( Second writing ) I decided to keep them both, so there is greater a chance of my awful language could being understood by some brave reader !
Elementary, my dear xarax !
Just after a night s sleep ( and even if I was awaken at one moment by a nightmare involving a certain roodent making grunting noises and trying to bite me 
), I saw the light of the day ! My mood has changed : No more questions, just some (tentative) answers.
I just made a “thought experiment” of the exact opposite. Instead of starting from a hitch, and trying to see if and how we can define it as a knot, we better start from a “knot”, and see if and how we can define it as a hitch. In a split second, everything became much more enlightened !
Imagine that we have our local, tight, dense tangle, the “nub”, which I have defined as the “knot” proper. How we can make a hitch out of it, literally ? Simply by penetrating it, from side to side, by a tensioned main line ( in the case of a hitch around a rope ) or by a rigid pole ( in the case of the common rope-to-object hitches ). What will happen ? The ex-knot will become a hitch, wrapped around the main line ot the pole, and bound to be knotted, literally again, by the mere presence of the main line or the pole inside its core. In short, a “knot” transformed into a hitch, which can not alter its existence, so it was, it is and it will be a knot, ever.
Now, imagine the diameter of the cross section of this main line or pole becomes zero - and, also, the straight, rigid line of the axis of the main line or the pole becomes curvilinear and flexible. What would have happen ? The ex-hitch would become a knot, again, wrapped around a curvilinear line of zero area cross section. In short, the hitch would become what it never ceased to be : a “knot”.
Why this knot-hitch-knot is a “knot” condemned to remain knotted ? Because of its topology, the fact that no segment of it is allowed to cross this imaginary mail line or pole of zero diameter, so, if it was knotted when it was wrapped around a real main line or pole, it would remain knotted, in theory, and so it should be considered as knotted, when it will be wrapped around this invisible, imaginary path.
I had discovered that I had already though of something like this, albeit in a slightly different context, which was dealing more with the (secondary) issue of a TIB or not-TIB hitch. However, to my surprize, I had not made the small step needed to be able to define what a hitch is, before examining if there is a difference between TIB and not-TIB hitches. The interested reader is kindly requested to read the following paragraph, thinking of main lines or poles instead of “neutral elements” - and realizing, as I did, today (!) , that what I am actually describing is the relation between the set of knots and its subset, the hitches.