... the most general answer belongs in realm of physics, not semantics, math of informatics. This is because these sciences are abstract, which means, they're operating with concepts, which are itself a generalization (i.e. abstracts) of many entities of reality.

I think what you have written here is quite true, and I do not think that the analogy of semantics as a phase shift in the way it is meant in Physics is a truly helpful way of looking at it. Language is fluidly responsive it is true, but it also has to conform to rules (even recognised exceptions acknowledge the existence of these rules). When languages absorb new words and meanings they do so under their own rules regarding spelling and usage. For instance the present evolution of texting (SMS) is actually not the seemingly random mess we may think it is at first sight. In fact, in order for us to decode the message certain syntactical rules have to be followed otherwise the message is unlikely to make sense.

Actually I think that the flexibility of maths to respond to change far outweighs that of language.