At the LL1 page they define a lightweight language as one which is easy to learn.

Following the analogy of heavyweight vs lightweight threads, I initially thought it would be about languages which are either very small once compiled, or that have simple interpreters.

But who am I to question some guys at MIT?