle is, in fact, a central feature of human language and reason. It, thus, seems natural that intelligent machines will also have to deal with the issue of self-reference. We discuss some of the formal problems, and potential solutions and applications. Portions of this essay are descriptive in nature, portions prescriptive. We are involved in the development of some of the ideas in the relevant literature, and make no apology for injecting a certain subjective note into the text, as opposed to forcing a false objectivity. We have also freely drawn on portions of essays written by one of the authors.