Thanks for you explanation, Paul, but it seems to me that the complexity you are talking about derives from the ambition to control playback simply from the player instructions used in scores.
You describe a situation in which (presumably) we have a instrument being played back by a suitable long-note articulation. We now add the instruction con vibrato (presumably previously the player was instructed to play senza vib.) In order to hear this in playback I may want to either change the actual articulation in which case I would need to add a key-switch, program change, or even (in the case of the Spitfire libraries) a C32 value - which I guess I should now be able to do using a midi controller lane. Whatever the chosen method this would result in the previous articulation being replaced.
Alternatively, if the current articulation includes controllable vibrato I might want to use a midi controller lane to draw in a suitable envelope. Yet another way might be to key-switch into play two articulations - one with vibrato and another without, and use a midi CC to crossfade musically between the the two. Because I might want to do any or all of these things in a score - the last thing I want the ‘con vib’ instruction to do is trigger a pre-determined one-size-fits-all ‘solution’.
Next, you add the instruction ‘legato’ this would not typically require a change of articulation or indeed the change of a CC value - it’s only a question of note lengths.
I know that many sound libraries have articulations labelled ‘legato’ but this means something different in the library context. You wouldn’t necessarily want the instruction ‘legato’ to change to this, or any specific articulation - legato is possible with any sample longer than the note required.
So, as far as I can see, the ultimate in flexibility derives from human control over a few basic midi commands - easy for the computer and easy for the human.