Far from being sleep inducing, such discussions can be eye opening. I must say, however, that Chicagos's generalizations (if quoted correctly) seem drug influenced to me, at least they leave me wide eyed and open mouthed. Put into practice, the results can be mind boggling.
I may be grammar obsessed, but all of the above doesn't look right to me.
Bob, you say: Anyway, where is the ambiguity in them? The ambiguity is in the grammar, it's not a semantical ambiguity. A coat that is moth eaten is not a moth being eaten, but when you encounter a noun in a sentence you rather expect it to be modified by a modifier, be it an adjective or a relative clause, not to modify something else. I think that was what MikeE was aiming at. When I said, further above, that the grammar in that product is the best known versus that product is the best-known is completely different I was trying to point out something similar: Of all detergents [available = that are available], X is the best [product that is] known [to us = that we know of] / of all those detergents, X is the best-known [detergent = the one most people are familiar with]. You are certainly right in saying that in moth-eaten, horse-drawn etc., there is no such ambiguity, but it seems rash to me to conclude that you can leave out the hyphen. What remains is the fact that, when you read such a sentence, you will have to 'recalculate' the syntactical role of the noun as you go along. In predicative position, the noun is not modified, it is a modifier. It makes you stumble.
And how in the world can a noun be "like an adverb," whether modifying a participle or anything else?
Easy. A chauffeur-driven car is a car driven by a chauffeur, a moth-eaten coat is eaten, but not by you and me for dinner, but by moths. The noun says how the participle must be understood, and as such, it's a modifier. In addition, the participle eaten on its own wouldn't make sense, moth-eaten is a true compound adjective and remains an adjective both in its attributive and predicative positions.
The editors were thinking of nothing other than the examples they provide.
Really? Then what does 'such as', 'usually', 'in general' mean?
Take the sentence you quoted above: When compound modifiers such as open-mouthed or full-length precede a noun, hyphenation usually lends clarity. ... When such compounds follow the noun they modify, hyphenation is usually unnecessary.
Open-mouthed and full-length are not even in the same class, so it's hard to imagine what such as means in this case.
You go on to say (in your own words, I believe), Compounds involving adjective + noun, ... noun + adjective, noun + participle, and a few others should generally be hyphenated before a noun but not after it.
I don't have Chicago and can't look it up, but do they really cover the case noun + participle in a predicative position? Or is that your conclusion?
Everytime (in #8) quoted the following problematic advice: Consult a dictionary. If not listed the compound should be open.
As you said above, you found results-driven in MW, but not reward-driven. You say: Since the latter is hyphenated, no doubt the former should be as well. If you follow Chicago's advice, it should be open.
Dictionaries list words. They specify parts of speech, but they usually don't specify syntactic functions. Can you tell if the editors had the attributive adjective in mind or the predicative use of the same adjective? If you follow Chicago, they would have to list both, as the syntactic function affects spelling.
Unless, of course, the adjective reward-driven ceases to be an adjective when you drop the hyphen. But then, what does it become? Which brings us full-circle to the initial grammatical problem. Either it's a compound adjective or it's a mere succession of words.
Sorry for rambling on for so long.