It is the specified white point for sRGB and Adobe RGB. Okay. That makes sense. When we're talking about a screen to print match though there's so many variables involved including the color temperature of the paper, and the color temperature of the light hitting the paper. I've found that in order to get an accurate match I really need to tweak the monitor's color temp to match my print accurately, otherwise it's difficult to accurately judge things. Not everyone uses the same light source to view prints. Viewing booths with 6500şK fluorescent light sources are very popular but Solux's 4700şk and 5000şk lights are popular as well. Another huge variable is the rather LARGE (I've observed differences for what a puck report a white point to be of over 1800şK for the same white point) discrepancy between what colorimeters report for the same white point.
When you consider how inconsistent colorimeters are in reporting color temperature, and the differences in color temp from paper to paper, plus the color temperature of the light used to illuminate a print it almost seems silly to specify a value to profile a monitor to, yet 6500şK is constantly preached again and again and again. Why is that?
You know when it comes down to it this color stuff is pretty complex, and the reality is that a lot of information gets "distilled" into what is hopefully a more comprehensible term or concept. In some cases that helps and in others it hurts. As you point out, it should not be assumed that viewing Adobe RGB tagged images on a monitor calibrated "K6500" will match a print viewed under a fluorescent with a 6500 K white point. To really explain this to someone who hasn't invested a fair amount of time understand basic color science is kind of complicated. Ideally we would not let people get away with using D65 and 6500K interchangeably, and people would understand the variables associated with trying to match emissive vs reflective sources, but not everyone has the time to learn all that. So in the end if you espouse the virtues of Adobe RGB and tell people to calibrate the monitors to 6500k(which many of the better monitors use as the target native white point) and it's somehow understood that viewing them under a 6500K source is ideal then it's at least giving them a fighting chance There's a whole separate debate about profiling the monitor to the "native" WP vs 6500k but I won't even go into that.
In the offset world everything is D50 based. Trying to get designers to calibrate their monitors to 5000k(hey this looks yellow) and proof their job using the press profile(hey it looks like crap now) will give them a much better idea of what the printed piece they are designing is going to look like, and yet it's not always easy to implement. Still if we can get the designers working in a color space that is somewhat close to the gamut of the press and if we can get the pressman to look at jobs in the light booth and if all the designers remember is D50 and SWOP then we've made some progress.
It's taken a long type for people to move from the density paradigm(I'm sorry for using that term) to colorimetric terms like L*a*b*. RGB is new for many people. Eventually I think true spectrally based color models will take hold, and fully account for all the issues you mentioned in your post, but I'm not holding my breath.