There are many different reasons why people fear Lisp. Nearly every single one of them is wrong in one way or another, or in many different ways. Some are simple notions which are quite easy to disabuse someone of, for example the Inefficiency of Lisp systems. This is simple -- just show them some performance figures from a Lisp vendor's website, or have them read papers comparing Lisp performance to other languages - Lisp Vs. Java.
But some fears are not so simple, such as the parenthesis fear. This is somewhat related to the problem of introducing Lisp's idiosyncratic terminology, such as CAR, CDR, MAPCAR, and RPLACA, to name a few odd ones. An excellent defense (or rather, explanation and plea for patience) of Lisp's oddnesses in terminology and parenthesization is in Pascal Costanza's Highly Opinionated Guide to Lisp, in section 4 "Some first obstacles".
One particularly intractable problem is the complaint in professional situations that Lisp programmers are hard to find. You may suggest using Lisp as the development language but the answer always seems to come back from the boss sounding like "Yeah, but where will I find someone else to help maintain it?" This is indeed a hard problem, since it's really the truth that Lisp programmers are not a terribly common breed, despite their use of a Common Language.
(Add other examples, discuss, etc.)
Is it just me, or is this (a) completely information-free, and (b) basically irrelevant to CLiki anyway? I see no figures, reports or references to provide evidence for the author's hypothesis, and I can't see much about free software or Unix either.
(I don't even see anecdotes. What is this place coming to when even anecdotal evidence is considered too much like hard work?)
Here's a anecdote, written by Erik Naggum: Choice of Language and Programmer Availability.
And furthermore, I don't think this kind of whining and "people who don't use Lisp are stupid"-attitude does Lisp a favor.
As far as syntax goes it is my opinion that what makes people go 'yuk!' is the absence of distinct textual landmarks. CL is very positional, context-dependent and full of quirks. Consistent indentation only goes so far in revealing structure. It takes a lot of practice to be able to really read lisp rather than merely to decode it. It is, in many ways, the same problem as with standard music notation.
On the other hand, I would love to find some people who could actually read C/C++ rather than "understanding" an application based on some rather arbitrary picture.
When you have a powerful language people will not tolerate anything less than they're used to, so when you've got lisp hackers they are easier to get because of the fact that they can't use anything less powerful than lisp