Though you may not accept it, this is really simple. You have a known signal for input. You can monitor the output say at the binding posts of the speaker. Transparent is when the signal at the end of the chain matches the signal input. It cannot be perfect at a bare minimum due to thermal noise. Other sources of noise can be measured and perhaps tamed or eliminated.
If your cable makes glaring differences the cable or equipment has an issue. While sources of this may be outside the audible band of frequencies it is only the change in audible sounds which can be heard as different. When done well, which isn't terribly expensive, good ole wire of the right type can transmit signal with darn little alteration of them. Certainly not glaring differences by any means. So no it isn't circular reasoning it is knowledge of how signals can be transmitted which leads one to the idea competently designed cable would not make a lot of difference.
That is completely circular. What variables are you measuring at the output and how does that specific set of variables correlate with acoustic output? That is why something like the Nordost/Vertex software would be great as it would provide more meaningful information beyond RadioShack meters, freeware, and 1950s engineering....
People use the same arguments regarding tubes vs transistors, vinyl versus digital, etc and round and round people go because the measurable variables that actually correlate with acoustic phenomena are missing. There is a lag. What really makes me laugh is the number of cheeseball engineers who hold court on forums and patronize audiophiles and yet own systems that will make animals and small children flee. The dichotomy is just too rich for words.....