In practice

In practice, however, the type system constrains the programmer in many ways that are unacceptable to Lisp programmers: data cannot be interpreted as programs, lists must be homogeneous, and functions are not easily redefned without recompiling other dependent functions.

>>17
That argument would only work if a lot of languages didn't already have interpreters for them. Some of those interpreters may very well be huge in code size, but they work, which means not only that it's possible, but usable.

Also, eval in some CL implementation is just (defun eval (code) (funcall (compile nil `(lambda () ,code)))), which in other languages can be accomplished by calling the compiler and loading the compiled module dynamically in memory and calling the functions (I can't say it's as nice as in CL, but it does work just fine).

And no, I'm not of the nationality you mentioned. Most languages have eitehr compilers or interpreters or both as viable implementations, which shows how possible this is. I did not claim this is something you could do in your afternoon, but you can always use code others have written.