This was all lots of fun, and the systems were successful as academic projects go. But it wasn't leading me to the Grand Insights I thought I should be having. The implicit vision behind these efforts was something that could scale up to something more like Marvin Minsky's Society of Mind, which was a mechanical model not just of animal behavior but of human thought. I don't think that ever happened, and while I might blame my own inadequacies it might be also be that Minsky's theories were not very language-like. A good language like Lisp is built around basically a single idea, or maybe two. Minsky's theory was a suite of many dozens of ideas, each of which was at least in theory mechanizable, but they didn't necessary slot together cleanly as concepts do in a pristine language design.
Structure and Interpretation of Computer Programs by Hal Abelson and Jerry Sussman. In olden times, used as the text for the first course in MIT's CS program. Uses Scheme (a Lisp dialect) which gives it a certain timeless feel.
A big influence and so far the only one that isn't a person (John McCarthy invented the first version but it's grown quite beyond his original concept). I learned Lisp by hanging out at the MIT AI lab with some of the Symbolics founders, and have been using it as my main tool ever since. The Lisp Machine remains at the pinnacle of environments for programming as a design activity; I don't know why the state of the art hasn't advanced since then.
Working in Lisp gives you a kind of feeling that is hard to describe; its almost as if abstractions take on a tactile quality; there is very little boundary between thought and its realization. Lisp is not the only computational system to have this quality but it's been that for me. Roam has some of that quality and it's not a coincidence that it is implemented in the Lisp dialect Clojure.