Sure, bolts have extractors, but on blowback actions, nothing touches the bolt other than the case. The case pushes the bolt backward, and the case is pressurized by gas in the barrel.
On a submachine gun using advanced primer ignition, the firing pin is dimensionally optimized to ignite the primer before the bolt is fully seated. Momentum - or the moving inertia - of the bolt is used to keep the bolt closed until the bullet leaves the barrel. Once gas pressure has built up enough to overcome bolt momentum, it then pushes the cartridge face against the bolt to cycle the action.
Extractors are part and parcel of unloading unfired ammo, and controlling the way the case is propelled out the ejection port. Cases that stick in the chamber due to excessive expansion will not extract and sometimes suffer case head separation - if the extractor is sharp and maintains a grip.
Here's where things get confusing. Systems that impart energy to the bolt by DI or piston react differently than systems that use the case to move the bolt. DI and piston systems start a part of the bolt moving to unlock it. Blowback systems use the case to impart movement to the bolt.
In some sub gun systems, the chamber is even designed to overexpand the case to delay bolt movement by absorbing energy to literally resize the case - and delay or prevent premature extraction of the cartridge that could cause case wall failure. The stainless 9mm Pocket Pal also used the technique. It's also a good example of the force vectors involved. Rearward propulsion of the case out the chamber was so high delay was necessary to properly time the action cycle.
It's not the extractor doing that, it's a gas pressurized case cycling the action.
In a short action piston conversion for the AR - short meaning the travel of the piston is less than cartridge length - the piston rod strikes the lump with enough force to propel the bolt carrier group the full length of the action, compressing the spring. In a submachine gun, the case does that. In a DI gun, it's entirely possible the rearward movement of the bolt carrier group could be doing that. One thing that leads me to think about the case still pushing on the bolt face is that once the bolt is unlocked, the pin moves into alignment with the gas key slot - which prevents the bolt from rotating back into the locked position while cycling.
Some would have it that there is a lot of pressure on the bolt face during cycling trying to collapse it - and even sell a roller bearing to relieve the pin pressure in the upper. I conclude Stoner knew the case was pushing pretty hard on the bolt face and not just getting pulled along for the ride.
Haven't even touched on recoil operated mechanisms - which is the whole barrel moving backward against the gas pressure pushing the bullet forward.
