A quick thought:

IMO, the goal of neutral playback from a music listener's perspective isn't to get rid of the recording's coloration, it's to get rid of the music playback systems's coloration.  So you don't need to use the same system they used.  Hearing the recording's coloration is part of what makes different albums unique and interesting.  Most music playback gear is so colored that all you have to do is get close to neutral and suddenly, you're not listening to colored gear anymore, you're just listening to colored recordings.  The recording's colorations are interesting, beautiful, and IMO almost never harsh.  It's the music playback's colorations that are bothersome. 

The recording's colorations are interesting, beautiful, and IMO almost never harsh.

Well....most recordings I hear have something wrong with them one way or another.  I don't just mean treble and sibilance either so let's not blame everything on 6khz peaks.      Reminds me of a track someone linked in the SS forum for ABing FLAC v. MP3.  It was recorded/mastered worse than 95% of what you hear using Youtube. 

 

A quick thought:
 
 IMO, the goal of neutral playback from a music listener's perspective isn't to get rid of the recording's coloration, it's to get rid of the music playback systems's coloration.  So you don't need to use the same system they used.  Hearing the recording's coloration is part of what makes different albums unique and interesting.  Most music playback gear is so colored that all you have to do is get close to neutral and suddenly, you're not listening to colored gear anymore, you're just listening to colored recordings.  The recording's colorations are interesting, beautiful, and IMO almost never harsh.  It's the music playback's colorations that are bothersome.
 

^ This.

By neutral, do you mean neutral so not to favor or harm any particular type of consumer/end-user playback system?  Or neutral as in a faithful 1:1 reproduction of the original performance?  Because I think the two would be very different things.  I guess the chances are that if you listened to reasonably well-recorded material, such engineered compromises would be fewer if any.  But it would still leave variables like:
 - Studio and hardware (speakers, amplification, room treatment, etc.)
 - Engineer's tastes and personal hearing capabilities
 - Engineer's and artist's artistic intent
 
 How can we account for all of these things?  To me, the cumulative effect of all of these would far outweigh the difference between two similar DACs, for example.

You have to make a distinction between accuracy and precision! Does the client want as close to real life as possible (accuracy) or do they want a performance that will give them a sonic landscape they can closely control and manipulate (precision). This should be the very first question on an artists mind when they begin recording. By the time it gets to the mastering engineer, this decision has been made. It is then up to the mastering engineer to make sure the resulting product will be as polished as possible. We use a neutral chain so that we can hear the performance and manipulate the infinite parameters of sound in a finite way so as to produce the best possible sound in various circumstances.  There is an art to this science and it comes in knowing how far you can push your gear, your ears and the consumers expectations.

I agree with the "more likely" part.  But at best, isn't it just a very rough approximation?  Maybe there needs to be a distinction between what is neutral and what the artist intended?  I think neutral can for the most part be measured and/or identified.  Most arguments over what sounds better seems to really be over what sounds better to that person- not what is more neutral.  But to identify what the artist intended would seem to require one to get inside of another person's head.  It just seems that in a game where people are swapping opamps or upgrading power supplies and hearing differences, there is a much larger variable that need to be accou nted for.  Otherwise we might be taking ultra-precise aim at the wrong target.
 
 (BTW, I don't mean to sound argumentative throughout this… I hope I don't come off that way.  I'm getting a slight headache thinking this through myself.  It took me over three hours to figure out this reply!!!)

Read my first reply. By the time the mastering engineer gets his hand on the album, the artist intent is well established and should be clear to the mastering engineer. We then polish the recording so that it can shine in the best light possible.  I never master something thinking “Oh….this ‘Scar Tissue’ guitar riff must sound neutral!” I ask myself…what should the spot light be on? Should there be multiple spot lights? Can I make these spot lights shine in a way which brings out the best without washing it out in brightness or making it too dull? Can I make these spot lights dance with it each other in a blend that will not harm the music? Once all that is answered, I aim at producing a natural sounding master.
That is what the artists hires me for. If they are second guessing my merit/judgment then I send them elsewhere. They must have faith in my abilities to properly polish a recording up. This is why I always try to have an in depth conversation with the artist to make sure he understands what he/she can attain using my skills.
You don’t sound argumentative btw.
Your other concerns are pretty much answered above. It comes down to dealing with and fixing what you are given. The decision of multi-tracked vs. minimalist approach is a decision that is made by the producer, artist and recording engineer. It depends on their goals, what they have available and what they can afford.
As far as a mastering engineer is concerned, a neutral system will reveal the flaws the mixing and recording engineers mixed. Sometimes you get a recording with low level hum or a high pitched noise which was missed by everyone else. We don’t aim for neutral recordings, just a neutral playback of what we are given so we can properly judge what needs fixing/enhancement. Sometimes our job is to simply do nothing. 

Tonal neutrality is difficult enough to attain, but with headphones I've found that one of the more important tricks is to catch on to how to adjust crossfeed to compensate for different approaches to recording/mastering. Some recordings are just anathema to headphone listening and by default sound lifeless if not coming through a speaker and at that point I think it's fair to do a bit of editorializing, through eq & crossfeed, to get to some semblance of a natural sound.

Then again there's also genre. As far as I'm concerned the usual audiophile pretensions of purity go out the window when listening to something like Aphex Twin. Modern electronic music is often layered with distortion for artistic effect and it becomes silly to sit back and worry about maintaining the neutrality of nintendo synths and sped-up samples of old vinyl.