Python is, semantically, pretty similar to Common Lisp - they have far more in Common than they do with C++, for instance. So, even if it's not the syntax, it's clearly not the semantics or language behavior, either.

I'm pretty sure it's the syntax, though. You used the word "unappealing", which often refers to a surface attraction or draw. I think that this is absolutely correct - Lisps are unpopular in large part to their very weird syntax, and indeed, one of the most common comments that you'll hear when you bring up Lisp is "Is that the weird one with all of the parentheses?"

This is further reinforced by the fact that most programmers have never even tried to run a single line of a Lisp - and the vast majority haven't invested enough time to learn enough to write a non-trivial program with it, and therefore haven't had enough exposure to even judge whether they like the language or not. Therefore, any reason for unpopularity must be superficial.

Mainly, Lisp was developed on institutional hardware in the era of big iron. It became larger and more complicated, keeping up pace with the capabilities of those machines, which Lisp could easily strain.

Suddenly, microcomputers showed up in the 1970's. The industrial strength Lisps just wouldn't go down from hundreds of kilobytes or megabytes of RAM required down to two-digit kilobytes.

The situation didn't start to improve until probably the late 1980's.

Some developers used Lisp for prototyping, and had the expensive hardware; then to create a product to ship to users on consumer hardware, they rewrote the prototype. That story was exemplified by the CLIPS expert system.

An entire generation of programmers who started on microcomputers in the 1970's and 1980's wouldn't have even had opportunities to use Lisp.

When I started programming in 1982 or so, I inherited a stack of Creative Computing magazine back issues, so I read about all sorts of interesting stuff, including Lisp; but I wouldn't have been able to get my hands on it anywhere: and not for the lack of knowing about it or wanting.

Microcomputers steam-rolled a vast amount of software that preceded them. Software that didn't make the jump from big iron to microcomputers either died or was relegated to niches. Fortran, Cobol, operating systems like Multics, you name it.

We use Unix-like systems today because Unix started on minicomputers that were close in capability to microcomputers. There was only a small lag time between Unix being "bigger iron only" to running on microcomputers. It used pretty modest resources; e.g in 1980, /vmunix was maybe 50-something kilobytes. But even that was a little bit large, and the microcomputer jump did hurt Unix; it allowed disk operating systems like CP/M, PC-DOS and MS-DOS to gain a foothold. Unix never quite recovered; we now have free versions that are quite popular, but mostly in serving and embedded.

Speaking of Unix, I read a book on using Unix in the mid 1980's. As a kid. I had nowhere to actually access it for hands-on experience, but was exciting to read about. You had to be a university student, or else a grown-up working at some company or institution. (Or else, dial into something and break in.)

Anyways, the time lag between the start of the microcomputer revolution, and Lisp being able to run well like on 1970 big iron, was quite long! Long enough to see people retire. There is nobody out there to do mentoring. If millions of programmers started to be interested in Lisp, there would be nobody available for mentoring. Everyone would have to use just books and such. You can hardly find a course to take or anything like that. Some universities use some Lisp or Scheme dialect in a few courses, and that's about it.

Even if every great Lisp hacker who ever lived was still alive and interested in mentoring, it wouldn't be enough; there simply weren't that many people in computing 30, 40, 50 years ago.