Prolog is not a new language, it's been around for 53 years.
> Why not instead do it via teaching about constraints solvers?
Prolog is used for more than just constraint solving. If that's all you think it does, perhaps you could consider studying it and finding out what it can do.
I have no interest in studying prolog to extract what knowledge from it would be practical.
Learning principles in the context of a language is learning in the context of application rather than pure theory. For some, this helps ground ideas by connecting them to others they already understand. Of course, no advice is universal, and each person needs to decide if it applies to them.
This advice most applicable in the context of deciding what language to learn next. Assuming you want to improve your ability to tackle more kinds of problems, rather than fulfill a particular checkbox. Like learning Java when you already know C# is going to contribute to personal growth less than a LISP or prolog or Rust.
My comment was that there are more practical alternatives than just programming languages.