Yes, physical modeling is actually TAMPERING with the basic acoustic nature of a sound, by altering the waveform. Layering dynamics is not the same thing. You are still hearing mainly a recording of a performance, not a newly created tonality.
VSL, unless I am mistaken, is fundamentally opposed to the alteration of waveforms that is seen with modeling. Because at least up till now, their philosophy has been similar to that of classical record companies - to represent as pristinely as possible the original sound created by extremely good orchestral musicians, and under no circumstances mess with its basic timbre. Of course legato or alternation instruments dissect the performances, but they are still heard purely though in a new time frame.
One thing that is good about this is it preserves the original musicianship of the players. What happens with some poor sap who plays an instrument that is then altered by ripping it to shreds and then reconstructing it with physical modeling? He is a non-entity, that is what happens, at lest in reference to his musical performance. But the VSL musicians are highly recognizable - for example the solo trumpet, flute, even the ensembles have very distinctive characters.
Not that I am opposed to physical modeling. It will probably become normal in the future but right now it has a very artificial sound compared to highly detailed sampling.