FLingfrog
I have certainly used muscle memory in gliders. Last year I was flying the club PW5 and trying to core a very narrow thermal. To do that I was flying at minimum speed and high bank angle to minimize my turning circle. I was momentarily distracted by a radio call and inadvertently let the airspeed drop a few knots which caused the glider to stall. The stick went forward without conscious thought and then stepping on the rudder stopped the yaw. I flew out of the thermal under control after losing less than 100 feet.
I do believe that the instinctive pitch down if the airplane surprised you with an inadvertent stall can be taught. Finally most airplanes will talk to you when they are getting slow and at high AOA. I make a point of demonstrating the so called “soft stick” and get pilots to maneuver in slow flight so that they recognize the change in control forces.
I guess I have hijacked the thread so back to the OP’s question. As a previous poster noted the effect of aft C of G can be significant in gliders but less so in your typical powered trainer. However most trainers that are cleared for spinning have a more restrictive C of G envelope for spins. This will result in the aft C of G limit being moved farther forward in order to eliminate undesirable or even dangerous spin characteristics.