Why does my Prolog S-expression tokenizer fail on its base case?

Question

To learn some Prolog (I\'m using GNU Prolog) and grok its parsing abilities, I am starting by writing a Lisp (or S-expression, if I\'m being exact) tokenizer, which given a