"We really need to get away from arguing for the goodness of languages based on the success of the people using them."

I think it devolves into this when you get into a Blub situation.

So you want to tell someone: "But it has closures! Macros! Object orientation based on multiple-dispatch generic functions! First class functions!"

And you see this stunned, glazed look come across your listener's face.

"Well, Paul Graham made a ton of money on it. Peter Norvig wrote great books about Lisp and now runs Google research. Richard Stallman made it the macro language for the greatest editor ever invented [EDITOR'S NOTE: only try this one with committed Emacs users.]. The Guy who wrote much of the spec for Java after James Gosling wrote famous compilers for Lisp dialect and authored the document that defined the Common Lisp standard. The Greatest Computer Science Course Ever taught at MIT uses a dialect of Lisp. A company who is cleaning their competition's clocks in the airline reservation business uses Lisp."

Etc. etc.

It would be much more compelling to say, "Joe Schmoe picked up this language in a weekend and used it to create a dating site that got 3mil users in a year" than "a few superstar genius programmers who built things you'll never ever build swear by it."

Most people are convinced by examples they can relate to and feel they can achieve, not by the feats of the top experts in a field that feel outside of most people's reach.

It's like selling a religion because John Travolta and Tom Cruise are in it...

well said. well, the having chose Lisp part might not be the most interesting part, but deciding to write an application as an Emacs plugin is pretty interesting.