About taking off and landing: If you take as the goal 1 g downward --
that's an integer, not 1.0000 plus or minus a little bit, you can't do
it. There is acceleration. The same thing is true for a loop, there
just are not enough degrees of freedom to allow the pilot to see the
horizon drop down as he climbs, then reappear inverted at the top of
the loop, without experiencing some incremental (even if small) g
forces.

I think a roll adds the additional variables one might need. Consider,
for example, a plane about 45 degrees into a roll. At that moment, in
coordinated (pilot talk for keeping the pilot's weight centered on the
seat) level flight there's a g toward the center of the earth, and
another along the radius of the turn. The pilot experiences 1.414 gs
into his seat. If, however, the airplane is also pitched down 45
degrees accelerating, and coordinated, you could choose numbers that'll
resolve to 1 g into the seat.

Take now a bank of 90 degrees. If the airplane is pointed straight down
and accelerating at 1 G, that is, in free fall vertically, there'd be
no fore and aft weight component. The pilot would, however, have to be
pulling back on the yoke hard enough to accelerate in the nose up
direction at 1 G.

At inverted if level he'd be experiencing 1 g "up", so he'd have to
have the yoke back far enough to accelerate in the nose up direction 2
gs worth.

You can, at each point point in the "roll", calculate how the airplane
must be accelerating in the nose up direction and what direction the
nose must be pointing for the pilot to experience 1 g down.

I don't know if it's a realizable manouver -- it'll take some serious
elevator "authority" to provide the nose up accelerations that are
needed. Some insightful person in this thread made the observation that
if the airplane was during the roll just accelerating downward at 1 G
-- in free fall, if you will, one need not worry about the gravity
effects and the pilot would just have to pull back on the yoke hard
enough to keep the nose accelerating up at 1 g.

It's an interesting problem. I think the airplane, as seen from
outside, would look like it was in a death sprial. The pilot, however,
would see the horizon rotate through 360 degrees, so he'd say he rolled
the airplane. At the end of the roll the airplane would be in a serious
nose down attitude, and going pretty fast. I don't think you can get
from there to straight and level while keeping a local 1 g down weight
component.