When did you first try Lisp (meaning here and throughout the survey "any member of the Lisp family") seriously, and which Lisp family member was it?
I started using clisp and cmucl seriously in university. I wanted to apply the power of Common Lisp to physics problems. It worked rather well, of course FORTRAN can be optimized better, but the higher level abstractions gave me a advantage. The other students or professors where looking at C++ or Java as an alternative to F77. Some went back to FORTRAN 2000 I've heard.
What led you to try Lisp?
Lambda caculus, macros, emacs and the Hackers dictionary. Using Mathematica already convinced me that Lambda caculus was an ideal tool for physics.
If you were trying Lisp out of unhappiness with another language, what was that other language and what did you not like about it, or what were you hoping to find different in Lisp?
I found FORTRAN 77 to primitive, FORTRAN 90 to immature (at that time), C++ a plain horror of possesiveness, Java just wrongly OO'ed, Smalltalk useless without a GUI, Mathematica just too slow.
How far have you gotten in your study of Lisp? (I know, that is hard to measure)
Pretty far, but still I find hidden treasure almost every time I managed to write CL code.
What do you think of Lisp so far?
The final language. Other languages might be more adapted to small quickly solved problems (I am a AWK fan too) but if the problem becomes too big or you don't know the problem yet, then the only rational choice is Common Lisp. No other language lets you explore the problem so fast and most of the time, buy the time you've described the problem you've solved it. The solution might be slow, hacked together and brittle, but it can be made fast, well designed and stable in a incremental way, having a working solution almost all the time and changing only those parts that need work. I know of no other language (except Common Lisp lookalikes) that offer that much support for this kind of work.