Isn't that what Va is all about?

Conventional wisdom says so, but there is no requirement in Part 23
that says this must be true. Part 23 only uses this speed in its
requirements for control surfaces.

In my view, the most correct definition of Va will be it's the speed
above which you cannot make full or abrupt control movements, due to
control surface integrity.

New airplanes are supposed to come with a new Vo speed, which DOES
require that the airplane stall before exceeding the load factor.

Here's a copy from a draft copy of an AC 23.something that I found.
The AC was intended to make this clear to test pilots, but I don't
think the draft was ever finished:

------------snip-----------------
VA should not be interpreted as a speed that would permit the pilot
unrestricted flight-control movement without exceeding airplane
structural limits nor should it be interpreted as a gust penetration
speed. Only if VA = Vs sqrt(n) , will the airplane stall in a nose-up
pitching maneuver at, or near, limit load factor. For maneuvers where
VAVS n , the pilot would have to check the maneuver; otherwise the
airplane would exceed the limit load factor.

Amendment 23-45 added the operating maneuvering speed, VO in §
23.1507. VO is established not greater than VS sqrt(n) , and is a
speed where the airplane will stall in a nose-up pitching maneuver
before exceeding the airplane structural limits.

------------snip-----------------


Va would be the same at any aircraft weight, which it certainly
isn't.

It is in some airplanes. My Piper arrow doesn't scale it with weight.

Moreover, Part 23 says that Va is *only* defined at max gross. Some
manufacturers do publish Va's at lower weight, but that appears to be
at their option. As written, it doesn't match Part 23 definition.