HOW CAN WE EXPLAIN A SOUND?

To any exacting measurement? Forget it. No way.
TECHNICALLY? NO, NOT REALLY:
You're hearing a lightbodied, thin to medium jazz chorus Riki based sound with minimum twin layer chorusing about ½ effected to it that is light but noticably modulated (about 1-3 cents, pitch depth) with a linear fall rise rate completing the slewed chorusing cycle at about ½-1Hz. Hmm I don't think so.
Technically may approach the effects methods, but "a lightbodied, thin to medium jazz chorus Riki based sound" - that's rather coarse and vauge wouldn't you say?

Now if I told all that to a female guitarist, she'd slap me. A dude guitarist, he'd throw his drink in my face (well maybe just the ice LOL) But later after he got drunk enough he'd probably pee on me.
Even if the given guitarist was technically minded enough to know what I said, that still does not come close to pinpointing the sound. There are harmonic charts and spectral sound plotting methods, but it's so sureal and way involved; far too much for dealing with and better saved for it's intended purpose of alogrythms.

The closest, shortest and most understandable method is by utilizing pre-defined music itself in descriptions.
The basic template for that is...
From: {Pre-defined sound} of genre,artist,title,instrument / To: {Pre-defined sound} of genre,artist,title,instrument
EPU, Your sought sound might be similarly described as follows:
From: {Chorused Clear Guitar-thinner sound} of 80'sRock,Sting,"SOS",BGnd-2nd Guitar /
To: {Chorused Clear Guitar-thicker sound} of 80'sRock,Tom Petty,"Best Friend's Girl",2nd Guitar
This was a hands on thing taught to me when I was a producer's squaw. It helps the musician understand or "hear" hopefully in the neighborhood of the sound that is sought. It's exactly as accurate as the distance between the two pre-defined sounds that are given. It usually communicates a description good enough to put the musician on the same page as the person describing it.
They're right, you cant just pack it and ship it. The tracks you listed,5of6 use multilayer chorusing effects applied to the instrument's signal. We are dealing with multiple sets of infinite combinations.

If Roland Corp. had structured their Guitar C.O.S.M. method in any sort of understandably formulated way, it might be possible to finer that resolution, but they didn't.

Are there no areas among the pieces that you mentioned whereby you can sample a skeletal signal for a basis to start from? I heard some. It's just an idea. Depending on your final intentions of this sound, that method might be the premium choice, or could be a waste of time and effort and the worst thing toward your goal since sound samples have some serious limitations, even when near multiples are used.
If you know how to play the guitar and have a digital recording device I would go to Guitar Center, grab a Paul Reed Smith with a rear split/reverse coil pickup, a Rickenbacker, a stool, a set of headphones and take my chances on a Sholtz/Rockman Chamealion first. Next I would try Sony effects believe it or not if the Rockman preamp didn't grasp the sound. Once I got the exact sound, I'd bypass to retune, then go for it right out of the headphone jack and play what I am out to do in just about every way, position, and style I know, slow to fast. Just keep recording. They are cool with it as long as you put the gear all back where you got it from.