[_dps]

I share your dislike for a central language authority, but I think there are non-authoritative solutions that are consistent with the missions and practices of dictionaries.

One possibility: dictionaries could report that a large fraction of English experts or teachers consider a particular usage incorrect. This would be a dispassionate reporting of people's behavior as well.

You may ask "How do you pick the experts?" Well, how did we pick the people whose diction the dictionary supposedly represents? :) We make do with approximations. I think that the OED could easily afford to sample 1000 English teachers ranging from professors to grade school, and ranging across many geographical regions and social strata. It would actually be fascinating data to see alongside a definition!

Edit: I had meant to add that the sampling for the words need not happen for every (word, expert) combination; giving each instructor 1000 words chosen uniformly from a pool of 10,000 words should get you reasonable coverage for almost all of those 10,000 words. So the costs of this sampling experiment on both OED and the expert volunteers would be quite low.

[lutusp]

> I share your dislike for a central language authority, but I think there are non-authoritative solutions that are consistent with the missions and practices of dictionaries.

There are well-tested solutions. One is to have a language authority such as you suggest above, like the Académie française, which struggles against what it regards as erosions and distortions of proper French:

http://en.wikipedia.org/wiki/Acad%C3%A9mie_fran%C3%A7aise

Another is simply to list definitions in the order of their contemporary preference, as dictionary.com does. Interestingly, Merriam-Webster uses the reverse order, listing the oldest definitions first, as in the case of "Decimated", which once meant reduced by one tenth but now means substantially destroyed:

http://www.merriam-webster.com/dictionary/decimate

http://dictionary.reference.com/browse/decimated

Note the difference in listing order.

> One possibility: dictionaries could report that a large fraction of English experts or teachers consider a particular usage incorrect.

Apart from the problem of choosing experts, I think this would only slow the natural rate of language evolution. I think the problem is deeper -- it asks whether language is (or should be) susceptible to evolution by means of natural selection as people's tastes and needs change.