I was thinking something similar but along a different line of thought using English alone. Possessive modifiers behave differently than other "pure" adjectives and whether it's anaphors.
1) Jim's red barn
1a) My red barn
2) Red Jim's barn (Not awkward like in the headline example, but the meaning has changed completely)
2b) * Red my barn (invalid, period. even if my -> mine - I can do this word order in Korean or Japanese easily due to more explicit case marking)
We have to start using some syntactic movement to make the construct syntactically acceptable and semantically natural and equivalent, but doing so makes it possible to be free with word order:
3) Jim's barn that is red
3a) My barn that is red
4) Red barn that is Jim's
4a) Red barn that is mine (my -> mine)
5) Barn that is red and Jim's
5a) Barn that is red and mine
6) Barn that is Jim's and red
6a) Barn that is mine and red
The preferences aren't as clear-cut to me as in the headline clause, but we'd rather keep the possessive separate from the pure adjective (3-4 > 5-6, higher is more natural). Using only "pure" adjectives, it looks like the semantic binding preference still sticks though:
7) Big red barn =>
7a) Red barn that is big
7b) Big barn that is red
7c) Barn that is red and big
7d) Barn that is big and red
Naturalness: 7b > 7a and 7d > 7c. And since 7d > 7b, we have 7d > 7c >= 7b 7a. I'm thinking 7b > 7c but wasn't quite sure because both sound awkward and comparing two mildly awkward clauses is probably not valid here.
But hopefully other native speakers roughly agree on the precedence of awkwardness with me and understand why this article is just the tip of the iceberg.
1) Jim's red barn 1a) My red barn 2) Red Jim's barn (Not awkward like in the headline example, but the meaning has changed completely) 2b) * Red my barn (invalid, period. even if my -> mine - I can do this word order in Korean or Japanese easily due to more explicit case marking)
We have to start using some syntactic movement to make the construct syntactically acceptable and semantically natural and equivalent, but doing so makes it possible to be free with word order:
3) Jim's barn that is red 3a) My barn that is red 4) Red barn that is Jim's 4a) Red barn that is mine (my -> mine) 5) Barn that is red and Jim's 5a) Barn that is red and mine 6) Barn that is Jim's and red 6a) Barn that is mine and red
The preferences aren't as clear-cut to me as in the headline clause, but we'd rather keep the possessive separate from the pure adjective (3-4 > 5-6, higher is more natural). Using only "pure" adjectives, it looks like the semantic binding preference still sticks though:
7) Big red barn => 7a) Red barn that is big 7b) Big barn that is red 7c) Barn that is red and big 7d) Barn that is big and red
Naturalness: 7b > 7a and 7d > 7c. And since 7d > 7b, we have 7d > 7c >= 7b 7a. I'm thinking 7b > 7c but wasn't quite sure because both sound awkward and comparing two mildly awkward clauses is probably not valid here.
But hopefully other native speakers roughly agree on the precedence of awkwardness with me and understand why this article is just the tip of the iceberg.