Why do we want to distortions?

This growing consensus about the appealing nature of low and even order harmonics and repellent nature of high an odd order harmonics is fascinating. From a biological standpoint it suggests the question, “how did this come to be“? If we step outside the arena of audio reproduction and look at acoustics and hearing by themselves I would say from my own experience that certain environments, rooms, auditoriums have better or worse “acoustics“. So do we encounter pleasing and displeasing harmonics naturally and could this be a foundation to explain how evolution tunes our hearing to prefer one over the other. I don’t know. I offer it as a hypothesis. Maybe others here with more expertise in acoustics can comment. I do know that if you take an acoustic instrument outdoors it becomes a mere shadow of itself sonically, likewise in an anechoic chamber. However in certain spaces the same instrument becomes much more pleasing. Could it be possible that this is the effect of these various types of harmonic distortion occurring naturally? Tying it back to evolution could the more pleasing acoustic environment also have been the better one for us to function in both predators and prey? @anders65 thanks for your comments which spurred this line of thought.

There are different types of feedback and I’m referring to traditional "loop" feedback. Degeneration is a bit different. But I think error correction is much better if that is possible in the design.

Correction is good. Degeneration does not tend to cause higher ordered harmonic distortion the way that loop feedback does. But one thing about loop feedback- it lowers the output impedance while degeneration raises it. So loop feedback is often used for that characteristic alone. But if you use it, you have to understand that loop feedback makes distortion of its own thru bifurcation, so you'll need a lot to minimize this problem. That's why I say 35dB is a good minimum amount. Any less than that (on a bell curve) and the distortion generated is audible as brightness and harshness added to the input signal.

I have owned just one class D and I've heard a few in my time, just nothing that was really really good sounding. I'm not trying to pick a fight, I'm just saying that I haven't been introduced to a class D that sounded as good as many of the solid state A, AB and valve amplifiers I have been able to listen to over an extended period of time. 

For a long time that was my experience too. The first class D amp I heard had me incredulous that anyone would buy it! But that was a long time ago and times have changed.

From a biological standpoint it suggests the question, “how did this come to be“?

Sine waves don't exist in nature. So its logical to expect that to detect the sound pressure of a sound that the higher orders would be the ones that the ear is tuned to listen for; they are far enough away from the fundamental frequency that they stand out much easier. In addition, the ear is tuned to be most sensitive at birdsong frequencies, likely owing the fact that birds will pipe up if they find a predator in the area- if you can't hear the birds easily you could wind up dead. Today we know the birdsong frequencies as the Fletcher-Munson curve.

Thanks @bruce19 for opening the door to the acoustic aspects of encountering pleasure in music. I'm not an expert in this subject but I can at least give my view. I see the room is an extension of the instrument and one reason is room reflections.

We don't like listening to music in sound-wise dead rooms. We usually say, and I agree on that its much harder, if even possible, to have a good musical experience in such room. The reflections in a room increase the time a harmonic or any tone live - reverbing. Short pulses of tones of <100ms are less audible for us than longer pulses. Room reflections will thus have an effect on how loud we hear frequencies with short duration and maybe this is one part of musicality. I have good experiences by using sound diffusers to make the decay times a bit smoother over the frequency range of room reflections. 

Another interesting part of our hearing ability is localization of sound sources. This very developed ability may be a part of our evolution as both hunting and dangerous animals require very precise localization - sometimes in darkness or in high grass for example.

Our sound localization ability is made of a mix of detecting sound level differences between our ears together with the timing difference when the sound hit our ears. I did an experiment on myself using headphones that showed that it was possible to detect a time difference between the ears of as little as ~6-8 microseconds. In a stereophonic playback system, maybe this can be relevant when thinking about the holographic image?