Grading curves set median grades based on departmental or administrative preferences to ensure consistent marking between instructors at a given university or department in classes with a threshold number of students (Harvard uses 30) to apply statistical significance.
In fact, grading curves are a common method to prevent grade inflation. The curve does not have the purpose you imply it does; it lowers opportunities for high grades as well as increases some low grades.
Grade inflation is a continuing problem, and has been for decades. See:
http://www.azstarnet.com/grades/1928.html
As you can see, that occurred 70 years ago, largely because instructors ignored the curve and its limits on upper level grades.
Social promotion is an interesting case. You imply that students are simply upped into he next grade, but this isn't actually the case. Different boards use different strategies. Age-grade coincides with supplemental studies when the program is properly implemented, and on the flipside, grade skipping tends to also be limited. In both cases, supplementary material is applied to the individual student's curriculum, either as a remedial method or to provide additional challenges.
Academic pactices are not generally based solely on sentiment or tradition. They are based on academic analysis linked to the issues that come up in actual institutions. Casually judging them is problematic, and likely to come up with inaccurate generalizations. Applying them to martial arts is likely to be even more problematic.
While a veneer of academic clout may be tempting, martial arts knowledge is primarily demonstrative -- i.e, we know somebody is skilled because they can actually do something and practically apply their knowledge. In education, this is actually considered a viable third alternative to rigid testing and individualized study, both of which have problems. Fortunately, martial arts schools are not bound by the time constrainst and scale of traditional education.
I think that usually, it's the arts that emphasize doing that maintain consistent standards. I note, for instance, that BJJ belt and Judo belts rarely experience questions about their competency. A BJJ blue has demonstrated what he knows repeatedly against live opponents. If a practitioner from demonstration-based art knows his stuff, *you* also know.
The only problem this creates is that demonstrations must be broad enough to encompass the full curriculum of a given art. I would go further to say that testing reflects the values of that art. In my case, I would trust the guy who can apply principles in an effective way and share them with another to the extent that they can do the same, over someone who studies something with more theoretical depth, but whose primary demonstrations have been static scenarios, essays and the use of art-related terminology.
