As a lisp novice, I found this to be a principal challenge. Most of the docs I find will have either function signatures or examples, but the examples rarely cover the bases. I think it's pretty common for dynamically typed languages to have very hand-wavy documentation, but lisp is big on the arcane and powerful. I'd often find answers to my questions in the entirely wrong places, e.g. pearls of wisdom tucked in an example that I stumbled across looking for something entirely different.

I tried to get into SBCL a few months ago and I found all of the documentation I encountered to be really good (really good) but there was just too much to learn. I'm not complaining, I just didn't have the time personally to climb the learning curve (despite all the available lore from previous climbers.)

When you start with a small system and reinvent things, you know them from the inside out. Often that's easier and more fun than reading up on an existing large system. The downside, of course, is that we wind up with incompatibilities and myriad idiosyncratic ways of doing things.

I'm aware! Although Common Lisp as specified does differ somewhat from the language he describes in his book.

But you kinda implicitly assume my point: Scheme and Lisp are different things. That doesn't make Scheme bad or Lisp good, although I will assert that Scheme is underspecified and Lisp is pretty decent.