Actually, all the energy of recoil is still absorbed by the soldier's shoulder. The spring just serves to spread the energy release out more so that recoil occurs over a quarter of a second instead of an eighth or something like that.
Apologies in advance if I screw the math/physics here up, I'm out of practice.
I think you point is correct, but the time is shorter, in the realm of hundredths of a second. I think the formula to figure the time is d = 1/2 (v1 + v2)(t), which would give 1/400 seconds assuming a 1m gun barrel, constant acceleration and 800m/sec muzzle velocity.
Another factor in why it's better to be on the butt-end of a gun rather than the barrel end is the gun won't have the kinetic energy that the bullet does because it has higher mass.
take a 30-06 round, muzzle velocity is about 850 meters/sec and has a mass of roughly 10grams for a momentum of 8500 g*m/sec (not converting units, I'm lazy)
A gun will weigh something like 2.5kg or 2,500grams
Given conservation of momentum, M1V1=M2V2, solve for gun velocity (going backwards) it comes out to about 3.4meters/sec (or negative 3.4 if you want to get picky). It won't really be like that in real life as the gun is coupled to the shooter...
Now figure the kinetic energy (capacity to do work). formula is 1/2 mv^2
Bullet: 0.5 * 10gr * (850)^2 = 3,612,500
Gun: 0.5 * 2,500gr * (3.4)^2 = 14,450
Hopefully I didn't hit a wrong button on the calculator.
My physics is too fuzzy to remember what the units are here (newtons or centinewtons or ergs or something), but I think the general point is evident. Other things mentioned (impulse, spreading the energy over a larger surface area) are also important of course.