I've been trying to stay out of this...mostly it involves folks talking past each other, and there does seem to be something of an anti-science bias, or at least a misunderstanding of how modeling is done, on this site. What finally got me was Dillbag's insult to poor old Hooke,
Hooke's Law was never a good or even slightly ok approximation of anything involving a dynamic climbing rope.
This doesn't seem to be true, but in any case the fact that a climbing rope is constructed of a core and sheath provides no insight into whether or not Hooke's Law should be a good approximation.
Here is my current understanding.
1. Hooke's law provides an good model for the extension phase of a fixed length of dynamic climbing rope dynamically loaded. It does not seem to model a dynamic climbing rope under slow pulling, nor does it model the behavior of static ropes. These ropes have load-extension curves for slow pulling that have experimentally fitted quadratic equations. Whether or not the load-extension curve is "really" quadratic awaits an appropriate physical argument.
2. No one ever thought that Hooke's law should apply to the contraction phase, but the fact that a rope is neither a spring nor a rubber band does not mean that its load-extension curve might not be approximately linear. It follows that arguments whose only content is that a rope is not a spring are far from decisive.
3. Strat mentions the viscoelasticity of the polymer materials. This goes beyond anything I know about, but one has to remember that the physical construction of the rope has an enormous influence on its dynamic behavior, and the twisted core of a dynamic rope (as opposed to the straight core of a static rope) acts physically somewhat like a spring. So one cannot rule out Hooke's law entirely on the basis of a materials argument.
4. An ingredient that is not modeled by Hooke's law is the internal friction of the rope. There are, however, various ways to introduce this into the model. The Italian Alpine Club uses the classical model for a damped oscillator and claims to get very good agreement with experimental evidence. However, the damping term in the classical model is for so-called viscous damping, which is proportional to velocity, and there doesn't seem to be a physical justification for this assumption.
My guess (unfortunately, I don't have time to work this out now) is that frictional damping proportional to extension is a better bet, since as the sheath extends, it narrows in cross-section and so binds the core fibers more tightly.
5. Frictional resistance may explain the discrepancy in results for dynamic and slow-pull loading, since sliding friction is typically less than static friction. Indeed, it seems that a dynamically loaded rope whose tension is suddenly released does indeed spring back like a rubber band, although for slow pulling the recovery time is much slower.
If one forgot about internal friction, one would think that a stretched rope was somehow still under tension, even though the weight had been removed; an impossibility. What happens is that the tension force is nearly balanced by friction so that there is very little restorative force acting to contract the rope.
In view of all this, squatting or other fairly static methods of stretching the rope ought to provide the belayer with some slack to take in since recovery is slow. Once taken in, the rope's internal tension, though balanced by internal friction, will result in less stretch in subsequent falls. Of course, Dave already knows this and has known it for years, my only point being that this observation does not contradict theoretical considerations.