I think most speakers sold these days produce a semblance of the real thing in a reasonably good way. So what separates a speaker that is just OK from one that is considered good by most people to one that makes things sound more real? I think that is the critical question and probably at the meat of this thread.
The ultimate goal is "realness" of reproduction, with all it's textures, nuances, dynamic variations, boundary interactions, phasing, timber, pitch, pressurization and airflow, clarity (not necissarily excessive detail here though - that can be artificial), etc. This is what professional musicians work on to perfect their craft. A speaker has to be able to reproduce all of these from all types of instruments, including voice (which is probably the most complex of all). Not an easy job, and each designer tends to focus on a limited number of these parameters most important to them and most easily solved by the type of speaker being designed - cone, planar, electrostat, horn, etc. The odds of getting them all fairly right are worse than winning the lottery.
So, my order of preference is surely:
tonality
continuousness (top to bottom balance)
texture (part of getting the clarity right and not overdone)
ability to disappear and image (clarity without haze)
inner dynamics
outer dynamics
range (top to bottom extension)
Then of course, there's the obvious dependency on upstream components to pass these characteristics along to the speakers. So the speakers may be capable of reporoducing the characteristic, but may not be receiving the information or the power to do so. I think each component needs to be able to have the same issues addressed as the speaker. That's the magic of synergy.
Enjoy,
Bob