AH, but have a think about what would happen if the jaws of the vice had a thread of the same diameter and thread form in them as the bolt, and the bolt was clamped into this thread.
Now assume that the vice is tensioned up to clamp onto the thread so now you have the frictional resistance caused by the pressure of the vice on the thread and also the shearing force required to strip out the thread, to be applied before you can pull the bolt from the vice.
I am neither an engineer nor a LAME but I suspect that, in the case of the cylinder heads under discussion, both the thread and the shrink fit are required to ensure that the cylinder head remains on the cylinder barrel. I always believed that the thread was the primary way that the head was secured with the shrink fit ensuring that the thread both on the barrel and in the head remained very tight so that there was no tendency for the head to fret or move during operation, especially as they are subject to such varying temperatures.
Again see my disclaimer above and I certainly will be happy if I have this wrong and someone with accurate information (data) sets me straight.
The 20000 pounds of force mentioned by Walter is just an approximation of the measure of the surface area of the cylinder in square inches multiplied by the maximum cylinder pressure in PSI and is the force being applied to the cylinder head during each power stroke. This is the force that the head to barrel connection must resist, with the appropriate safety margins to allow for changes in temperatures etc.
Edit: I posted this while Walter was posting. I have just learned some more. Thanks Walter.