Norvig wrote an article or post where he deemed Python to be “an acceptable lisp” (or something like that, I may not have it exactly right).

Hardly a ringing endorsement! Python is certainly succinct and expressive like Lisp (when his alternative is Java). ML was designed to be a better Lisp.

ML wasn't designed as a better Lisp. It was designed as a 'better' (and statically typed) functional programming language for writing proofs in a theorem prover. The theorem prover and then also the first ML version was written in Lisp. That was the motivation for the first version of ML.

I did not mean to imply it was a Lisp. It was designed to solve the same problems as lisp, i.e. symbolic manipulation; and was used to rewrite code originally written in Lisp.

ML was at first a domain specific language for writing/assisting proofs: a) maths-like notation, b) functional, c) statically typed, ...

Lisp did only do b) to some extent. It was and still is common to develop languages as internal or external domain specific languages on top of Lisp. ML was such an example. Often these languages then are/were ported away from Lisp to a specialized (and possibly smaller/more efficient) implementation. Thus Lisp served as a prototyping environment.