Universal 1645: head-marking anywhere ⇒ head-marking at the clause level; dependent-marking at the clause level ⇒ > dependent-marking at the phrase level
Universal 1645: head-marking anywhere ⇒ head-marking at the clause level; dependent-marking at the clause level ⇒ > dependent-marking at the phrase level
Original
If a language has major, salient, head-marking morphology anywhere, it will have it at the clause level. If a language has dependent-marking morphology at the clause level, it will have it at the phrase level.
Standardized
IF there is major, salient dependent-marking morphology at the clause level, THEN there is dependent marking morphology at the phrase level.
Nanai (Southern Tungus), in which phrases are head-marked, but clauses are mostly dependent-marked. Two potential counterexamples are Ket (isolate) and Ewenki (Northern Tungus). Ket uses subject and object cases, but is otherwise almost entirely head-marking: its sole phrasal dependent-marked pattern is partial agreement of adjectives with head nouns. Ewenki is much like Nanai, except that its adjectives agree. Ket and Ewenki thus support the second statement; but this is poor support, in that agreement of adjectives with head nouns is cross-linguistically not well correlated with over-all language type (Nichols 1986: 76).
Nichols (1986: 106) uses the term GOVERNED to refer to dependents whose presence is required, and whose morphological form is determined, by their heads; and the term SUBCATEGORIZED for those whose presence is required, but whose form is not determined by the head.
Nichols (1986: 106) uses the term GOVERNED to refer to dependents whose presence is required, and whose morphological form is determined, by their heads; and the term SUBCATEGORIZED for those whose presence is required, but whose form is not determined by the head.