Ah, but the paper that you cite is about Adjectival hierarchies (something I've noticed decades ago, and also wondered about, ever since I saw a sign written by someone who was obviously not a native speaker because of the misordering of adjectives. Glad to see stuff written about it). This is not about grouping at all. Quite the contrary. The examples given in the paper are, almost without exception, about the order of adjectives that are all at the same level (that is, modifying the noun), and not their grouping. e.g. "a small green bottle" (lo cmalu je crino botpi) and "a vegetarian Russian lawyer" (lo nalre'ucti je rusko fladju) What associative grouping is really talking about is precisely the examples I did give, which you dismissed as words that we might hyphenate or adverbs. But that is precisely the point. Because remember, since we don't have in lojban the traditional parts of speech of noun, verb, adjective, and adverb, what brivla function like can only be determined by their position in a bridi or clause. "broda brode" in isolation is the equivalent of either a noun adjunct or an adjective modifying a noun, or an adverbs modifying a verb (which in English is usually the other order) . "broda brode brodi", otoh, is either adverb adjective noun (dull gray house, frighteningly fast ride), (adjective/noun adjunct-noun) noun (stretch limo driver) or adverb adverb verb (very surprisingly caught), the latter construction not appearing often in English, and in those constructions, English is left grouping just like lojban. Contrariwise, constructionslike the ones you cited, adjective adjective noun, don't really group at all (usually). They are "adj. AND adj. noun". (otoh, verb adverb adverb constructions in English, DO right group (run really quickly)
--gejyspa