One of the features I like the most in LISP is to be able to change the code in the middle of the running process and to be able to keep evaluating the functions and variables with the new behavior on the fly. This is an extremely powerful tool when you are trying to figure out the right structure for a given problem. You can play with the different structures and still be able to keep the application running and see how it reacts to the changes you make. It is like running in a self-evaluating forever debug mode. But that is exactly what LISP is, isn't it?
I can say that when I think about prototyping at the model designing level I am more and more choosing to do it in LISP and then, when I feel more confident about the model port it to c# than to start straight in C#. It is simply more productive in my opinion.