For the most part, Lisp people will never like any modification to the language that changes their pure abstract computation engine into something that more people would want to use.
Behold, a Usenet post from 2000 on the notion of a "lispscript" that could compete with perl or awk:
That said, I personally have no idea if Arc is all that good or not. I get it, but it doesn't inspire me the way Haskell or Erlang do right now. But Arc seems to be successful by the designer's own standards.