Rodman,
Thanks for your thoughts about capacitance. When you say that the dielectric material "charges" (so to speak) over time, it seems to me as though the cable should sound worse over time. And the more charged, the worse it should sound.
The reason I say this is because air dielectric sounds best. And air is less chargeable than the worst sounding of the dielectrics, namely PVC. Electrostatic build-up is a very bad thing for audio, which is alternating current. Hence the many products on the market to avoid this.
So I don't agree with the "charging up" theory for dielectrics as the reason for cable burn-in resulting in a more natural sound.
I presently believe that the burn-in process could be due to something we can learn from what we know about magnetic domains.
When magnetic domains are large, or else there are several of them charged magnetically in a similar direction, the result is magnetism on a larger scale. Magnetism is a flux, a movement and direction of force.
Indeed, scale is all there is to magnetism. The stronger the magnet, the more uniform the domain orientation.
It could be that an amount of a different type of magnetic domains are present in wire such as copper or silver, which do not display ferromagnetism, as iron does. Perhaps "magnetic domains" is not the proper term, since there is no memory to speak of. Hence, no lasting magnetism. The domains might be present but not lasting, as they are in magnetized iron.
I postulate that during instantaneous applications, such as that of an electromagnetic alternating current running in real-time, perhaps these types of "magnetic domains" I am imagining are indeed active and influence instantaneously the electromagnetic signal in some small way.
So, getting rid of these gets rid of the memory of the cable's metal. Even if it is copper or silver.
A speculative analogy, crossing the fields of magnetism and psychology:
Magnetic memory in iron would be like the recollection of something fixed in your own mind. You can draw this thought up at any time (= detect the magnetism at any time). It "stays put".
Memory in copper or silver would be like associative thought in your mind, brought forth by the similarity of one thought to another. Thoughts in this state of flux are intertwined and depend on the preceding thought. It must "move" to be awakened.
So, if my theory is correct, if we can get rid of the "associative" type of memory in copper or silver, we've achieved the ultimate burn-in we can possibly achieve. The signals should pass without awaking associatively operating domains in the non-ferrous metal of the wire.
The reason I think this is true is that even cables with 99,999999 (add as many nines as you please...) pure silver with air dielectric still burns in.
Louis Motek