He created lisp-stat in the late 80's
https://www.jstatsoft.org/article/view/v013i09
"While R and Lisp are internally very similar, in places where they differ the design choices of Lisp are in many cases superior. The difficulty of predicting performance and hence writing code that is guaranteed to be efficient in problems with larger data sets is an issue that R will need to come to grips with, and it is not likely that this can happen without some significant design changes."