The semicolons look like operator sequencers (not terminators) in Algol68.
LISP 1.5 was a punched card, so comments would probably have been written on the cards themselves, I think? The manual does not indicate that there is a mechanism for commenting in the language.
The Stanford Lisp 1.6 manual shows semicolon comments.
As for the origins, I would look that if early assemblers used something to indicate comments. Of course, semicolons are common enough in current ones.
source share